MONEY IN MOTION THE JEROME LEVY ECONOMICS INSTITUTE SERIES General Editor: Dimitri B. Papadimitriou, Levy Institute Professor of Economics, Bard College, Annandale-on-Hudson, New York Ghislain Deleplace and Edward J. Nell (editors) MONEY IN MOTION: The Post Keynesian and Circulation Approaches Geoffrey Harcourt; Alessandro Roncaglia and Robin Rowley (editors) INCOME AND EMPLOYMENT IN THEORY AND PRACTICE Dimitri B. Papadimitriou (editor) ASPECTS OF DISTRIBUTION OF WEALTH AND INCOME STABILITY IN THE FINANCIAL SYSTEM Dimitri B. Papadimitriou and Edward N. Wolff (editors) POVERTY AND PROSPERITY IN THE USA IN THE LATE TWENTIETH CENTURY Money in Motion The Post Keynesian and Circulation Approaches Edited by Ghislain Deleplace Professor of Economics University of Paris 8 - Saint Denis and Edward 1. Nell Malcolm B. Smith Professor of Economics New School for Social Research, New York First published in Great Britain 1996 by MACMILLAN PRESS LTD Houndmills, Basingstoke. Hampshire RG21 6XS and London Companies and representatives throughout the world A catalogue record for this book is available from the British Library. ISBN 978-1-349-24527-7 ISBN 978-1-349-24525-3 (eBook) DOI 10.1007/978-1-349-24525-3 First published in the United States of America 1996 by ST. MARTIN'S PRESS, INC., Scholarly and Reference Division, 175 Fifth Avenue, New York, N.Y. 10010 ISBN 978-0-312-12543-1 Library of Congress Cataloging-in-Publication Data Money in motion: the post Keynesian and circulation approaches I edited by Ghislain Deleplace and Edward J. Nell. p. cm. - (The Jerome Lcvy Economics institute series) Includes bibliographical references and index. ISBN 978-0-312-12543-1 I. Money. 2. Keynesian economics. II. Nell, Edward J. III. Series. HG221.M8142 332.4-- M. Money, M, is invested in capital assets, 'C', and over time these capital assets are expected to generate money flows, 'M", which are sufficiently greater than M so that the fear and loathing of uncertainty is overcome and investment is both demanded and financed. The model for understanding an M ~ C ~ Ex M' economy must include bankers, businesses that finance activity and positions in capital by borrowing from banks or through the good offices of bankers, and households that directly or indirectly own the instruments bankers market and that finance the position of dealers in financial instruments. 24 These classes of agents are to be in the formal analytical structure at the beginning of the analysis and are not to be introduced into the argument in order to transform an unwanted result into a predetermined desired result. Thus, as a minimum, the agents in the analysis have to include consuming units which are the ultimate owners of wealth, business firms which are the proximate investing units and banking units.2s M' is a flow of money incomes through time that is available to validate the investment of M (of money) in C (an investment output): M' is the capital income or profit derived from production. The pro forma's of the negotiations between bankers and business men are conjectural (or ex ante) income statements in which revenues and costs are estimated. The income and spending statements of businesses, households, government units and classes of financial organizations are ex post summaries of what happened in the economy. The structure of an economic model that is relevant for a capitalist economy needs to include the interrelated balance sheets and income statements of the units of the economy. The principle of double entry bookkeeping, where financial assets are liabilities on another balance sheet and where every entry on a balance sheet has a dual in another entry on the same balance sheet, means that every transaction in assets requires four entries. Furthermore, the instruments of balance sheets include promises to make payments in some agreed upon currency either on demand, at some time, or if some contingency occurs. The M' states that the means to make such promised payments is normally derived from either realized or expected future income flows. 26 Looking at a capitalist economy as a set of interrelated balance sheets and income statements, recognizing that items in balance sheets set up cash flows through time, and integrating balance sheet changes with income flows, leads to an alternative structuring of the economy to that which takes the technical conditions of production, a construct called preferences and maximizing behavior as primitive economic concepts. In this modeling maximizing behavior remains 78 Characteristics of Post Keynesian Economics important, but the maximizing behavior of critical importance takes the form of present decisions that Ms over time will exceed M with an ample margin of safety. The appropriate construct to use in modeling such relations is the family of short- and long-term cost curvesY The maximizing decisions that lead to M ~ C action (financing M of spending on C of investment output) cannot be divorced from uncertainty. Of course, the importance of the linked balance sheets and income statements depends upon the economy being capitalist, but capitalism is the economic society in which we actually live and whose behavior we seek to understand. Two Price Levels For non-neutrality of money (and finance) to be a deep part of a model of the economy and not an afterthought, the monetary (and financial) variables of the model need to enter in essentially different ways in different parts of the system. Following Keynes, an increase in structure is obtained by separating aggregate demand into investment and consumption demand and allowing for two sets of prices to be endogenous to the model: the prices of current output, including investment output (the CPI), and the prices of capital and financial assets (the Dow Jones). The proximate determinants of these two price levels are quite different. Output prices are the vehicle by which businesses recover their direct costs, acquire the funds that enable firms to pay overhead costs and earn profits. Money wages are the main component of direct costs. The mark up is the way firms capture the funds that pay overhead costs and earn a profit. As direct costs are reduced to labor costs in a closed economy, the price level of current output can always be decomposed into labor costs per unit of output and a mark up. Part of the mark-up is used to pay the overhead labor, the rest in the integrated economy are gross profits. The aggregate of the mark-ups that can be earned by the producers of consumption goods is determined by gross profits available for the producers of consumption goods, which in tum are determined by the composition of aggregate demands. In particular the gross profits available for the production of consumer goods is not determined by an interaction of production functions and preference systems which would have it that the gross profits are equal to the marginal product of capital times the quantity of capital. Indeed, under heroic simplifying assumptions, the total sum of the mark-ups that can be earned in the production of consumer goods is determined by the wage bill in the production of investment outputS. 28 In this way of looking at the economic process, profits in the production of capital assets are determined in the negotiations between the purchasers and producers of investment goods, where the weight of an agent in determining this mark-up is determined by its market power (access to financing is one attribute of Hyman P. Minsky 79 market power). Aggregate profits in the simple skeletal case are equal to the wage bill in the production of investment output, plus the profits earned in the production of investment output: aggregate profits equal the total spent on investment goods. Thus the outcome of the negotiations between business men and bankers leads to a money amount of financed investment, which is the savings of all units even though some capital income finances consumption spending. Aggregate financed investment over an accounting period, as determined by the decisions of 'bankers and business men' , determines the mass of profits available to businesses. Once this mass is determined, then the competition among 'capitals' for profits determines the details of what is produced and employment. Hicksian Non-neutrality All economists are familiar with one model in which money is not neutral. It is the fixed money-wage IS-LM model of Hicks. The IS-LM model without a labour market reduces to the familiar graph as seen in Figure 2.1. In this graph an increase in money from Ml to M2 leads to a lowering of interest rates and an increase in income. Note that in the 'liquidity trap' an increase in money from M3 to M4 does not change either income or the interest rate: money is neutral when the economy is in a low-level 'equilibrium'. Now for some of the history of thought. Pigou had a labor market determination of aggregate output prior to Keynes's General Theory.29 One objective of the General Theory was to create a model of the economy in which the standard labor market equilibrium does not determine an economy's normal state or center of gravity. Most well-trained economists of the time (and ours) are unwilling to give the monetary-financial sphere the full partnership in determining the behavior of the economy, which is central to Keynes's vision of the capitalist economy. ---Is Figure 2.1 Graph of Hicks' model 80 Characteristics of Post Keynesian Economics As a result economists grafted a labor market determination of aggregate income upon the exposition designed by Hicks. In this view, in an unemployment position of the standard Hicks diagram, market forces determine two levels of income: one determined by the Keynesian aggregate demand relations and the second determined by the labor market. These two have to be reconciled. The reconciliation begins by noting that unemployment is like any other excess supply situation: excess supply of labor leads to the lowering of wages and thus of prices. This increases the price level deflated value of the presumed exogenously determined money supply and this, in turn, increases the level of consumption at every level of income. 3o Such an upward drift of consumption due to price level deflation shifts Hicks's IS function to the north-west. With an unchanged quantity of money both income and interest rates increase. This wage rate-price level reaction ends when the excess supply of labor is eliminated. As there is no eqUilibrium outside full employment which is determined by productivity and thrift, money is neutral. Recent work by Caskey and Fazzari, and De Long and Summers,3) tend to validate the Keynesian theorem that if price level flexibility exists then an initial condition of unemployment is likely to raise, not lower, unemployment. In this work, if dp/dt < 0 then either or both of the burden of private debts increases or the real (price-level adjusted) interest rate increases. This means that in an economy with private debts and Yf > Y, a fall in money wages and money prices will lead to a fall in Y that swamps whatever increase in Y that is triggered by the effect enunciated above. This is so because the fall in wages and prices increases the already heavy burden of debts, which has adverse effects on the viability of financial institutions and the financing of investment. Keynesians and macroeconomists in general need to distinguish between relative price flexibility and price-level flexibility. Relative price flexibility serves a useful purpose in resource allocation, whereas the usefulness of price-level flexibility in response to excess supply of labor is questionable. Liability Structures The burden of debt is a useful concept for macroeconomic research. We distinguish classes of units in debt: business, households, government and international. During each accounting period a portion of the revenues of each economic agent has been prior committed by debt, equity and lease contracts: these prior commitments are on account of both principle and interest. In the stripped Keynesian model which draws upon Kalecki we have for business firms: Pi = I, Profits equals investment. In the more complete statement we have Pi = I + Gov Def - Bal Tr Def + C(Pi) - S(w). Hyman P. Minsky 81 Internal finance is: Int Fin = Pi - Tx(Pi) - (lnt + Prin) Bnds - (Int + Prin) Lns - Cust Div. Note that over any period the greater the interest and principle paid on bonds and loans and the qreater the customary dividend, the smaller the amount available for internal finance. Furthermore, the greater these prior committed payments due to the liability structure, the smaller the fall in Pi that would lead to an inability to fulfill financial commitments out of the proceeds from market revenues. Aggregate internal funds is a rectangular hyperbola in the price investment plane. For a fixed aggregate profits (Pi) (Figure 2.2), the greater the tax rate on profits, the level of indebtedness, the interest rate and the traditional dividend, the smaller the aggregate internal funds. Note that if debts are either short term or long term but rates are adjustable to market rates (a prime + convention for the indexing of rates), then a rise in interest rates will draw the internal finance rectangular hyperbola towards the origin, reducing the amount of investment that can be financed with internal funds. A sufficiently great rise in interest rates may make the carrying costs on outstanding debts greater than profits. Firms in such a position will be unable to finance any investment internally and may be forced to sell assets to meet payment commitments on debts. Lenders and borrowers risk enter into the determination of investment. The Pk depends upon expectations of future profits, which in tum reflects the maintained model of the economy of the agents whose expectations are relevant to investment. The PI is a mark-up on labor costs. The intersection of the internal finance hyperbola with the PI line yields the maximum internally financed investment (Int on Figure 2.2). Greater investment depends upon the willingness of external financiers (bankers) to lend and business to borrow. The extent of borrowing depends upon the risk assigned to leveraging by borrowing firms and the risk assigned to lending to levered firms by financiers. The result is that at some point beyond the investment that can be financed internally the 'demand' price for investment goods falls away from the P k line because of borrowers risk, and at some point the 'supply' price of the investment good rises above the PI line because of lenders risk.32 (I, reflects greater borrowers and lenders risk premia than 12,) Investment finance depends upon internal funds and the availability of external funds. The liability structure that results leads to a prior commitment of profits earned as the future unfolds. The aggregate burden of the debt depends on the size of these prior commitments relative to the gross profits of the firms in the economy. In a capitalist economy where positions in capital are financed in part by debt, a collapse of aggregate profits will lead to an explosion in the burden of the debt. 33 82 Characteristics of Post Keynesian Economics P R Pk ~----~--------~--~------ E It It h INVESTMENT Figure 2.2 The Minsky diagram 3 THE SPECIAL MINSKY VIEW The above identifies Post Keynesian economics with a theory of investment which emphasizes the cash flows that firms earn and expect to earn, and the need to finance investment programs by a combination of internal and external funds. The emphasis upon expected cash flows as the motivation for investment and realized cash flows as the source of funds, which validates or does not validate the payment commitments embodied in the liability structure, means that a market rather than a technological determination of these cash flows is needed. This weds the Keynesian analysis which emphasizes the capitalist nature of the economy being analyzed with the Kaleckian analysis of the determination of gross profits. 34 It was pointed out that the internal funds available for investment are determined by the investment plus investment surrogates through government deficit financing and the international trade balance minus the funds committed prior to the validation of outstanding debts and to a customary dividend. A growth of debt relative to profits or a sharp rise in interest rates can decrease the internal funds available to finance investment. A special 'Minsky' hypothesis states: 'Over a run of good times borrowers' and lenders' risks are attenuated and the ratio of debt financing to internal financing increases.' In as much as the circuit runs from total investment to profits, a rise in the ratio of total investment to internal funds leads to a rise in profits, which in the aggregate makes the validation of the increased debt financing 'easy': furthermore, those who leveraged prosper. The interest rate Hyman P. Minsky 83 structure during periods when the business debt burden is light provides an inducement to business men and their bankers to not only leverage investment but also to (i) use short rather than long-term debt financing, and (ii) refinance lightly indebted positions in capital assets so as to increase indebtedness. I characterize payment commitment - cash flow relations as being either hedge, speculative or Ponzi finance, depending upon whether the cash flows can readily pay interest and the principle due on maturing debts during the period of reference (hedge finance), whether there is a need to refinance part or all of the principle falling due each period even as interest can be paid out of earnings (speculative finance), or whether there is a need to capitalize interest that is due (Ponzi finance). A special Minsky conclusion is that over a period of good times the weight of units in first a speculative and then a Ponzi plus speculative finance posture increases. This progression transforms the financial-economic system from being robust, in the sense that small changes in indebtedness, interest rates or profit flows are damped out, to being fragile, in the sense that such small changes will trigger large responses. It also follows that in a small government-constrained central bank capitalist economy this progression of financing relations makes the economic environment hospitable to a debt deflation process. A not unusual event, which in a robust financial structure has no significant effects, will have large consequences in a fragile economy. I believe the events of the past decades more or less conform to the special Minsky case, with government plus central banking containing the thrust to incoherence. In this view, the large government deficit and the so-called bail-out of savings institutions and banks aborted developments which would have led to a large depression. One aspect of this view is that what happens during business cycles is determined by a combination of endogenous developments and policy interventions. In the abstract even a debt deflation without the constraints due to government intervention has a natural ending, when all debts are wiped out and all capital assets are held in simple ownership. It is also noted that unless there is a breakdown in the authority of government, intervention will take place to halt the final liquidation of the banks and of all debts that this implies. 4 CONCLUSION: THE POST KEYNESIAN VIEW The subject of Post Keynesian economics is capitalism. As a result of the demise of the Lenin-Stalin model of socialism, the problems raised by Post Keynesian economics is of even greater relevance than hitherto. From a Post Keyensian perspective the problem of the creation of capitalism in what had been the Socialist economies is the problem of creating a financial structure. In particular there are capital assets in industries that would be best private but there is no wealth in the population which can purchase them. 84 Characteristics of Post Keynesian Economics This is also, but to a much lesser extent, a problem in those rich capitalist economies, such as Italy, where privatization may be the order of the day but there seems to be a shortage of domestic private wealth which is willing to take positions in these privatized firms. It seems as if it is no simple matter to get off the back of the tiger of a Socialist scheme, for the financial prerequisites of capitalism have to be created which includes the existence of private wealth which, of course, does not exist in these countries. The problem with the theory that allows for I. 2. 3. 4. 5. 6. capital values to be endogenous as the capitalization of income; that has the capitalization of future incomes determine the demand price of investment; which has the supply price of investment as a function of the money wage; which allows for the internal and external financing of investment and positions in capital assets; which has debt structures that are prior claims to the income imputed to capital assets; and in which gross capital income, as well as the overall level of income, is determined by investment and investment surrogates is that the nice equilibrium solution does not exist. The modeling leads to complex non-linear time-dependent relations which, as attempts to model and solve or simulate ideas has shown. lead to complex time series which exhibit what can be considered periods of chaotic behavior. All however is not lost. If we refer back to work done in the 1950s and early 1960s. ceilings and floors. which could be manipulated by the Federal Reserve System. transformed explosive accelerator multiplier processes into rather wellbehaved constrained cyclical processes. 35 Analytically. the ceilings and floors of these models are the imposition of new initial conditions and the time series that 'run' are the result of the structural interdependencies, the reaction coefficients which determine the coefficients that reflect interactions and the initial conditions: when a dynamic model is in an unconstrained run then the initial conditions for time t are the model determined values of prior periods. Policy. including the policies that. for example, are built into a tax structure, can make the initial conditions for the subsequent run of the economic process other than what the model would determine. In this interpretation a policy regime, which includes much of the institutional structure of an economy, is a device to prevent or contain the endogenous disruptive forces in a complex capitalist economy. The essential conclusion of this variety of Post Keynesian economics is anti laissezJaire. that market economies, which are as convoluted as modern capitalist economies, require an apt policy regime if the endogenously determined thrusts to incoherence are to be contained. Hyman P. Minsky 85 Notes and References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. Introduction to the French edition of John Maynard Keynes, The General Theory Of Employment Interest and Money as reprinted in Collected Works of John Maynard Keynes, Vol. VII pp. xxxiv-xxxv. Henceforth I will refer to this work of Keynes as the General Theory. I have taken the liberty of breaking what Keynes wrote as part of one paragraph into these sound bites. See A. Klamer, Conversations with Economists (Totowa, NJ: Rowman & Little, 1983) for the disdain that many of the successful mainstream economists hold for Post Keynesian views. Post Keynesian economists have done better at analyzing and explaining the progress of advanced capitalist economies, over the entire post-war period as well as over recent years, than a representative orthodox or mainstream economist. This is so because Post Keynesian economists reject any formulation of the economic process th'at splits the monetary and financial aspects of the economic process from so called 'real' sectors. The distinction between the Economics of Keynes and Keynesian economics was made by Axel Leijonhufvud, On Keynesian Economics and the Economics of Keynes (London: Oxford University Press, 1968). J.R. Hicks, 'Mr Keynes and the Classics: A Suggested Interpretation, Econometrica, 1937: 147-59. The introduction of a simplistic supply and demand for labor as determining the equilibrium employment and real wage, which dominate the aggregate demand determining relations in setting the equilibrium of the economy, is the key step in forcing the Keynes structure into an equilibrium framework. The result has been a research program which requires that macroeconomic relations conform to what are taken to be the microeconomic conditions. These forecasting models were initially derived from the tools used in macroeconomic planning during World War II. In the war and early post-war era the close interrelations that are normal for a capitalist economy among investment and the banking and financial structure were attenuated. It is not at all surprising that models which ignored finance were adequate for forecasting purposes in the first decades after World War II, and lost their power as instruments for forecasting when money and finance became of increasing importance. William C. Brainard and James Tobin, 'Pitfalls in Financial Model Building', American Economic Review, May, 1968. This means that the equilibrium of the economy is not only independent of the values of the monetary variables, but that the equilibrium is independent of the path by which it is achieved. This separation is particularly clear in Keynes's rebuttal of Prof. Jacob Viner's review of the General Theory. J. Viner, 'Mr Keynes on the Causes of Unemployment', Ouarterly Journal of Economics, November 1936; J.M. Keynes, 'The General Theory of Employment', Quarterly Journal of Economics, February, 1937. Keynes had the prices of assets, both financial and capital, depend upon the Q's (Quasi rents), c's (carrying costs) and L (liquidity) they yield - General Theory: 225-7, see also Hyman P. Minsky, John Maynard Keyne.v (New York: Columbia University Press, 1975), 4. At several points in his argument Keynes assumed that there is an equilibrium implicit in a set of initial conditions and adjustment processes. However, the changes that the search for this equilibrium brings ahout change both initial conditions and the search process. Therefore at best the implicit equilibrium is changing 86 II. 12. 13. 14. 15. 16. 17. 18. 19. Characteristics of Post Keynesian Economics even as adjustments are made on the basis of a prior observed disequilibrium. There is no need for such a process to converge to an equilibrium, let alone a unique equilibrium. I know of three articles that are still in the development stage that add financial concerns to aggregate models and derive time series that endogenously generate chaotic conditions. These three are: S. Keen, 'Goodwin + Minsky::: Chaos', School of Economics, University of New South Wales; T. Mott, 'The Effects of Changes in Income Distribution, Financial Conditions and Tax Policy on the Dynamics of a Kalecki-Minsky Macroeconomic Model', University of Denver, Edward Slattery, Southwest Missouri State University and Grainger Caudle, Mars Hill College; D. Delli Gatti, M. Gallegatti and H.P. Minsky, 'Financial Institutions, Economic Policy and the Dynamic Behavior of the Economy'. Keynes was not without responsibility for the integration of his theory into the equilibrium structure of neo-Classical theory. He speaks of the equilibrium of the orthodox theory as one of a multitude of possible equilibria and he accepted, however mistakenly, the J.R. Hicks interpretation of his theory. It will be pointed out in the discussion of the Hicks diagram that money is neutral in 'the liquidity trap'. 'The objectives of agents that determine their actions and plans do not depend upon any nominal magnitudes. Agents care only about real things such as goods, (properly distinguished by states of nature) leisure and effort. We know this as the axiom of the absence of money illusion, which it seems impossible to abandon in any sensible analysis', Frank Hahn, Money and Inflation, (Cambridge, MA: MIT Press, 1983); Gerard Debreu, Theory of Value, (New haven: Yale University Press, 1959); Kenneth Anow and Frank Hahn, General Competitive Equilibrium, (Oakland, CA: Holden Day, 1971). R. Lucas, Models of Business Cycles, Yrjo Jahnsson Lectures, (Oxford: Basil Blackwell, 1987) p. 20. Ex signifies expectations and U(c) is a utility function with consumption as the sole operative variable. J.G. Gurley, Review of Milton Friedman, 'A Program for Monetary Stability', Review of Economics Statistics, 43, 1961: 307-8. Money illusion occurs when some agents interpret an increase in the price of what they sell that is the result of inflation as an increase in their real income, while others make the correct inference. 'Despite the importance of enterprises and money in our actual economy, and despite the numerous and complex problems they raise, the central characteristic of the market technique of achieving coordination is fully displayed in the simple exchange economy that contains neither enterprises nor money', Milton Friedman, Capitalism and Freedom (Chicago: University of Chicago Press, 1962) p. 14. The 'real' economic results such as employment, its distribution among outputs, relative prices, including wages and interest rates are' ground out by the Walrasian system ()f general equilibrium equations, provided there is imbedded in them the aCll/al structural characteristics of the labor and commodity markets, including market impel/ections, stochastic variability in demands and supplies, the costs of gathering ;,!formation about job vacancies and labor availability, the costs of labor mobility and so on', according to Milton Friedman, 'The Role of Monetary Policy,' American Economic Review, 58 (I), March, 1968. Hyman P. Minsky 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 87 G. Akerlof, The Market for Lemons: Quality Uncertainty and the Market Mechanism, Ouarterly Journal of Economics, August, 1970, pp. 88-500 It is now well known that the existence of a general equilibrium depends upon assuming that agents have perfect foresight; the equilibrium, if it exists, is not unique and in general the equilibria are not stable. See B. Ingrau and G. Israel, The Invisible Hand: Economic Equilibrium in the History of Science (1990) transtated by Ian McGilvray, (Cambridge, MA: MIT Press). J. Stiglitz and A. Weiss, 'Credit Rationing in Markets with Imperfect Information', American Economic Review, June, 1981. Private or asymmetric information implies a monopoly. Each agent knows what she knows and others need conjecture what she knows. Instead of dumb and smart agents, the asymmetric information argument postulates that markets are rife with possibilities of adverse selection: only lemons will be offered on the market whether it be for used cars or financing. J.M. Keynes, General Theory, p. 3. The wording allows for complex financing layerings where there are 'bankers' (money managers) such as mutual and pension funds which take positions in assets created by 'bankers', as for example the packaging of mortgages into securities. Of course, the real world includes governments as well as other countries. Furthermore the 'banking' component of the model opens up to include a wide variety of financial institutions in addition to the banks which issue liabilities that are counted as money. If a capitalist economy is functioning normally then expectations of future Ms will enable a unit to obtain at least some of the funds currently due on liabilities by rolling over the principle part of the payments that are due. The primacy of costs and the division of costs into out of pocket costs, overhead costs and financing costs is part of the additional structure needed for the modeling of a modern capitalist economy. . Whereas the labor and material costs in the production of investment goods are well determined, the mark-ups in the production of investment goods depend upon bargaining among the producers of investment goods and the purchasers of investment goods. As far as profits in consumer goods production is concerned, 'the mark-up' per unit of output is an ex ante variable determined by producers for those consumer goods whose producers are able to exercise market power and is an ex post variable for those consumer goods that are produced and sold under competitive conditions. The aggregate of profits over consumer goods sold under competitive and market power conditions adds up to the amount determined by the structures of demand. A.C. Pigou Theory of Unemployment, 1933. London: Macmillan. There is an ambiguity in the mechanism that leads to the upward drift of the consumption function: whether it is due to accumulation or to a rise in the price level deflated value of money balances. A further ambiguity is whether money is outside or inside. John Caskey and Steve Fazzari, 'Aggregate Demand Contractions with Nominal Debt Commitments', Economic Inquiry, October, 1987, pp. 283-97; J. Bradford De Long and Lawrence H. Summers, 'Is Increasing Price Flexibility Stabilizing?' American Economic Review, December 1986, pp. 1031-44. Lenders risk can be observed in higher interest rates and tighter covenants on financing contracts. Borrowers risk does not have such observable measures but takes the form of firms that have not used all their 'borrowing' power. Albert G. Hart, 'Debts and Recovery Twentieth Century Fund', 1937, noted that the burden of indebtedness of railroads increased between 1929 and 1933 because income fell. 88 34. 35. Characteristics of Post Keynesian Economics This is also the model of the determination of profits of Jerome Levy. See S. Jay and David Levy, Profits and the Future of American Society (New York: Harper & Row, 1983). Hyman P. Minsky, 'Monetary Systems and Accelerator Models', American Economic Review, December, 1957; Piero Ferri and Hyman P. Minsky, 'A Linear Model of Cyclical Growth' , Review of Economics and Statistics, May, 1960. 3 The Money Supply Process: A Historical Reinterpretation Basil J. Moore INTRODUCTION There is currently profound disagreement about the nature of the money supply process. I The mainstream view, as presented in most textbooks, takes for granted that the money supply should be viewed as under the exogenous control of the monetary authorities, in principle at least, via the base-reserve-multiplier process. The money supply bears an empirically stable relationship to bank reserves and to the base. These magnitudes in tum, which constitute the liabilities of the central bank, are directly controllable by open market operations. The money supply function should therefore be modeled as vertical, or at least sharply upwardsloping, in interest money space. The money supply is an exogenous variable determined by the monetary authorities. The Post Keynesian challenge insists in contrast that the money supply should be viewed as endogenous, credit driven and demand determined. The monetary authorities' power is confined to administering exogenously the supply price of additional bank reserves, which in tum governs the general level of short-term market interest rates. The money supply function should therefore be modeled as horizontal in interest money space, at the level of interest rates administered by the central bank. The money supply is an endogenous variable determined by market forces. There has so far been little meeting of minds in this controversy. Each group unfortunately has largely ignored the other. This chapter attempts a reconciliation of these two positions which hopefully will be acceptable to both camps. It will be argued that there are two conceptually distinct processes of monetary creation, the portfolio change process and the income generating finance process. Mainstream theory - that the money supply is exogenously determined by monetary authorities - focuses on the first process, while the Post Keynesian position - that the money supply is endogenously determined by credit demand - focuses on the second. It is shown that both views are correct, but that each is applicable to a different historical period. Circumstances change models. Each view captures correctly a central truth of the historical money supply process. Advocates of each camp can find ampleevid-ence strongly confirming their position. It is probably for this reason that the battle lines are so defiantly drawn up, and each side is so intolerant of the 89 The Money Supply Process 90 other for failing to see the 'simple truth'. But in fact truth is seldom simple. Each view is correct, but each can be shown to apply to a different historical period. Circum-stances change models. Economics, like historians, must take history seriously. As Hicks insisted, 'economics is on the edge of science and on the edge of history' . 2 TWOVIEWS The critical point of departure for understanding the money supply process is the recognition that the supply of credit money can increase or decrease via two quite distinct processes, one initiated by the monetary authorities, the other by bank borrowers. 2 In the first case, which has been termed the 'portfolio change process',3 a monetary increase is initiated by the monetary authorities' purchase of previously existing marketable securities in the open market. The banking system, finding itself with a surplus of actual reserves over required reserves, will if available purchase positive income-earning marketable securities, rather than hold barren excess reserves. In security markets, banks are price takers and quantity setters. Nevertheless, bank security purchases are only a residual use of bank funds. Banks are basically credit retailers. Their primary business is selling credit, i.e. making loans.4 Individual banks purchase securities only when they find themselves in a surplus position, that is when their deposits minus their required reserves exceed their loans [(D - R) > LJ. This first case is the process described in the traditional textbook explanation of money creation: 'Excess reserves make deposits.' Since banks ordinarily maintain a stable reserve ratio, and the change in banks' reserves originates at the initiative of the central bank by an open market purchase of securities, it appears obvious that the money supply should be viewed as determined exogenously. It is ordinarily a stable 'multiple' of the high-powered base. In the second case, which has been termed the 'income generating finance process',5 a monetary increase is initiated by bank borrowers demanding additionalloans from their bankers. In their retail loan and deposit markets, banks are price setters and quantity takers. Providing borrowers have sufficient asset and/or income collateral, they will be granted formal lines of credit up to some predetermined amount. The vast majority of business loans are in practice granted under outstanding credit commitments. Since the utilization rate for lines of credit averages only about 50 per cent, unutilized off balance sheet credit commitments are approximately as large as total loans outstanding. The development of liability management has enabled banks to guarantee virtually unlimited credit lines, since they are always able to borrow additional funds simply by issuing additional CDs. Basil J. Moore 91 This second case is the process of endogenous money described by Post Keynesians. 6 'Loans make deposits'. Deposits are created in the act of bank borrowing, as banks credit the loan proceeds to the borrower's account. As deposits rise, increased reserves must necessarily be provided by the monetary authorities in order to enable banks to satisfy their legal reserve requirements. The stable ratio of reserves to deposits is maintained, but the direction of causality between reserves and deposits is then reversed. In the second case, the central bank merely administers the supply price at which it makes additional reserves available. It does not have a choice as to the total quantity of reserves it provides, but only as to the manner of their provision. Open market purchases increase the volume of non-borrowed reserves. Any unmet required reserves must necessarily be borrowed. Borrowed reserves may be provided freely in unlimited amounts at the discount window at the central bank's official discount rate. This is in fact the practice followed in most countries. Alternatively, reserve provision may be rationed and subject to discretionary 'frown costs'. The latter describes the current administration of the discount window by the Federal Reserve, where 'discount window borrowing is a privilege and not a right'.1 In this case, as borrowing increases, the total costs of borrowing from the discount window rise progressively above the pecuniary nominal discount rate charged, as borrowing banks are subjected by the Federal Reserve discount officer to increasing surveillance and other implicit nuisance administrative costs of borrowing (Frown costS).8 Banks always have the option of borrowing the excess reserves of other banks in the federal funds market, as a simple market transaction without any 'frown costs'. The differential of the federal funds rate over the discount rate represents an indirect measure of the degree of reserve restraint imposed on the banking system by the monetary authorities. The modem post-1979 process of directly targeting the degree of reserve restraint, and so indirectly targeting the differential of the federal funds rate over the discount rate, has appositely been termed 'dirty' interest rate targeting. 9 It is undertaken to disguise the linkage between the monetary authorities' actions and the level of short-term interest rates, so as to provide the authorities with greater freedom and shield them from political criticism. 10 In the second case, the money supply is endogenously credit driven and demand determined. The level of short-term interest rates is the exogenous control variable administered by the authorities. To what extent are the monetary authorities and the banking system responsible for the quantity of credit money supplied, and to what extent is the money supply endogenously determined? The mainstream view focuses exclusively on the first case, and concludes that the money supply is in general properly viewed as if it were under the exogenous control of the monetary authorities. The Post Keynesian view focuses exclusively on the second case, and concludes that the money supply is in general appropriately viewed as if it were not a control but rather an endogenous variable. Which explanation and which modeling procedure is more correct? 92 The Money Supply Process Both sides appear closed to persuasion by the other, wrapped in the obvious certainty of their own respective vision. It will be argued that both views are correct, but that each is applicable to a different historical period. 3 A HISTORICAL RECONCILIATION: THE PERIOD 1930-60 Consider first the historical situation when the General Theory appeared - the great depression of the 1930s. As is well known Keynes in the General Theory explicitly assumed that the money supply was exogenously determined by the monetary authorities. This was in sharp contrast to the position he had earlier developed at length in the Treatise, where he had maintained that central banks control the level of interest rates rather than the supply of money (Moore, 1984). As the great depression persisted, private loan demand was low and falling, and banks held ever larger amounts of excess reserves. These excess reserves arose primarily as a result of gold inflows and increased open market purchases of securities by the monetary authorities. The banking system as a result was generally in a cash surplus situation, so that (D - RR) > L. Due to the existence of substantial excess capacity, flagging animal spirits, high collateral requirements, and uncertainty over the possible future continuation of the recent substantial fall in the price level, private loan demand remained very weak throughout the depression, in spite of low nominal lending rates of 2-3 per cent. The bond markets on which marketable securities are bought and sold can be regarded as approximating the necessary conditions for perfect competition. In response to their huge excess reserve holdings, banks purchased large amounts of marketable securities. As a result the supply of credit money increased as the conventional view maintained. The only anomaly was that the banks did not choose to keep fully invested in securities, but permitted their excess reserves to grow. This occurred for two distinct reasons. First was the fear of capital losses on holdings of longer term securities, should interest rates rise in the future. The prices of longer term government securities had been bid up to historically unprecedented levels, as yields had been pushed down to 2-3 per cent. Even a small rise in future market interest rates during the maturity of the securities would have created capital losses that could substantially more than offset the annual interest return. Second, the yields on short-term securities, where fear of future capital losses was not applicable, had been driven down to unprecedented low levels. It is not widely known that by the end of the 1930s, the return on 90 day US Treasury bills had fallen to onetwentieth of one per cent, i.e. 55 basis points. At these minuscule yields, shortterm bills barely covered their transactions costs, and were well below the marginal costs of deposits. In the face of such low opportunity costs of foregone lending, most banks were willing to regard any accumulation of excess reserves with equanimity. II 93 Basil 1. Moore Under such circumstances, increases in the money supply need not be associated with an accompanying increase in aggregate demand. They may result solely in an increase in the liquidity of wealth portfolios, a rise in security prices, and a fall in yields. If aggregate demand is not stimulated, the result will be simply a fall in the income velocity of money. This will necessarily be the case whenever there is no induced deficit spending as a result of the lower rates and higher liquidity. The monetary authorities and the banks together then purchase only previously-existing securities held in the asset portfolios of households and firms in the private sector of the economy. This situation is illustrated in the following T-accounts; where S = Securities, R = Reserves, L = Loans, D = Deposits, and d denotes 'change in'. Banking System Central Bank 85+ IM+ dR+ Public dD+ dD+ dS- dS+ dLO dSO dLO Only if it induces additional deficit spending, financed by the sale of previously existing or newly issued securities, will a portfolio change increase in the money supply generate an increase in aggregate demand for current output. Throughout the 1930s depression, additional deficit spending was confined predominately to the government sector. Whenever newly created government securities were purchased by the banking system, the banks monetized the government sector's IOUs. In this case aggregate demand increased above the level of aggregate income in the previous period. This is shown as follows, where NW - Net Worth and G - Government Expenditure on goods and services. Central Bank 85+ I AN+ Banking System 6.R+ 6.S+ dLO 6.D+ Public 6.D+ 6.LO dSO dNW+ Government 6.D+ 6.S+ dDdG+ Even in this situation, where excess reserves create deposits to finance additional deficit spending, banks remain quantity takers and price setters in their retail loan and deposit markets, so that total loans and deposits are always demand determined. As a result security purchases, although discretionary, since banks are quantity setters and price takers in security markets. are also residual. Bank holdings of securities are equal to the excess of deposits minus reserves over loans: [S = (D - R) - LJ. This situation is illustrated in Figure 3.1. This excess reserve case makes crystal clear the fact that since additional bank purchases of securities result for the banking system as a whole in the creation of 94 The Money Supply Process DD % ~ f_--~~~-----------~--- rlV f_---_t_-----=-------r----DL ~f_---_t_------r_-~~------ L Figure 3.1 S R $ Determination of bank security holdings additional bank deposits, and bank deposits are always demand determined, the quantity of deposits demanded cannot be solely a function of income, wealth and interest rates, as the conventional view maintains. The quantity of deposits demanded is a function of the quantity of deposits created and supplied by the banking system. This is the peculiarity of money. Due to its general acceptability as a means of payment, it is always accepted, i.e. demanded, in exchange for goods and services, without the necessity of any change in its price. An increase in bank assets and bank deposits implies an increase in 'convenience lending' to the banking system (Moore, 1988b). Thus, while it remains true that the money supply is always demand determined, the demand for and supply of credit money are necessarily interdependent. 12 In depression periods, when excess reserves induce additional bank security purchases and so deposits, the monetary authorities exogenously administer the level of short-term interest rates in establishing the supply price of reserves. As the experience of the 1930s clearly demonstrates, while the authorities may be able to reduce nominal short-term market interest rates to extremely low, though always positive, levels, they will be unable to reduce nominal long-term rates below the market's collective expectations of the average level of future short rates. They will also be unable to reduce bank lending rates below some floor level, set by the spread required to cover the average total costs of bank intermediation, plus some small positive deposit rate which must be paid to induce depositors to forego liquidity by holding time and savings deposits. These intermediation costs were explicitly recognized by Keynes as posing a difficulty for central banks in bringing the effective nominal rate of interest below a certain level, but have nevertheless been overlooked in the literature: As the pure rate of interest declines it does not follow that the allowances for expense and risk decline pari passu. Thus the rate of interest which the typical Basil 1. Moore 95 borrower has to pay may decline more slowly than the pure rate of interest, and may be incapable of being brought, by the methods of existing banking and financial organization, below a certain minimum figure. (Keynes, 1936, p. 208) This is surely the simplest and most direct explanation for the existence of a positive floor level of borrowing rates, Keynes's so called 'liquidity trap'. In the US, during World War II, the unprecedented government deficit spending was financed by the issue of securities, at a maximum coupon rate of 2.5 per cent. To the extent such newly issued government bonds were purchased by the banking system, the result was the monetization of such debt, the creation of additional bank deposits, and an increase in aggregate demand. In contrast, when such newly issued government bonds were purchased by non-bank units, they were financed by increased saving out of current income, as manifest in the accumulation of previously existing deposits. Since non-bank financed government deficit spending was financed by additional volitional saving, it did not result in an increase in aggregate demand above the previous period's aggregate income. As a result Treasury and central bank policy was explicitly to attempt to sell as much newly created government debt as possible directly to the private sector, at the going low rate of interest. Central bank and banking system purchase of bonds were relied on only for any excess debt issue that was not taken up directly by the private sector. In so far as government securities are marketable and can be resold on secondary markets, banks are price takers and quantity setters in securities markets. Increases in the money supply could be viewed as regularly related to net central bank open market purchases, which resulted in the provision of net new excess reserves to the banking system. Since the enormous bank excess reserves of the 1930s had been largely absorbed by mandated increases in legal reserve requirements, the monetary authorities could correctly be viewed as determining the rate of growth of the money supply, by the rate of net reserve injection through open market operations. It was still a period which could be described as 'reserves create deposits'. Banks would buy securities whenever they found themselves with excess reserves (llD - llRR) > llL. Under such conditions the money supply function is still properly regarded as horizontal in interest money space, at the level of short-term interest rates administered by the monetary authorities. Although the money supply is demand determined, and deposits are a non-discretionary variable from and individual bank's 'point of view', the quantity of money demanded is always a function of the quantity of money supplied, which in turn is governed by the amount of securities purchased by banks and the amount of reserves provided by the central bank. By the end of the war government securities dominated bank earning asset potfolios. As loan demand gradually revived, the banks initially were able to finance loan expansion by the sale of securities from their portfolios at a price that was 96 The Money Supply Process guaranteed by the monetary authorities. Since long-term as well as short-term rates were being pegged by the central bank, long-term bonds became equally liquid as short-term bonds, but with a higher yield. As a result the portfolio demand for short-term securities evaporated, and the authorities lost control of reserve growth. After the Treasury-Federal Reserve Accord of 1951, the Federal Reserve was no longer obligated to support short- and long-term security prices at predetermined levels. Interest rates, particularly longer term rates, drifted upwards gradually, as the Fed raised the current supply price and the degree of uncertainty concerning the future supply price at which it would be willing to provide additional reserves. The much discussed 'locking in' effect, of anticipated capital losses on security resale, did not in fact prevent banks from continuing to dump their security holdings as needed in order to finance new and more profitable loan demands. As a result in the immediate post-war period the ratio of securities to total assets in bank portfolios fell sharply. As banks sold their previously existing securities to the public in exchange for previously existing deposits to finance net deficit spending associated with newly created loans, the income velocity of money increased. Money supply growth was always demand determined. But money supply and demand growth was less rapid, since bank loan acquisition was partly offset by bank security sales. 4 HISTORY CONTINUED: THE PERIOD 1960-80 By the end of the 1950s the banks' government security holdings had been largely depleted. In order to continue to be able to finance outstanding unutiIized credit lines, in 1961 banks initiated the process now known as liability management. By issuing marketable certificates of deposits, banks were able actively to bid for funds as their credit needs dictated. 'Spread' banking developed, as banks commenced to administer their lending rates at some stable mark-up over their costs of acquiring wholesale funds. Gradually loans came to dominate bank portfolios. Since banks were able to bid actively for external funds by the issue of marketable CDs and other managed liabilities, bank portfolio demand for marketable securities to be held as secondary liquidity reserves declined. In the process bank reserves increasingly came to be provided endogenously by the monetary authorities. Banks had large outstanding credit commitments, so that bank loans were granted increasingly at the initiative of bank borrowers, not the banks themselves. New loans then created new deposits. The monetary authorities had no choice but to make the reserves available as required for the now larger deposit totals. With lagged reserve requirements, required reserves infact became a predetermined variable. The authorities' only choice was the manner in which the required reserve volume would be provided, by nonborrowed reserves through open market purchases, or by borrowed reserves Basil 1. Moore 97 through discount window lending. The process of keeping the banking system 'in the bank', so as to make the authorities' lending rate the effective determinant of the level of short-term market interest rates, or what was called net free reserve targeting, was generally followed. In 1979 reserve operating procedures were radically changed, as the authorities attempted to target non-borrowed reserves. Since 1982 a process of 'dirty' interest rate targeting has gradually evolved, in which the authorities now directly target not interest rates but rather borrowed reserves or the level of 'reserve restraint' .13 Over the post-war period it gradually became increasingly clearer that the supply of credit money was endogenously credit driven and demand determined. 14 So long as loans were not marketable instruments traded on secondary markets, and banks were price setters and quantity takers in loan markets, loan quantity could not be regulated directly, either by the banks themselves or by the central bank, other than by direct quantitative credit controls. Traditional monetary policy was now clearly confined to administering the supply price of additional reserves. The forces determining the volume of bank intermediation were in this manner gradually taken away from the banking system and the monetary authorities, and put in the hands of bank borrowers and bank depositors. It is true that US banks on balance even now maintain substantial security holdings. But the decision as to the quantities of securities held are in many cases no longer made independently by the banks themselves. First, a substantial proportion of security holdings are required by individual depositor and borrower agreements. This is the case for federal, state and local government security holdings. IS Second, the residual level of securities held by the banking system now largely reflects the distribution of loan and deposit demand among individual banks. Banks are quantity setters and price takers in wholesale markets. Individual banks acquire additional securities only when they are confronted with a cash surplus, i.e. an excess of new deposits minus required reserves over new loans [(6D - 6RR) > 6L)]. Similarly, banks issue CDs or other managed liabilities whenever they are faced with a cash deficiency, i.e. their changes in loans exceed their deposits minus required reserves [6L > (6D - 6RR)]. In practice smaller banks are primarily in 'surplus' positions, while larger money market banks are characteristically in 'deficit' positions. As a result, as is the case with the federal funds market, the total volume of bank security holdings and bank managed liabilities have become increasingly an interbank phenomenon. The volume of securities and CDs reflect primarily the degree of mismatching in the distribution of loan and deposit demand among individual banks. Wholesale assets (securities) and wholesale liabilities (CDs) are approximately equal for the system as a whole, and so cancel out, leaving total loans approximately equal to total deposits minus reserves. 16 The volume of total bank assets and liabilities thus considerably overstates the true volume of bank intermediation with the nonbank sector. The Money Supply Process 98 5 CONCLUSIONS Each of the two opposing views are accurate portrayals of the money supply process in different historical episodes. During periods of deflation, when loan demand is weak, the mainstream view is correct. Central bank initiated open market purchases result in a multiple expansion of bank deposits, providing only that banks invest their excess reserves in marketable securities. During periods of inflation, when loan demand is strong, the Post Keynesian view is correct. Borrower demand for bank credit determines the rate at which deposits expand. The central bank, so long as it avows non-market quantitative controls on bank credit, is powerless to control directly the rate of money supply growth by quantitative reserve restriction. All it can do is attempt to restrain the quantity of credit demanded by raising the supply price of reserves to the banking system, and so the prevailing level of short-term market and bank administered interest rates. The issue whether the money supply is more appropriately viewed as endogenously determined by market forces, or exogenously under central bank control, thus depends on the historical circumstances. However, as shown in all situations the central bank is responsible for directly administering the supply price of reserves to the banking system, and so the general level of short-term interest rates. In all situations banks are price setters and quantity takers in their retail loan and deposit markets. In all situations the volume of bank credit and bank deposits are demand determined. As a result the money supply function is always appropriately viewed as horizontal, and the money supply is endogenously credit driven and demand determined. Nevertheless, due to the peculiarity of money, it is also necessarily the case that the demand and supply of money are interdependent. Money is different from all other commodities in that, in so far as it retains its money ness, it is always generally accepted and so 'demanded' in exchange for non-monetary goods and services, without a change in price. Whether the money supply increases, due to an increase in bank loans, or due to an increase in bank security holdings, an increased supply of money will necessarily always be demanded. Economists conventionally distinguish both a transactions and an asset or portfolio demand for money. Like any other good, the quantity of money demanded is treated as a single-valued function of income, wealth and interest rates. But money is generally accepted in exchange for all other goods, and the supply of credit money is not bounded by any production function relationship. The quantity of money created is bounded only by the price that must be paid to borrow it. All increases in the money supply will always be demanded. Since there is no unique 'equilibrium' quantity of money demanded, it is not meaningful to attempt to distinguish 'disequilibrium' money holdings. An increase in the demand for and the supply of bank credit implies an increase in the supply of and the demand for credit money, not as an asset or portfolio demand for money to hold, but as a demand for the goods and services that money can command. Basil J. Moore 99 An increase in total deposits represents an increase in 'convenience lending' by bank depositors to the banking system, and indirectly through the banks as intermediaries to bank borrowers. It is increases in such 'convenience lending' which finance net deficit spending by the economy, i.e. an increase in aggregate demand above the previous period's aggregate income. As a result an increase in investment expenditure financed by bank credit will always call forth an equal increase in 'convenience' saving, quite apart from the Keynesian multiplier process of income change. Nominal investment thus determines nominal savings (Moore, 1988b). A horizontal money supply function implies that the money supply is endogenously demand determined. This is necessarily always the case. Nevertheless, since the quantity of money demanded is always identical to the quantity of money supplied, monetary endogeneity need not imply the non-controllability of money by the central bank. For the case of monetary expansion, the monetary authorities are always able to increase the money supply at their initiative by the purchase of securities in the open market. Insofar as they are able to induce banks to purchase additional securities through the provision of excess cash reserves, the money supply can always be increased, by some multiple of the growth of bank reserves, at the initiative of the central bank. But there is a very important asymmetry. Whenever banks are 'loaned up', the monetary authorities will be unable to reduce the money supply at their discretion. They can then only raise the supply price of reserves. The supply of credit money is always endogenously demand determined. But since the demand for and supply of money are interdependent, the central bank is able to control the quantity of money demanded whenever it is able to control the quantity of money supplied. Whenever banks acquire additional assets, their total liabilities will likewise expand, since the bank balance sheet constraint necessarily always holds. 17 The monetary authorities are thus usually correctly viewed as being able to increase the supply of credit money. IS Endogeneity of the money supply does not imply the inability of the authorities to increase or decrease the money supply in an upward direction. Monetary policy is fundamentally asymmetrical. 19 Notes I. 2. 3. 4. 5. 6. 7. 8. See Dow and Saville (1988); Jarsulic, (1985); Kaldor (1986); Lavoie (1984); Melton (1985); Miles (1984); Moore (l988a); Rousseas (1986). Davidson (1972), p. 226. Ibid. Moore (1988, b,c). Davidson (1972), p. 226. A better term would be 'expenditure generating'. Lavoie (1984); Kaldor (1986); Moore (1988b, 1989). Goodfriend (1983); Mengle (1986). Greider (1987); Melton (1985); Mengle (1986); Moore (1988a). The Money Supply Process 100 9. 10. II. 12. 13. 14. IS. 16. 17. 18. 19. Goodfriend (1983); Melton (1985). Greider (1987); Moore (I 988.b); Wallich (1984). This was the origin for the expression 'you can't push on a string' to describe the difficulty the monetary authorities faced in further expanding the money supply. This was the real issue addressed by Tobin in his well-known 'New View' article. 'Commercial Banks as "Creators" of Money', Tobin (1963). Goodfriend (1983); Melton (1985); Moore (1988,b); WalJich (1984). Kaldor was the first economist of prominence to have publicly recognized this phenomenon. 1970. Silber (1978). In terms of Figure 3.1. the excess of deposits minus reserves over loans of surplus banks is equal to the excess of loans over deposits minus reserves of deficit banks. The supply of money would only not increase in response to the banks' acquisition of additional earning assets if all new deposits were used immediately to purchase securities from bank portfolios or to repay outstanding bank loans. For the case in which all firms had bank overdrafts. this latter constraint could well be operationally binding. The sole exception is when short-term rates have been reduced to such low levels that banks prefer holding excess reserves rather than purchase bills. The central bank is always able to increase the money supply in a boom. when it wishes to restrain it. It is also able to reduce the money supply in a slump. when it wishes to expend it. References Davidson. P. (1972) Money in the Real World (London: Macmillan). Dow. 1.C.R. and 1.D. Saville (1988). A Critique of Monetary Policy (New York: Oxford University Press). Goodfriend, M. (1983). 'Discount Window Borrowing. Monetary Policy and the Post October 6. 1979 Federal Reserve Operating Procedures', Journal of Monetary Economics, 12. pp. 343-56. Greider. W. (1987). Secrets of the Temple: How the Federal Reserve Runs the Country (New York: Simon & Schuster). Jarsulic, M. (1985). Money and Macro Policy (Boston: Klewver-Nijhoff). Kaldor. N. (1970) 'The New Monetarism'. Lloyds Bank Review. July. Kaldor. N. (1986). The Scourge of Monetarism. 2nd edn (Oxford: Oxford University Press). Keynes. 1.M. (1936) The General Theory of Employment. Interest and Money (London: Macmillan). Lavoie. M. (1984). 'The Endogenous Flow of Credit and the Post-Keynesian Theory of Money', Journal of Economics Issues, September, pp. 771-97. Melton, W. (1985), Inside the Fed: Making Monetary Policy (Harrison, ILL.: Dow Joneshwin). Mengle, D. (1986), 'The Discount Window', in Instruments of the Money Market, 6th edn (Federal Reserve Bank of Richmond). Miles, M. (1984), Beyond Monetarism (New York: Basic Books). Basil J. Moore 101 Moore, B.J. (1984) 'Keynes and the Endogeneity of the Money Stock', Studi Economici, 22. Moore, B.l. (1988a), 'The Endogenous Money Supply', Journal of Post Keynesian Economics, X (3), pp. 372-85. Moore, B.J. (1988b), Horizontalists and Verticalists: The Macroeconomics of Credit Money (New York: Cambridge University Press). Moore, BJ. (1989), 'A Simple Model of bank intermediation', Journal of Post Keynesian Economics 12(1), Fall, 10-28. Rousseas, S. (1986), Post Keynesian Monetary Economics (Armonk, NY: Sharpe). Silber, W. (1978), Commercial Bank Liability Management (New York: Association of Reserve City Bankers). Tobin, J. (1963), 'Commercial Banks as Creators of Money', in D. Carson (ed.), Banking and Monetary Studies (Homewood, III.: Irwin). Wallich, H. (1984), 'Recent Techniques of Monetary Policy', Economic Review, Federal Reserve Bank of Kansas City, May, pp. 21-30. B. Circulation Views 4 A New Paradigm for the Determination of Money Prices Bernard Schmitt How are goods measured? Are the relevant measures dimensional or purely numerical? Classical economists thought that goods had a value-dimension and were accordingly measured in units of value, e.g. in units of labour-value. Physicists measure space, time, energy and mass. In Classical economics the fundamental problem of the definition of economic value is already raised: are goods measured in some objective dimension, like energy or mass? Within the Classical school it was possible to make direct statements about quantities of value, in a given system of numeration, because each indivisible commodity was regarded as embodying units of cost defined in terms of its production by labour. When David Ricardo refers to a correspondence between two classes of objects, namely commodities and numbers, he never posits an equivalence of the following type: 1 apple = number 3 (1) In Ricardian economics a pure number can have no economic value and can therefore never be equated with a commodity. In equation (I), 3 is not just a number but a number of money units or, more stringently, a number of units of value. The two terms of equation (1) have the same objective value in the same way as two physical quantities have the same mass. Equation (1) states the equivalence between the value, or economic mass, of an apple and the value, or economic mass, of 3 units of money (gold or silver). In this chapter we shall leave Classical economics aside. We shall first take an interest in the neo-Classical paradigm, or paradigm I (PI), which may still be considered today as defining mainstream economics. Paradigm 1 is concerned with replacing absolute values by relative values. Is this radical step to be interpreted as the abandoning, in economics, of dimensional value units? Certainly not. If goods were not scarce and if they were of no use to mankind, they would have no value. In particular the monetary value of goods, i.e. their price, would then be nil. True, the whole emphasis with Leon Walras, who founded PI, is on prices stated in a numeraire. Since any chosen 104 Bernard Schmitt 105 numeraire is expressed by a number it would seem that numeraire-prices are direct numerical measurements of physical goods, whereas 'classical' money prices were meant to measure the economic value of physical goods - not the physical goods or services considered in themselves. In fact no such distinction can legitimately be made with regard to Pl. In general equilibrium theory PI no agent exchanges a good against a pure number; on the contrary, only goods (commercial or financial) are exchanged in the various markets. Describing the Arrow-Debreu-McKenzie (ADM) model, E. Roy Weintraub defines equilibrium in the following terms: An equilibrium can be thought of as a set of non-negative prices, one for each of the J goods, such that if consumers and producers were each individually to optimize taking those prices as given, the resulting market demand and supply quantities would just balance and would yield market prices identical to those taken as 'given'. Existence of equilibrium is thus equivalent to the logical possibility of pre-reconcilable choices. (Weintraub, 1979: 29). Clearly the condition for equilibrium is that the supply of goods must be equal to the demand for goods in every market; numbers enter into the picture only axiomatically and for one precise purpose, namely to pave the way for the implementation of mathematical techniques. We shall attempt to show that a logical flaw undermines the whole body of general equilibrium theory (GET). Unless money is granted its true status, which it is denied in GET, the national economy cannot function under the rules of a single price system but is pushed instead into a state of anarchy, where the existence of a complex of price systems, none of which can be elected as being the 'right' one by any rational law, prevents even the most skilful mathematician from ever arriving at the determination of prices. To repeat: in our endeavour we shall use one tool only, that is logic. Mathematics is only one branch of logic. Our purpose is to prove that GET is based on a crucial illogicality even before it reaches the hands of mathematicians. If misconceptions of money are the 'culprits' in the demise of GET, we may still nurture the hope that a direct study of money expenditures will be able to salvage GET. Now, in his scientific papers Paul A. Samuelson develops to a considerable degree the theory of money expenditures which occur in real time. A close study into the logical consistency of the theory of expenditures advanced by Samuelson results in a disappointing conclusion; like GET proper, Samuelson's theory, despite its unquestionable mathematical excellence, falls prey to fundamental illogic. All the harm befalling GET and Samuelson's theory of expenditures derives from a single cause, i.e. the falsehood of paradigm 1. 106 Determination of Money Prices In the second section a new paradigm will be defined, in compliance with the principle of excluded third. In other words, PI and P2 exhaust all logical possibilities; if there is a truth for every error, then P2 must be true since PI is erroneous. THE DEMISE OF GET 'What must strike the naive beholder above all,' said Wittgenstein, 'is that mathematicians are continually terrified of one thing, which is a sort of nightmare for them, namely contradiction.' GET is highly mathematical theory and it is built on a pervading contradiction. Our purpose is to prove that GET rests on a hidden assumption, namely the supposition that goods occupy an invariable extent of space in the set of real numbers. In physics the concept of mass relates to quantities of matter. Physical mass (or energy) is a quality of goods in so far as they are physical objects. Not being a branch of physics GET takes no interest in the physical mass of goods. It does not follow that GET ascribes no mass whatsoever to goods. Besides its physical mass, which is irrelevant to economics, each good belongs, axiomatically at least, in the set of real numbers. In that sense we shall say that each good has a numerical mass. If that were not the case, economics could not aspire to become eventually a science. But GET goes one step further, albeit surreptitiously, for the validity of its theorems depends on the invariability of economic masses. The contradiction we propose to uncover runs in the following terms: (a) (b) GET can determine prices only if each good has a constant numerical mass; but if each good has a constant numerical mass, prices cannot really be determined for they are predetermined, i.e. given in a petitio principii. III GET the determillacy of prices is founded all the hiddell assumptioll of the constant numerical mass of each good. Three agents, A, Band C are each endowed with one good, a, band c respectively. GET fails even in those simplified circumstances. Consider in tum direct exchanges between goods, exchanges with the help of a lIumeraire and exchanges effected via a medium (money). Direct Exchanges Between Goods Suppose that at general equilibrium each agent parts with a quantity of the good with which he was initially endowed. In the absence of units of account (specified in terms of a lIumeraire) and of units of payment (money) exchanges are ratios of purely physical quantities. Thus xa/yb, where x and yare physical quantities, is a price defined for goods a and b: the price of xa is yb; conversely, the price of yb Bernard Schmitt 107 is xa. Likewise, yblzc is the price of goods band c relative to each other. Only two relative prices are to be detennined between three goods. The Method of GET The unknowns are the values of x and y and z which reconcile the choices of all agents. Let us follow a concrete example. The initial stocks of goods in the hands of each agent are 10 pens (with A), 12 reams of paper (with B) and 60 bottles of ink (with C). Assume that all choices are reconciled when agents exchange, directly or through an exchange involving the third good, 2 pens for 1.5 reams of paper. equilibrium price of 1.5 reams of paper in terms of pens 2 (in reams) equilibrium price of 2 pens in terms of reams of paper 1.5 (in pens) By which method can these equilibrium prices be determined? GET thinks it knows the answer: general equilibrium can be determined mathematically. It suffices to consider one unknown, say the equilibrium price of pens in tenns of paper. In order to find the relevant solution we take any given number of pens as fixed, two for instance, in relation to all conceivable values of quantities of paper, a series of values defined as coextensive with the set of real numbers, from a zero-quantity to an infinitely large number of reams. If the solution exceeds initial endowments it is still mathematically determinate and lacks feasibility on purely practical grounds. units of paper: 0 2 pens ~ 00 (2) It would seem that ratio 2 provides agents with the set of the infinite number of all possible prices of paper relative to pens. It might happen, of course, that none of these prices reconciles the choices of all agents; but then no exchange between paper and pens takes place in the relevant market session. But if one of the stated prices is agreed upon by all agents as defining the generally accepted terms of a desired exchange, such as 1.5 reams of paper for 2 pens, then the solution We are looking for is attained and 1.5 reams of paper are actually exchanged for 2 pens. More generally, let us choose as a reference x = 2 units of good a; the problem then is to detennine the quantity of good b which it would be agreeable to all agents to exchange for x units of a. With that purpose in mind, starting from a 108 Determination of Money Prices zero quantity of good b we scan the set of real numbers until we find the equilibrium value of y. y units of b : 0 ~ 00 2 units of a (x = 2) (3) If in the series of the infinitely numerous values of y, y = 1.5 (meters, pounds, ... ) units of commodity b relative to 2 units (specified in any physical dimension whatever) of commodity a reconciles the choices of all agents, then 1.5 (say meters) of a and 2 (say pounds) of b define the terms of an equilibrium price. Whether the solution is unique is another problem. The process by which agents scan the set of real numbers in the fashion we have just described is called tatonnement. Other methods may be used. The essential point is that all possible values of physical units of good b should be 'testable', before the eyes of all agents, against the given physical quantity of gooda. Although it is deeply embedded in mathematics, the method set out by GET is logically flawed. The Logical Flaw in GET We could choose to base our critical analysis on the fact that goods are indivisible while 'taken together the assumptions [of GET] suggest that goods are finite in number and completely divisible' (ibid., p. 28). Pens, like any other good, belong in the set of integers, or natural numbers and not, as GET would have it, in the set of real numbers. Nevertheless, for the sake of argument we accept that any commodity (even a Boeing 777) remains true to its definition down to its smallest, infinitesimal parts. GET is weakened, not to say destroyed, by a much more serious deficiency, of a purely formal nature. The case where two physical units of commodity a are chosen as a point of reference is arbitrary. All other conceivable cases must be considered. If quantity y of commodity b can vary from zero to infinity, it must follow that quantity x of commodity a is likewise a number (of physical units) which it is up to the theorist to determine within the whole set of real numbers. For each value of x there is a corresponding 'price system'. Agents are therefore encumbered with a whole complex of price systems, infinite in number. y units of b : 0 ~ 00 x units of a : 0 ~ 00 (4) In ratio 4 an infinite number is divided by an infinite number, a perfect instance of a complete indeterminacy. Bernard Schmitt 109 There must be some 'good' reason why GET eschews ratio (4), replacing it by ratio (3). In a way it is a wise decision whose purpose and effect is to create the necessary condition for price determination. Then everything is simple; no profound puzzles exist any longer and the construction of economic science can be entrusted to the mathematician. But the reader is now well aware of the fact that the question cannot be settled to the joint satisfaction of logic and mathematics. 'Whereas logicians have often been said to take no more interest in reality than do mathematicians, for Wittgenstein, young and old, logic was "interested only in reality'" (Hallett, 1977, p. 169). And can there be any doubt as to the correct interpretation of the relevant reality? Hardly, for goods a and b, and c for that matter, are freely interchangeable, at least to the extent of available stocks, so that there can be no justification at all for arresting one's choice on an invariable quantity of one of the goods, 'choisi au hasard' (Walras), in order to determine the price of the others against it. But there is another factor explaining why GET unwittingly begs the question. Mathematicians tend to read their ways into any discipline which is apt to alert their mind. In mathematics x and yare dimensionless or pure numbers; in consequence one of the terms of ratio (3) can be maintained constant at any numerical value (such as x = 2) while the other term varies from zero to infinity: ratio (3) then assumes all possible values in the set of real numbers. Replacing ratio (3) with ratio (4) then serves no clear purpose. Things are fundamentally different in economics where x and yare physical quantities (or dimensions) instead of pure numbers. Each value of IX (see below) commands an entirely different price. IX x 1.5 reams of paper IX x 2 pens GET treats physical quantities as if they were pure numbers, a device which points to the hidden assumption of constant numerical masses. GET rests on the hidden assumption of numerical masses Let us associate number 20 to the initial endowment of 10 identical pens. Can we determine the number corresponding to one pen? If we define the numerical value of one pen as one-tenth of the numerical value of ten pens we posit the existence of numerical masses in goods. We could choose to follow a slightly more complicated, non-linear method of computation. Take a given proportion (q) of the initial stocks; the assumption of constant numerical values obtains if (and only if) the numerical value of that proportion of each endowment is the same percentage (q') for each good of its initial global numerical value. Take q = q'. Even before the process of adjustment (by ttltonnement or otherwise) has got under way, we know, a priori, that it would be odd to question its 110 Determination of Money Prices logical validity. There is no logical law preventing us from creating a conventional link between a commodity and a number. A convention requires no justification. We cannot possibly err if we posit that the stock of 10 pens and 7.5 reams of paper occupy an equal space in the set of real numbers. In other words. 10 pens are objectively equivalent to 7.5 reams of paper. Under the constraint of the assumption of constant numerical values. agents have no choice but to accept the predetermined price ratios; all the same. they are left to enjoy a degree of freedom for they can choose the number of pens. for instance. and by the same token the number of reams of paper. which are to change hands. In conformity with constant numerical masses. the comparative 'weights' of the initial stocks of goods are predetermined in the set of real numbers. Then only two tasks are left for the adjustment process to accomplish; i.e. firstly to reveal the precise measures of those weights. and secondly to take into account the contractual choices of agents. who freely determine the proportion by which the initial stocks are actually exchanged. In the end the evident fact is that GET is affected either by cholera (existence of a complex of an infinite number of price systems) or by plague (where price ratios are predetermined by the assumption of constant numerical masses). This very same unrewarding conclusion unremittingly persists after the introduction of a numeraire. Analysis of Real Exchanges Measured in a Numeraire Nothing bewilderingly new occurs with the introduction of a numeraire and there is no need here for protracted demonstrations. After all the chosen numeraire is just an ordinary commodity serving as a benchmark to measure prices. The crucial point is to determine the price of the numeraire or. more precisely. the price of the good chosen as a numeraire. It is most remarkable that all protagonists of GET make the same chief blunder; as one man they construe the price of the numeraire as being a pure number and. as if that were not enough. as identically equal to ~ very particular number. namely the number 1. Leon Walras was the first to do so. starting a continuous tradition in which Wald is a foremost member. 'Wald's theorem then stated that the exchange equations had at least one solution for the relative prices P2. P3 • .... Pm (PI = 1 as numeraire') (Weintraub. 1979: 21). If we again take three goods and a three-agent economy. do we really err when we decide that one of these goods is to serve as a numeraire in the determination of the equilibrium price of the other two? We would be proffering an untenable thesis if we held that view .. Thus we are not supposing that GET plays against the rules by converting a physical good into a pure number. Although it is highly convenient for a mathematician to take such a step. it would be odd to question its logical validity. There is no logical law preventing us from creating a conventional link between a commodity and a number. A convention Bernard Schmitt tIl requires no justification. We cannot possibly err if we posit that the stock of 10 pens is henceforth 'signified' or denoted by the number 1. According to an axiom lying at the heart of GET, goods are numbers anyway. Moreover, there is no point in trying to disprove an axiom. But there is more to the concept of the numeraire. As soon as it has been agreed that IO pens are represented by the number l,'t is henceforth the purely numerical price of IO pens. To realize this we must remember that in GET prices are relative; clearly, then, the numerical price of any commodity stated in pens inversely defines the price of pens stated in that commodity. Now that we can rest assured that we have made no logical slip as yet, we can boldly build on the basic information that we have just gained, namely the certitude that one of the three prices is known from the start: unquestionably the price of the available 'bundle' of pens is the first integer in the set of real numbers. Thus, only two prices remain to be determined in our three-commodity economy. GET suggests a variety of solutions in the much more difficult case of an ncommodity economy, where n is any finite number. The only apposite question is whether such solutions are formally acceptable. If so the awesome efforts in building recondite mathematical models made by great minds such as Walras, Wald, Arrow, Debreu and McKenzie are fully justified and fruitful. It then follows that economics, to a greater extent even than modem physics, is a branch of pure mathematics. We are reminded here of Wittgenstein's remark on the scope of mathematical investigations. 'We do not interfere with what the mathematician does; only when he claims he is doing metamathematics do we check on him' (in Hallett, 1977: 219). Whenever a mathematician annexes economics, on so-called intellectual grounds, it is right and proper that we should check on him. But we cannot hope to be at all convincing unless we avail ourselves of only one trenchant 'weapon', that is bare logic. E. Roy Weintraub asserts that 'anyone who argues that "neo-c1assical economics will be non-sensical as long as rational economic agents are assumed" is committing ... [aJ methodological blunder' (1979: 37). If, looking for the truth of the matter, or more modestly, if striving after the sense of GET we eventually find that it is indeed non-sensical, then we can be vindicated only if our argument is logically compelling. Paraphrasing Weintraub we may say that 'scientific methodology is more than obeisance to the mathematical method'. Even if it were true that macroeconomics is founded on microeconomics, or should be thus founded, mathematics is not founded on itself. But there is no need here to probe into the foundations of mathematics; we are mere economists and in that capacity we should be content to examine the economic foundations of the mathematical method which is used in our discipline. Hereby we do not refer to experiential or descriptive, but strictly logical, foundations. If the logic of mathematics can in no way be coupled with the logic of economics, then pure economics can only be purely non-mathematical. GET is 112 Determination of Money Prices the sudden and overwhelming penetration of mathematics into the preserve of economics; is its domineering power legitimate? The answer must be in the affirmative if the numeraire provides a reference for the determination of n - 1 prices (2 in our example). But this is precisely what the numeraire is incapable of doing. In an age where mathematical techniques were highly developed already, Classical economists, who get short shrift from the authors of GET, had it right after all: only an absolute reference is capable of providing the logical foundation for price determination. The numeraire is a benchmark of sorts but one which affords no more than a relative reference. Neo-Classical economists would readily go along with that assessment but they labour under the illusion that a relative reference is capable of rendering the same services as an absolute standard of the Classical type; although a commodity conforming to the doctrine of GET is devoid of any axiological dimension, such as labour-value, it is nevertheless reputed in that school to be a genuine yardstick for measuring prices, under the sole proviso that it be chosen as a numeraire. Our purpose is to oppose all such 'thinking', with its disdainful disregard for 'ordinary logic'. Let us give a central place to a numeraire exemplified by the number 1 as the price of 10 pens. Is the question of the equilibrium price of pens thus fully settled? So it seems. Since we allot one specific numerical value (1) to the available pens, their price is 1; not 1 dollar or 1 franc, but just I, a pure number. Now, nothing can be more constant at a moment of time than a single value. The price of pens is identically equal to 1. From the moment we elect pens to the status of a numeraire, we pinpoint their position in the set of numbers while each of the other goods is left to find its own equilibrium in the maze of possible values. Any price must be an equilibrium price when there is only one possible price. In short, t is the equilibrium price of the given pens. The remaining unknowns are the equilibrium prices of the two other commodities (paper and ink) stated in the chosen numeraire. Should we succeed in determining the price of paper in terms of the numeraire, then the determination of the price of ink would result from the same pattern. All possible prices of paper in terms of the given stock of pens are included in the following ratio. quantities of paper: 0 ~ 00 t (pure number) (5) Completely divisible reams of paper are made to vary in ratio (5) from zero to infinity against the constant 'numeraire-commodity', represented by a pure number, I. If none of these (feasible) values reconciles the desires of all agents, no paper can change hands. But if, in the process of continuously scanning the increasing quantities of paper relative to the numeraire, starting from a zero- Bernard Schmitt 113 quantity of paper, we hit upon a value which all agents are ready to agree on, then that precise quantity of paper defines the equilibrium price of paper expressed in the numeraire. How could a simple argument of this kind lead us into serious difficulty? But it does. In fact two critical difficulties are implicit in the definition of the numeraire: first, if an arbitrarily chosen number (1) stands for the given stock of 10 pens, does this suggest that we know which precise dimensionless number corresponds to, say, 5 pens? Secondly, if the numeraire is initially associated with 10 pens, are we supposed to derive from this definition the information as to precisely how many pens will actually change hands when equilibrium is reached? Not being aware of these difficulties GET leaves them hopelessly unaddressed and keenly resumes the catharsis of pure mathematics. GET is a profoundly illogical research program, which fails to determine the price of the good that is chosen to play the role of a numeraire. But are the prices of the other commodities determinate in GET? In fact, all prices remain indeterminate. If some coherent method led to the determination of the quantity of good a (the numeraire) which it is the choice of all agents to see injected into the general array of exchanges taking place in a market session, then the same determining process would extrude the corresponding quantities of goods band c. But we have no indication whatsoever concerning the quantity of the 'numerairecommodity' which changes hand at general equilibrium. The mathematical body defining GET is truly admirable. Unfortunately, it is based on an Achilles heel. Before they can hope to lay their hands on economics, mathematicians must, for a brief but daring moment, step into the open in order to prepare the ground on which their trade is to prosper: at that moment mathematics is still in the wings but logic is already on the scene. The complete certainty of the a priori sciences, logic and mathematics, depends mainly on the fact that in them the path from ground to consequent is open to us, and is always certain. This endows them with the character of purely objective sciences, in other words on sciences about whose truths all must judge in common, when they understand them. (Schopenhauer in Hallett, 1977: 737) Is the quantity of the 'numeraire-commodity' which will actually be exchanged at general equilibrium predetermined in the definition of the numeraire? This question is of a logical, albeit not of a mathematical, kind. Any answer to this question, if meaningful, is either true or false, exactly as is the case in mathematical reasoning. Furthermore, it cannot be all that easy to provide the correct answer since hosts of theorists have missed it. If we posit that 10 pens are identically equal to the number 1, we have as yet no clue as to the numerical price of any part of that 'bundle'. Furthermore, we know nothing aboul the number of pens which will be exchanged at general 114 Determination of Money Prices equilibrium. The correct inference, then, is to say that the price of the' numerairecommodity' is just as indeterminate as the price of any other good. For each and every arbitrary quantity of the 'numeraire-commodity' there is one Waldian price system (based on that quantity of the numeraire), yielding a set of prices, 'P2' P3' ... , Pm (PI = I, as numeraire),. What chance is there for precisely that haphazard quantity of the 'numeraire-commodity' to be the quantity of that good determined at general equilibrium? Hardly any, for an infinite number of distinct instances corresponding to PI = 1 exists in relation to one and the same numeraire. Again GET may resort to the assumption of constant numerical masses; then GET is self-contradictory; otherwise, it is vacuous. Analysis of Exchanges Between Real Goods Effected by the Intermediation of Money A new situation is created by the introduction of a concrete means of payment, money proper, in contrast to an abstract unit of account or a numeraire. But let it be clear from the start that GET cannot hope to derive any benefit from the introduction of a sum of money units unless each unit of money functions as a pure number. There is no reason why we should not accept the supposition that exchangeable commodities (whether commercial or financial) are always real goods in contrast to sums of money which are purely instrumental in facilitating economic transactions. However, there is always the danger that we might use the word money in a different sense. How can we make absolutely sure that we will never allow money into our reasoning save in a strictly instrumental capacity? GET offers the correct criterion to that effect: at general equilibrium each agent spends the exact sum of money which he simultaneously earns. By introducing money, even if only in that strict sense, we fundamentally change the nature of price. Expressed in money units, price is a purely numerical entity, closely analogous to the mathematical concept of number. One is inclined then to say that when money serves the purely catalytic function of an instrument, in contradistinction to an object of exchange, it does so in conformity with the rigorous laws of mathematics. On closer scrutiny we will have to suggest otherwise. The problem of the inconsistency underlying GET remains, unabated. All prices, preceding general eqUilibrium or concomitant with it, are specified in money units. Ratio 7 below therefore comprises all possible monetary prices of good a (reams of paper) relative to good b (pens). monetary (or numerical) price of paper: 0 ~ 00 the given monet. or num. price of 1 pen (7) Bernard Schmitt 115 We no longer have any right to use the general consideration, which we have made abundantly clear, that a ratio involving physical goods instead of numbers logically contains not merely one but two series of values, both infinite in number. As soon as prices are expressed in money, every price ratio can be determined relative to a 'fixed point'; thus, in ratio 7, we can fix the monetary value of one pen and then proceed to look for the corresponding equilibrium price of paper in terms of pens; the whole range of possible prices is covered when the magnitude of the numerator is made to vary from zero to infinity while the denominator is kept constant. Does this entail a new line of solution? The criterion of equilibrium now consists of the instantaneous equality of money inflows and money outflows established for each agent. Unless this condition is strictly met, money cannot be prevented from encroaching on the class of final goods, a state of affairs which would hark back to physical price ratios (see ratio (4) above) and their noxious consequences. The only adequate foundation of the new approach educed by the emergence of money consists of a series of equations, by which each agent equalizes two magnitudes, namely the value of his sales and the value of his purchases. We have just reached the exact point where a 'law' makes its entry, known as Walras's law. Given that even in a monetary economy exchanges occur only between real terms, in each market session the sum of all sales must necessarily be equal to the sum of all purchases. In a two-agent economy it follows that the equalization of A's sales and purchases, SA = P A, entails the equalization of B' s sales and purchases, SB = P B • Not a very powerful law, one must admit. But nevertheless sufficiently cogent to provide GET with a safe foundation? Definitely not. Nothing that is implied in Walras's law is in the least objectionable; but it takes us nowhere. To the extent that we read any positive meaning into it we fall into error. Time and again we are led to believe that Walras's so-called law really is a law, like the law of faIling bodies; in fact, it is a mere tautology, not to say a truism. Considering the crucial importance of the point here at issue, which again threatens the very foundation of GET, let us reason with extreme caution; repetition in argument may be preferable to a lack of clarity. We remain within the boundaries of a two-agent, two-commodity economy. Prices are objectively stated in money units, by an 'auctioneer'. When a random price is 'called', it may well be that neither A nor B agree to exchange any quantities of a and b at that price; it is more likely that a ratio (P) which is not an equilibrium price will be favourable to one agent to the detriment of the other: in that case one of the agents is prepared to exchange x units of a for y units of b, where xa yb =p (6) Determination of Money Prices 116 At the same time, p may be a 'prohibitive' price in the eyes of the other agent, who refuses accordingly to surrender any positive quantity of his own endowment in order to acquire even a fraction of the good which he finds too costly. It is clear then that at price p the 'other agent' refuses to exchange units of a for y' units of b, where x' and y' are any amounts of goods a and b respectively. In other words, while one of the agents accepts equation (6) for positive quantities of goods a and b, xa and yb, the other agent rejects equation (7) for any positive amounts x' and y' of goods a and b. x x'a -*p y'b (7) Of particular importance here is that the first task we are confronted with is to reduce inequality (7) and replace it with a genuine equation. Nobody, presumably, would suggest that Walras's law provides the missing link; if that were true, equation (7) would be implicit in equation (6) and there would be no need to search for an equilibrium price. All conceivable prices would be equilibrium prices. It would be somewhat unfair if we chose to play havoc with GET without even bothering first to look a little more deeply into the matter now that the economy has been granted an instrument of exchange (money). In fact, no 'serious' author propounds the 'theory' that at any given monetary price, p, equality xa/yb =p paves the way for equality xa/y'b =p. It is by no means certain that equation (8) below obtains and it is therefore interesting to search for the monetary price, Po, which is the equilibrium price precisely because it establishes (and is optimally the only price to do so) equation or equilibrium (8). x'a y'b = xa yb (8) To repeat: The equality of xa/yb and x a/y'b is neither axiomatic nor postulational; it is a condition of equilibrium; therefore the first task lying before us is to solve equation (8); and, methodologically, solving an equation is an truly determinative process. We set out with the intention of disproving GET; in the present case where the economy carries money as a means of exchange we still haven't made our point. On the face of it, GET is now unassailable. Starting from equation (8) if: Bernard Schmitt 117 x' = x then: y' = y Therefore, only one equation, either x' = x or y' = y, remains to be solved once equation (8) has been established. The need for a second determination would no doubt harm GET if Walras's law did not come to its rescue. If the said law has any meaning at all it can only be by making one determination redundant. We can assign two alternative functions to Walras's law, namely to replace by a simple definition anyone of two equations, either xa/yb = x'a/y'b or x = x' (and, identically, y = y'). Whichever choice we make, the determination of the monetary prices of n goods requires that we solve only n - 1 equations - that is, unless there is some error in our reasoning. And indeed there is such an error in the argumentation we have just pursued, and a really crude error at that. No law can supersede the freedom of agents A and B in determining the monetary price, Po, at which they both decide to exchange positive amounts, x, x', y and y' of goods a and b: it is entirely up to them whether at price Po they positively engage in trade, or otherwise. It would be quite absurd to claim that Walras's law tampers with the free choice of sellers and buyers. Furthermore, no law can give us any assurance as to the equality of x and x', or of y and y'. Again, agents are free to decide in what quantities they purchase goods at given prices. Like Say's law, the law formulated by Walras applies to a whole market and has nothing to say about the individual behaviour of sellers and buyers. In a monetary economy Walras's law simply states that in any given market session there can be no difference in value between the sum of purchases and the sum of sales. We are in a position now to submit GET to a litmus test: is Walras's law incompatible with the inequality of x and x', or of y and y'? If not, the equalization of x and x' (or, identically, of y and y') remains an open problem after equation (8) is solved and even though the 'law' of Walras is fully taken account of. Example: 2 units of good a for I unit of good b is the equilibrium price, Po. The corresponding monetary units are pure numbers. 2 - =Po 1 118 Determination of Money Prices At the equilibrium price each agent plans to exchange positive amounts of goods a and b, or else the stated price is not an equilibrium price. Suppose that A decides to exchange I unit of a for 0.5 unit of b, a ratio which conforms with the equilibrium price. On the other hand, B decides to exchange 4 units of a against 2 units of b, again a ratio which is in conformity with the equilibrium price. Is Walras's law trespassed under those conditions? Not in the least. In money units, A exerts a demand equal to 2 while B's demand is equal to 8 units of money; at the same time A's supply is equal to 2 and B's to 8. The two demands add up to a value of 10 units of money; the value of the supplies amounts to the same total of 10 units of money. In short, the 'law' of Walras is fully in place. Now both the prerequisites which GET makes use of in finding the equilibrium price are satisfied: each agent accepts one and the same price, Po, with respect to planned exchanges of positive quantities of goods a and b; and Walras's law obtains. What more could GET do? Absolutely nothing. Yet, no exchange can take place; granted, the subjective choices are reconciled; but the objective condition required for market exchanges to take place is still lacking: does the market displace I unit of good a, or rather 4 units of a (in both cases for an equivalent quantity of b)? Only if it were possible for the market objectively to merge numbers I and 4 into one identical number could GET make any logical sense. To find Po in the set of real numbers is already an awesome ('Waldian') feat. When, against all odds, Po is really known, A contemplates the exchange of x units of a for y units of b, where xa/yb is equal to Po; likewise B sees the exchange of y' units of b for x' units of a, x' a/y'b being equal to Po. But nothing can positively happen in the market unless we first make sure that we solve equation x =x', or, identically, equation y = y'. But how are we to be successful at all in solving one of those equations, as surely we must, unless we loose Po and make it vary in either direction, in positive or negative increments? In the end, Po both is and isn't an equilibrium price, a patent contradiction which swallows up GET. Physical bodies have gravitional mass; if goods had a numerical mass, price ratios would be objectively predetermined; GET could then set out to solve equation x =x' instead of searching for price Po. But then GET could no longer claim to be a theory of price determination. The assumption of numerical masses comes very close to assuming that each initial stock changes hands entirely in a single market session. There is no need then for a profound theory of price determination! In fact, the exchange of an entire stock is an exceptional case within GET (see Walras, 1889: 76; 1900: 123-33). In the general case the markets are cleared because the set of equilibrium prices mediates the choices made by the agents. 'That is, there exists a set of prices (a competitive equilibrium) that could logically reconcile the potentially conflicting choices of all the economic agents' (Weintraub. 1979, p. 89). Logically, there is no possibility whatever for such prices to exist. In conclusion, what exactly is GET capable of doing? And at which precise point does GET cease to have any meaning? Bernard Schmitt 119 The correct answers to these questions are unambiguous. Let us state them by means of an example. Consider the following price: 2 pens/l ream of paper. Two entirely distinct scalars can be applied to the given price. Physical scalar: price 4 pens/2 reams of paper is 'equal' to the given price, 2 penslI ream of paper, provided we multiply the 2 pens and the ream of paper by a pure number, namely coefficient 2. Purely numerical scalar. If 2 pens are measured by 2 units of money while the measure of I ream of paper is also 2 units of money, the given price, 2 pens/l ream of paper, defines the two equivalent terms of an exchange (virtual or realized), where each term is measured by 2 units of money. Now, if we change the definition of our 'yardstick' by multiplying each unit of money by a constant coefficient, say by the number 2, then 2 pens are measured by 4, instead of 2, units of money; the same applies to the ream of paper, which now 'exchanges' for 4 units of money, where its previous worth was 2 units of money. Clearly, money prices which differ only by a numerical scalar are really identical. Given that 2 pens and 1 ream of paper are the terms of an exchange (virtual or realized), it is immaterial which monetary values we attach to these goods so long as we leave their physical quantities (2 pens and 1 ream of paper) absolutely unchanged. We are now in a position to offer a final assessment of the logical validity of general eqUilibrium analysis. GET is in no capacity to determine equilibrium prices unless it goes against its own grain; in marginalism, of all theories, an exchange between 2 pens and 1 ream of paper can never be taken as being identical to an exchange of 4 pens for 2 reams of paper. Logically barred from having recourse to physical scalars, GET can never get anywhere near to determining equilibrium prices. And in a research program where no equilibrium prices can be determined, numerical price scalars are of no conceivable use. Something is deeply wrong with GET. Can it make good its essential weakness by giving more serious attention to the nature of money expenditures, particularly in relation to real time? By submitting Paul A. Samuelson's investigations devoted to the study of 'expenditures plotted against time' to a reductio ad absurdum, we will finally gain a clear insight into the nature of money expenditures. A new paradigm, encompassing the whole field of macroeconomics, will then emerge in its essential traits. 'It takes a theory to kill a theory.' Hopefully, the new paradigm will eventually kill general equilibrium analysis, but only to prolong its profound teachings in an entirely new form. 2 THE EMERGENCE OF THE NEW PARADIGM GET has been in existence for a period of over 100 years. A recent issue of The Economic Journal [1991] is entirely devoted to predicting what economics will be 120 Determination of Money Prices like 100 years out. Twenty-two authors were asked to look into the crystal ball. No wonder that what they saw was a projection of their own present beliefs. We are all of us secretly convinced that earlier writers had a transient message while science is receiving its definitive status at our hands. If we in tum may venture a prediction, it runs as follows: GET is an accomplished but also a finished 'science'; we may hail Debreu's demonstration that equilibria are finite in number; after all zero is a finite number; price equilibrium is a wilting concept; in the next 100 years economics will at long last be reconciled with the facts: relative prices are pure figments of our imagination; even axioms cannot overcome the fundamental illogicality consisting in measuring goods in goods (or, for that matter, in producing commodities by means of commodities); the principal canon of economics must be that goods are measured in numbers; the only prices which actually exist in the real world are identities established between physical goods and pure numbers. If no exchange whatever is concluded between distinct agents and if exchanges do exist, it then follows that all exchanges are clinched between each agent and himself. We are well aware of the fact that there will always be some theorists who will pretend he or she witnessed exchanges between distinct agents. But formal reasoning cannot be intimidated into silence by such empty 'arguments'. Exchanges between distinct agents are theoretical monsters which can only exist in a self-contradictory theory. In fact, nobody ever was able to observe an exchange effected between distinct agents, any more than a licorn. No exchange can be reduced to a mere barter. Each term of an exchange is measured by a number (GET is quite correct in this respect). By the pseudo-exchange in which a rabbit is traded for a carp, the price of the rabbit is a carp, not a number. No axiom can alter the fact that a number of carps is a collection of carps and not a number. The whole tradition founded by the French economist Uon Walras (Le. GET), a brilliant period in the history of our discipline, should therefore be considered as having come to an end. But is it not exceedingly odd to refer to an exchange taking place between an agent and himself or herself? We shall presently find that this 'new' definition of exchanges is really quite commonsensical. Anyway, there is no choice since the commonly accepted definition is fraught with formal error. From the moment we accept, as no doubt we must if we are to avoid contradicting ourselves, that all actual exchanges are 'absolute Hows', which can only be established from X to X, where in each conceivable experiment X is one and the same person, it must follow that each commodity, if it is exchanged at all, is exchanged 'for' or 'into' itself. Thus the relevant criterion is clear cut: any theory that purports to study 'exchanges' which it attributes to distinct agents and which are meant to involve Bernard Schmitt 121 distinct goods belongs to paradigm I, whatever its claims to heterodoxy may be in some other respect. It is therefore right to say that the theory of non-Walrasian equilibria and the theory of disequilibria both pertain to paradigm 1. Even Sraffa' s effort stems from the same paradigm. How else are we to interpret the assumed exchanges between iron and wheat, that is between two distinct commodities? In the short space of this article we can propose only to outline the main features of paradigm 2. But it seems much more important first to combat what we strongly feel to be the main errors we Post Keynesians share with those neo-Classical economists who deign to pay some serious attention to Keynes, like Paul A. Samuelson. Multiplier theory is a case in point. We shall therefore. attempt to show at some length that the theory of the Keynesian multiplier is a welter of contradictions. Samuelson, in'particular, comes to grief by failing to see that all incomes, not only 'marginal' incomes, are the object of a 'new creation' taking place in each period. Once multiplier theory has been reduced to debris it will still be found to contain an essential message, to the effect that in each period banks and entreprises newly create the monetary or numerical form of current outputs. We cannot hope to be able to conceive of these creations with any degree of clarity until we succeed in breaking through the multiplier barrier. We can then subject our minds to another crucial test by examining the pure theory of money expenditures in their relation to the flow of time. Here too Samuelson is at the centre of events; granted, again his analysis is infuriatingly illogical; but do we Post Keynesians do any better? 'The concept of money in its full sense can only be developed in a monetary eqUilibrium model which rejects Walrasian microtheory for Marshallian price analysis' (Davidson, 1982, p. 78). We shall find that Marshallian prices in turn must be relegated to the past for they too are relative prices. To say that concrete units of money, as they actually exist in the real world, are pure numbers may be a startling claim; nevertheless, no theory of truly absolute prices can be built on a different hypothesis; nor is the theory of money units defined as purely numerical units merely hypothetical: it is the true theory, perfectly suited to the experiential facts of the matter. We shall then briefly show that the fundamental reason why GET completely failed to come to grips with the main issue in economics, namely the determination of prices, lies in the curious conception which mathematically minded theorists entertain about the 'meaninglessness' of 'mere' definitions. In reality, true definitions are supreme pieces of information. In physics, the equivalence of mass and energy is a definition. Prices are determined by definitions or identities and not at equilibrium points. In conclusion, the theory of 'circular monetary flows' will be briefly tested against the principal teaching inherent in paradigm 2, namely the constant identities which hold in each period between the flow of newly produced physical goods and the corresponding 'streams' of money. 122 Determination of Money Prices Multiplier Theory and Beyond Capital Goods and Consumer Goods What happens when an additional sum of money, y, is 'injected' in the production of investment goods? The 'factors' of the production of investment goods spend their additional income on consumption. It follows (or so it seems) that expenditures to the amount of y elicit an additional production of consumption goods. Total production is then increased by twice its initial increment. This 'multiplication' is due to the fact that wages (and other incomes) earned in the production of investment goods (like roads) are spent in purchases of clothing, housing, food, furniture - and certainly not for the acquisition of a portion of a motorway. By increasing employment in one sector, the government triggers off an identical increase in the other sector. The multiplier is thus equal to 2. Total employment increases by an amount corresponding to the formation of additional incomes equal to 2 y. A Chain of Incomes This is only the first link in a whole chain of events (again, so it would seem). The original impulse, transmitted from the investment sector to the consumption sector endlessly continues its wave-like career within the consumption sector itself. The reason for such an income propagation is simply that the 'factors' employed in the production of an additional quantity of consumption goods (purchased by the workforce newly employed in the investment sector) in tum issue a new order for consumption goods, and so on. After n such 'rounds', it appears that if 10 men are newly employed for one day in the investment sector, 10 men are newly employed in the consumption sector for 9 consecutive days; the multiplier is then equal to 10. Propensities to Spend We have assumed that households spend 100 per cent of their income increments; their propensity to consume is then equal to 1. We are told that in 'fact' (?) the marginal propensity to spend is less than I; households save a proportion (say a third) of their additional incomes. But we are left in the dark as to what becomes of these savings; these sums of bank money must be borrowed by someone; now, whether it is households or firms which borrow the sums of income that are made available on the financial markets, total spending must always be equal to total incomes: logically, the total propensity to spend can only be equal to l. But the measure of the propensity to consume is only relevant in determining the magnitude of the multiplier, not its actual existence. Let us therefore suppose that the propensity to consume additional incomes is equal to I as if that were a special case. The multiplier is then in full force; if it fails even in that extreme case, the correct conclusion must be that income multipliers can never be greater than 1. Bernard Schmitt 123 The Multiplier in Either Sector It is noteworthy that the hypothetical 'propagation' implied in multiplier theory first occurs between the investment sector and the consumption sector but that it is then continued within the consumption sector. The thesis that the 'multiplying process' is rooted exclusively in the production of investment goods is therefore untenable. If no multiplication of incomes occurred inside the consumption sector, a marginal propensity to spend equal to I would yield a multiplier equal to 2 in the nth period and not a multiplier equal to n - I as is required by the theory. Once logic has been granted to take this first step, it is already clear that the original impulse can happen in either sector. A 'sudden' increase of employment in the consumption sector is followed by exactly the same multiplying process as if the increase had occurred in the investment sector. Hence, the multiplication of incomes can have nothing to do with the fact that wages earned in the production of investment goods are spent on consumption goods. The factors of production do not as a rule repurchase their own products from the firms which employ them. What is the difference to multiplier theory between incomes earned in producing neckties and spent on cars and incomes earned in building a railway and spent on food? Incomes and Additional Incomes Multiplier analysis starts from a given level of employment and income. At that level, no multiplication of incomes occurs; multiplier theory applies to increments of macroeconomic employment and income. Suppose that the national income is equal to 100 units of money. What is the corresponding level of economic income-expenditures? Multiplier analysis provides no response to this question; but the correct answer can be gained from pure reasoning: the given level of employment and income has been sustained in the periods preceding the new impulse that initiates the multiplying process; we know by inference that in each period where the given level of employment and income was sustained, final expenditures on the market for goods amounted to 100 units of money; if expenditures had been either below or above that sum, the level of employment could not have been sustained at its previous level. At this point a logician could not refrain from butting in: the 'multiplicand' isa positive increment relative to a given level of macroeconomic income; not only is the previous level of income repeated or reproduced in the relevant period but the multiplicand is a positi ve addition to it; multiplier analysis unjustifiably treats the two integral parts of total income defined in that manner in two entirely different ways: to the extent that the current income is equal to the previous income, no 'propensity' is applied to it; only the supplementary income gives rise to expenditures which are taken account of in the definition of the propensity to spend and, as a consequence, in the definition of the multiplier. Income is first sustained at the level of 100 units of money; the multiplicand initially adds 10 units of money 124 Determination of Money Prices to total income. Multiplier theory wantonly discriminates between the 100 'first' units of current income and the to units which are added to it; the 10 additional units will be multiplied in the subsequent periods, while the 100 other units will not. Truly baffling logic! The Multiplier Applied to the Economy as a Whole So let us consider the national income as a whole, including its increase, if there is one, in the given period; now the relevant propensity is the average - not the 'marginal' - propensity to spend. Firm 1 pays out 60 units of money to its factors or production. Suppose for a moment that each individual 'factor of production' has the same average propensity as the rest of the population, for instance a propensity to save one half of his current income. Under these circumstances firm 1 and 'business enterprises as a whole cannot possibly recoup in consumption sales their previous disbursements to the factors of production' (Samuelson, 1966: 1126). Zeno's Paradox Firm I can recoup only 30 units of money. We are here reminded of Zeno's paradox: if firm I cuts its workforce by half, the factors of production it employs will earn 30 instead of 60 units of money; out of these incomes they will save 15 units of money; eventually all firms must go into receivership; that absurd conclusion holds even if we avail ourselves of the 'law' of increasing propensities to spend out of diminishing incomes. As we already know, Samuelson suggests a less dismal diagnosis: firm t can avoid creating any unemployment provided a firm 2 concurrently produces an income of 60 units of money; out of a total income of 120 units. households spend 60 units of money (or less, if the average propensity to save has gone up - but never mind); firm I now recoups the total cost of its current production, that is 60 units of money. But what happens to firm 2? Firm 2 recoups a zero-sum of money. In the national economy the work force of one of these business enterprises is made redundant; the inevitable unemployment can be shared in any proportion between firms; the end result is always t~e same: if at full employment the total national income amounts to 120 units of money, considering that a sum of 60 units of money is added to savings, it must follow that out of every two factors of production one is made redundant. With a view of fending off this ludicrous result, Samuelson offers only one counterargument, which turns out to be specious: firm 2 produces investment goods. So what? Are firms producing investment goods dispensed from the obligation to recoup their costs of production? The Multiplier Logically Equal to 1 When the Propensities to Spend are Equal to 1 Consider first a national income of 100 units of money in period 0 which is sustained in period I. As a rule, firms pay their factors of production before they can' Bernard Schmitt 125 sell their current output; any production in excess of orders requires the entrepreneurs to foresee future sales; we are not concerned here with possible errors in anticipations; so let us suppose that firms have perfect foresight; it can then be inferred that the actual propensity to spend out of given incomes is equal to 1. Now, is the output of period I in any way induced from the output of period O? At first sight a positive answer seems to be called for; is it not because the national income formed in period 0 is entirely spent on the corresponding output that production can be repeated at the same level (100 units of money) in period I? Post hoc propter hoc. On second thought we must reverse that judgment. In itself propensity 1 applied to the output of period 0 yields no information concerning the propensity which will apply to the output of period 1. The Central Error in Multiplier Theory We can now deal with the central error underlying multiplier theory. The distinction between marginal and average propensities to spend covers up the much more important distinction separating the bulk of the national income formed in each period, which is taken to be 'self-reproducing', in conformity with the law of 'conservation of energy in economics?' (Samuelson, 1972, p. 450), and the increments to the national income which may occur in given periods and which, if and wheri they do occur, are due to 'money creations' or to 'income creations'. This distinction opens up an entirely new perspective. If some incomes are 'created' how can we explain that other incomes are simply maintained in 'inertia' through time, where they are deemed to be neither created nor destroyed? Where does the truth of the matter lie? Do incomes endure or are they ephemeral and indeed timeless entities? The only positive feature of multiplier theory is the multiplicand, a word which all authors employ to mean an income creation, often seen as being identical to a measure of governmental deficit spending. The/ormation of a new income constitutes a 'semi-transaction' which is only viable as a whole when the complementary semi-transaction is adjoined to it; simply put: the income created in the production of a commodity must later be spent in purchasing it; if the corresponding final expenditure is lacking, a given income creation brings no additional purchasing power to the economy as a whole, thus placing the improved level of employment in immediate jeopardy. A Crass Contradiction in Multiplier Theory When, hopefully at not too distant a point in the future, it is finally realized that incomes are quantum phenonema. theorists will look down on the Keynesian multiplier with some disdain. Whole generations of researchers have been led astray. From the beginning though an unprejudiced mind could have seen right through the multiplier theory. Inside one's own discipline it may be impossible to achieve any degree of objectivity; but suppose a pure mathematician, unpreju- 126 Determination of Money Prices diced by any reasoning in the field of economics, were to read the 'simple mathematics of income determination' (Samuelson, 1966: 1198-9): presumably he or she would not find much of it obscure but could not help rejecting the whole affair as being purely nonsensical. This is 'the heart of income analysis': 'By definition, national income (at market prices), Y, can initially be set equal to the sum of consumption expenditure, C, and net investment, I: Y= C+I 'If Keynes had stopped with this identity, we should be left with an indeterminate system. In his simplest model of income determination, he added the following two hypotheses: (a) consumption is a function of income, and (b) investment may provisionally be taken, at anyone time, as a constant. Mathematically, these relations may be written C = C(Y) and 1 = T 'When we substitute these two into our first identity, we come up with the simplest Keynesian income system (I) Y=C(Y)+[ 'This is a determinate system, being one equation to determine one unknown variable ... 'Equation (I) is crucially important for the history of economic thought' (ibid pp. 1198-9). The nominal definition of the national income is Y, where Y is a symbolic name; now Y is the definiendum. i.e. the 'thing' which it is the theorist's task to define; if the real definition of the national income where merely a matter of semantics or of words, then Y would not be a definiendum but simply a definitum. Scientists are indefatigably searching for definienda, not for definita, which are trivial. The corresponding definiens to Y correctly construed as a definiendum is, according to Samuelson, 'the sum of consumption expenditures, C, and net investment I' . Hence, provided we know what consumption expenditures are and also what Samuelson means by net investment, we can deduce the nature of the national income. It is the formal treatment which Samuelson inflicts on his own definition of the national income which is apt to alert the interest of the reader. Replace the definiens by the definiendum; equation 1 then reads Y = Y. Surely nothing is wrong with an equation whose terms are one and the same thing. However, Samuelson lays stress on the fact that Y is equal to Yonly at the point where the 'consumption Bernard Schmitt 127 function' yields an adequate value for C. In all other cases, infinite in number, Y is either greater or smaller than Y. Hitherto it had been believed that the terms of an identity or a definition are always equal to each other, under all conceivable circumstances. Certainly Keynes's theory is far from sterile if it be true that it establishes the existence of identities which are defined between non-identical terms: by definition, Y is the sum C + I; nevertheless, Y usually differs from C + I; that is, Y generally differs from itself; finally, the equality of two identical terms is a condition of equilibrium; for instance, when C = 112 Y, and when 1 is constantly equal to 10, Y is never equal to Yexcept in a single case, where Y is equal to 20. Samuelson then decides that Y =20 is the determinate value of the national income that we are looking for, the constant value of net investment being multiplied via the consumption expenditures! 'The intersection of C(YJ + 1 with the 45 0 line gives us our simplest 'Keynesian cross', which logically is exactly like a 'Marshallian-cross' of supply and demand' (ibid., p. 1199). The Keynesian cross is the point at which Y and C + I, that is Yand Y, are of an equal value! What we can be certain of is that the '45 0 line' is indeed a pons asinorum. We agree that 'equation (1) is crucially important for the history of economic thought' but under the express proviso that it be reduced simply to equation Y = C + I. If 1 is a constant, it then follows that the consumption function is C= Y-I. It is not always easy to find the reason why a theorist suddenly slips into logical error. In the present case the correct explanation is fairly obvious. In all probability Samuelson is well aware of the fact that an identity is an identity; but what matters here is whether a given identity is instantaneous or whether it has a positive timedimension. The supply of a commodity and its demand are Marshallian 'forces' which coincide at equilibrium; it can then be said that these two opposite 'forces' are merged into a single double-sided event. In Samuelson's view the national income is a global Supply while the sum of consumption and investment form the corresponding total Demand. 'Logically', says Samuelson, Keynesian global Supply and global Demand are exactly like the Marshallian supply of and demand for a given commodity. Now, who says that the supply and the demand relative to a given good are always identical 'forces'? Under the rule of paradigm I, which Samuelson accepts, supplies remain distinct from demands at all points in time with the exception of the single time-point where markets are cleared. It makes perfect sense, therefore, to study the variations over time of Y relative to C + I, in spite of the fact that C + 1 is the definition of Y. Or does it make sense? We can certainly be in no doubt about the correct answer, which is emphatically in the negative. If, like neo-Classical prices, incomes were instantaneous magnitudes, then the Marshallian and the Keynesian analyses might cross and deliver the same information. In fact, though, there is no such a thing as an instantaneous formation of income. Ex ante, ex post analysis is therefore irrelevant. It takes no time at all to conclude an exchange between commodities a and h. However, it takes a whole calendar year to produce an annual income. 128 Determination of Money Prices It is fair to say that Samuelson's study of the national income in relation to the flow of time is not quite masterly. This single fact accounts for most of the formal errors which profoundly mar his scientific work. Samuelson's theory of money expenditures; a reductio ad absurdum Samuelson's aim is to complement GET, not to reject it. The concept of money expenditure is hardly present in the ADM model (the modem version of GET) and when it is alluded to, it is reduced to the lesser status of the numeraire concept. These arguments force the conclusion that the kind of numeraire 'money' so easily introduced into the extended ADM model is intrinsically different from the 'money' which is the stuff of Keynes' economics. It is a credit to the perspicacity of various Post-Keynesians like Leijonhufvud, Clower, Davidson, Minsky, Shackle, S. Weintraub, Kregel, et aI., that they have recognized this point. (Weintraub, 1979: 96) Samuelson has great merit in probing beyond 'the 4,827th reexamination of Keynes' system' (ibid p. 38) in his endeavour to integrate money expenditures as they happen in the real world into the general equilibrium approach. Samuelson's arguments are of a general nature and can safely be said to represent the doctrine of both schools, neo-Classical and Post Keynesian. Once those arguments have been 'exploded' by the impact of a reductio proof, it may become more clearly apparent what precise fundamental breakthrough PostKeynesian economists have to convey if they are to convince mainstream theorists that the GET-ADM paradigm has finally become irretrievably obsolete. Small but Finite Periods of Time In the field of income determination, especially with regard to national income, there is no point in choosing a very short period of time, say a second (which is a considerable duration in quantum physics). But a devil's advocate might decide to reason in terms of a nanosecond. National income is correspondingly dwarfed but not annihilated. In a nanosecond a positive expenditure can take place; indeed it takes no time at all to spend a sum of money, however large. But it does not follow that an instantaneous expenditure defines the national income relative to a point in time. At any given instant no nation can produce a positive amount of goods. So, only a finite period of time, of positive duration, such as Dennis Robertson's 'day', can sensibly be chosen on the time axis as a correlate to the formation of national income. Bernard Schmitt 129 The following proposition, advanced by Samuelson, must therefore be rejected: 'there is no question that at every instant of time there will exist respective amounts of investment, consumption, and national income, but this instantaneous relationship will be anything but stable (Samuelson, 1966: 113 t). This proposition cannot be true. Investment and consumption are expenditures, each given in a single 'impulse' or in a series of successive impulses; but what can possibly differentiate those two cases if consumption and investment expenditures take place at every instant of time? Furthermore, if positive sums of consumption and investment expenditures took place at every instant in time, then it would necessarily follow that the national income computed over any finite period of time, however short (say a nanosecond), would be infinitely great. Samuelson's mind is prejudiced by the mathematical 'tool' which he wields with such obvious glee but which, far from really being a mere instrument in his hands, leads him to tamper with the facts and to jump to forgone conclusions. Does it take one year to payout the national income for one year? If so, 'the government is in a position to determine its own level of expenditure at each instant of time' (ibid.: 1127). Expenditure in Economics and Motions in Physics Imagine what would happen if the motion of the earth around the sun instantly came to a halt. Fortunately, the motion of the earth is continuously positive at each instant. True, the distance covered by the earth at any given instant is nil although the total motion of the earth is a complete revolution around the sun in a period of approximately 365 114 days. Why is it, then, that the earth covers a zero-distance at each point in time, defined as a zero-duration? The correct answer is quite simple: each infinitely small part of a year corresponds to an infinitely small part of the yearly distance covered by the earth. And if no positive time of any duration is allowed, then no distance is covered by any mobile unless its velocity were infinitely great. Samuelson seduces himself into thinking that expenditures are variables which quite naturally belong in difference or differential equations, a temptation which is strengthened by the fact that money expenditures have to do with numbers. Yet Samuelson has no misgivings concerning the existence of a peculiar class of expenditure, namely money creation (ibid.: 1158 and 1184) or, even, income creation (ibid.: 1207). There is an unbridgeable gulf between continuous streams of expenditure (which have no factual existence) and acts of money or income creation (which occur every day). If Samuelson had turned his back on the patent nonsense of streaming sums of money, outpouring from a mysterious source (presumably from a misconceived bank), he might have chosen instead to further his own idea of income creation, thus placing his collected papers on a firmer scienti fic footing. 130 Determination of Money Prices lnfinitesimals and lndivisibles Should we concern ourselves with the metamathematical question of the existence of infinitesimals and indivisibles? In so doing we could easily get lost. In fact the problem at hand is far easier to solve. If the relationship between expenditures and the flow of time is mathematical, then expenditures are afunction of time and should be studied accordingly. But if the logic governing expenditures in time is 'non-mathematical', analysis becomes more and more illogical the greater the use it makes of even the most rigorous mathematical techniques. The truth then lies with the logician while the mathematician is left with an 'empty box'. In fact, being strictly instantaneous events, money expenditures bear no relation to infinitely small spaces of time; on the other hand, all money expenditures span finite periods of time; for example, wages may be paid out every week; and weeks are divisible periods; what makes a period indivisible is the tautological fact that a given expenditure relates to that period as a whole. Physical Outputs and Money Expenditures An analysis of physical output as a function of time throws some light on the way expenditures are generally approached by economists (like Samuelson). Suppose cars are produced at the rate of 10 per unit time. Add the assumption that the production process is submitted to no acceleration (positive or negative) during the considered time period: production is constantly taking place at a rate of 10 cars per unit time. It follows that the production of cars is a constant function of the time which flows during the given period, a function which can be represented by a straight line. We can see no reason why the physical output of incommensurable or inchoate goods should not be comprehended as a function of continuous time. The real difficulty lies elsewhere, namely with money expenditures. Each instantaneous expenditure has a positive, finite result in contrast with instantaneous physical productions which invariably result in zero-outputs (see Mauro Baranzini (ed.) (1982) Advances in Economic Theory, 'Time as quantum': 115-25). Identities and Equilibria Form and Content The price of each commodity is determined by an identity and not at some point of equilibrium. Ludwig Wittgenstein rejected early 'the treatment of identity in Principia Mathematica, where one and the same identity sign was used in affirming a thing's self-identity and in denying the identity of two things, as though it made sense to suggest that two things might be one' (in Hallett, 1977: 294). Bernard Schmitt 131 Never certain to happen, equilibria are contingent events. Identities are logical certainties. Quite apart from the question of the determinacy of relative prices, which we have already dealt with, economics is therefore necessarily a more exact science if prices are identities. Wittgenstein's conundrum is resolved as soon as the distinction between form and content is applied to it. To the amount of x units, money is the nominal form 'enveloping' commodity a. Commodity a and x units of money are only one thing, material (physical) and formal (numerical). The Equilibrium Fallacy: or the Trojan Horse Which Introduces Mathematics into Economics All logical difficulties encountered by mainstream theorists are connected with the Equilibrium Fallacy. If it were true to say that the prices of x units of commodity a is y units of commodity b, then prices would indeed be determined at equilibrium points. We could then legitimately look for some equality between xa and yb; xa and yb would then contain the same number of 'utils' (Irving Fisher). The 'utility calculus' is unquestionably a branch of mathematics. When a mathematician is told that, in fact, prices are identities, he can have no further interest in macroeconomics. This is not to say that mathematicians turn a blind eye to the consideration of identities within their own science; on the contrary, it may be true to say that all mathematical theorems are tautological identities. However, identities in economics can be of no interest to a mathematician. So let us hope that Samuelson is correct in boldly asserting that Keynes really meant to further equilibrium-schedules when he clumsily put forward identities. 'Moreover, there is reason to believe that Keynes' thinking remained fuzzy on one important analytical matter throughout all his days: the relationship between "identity" and functional (or equilibrium-schedule) equality' (Samuelson, 1966, 1972: 1529). Paraphrasing Lichtenberg we may say that very few people have ever seen a pure identity. Samuelson is not one of them. Keynes simply means what he says when he lays down 'the definition of income, saving and investment' (Keynes, 1946: 52). Definitions are identities. Nothing of importance will remain to be discovered in macroeconomics once the scientifically correct definitions of income, saving and investment have been arrived at by 'a highly abstract argument' (ibid.: v). It serves no useful purpose - it is indeed outright counterproductive - to replace identities, which admittedly are very profoundly hidden in the depth of science, by superficial and artificial eqUilibrium schedules, whose only merit is to leave the scientist in his familiar surroundings (of difference or differential equations). A very serious confusion creates havoc with so-called modern economics: logic is the 'form' of all true science; mathematics is no exception; but mathematics is a formal science in yet another acceptance, for it is the theoretical study of numbers, which are 'forms' by definition. Mathematics is thus the fornial study of formal entities; economics is the formal study of substantive entities. When a the- 132 Determination of Money Prices orist sets out to study substantive entities by means of a science which is purely formal (mathematics) he is bound to come back empty-handed. 'So far as I can see, economic truths are not discoverable through the instrumentality of mathematics. If this view be unsound, there is at hand an easy means of refutation-the production of an economic truth, not before known, which has been thus arrived at' (John E. Cairnes in Samuelson, 1972: 1755). In answering this challenge, Samuelson cites the theory of general eqUilibrium: We may say of Walras what Lagrange ironically said in praise of Newton: 'Newton was assuredly the man of genius par excellence, but we must agree that he was also the luckiest: one finds only once the system of the world to be established!' And how lucky he was that 'in his time the system of the world still remained to be discovered'. Substitute 'system of equilibrium' for 'system of the world' and Walras for Newton and the equation remains valid. (ibid.: 1756) In the end John E. Cairnes might still be vindicated if it be true that general equilibrium theory is indeed sterile. Samuelson's own theory of expenditures cannot possibly salvage GET; on the contrary it brings along with it another, equally serious, logical error. Expenditures are quantum leaps and cannot therefore have any positive dimension in continuous time. The Paradigm of Repeated Creations and Destructions Versus the Theory of Circular Flows of Money It may seem strange to count Samuelson among the theorists who propound the paradigm of circular monetary flows. Yet, the following passage (which we have already cited) leaves no room for any doubt. 'For under the circumstances outlined, business enterprises as a whole cannot possibly recoup in consumption sales their previous disbursements to the factors of production' (ibid.: 1126). This is a clear statement of the heart of circular flow theory. Keynesian multiplier analysis is a perfect example of the same theory: apart from its possible (exogenous) increments the national income is continually disbursed (first half of a circle) and recouped (second half of a circle) by firms as a whole. The theory of circular flows is founded on the hypothesis of circulating incomes. More precisely, circular flow analysis is well founded if incomes change hands whenever they are spent. It can thus been seen that circular flow analysis is in the mind of all economists 'except those, hardly numerous, who claim that incomes are destroyed or completely wiped out as soon as they are spent. Paradigm 2 is the theory of instantaneous incomes, namely incomes which are destroyed' at the very moment they come to existence. No durable incomes exist anywhere in the world. This new paradigm, has been explained, no doubt imperfectly, in several books, written over a period of forty years (see mainly Schmitt, 1984). We can offer no further proof here; Anyway, if the argumentation Bernard Schmitt 133 developed in this article is correct, then it follows with logical necessity that no durable incomes can exist. GET needs a two-agent world to establish its theorems; the new paradigm can derive its principal teachings from a one-agent world. We shall first briefly uncover the definition of money, income and capital in a one-agent economy. In conclusion, we shall offer a sketchy outline of the new macroeconomic theory in relation to the national economy as a whole. Money, Income and Capital in a One-agent Economy The sole agent of the economy is A; A is employed by a firm whose payments are carried out by a bank; neither the firm nor the bank are agents; they are pure intermediaries. Leon Walras intuited the fact that money units are pure numbers. However, in themselves goods are distinct from numbers. It was overambitious to search for a theory purported to provide a linkage between goods and the set of pure numbers. The only logical way to introduce pure numbers into the economy as well as into the body of economics is to allow, on a temporary basis, the newly produced goods to be replaced by a number of monetary units. The only existing agent (A) would derive his payments in kind if he were to receive his own products directly. In fact, though, he is paid in money. In other words, we are considering a 'wage-economy'. The entrepreneur exchanges a sum of money (x units) for A's output. We shall presently see that it is more correct to say that the entrepreneur changes or converts A's output into a sum of money. Let us leave material money (gold or silver) aside. Today no material money is in currency anywhere in the world. All payments are initially issued by the bank acting as an intermediary between the entrepreneur and his employee. At the point in time when it is 'monetized', A's output is engulfed by a precise sum of negative money, -x units. From then on and over a positive but limited period of time, A's physical output is deleted and replaced by the positi ve sum of money, + x, defining A's current income. When, at some future date, A decides to spend his income, that is to destroy it (the positive sum of money, + x, being made to coincide with the equivalent negative sum of money, -x), he finally recovers his physical output, freed from its monetary mold. The definition of a negative sum of money is quite precise: even in the case where the firm owns a sufficient amount of 'circulating capital', the bank pays the factor of production, A, by debiting the enterprise. With respect to the bank, A holds a net credit (i.e. a positive sum of money) and, correspondingly, the firm holds a net debit (a negative sum of money). Coins and bank notes are no exception; whether the bank pays out units of 'book-entry' money, coins or notes to A, it is always true to say that the enterprise cannot escape the obligation to repay the bank. It can clearly be seen that commodity a is objectively changed or converted into a sum of x units of money since A's output, in as much as it is' defined in physical terms, has temporarily ceased to exist: it is nowhere to be found in the 134 Determination of Money Prices economy, not even with the entrepreneur who beholds only the grin of the cat, i.e. A's positive output deposited in an equivalent negative sum of money. We now know the meaning of money creation: the bank creates + x and -x units of money in one and the same 'impulse'. To create ex nihilo is a metaphysical conception; in fact, no positive sum of money can ever be created; the creation of +x units of money necessarily implies the simultaneous creation of -x units of money. We also know the meaning of a money income: A 'pays himself via the firm; the moment he receives his own products; income thus defines a physical output in monetary form. Most economists believe that money derives its purchasing power from the general level of prices. Purchasing power is thus given in a petitio principii: money expresses goods because goods are expressed in money. The basic and common error is the supposition that money and goods are dichotomous entities. In reality, A's money income is identical to A's physical output which is cast into a numerical (or nominal) form. It is meaningless to search for numerical expressions (prices) of physical goods except in production. In an exchange economy no prices can be determined. Production is the very foundation of macroeconomics, which is the science of values and measurements of all commodities (including financial assets). Only money can make productions and their results to cohere into homogeneous quantities. Figure 4.1 represents a complete instantaneous 'creation-destruction' of x units of money, in connection with the current production of factor A, employed by entreprise En. The bank creates x units of money in favour of the entrepreneur (En) who pays A via the bank; A in tum pays the bank. In this instantaneous process, A changes his physical output into a financial claim (certificate of deposit), that is into an IOU issued by the bank. Two identities are implicit in Figure 4.1; the current physical output of A is identical to x units of money, its 'numerical form'; the IOU issued by the bank is equivalent to its 'object', namely the x units of money which A deposits in the bank. In economics the only non-contradictory concept of value and price is to be found in those two equivalences. In GET 'entrepreneurs and banks are put to sleep for a while' (Morishima in The Economic Journal, 1991: 72); when they wake up GET will no longer be around. Pure theory will then again become possible and enjoyable, much to Frank Hahn's displeasure (see ibid.: 47) who decrees that the essence of pure thought is of necessity mathematical; for other types of thinking 'bring out the worst in us'. The income accruing to A is a macroeconomic value in a very strict sense: neither the bank nor the entrepreneur spends a positive income in the process from which A gains a positive income. No wonder theorists who believe that no incomes are formed unless equivalent incomes are spent, labour under the delusion that macroeconomics is founded in microeconomics. If the national economy produces no expenditures which are nil at one end and positive at the other, macroeconomics is deprived of any specific object. In the real world, the business enterprise (En) receives the new output, + x, deposited in -x units of money; En Bernard Schmitt 135 thus spends a zero-sum of income; this expenditure of a zero-income procures a positive income, + x, for A. The income earned by A is a macroeconomic magnitude, for the whole national economy gains an amount of net assets to the value of x units of money, which contain the newly produced current output. So paradigm 2 comes into its own; A engages in no 'exchange' with another person and A's output is not 'exchanged' for any other commodity; rather A objectively changes his own products into a sum of money; the essential point is that at no moment in time is A deprived of his own output which, as soon as it comes into existence in its monetary form or 'nominal container', belongs to him and to nobody else. This is tantamount to saying that even in a monetary economy current physical outputs are directly, if not immediately (banks and firms being intermediaries), owned by the factors of production. Above all - and this is the most significant point - units of money have no value in themselves; in fact, in the hands of En they are pure negative numbers which are made to absorb the newly produced commodities and in the hands of A they correspondingly define equivalent positive sums of money which function as surrogates to physical output. The new paradigm ensures that A's output - which is no longer arbitrarily given in initial edowments, that is, educed from nowhere (GET), but which is derived from A's production - is subjected, as a whole, to a 'quantum jump' that instantaneously changes it into a sum of numbers (monetary units); Consequently, the price of a is x units of money, where x is of necessity a constant since x is the only value of commodity a. Paradigm 2 yields the absolute price of commodity a, produced by A: x units of money is the price we are looking for, because the commodity produced by A is contained in x units of money. It would be nonsensical to ask at this point why A's output is contained in x, and not in c.x units of money; clearly, c is an arbitrary scalar. A sum of money is made up of pure numbers. What is the economic value of a number? A's output has value in one sense only: it is measured by its purely numerical 'envelope'. We are now able to understand the real reason why business enterprises have to recoup the disbursements that they make in favour of their factors of production. This constraint can have only one meaning: the physical output of each period has yet to be purchased when its 'factors' have already received their remuneration. How could this be true if prices were relative? To the extent that the 'factors' refrain from buying back their own products, firms are left to make up the deficit. Only at the moment when A disposes of his income by spending it does firm En recoup the + x units of money which 'fill in' the -x units of money where the physical output remained deposited with the firm, 'waiting' to be purchased. It would be a serious mistake to suppose that the current income survives after it has been spent; in fact, it is thus exhausted. Of the + x units of income created in the considered period not a single penny survives after the corresponding output has been retrieved from firms. The final purchase of good a gives rise to exactly the same flows as those depicted in Figure 4.1; the only difference lies in the permutation of the Determination of Money Prices 136 -x +x bank -x IOU=x En A -x +x output = x +x bank -x Figure 4.1 firm (which is now credited) and its employee (who is correspondingly debited). It is generally thought that in order to be transmogrified into capital, incomes only have to be maintained in existence for a certain (undefined) duration. That is nonsense. A positive event is needed to convert an income into capital. By purchasing an IOU from the bank A lends his income to the firm via the bank. An income lent, and nothing else, is a monetary capital. At this point two crucial pieces of information may be put forward. i) All incomes are necessarily destroyed at the moment of their 'birth'; and equivalent capital arises out of the destroyed incomes; such capital is the only wealth which the national economy gains without there being a corresponding loss of capital in the holdings of some other agent; the true definition of macroeconomic capital is thus uncovered. (ii) The usual insipid definitions of stocks versus flows are stated the wrong way around and are poised for a complete mutual reversal; in fact, monetary flows are instantaneous events while stocks last for positive durations. In the end, there is no such thing as a circular flow of money or of incomes; the correct description of monetary events is to say that incomes are instantaneously created and destroyed and thus transformed into an equivalent capital. Bernard Schmitt 137 A Sketchy Outline of the New Macroeconomics In an economy comprising any number of agents it remains true to say that each factor of production finally purchases his own products; if A purchases B's output, B purchases A's. It follows that prices are still strict identities. In a production economy general equilibrium theorists see two main markets, namely the market for productive services and the market for produced goods. We have just pointed out the fact that the financial market is implicitly and organically present 'between' the two other markets. Prices are thus simultaneously determined in three intimately interlinked markets. Another distinction is a 'prolegomenon to all future macroeconomic theories': before the discovery of paradigm 2 all monetary expenditures were taken to be purely arithmetical magnitudes; henceforth they are recognized for what they actually are, namely algebraic magnitudes. 'When we philosophize were are often tempted to represent things as the painter Klecksel did when, as a child, he drew human faces in profile and gave them all two eyes nonetheless (in Hallett, 1977, p. 147).' When we contemplate the facts in their true unbiased perspective, we find that expenditures lie on two opposite sides: money spent on the market for factors of production is income creating; money spent on the market for produced goods is income consuming. We are finally landed with an objective distinction between negative expenditures and positive expenditures. In GET one and the same event is a purchase for one agent and a sale for another; considered in itself such an event is neutral in general equilibrium theory. In the new paradigm a given money expenditure is either negative or positive in the concurrent judgment of all observers. Remunerations of factors of production are negative expenditures both for the firms and their employees; such expenditure are objectively negative, for purchasers and sellers alike, who know a positive income creation when they see one. Similarly, all expenditures which occur on the market for goods are objectively positive; irrespective of the observer, such expenditures are judged to be income destroying, for that is exactly what they are for the national economy and for the whole world. From now on economic theory will be endowed with a new dimension, negative and positive expenditures replacing the colourless conflation of all types of expenditure into a single unimaginative, ambiguous and misleading amalgam. The principle of excluded third notwithstanding, it is the presence of positive expenditures in the midst of negative expenditures which provides the only logically possible explanation for the existence of profits. A whole new vista opens before us, where real problems like unemployment, inflation and international debt will at last find their true scientific analysis and, consequently, their feasible remedies. 138 Determination of Money Prices References Davidson, P. (1982), International Money and the Real World (London: Macmillan). Economic Journal, The (1991), (Journal of the Royal Economic Society) (Oxford: Basil Blackwell), 101 (404). Hallet,G. (1977), A Companion to Wittgenstein's 'Philosophicallnvestigations' (Ithaca, NY and London: Cornell University Press). Keynes, 1.M. (1947), The General Theory of Employment Interest and Money (London: Macmillan). Samuelson, P.A. (1966, 1972), The Collected Scientific Papers of Paul A. Samuelson (Cambridge, MA and London: MIT Press). Schmitt, B. (1984),lnftation, chOmage et malformations du capital, macroeconomie quantique (Albeuve: Castella (1991); Paris: Economica). Walras, L. (1889), Elements d'economie po/itique pure ou theorie de la richesse sociale (Lausanne: F. Rouge). Walras, L. (1900), Elements d'economie politique pure, 4th edn (Paris: Pichon et DurandAuzias). Weintraub, E.R. (1979), Microfoundations: The Compatibility of Microeconomics and Macroeconomics (Cambridge: Cambridge University Press). 5 Money as Purchasing Power and Money as a Stock of Wealth in Keynesian Economic Thought Augusto Graziani 1 INTRODUCTION Ever since the publication of Keynes's General Theory, the definition of money as a stock of wealth has become dominant in economic theory. As all know, it has not always been so in the history of economic ideas. Originally the dominant character of money 'Yas considered as being endowed with purchasing power, and money was consequently defined as a means of exchange. The emphasis was moved away from the means-of-exchange function toward Ute store-of-wealth function as the result of a long debate concerning the definition and measurement of the value of money. It was then clear that money, if defined as mere purchasing power, could only be endowed with an indirect utility, originating from the utility of the goods money could buy on the market. On the other hand, if money had to yield utility on its own, it had to be useful not only when spent upon commodities but also when kept as a store of wealth. Money had consequently to be defined as a stock of liquid wealth. The transition from a definition of money as purchasing power to a definition of money as a stock of wealth was considered as a fundamental acquisition in the progress of economic theory. I Nowadays, supporters of the previous conception of money are few in the Anglo-Saxon world.2 Many more are active however in present-day France, most of them being the supporters of the so-called French school of the Monetary Circuit.3 In rigorous adherence to the principles of this school, Bernard Schmitt considers money as such to be in existence only in the very instant in which a payment is made and money itself passes from the balances of one agent to the balances of the next. When money stays idle in the balances of one single agent, Schmitt would consider it to be temporarily turned into wealth, something to be kept very distinct from money.4 The choice between one or the other of the two conceptions of money produces a number of analytical consequences, the main ones being summarized as follows: 139 140 1. 2. 3. 4. 5. Money as Purchasing Power and Stock of Wealth If money is defined as a stock of liquid wealth, it appears as an observable variable, the measurement of which poses no more problems than any other variable in the model. On the other hand, the notion of what Robertson called 'money which people acquire in order to use it' gets totally lost. 5 Clear evidence of this way of looking at money is given by Patinkin's statement that if money is demanded in order to spend it, this is not a demand for money but a demand for goods;6 If the definition of money as a stock of liquid wealth is coupled with a rigorous definition of eqUilibrium, according to which the budget constraints of all agents are simultaneously satisfied and there are no debts pending, bank money disappears altogether from equilibrium. In fact, a bank loan, being an unsettled debt of some agent towards the bank sector, would contradict the very definition of eqUilibrium. The only kind of money which can be in existence in equilibrium is money created by Government deficit - government debt being considered as consistent with eqUilibrium. In fact, in most macroeconomic models, the stock of money is defined as being limited to the cumulated fractions of previous government deficits not covered by debt issues. A further consequence of such a narrow definition of money is that the money stock can be defined as an exogenous variable and possibly as a given parameter; Thirdly, if money is defined as a stock of wealth, the definition of money can be extended to any kind of resource, provided it is accepted by agents as part of their liquid balances. In fact, it is now customary to speak of a multiplicity of monies, ranging from legal tender to bank deposits, to commercial paper, to government bonds as well as other resources. A similar way of looking at money allowed Kaldor to consider it as a totally endogenous variable,7 and nowadays allows Minsky to state that any agent would be able to produce money provided he were able to have it accepted in the liquid balances of some other agent. s A similar extention would be clearly rejected by those who define money as a means of payment. Actual means of payment do not go beyond legal tender and bank deposits, and the fact of paying a commodity by means of a private promise of payment would in fact be considered buying at credit in the absence of any true monetary payment; If bank loans and bank deposits are eliminated from eqUilibrium, interest paid by firms to banks (the so-called short-term interest of the money market) disappears as well, and any possible conflict between the world of industry and the world of finance is automatically eliminated. In fact, in most macroeconomic models, one single rate of interest is taken into consideration, namely interest paid by firms to savers (the so-called long term interest, or interest of the financial market); The disappearance of the money market gives rise to the further idea that the financial market might perform the role of bringing new liquidity to firms. If correctly understood, the role of the financial market is rather to make it possible for firms to get back their initial monetary outlays and to repay Augusto Graziani 141 their debts to the banks. Anyone who, deceived by the structure of the model, is led to believe that financial markets can be a source of new liquidity to firms as a whole, is bound to confuse liquidity with saving-two categories having little if anything in common.9 The counterpart of considering the financial market as a source of new liquidity is the old and popular idea that the role of banks is one of collecting savings and financing investment. 10 It should be clear instead that the reverse is true, namely that bank deposits are not a source of liquidity out of which bank loans are granted, since no deposit can exist if not as a consequence of a previous loan. If, on the other hand, money is defined as purchasing power, readily accepted on the market, the opposite consequences follow: (a) As we shall see later on when reverting to this problem, the first consequence of defining money as purchasing power not to be used as a store of wealth, is that if no uncertainty is present, no agent-if perfectly rationalwould keep idle balances. A puzzling consequence would be that money, while still being an intermediary of exchange, would disappear from the equilibrium position and would no longer be an observable variable; (b) If money is defined as purchasing power, the definition of money no longer needs to be limited to liquidity created by government deficit. In fact, even if in equilibrium no debts are pending between banks and firms, bank money may still be used as a means of payment. The only requirement of equilibrium is that any bank money which has been created and circulated in the market has also been destroyed when equilibrium is reached. Of course, those who prefer to define money as purchasing power tend also to consider the existence of a certain amount of debts among agents as being consistent with equilibrium. In this case, nothing prevents bank money from being in existence even in full equilibrium; (c) If money is defined as purchasing power, it only covers what is actually used as a means of payment, namely legal tender and bank deposits. The extension of the category of money to a multitude of liquid or semi-liquid resources is precluded (it is quite true that in principle any agent could print money and have it accepted in the market, but in fact this does not occur);· (d) If bank deposits are allowed to be in existence in equilibrium, two interest rates have to be considered and included in the model, the short-term rate (interest paid by firms to banks) and the long-term rate (interest paid by firms to savers); (e) Finally, when banks are included in the model, it becomes clear that their role is to create means of payments, while the role of the financial market is to return to firms their monetary outlays and to enable them to repay their bank debts. The financial market can no longer appear as a source of fresh liquidity to firms as a whole. 142 2 Money as Purchasing Power and Stock of Wealth THE DOUBLE RELEVANCE OF MONEY IN KEYNESIAN THOUGHT Keynes is usually considered as the economist who most explicitly and coherently pursued the idea that money matters only in so far as it is kept as a form of wealth. In fact, in the Keynesian model, the level of aggregate demand is influenced by the choice between real capital goods and cash balances, while sudden increases in liquidity preference may determine demand failures and unemployment. In this chapter an attempt will be made to show that common presentations of Keynesian monetary theory only reproduce that portion of Keynes's views which are developed in the General Theory. They do not reproduce the whole of Keynesian monetary thought, nor do they give a full account of the various roles performed by money in a market economy. As will be shown in what follows, in a capitalist economy money performs the fundamental role of determining the distribution of income. In fact, the very formation of profits is explained by the presence of money, as well as by the way in which money is created and introduced into the market circuits. The idea that money is a key element in income distribution is not explicitly developed in the General Theory. This is of little wonder, since the theoretical structure of the General Theory was developed in order to explain a totally different problem-namely that of permanent unemployment. However, other writings by Keynes, such as his early critical writings of 1911 concerning Irving Fisher's quantity theory, or the Treatise of 1930, as well as his articles of 1937-9 concerning the finance motive for holding money, describe in a most explicit way the role of money as a source of profits and power. A complete analysis of Keynesian monetary thought requires a richer basic framework than the one common to the best known macroeconomic models: a framework including as separate entities banks, firms, and households. It is no surprise that similar models are often considered with suspicion. Orthodox Keynesians follow closely the model of the General Theory, a model which ignores, or nearly ignores, banking activity, and which considers the stock of money as an exogenous, even if not necessarily constant, magnitude. Post Keynesians, on the other hand, consider money supply as being infinitely elastic, an assumption which causes them to neglect totally the problems connected with the creation of money. A more complete analysis of Keynes's writings before and after the General Theory will show that Keynes paid equal attention to money as a form of wealth and to money as an intermediary of the exchanges. It will also show that Keynes was equally conscious of the role of money in determining the level of aggregate demand, as well as of its role in determining the distribution of national income. 3 CREATION, CIRCULATION AND DESTRUCTION OF MONEY The process of creation, circulation and destruction of money has been analysed elsewhere,ll and it may be only briefly recalled here. Augusto Graziani 143 Let us imagine a pure credit economy, in which all payments are made by bank checks. Let us consider only consolidated macro sectors: banks, firms, and wage earners. Banks grant loans to firms, the outlays of which are exhausted by the wage bill. The moment wages are paid, firms become the debtors of the banks and wage earners the creditors of the same banks. A triangular debt-credit situation, typical of any monetary economy, is thus created. A picture of the economy taken immediately after the payment of wages would reveal that the whole of the existing money stock is a debt of the firms and a credit of wage earners towards the banks. If the government, along with the firms, were also to get credit from the banking sector, the nature of the process would not be altered. In this case, the central bank would probably be the source of finance, the government would become the debtor of the central bank, while the beneficiaries of the transfers would become the depositors (and therefore the creditors of the banks for the same amount). As soon as wage earners spend their incomes, money flows back to the firms. Under this viewpoint, it is immaterial whether wage earners spend their money on commodities or whether they buy securities issued by the firms. In both cases, the firms get back the money they have spent on wages and are thus enabled to repay their debt to the banks, or use the money again in order to pay new wages. In so far as the restoration of the firms' liquidity is concerned, consumption is equivalent to saving, provided saving does not take the form of hoarding. If no hoarding takes place, the proceeds originating from sales of commodities and from newly issued securities, enable firms to repay the whole of their debt to the banks. It need not be added that the firms, besides giving back the principal of their loans, are held to pay interest on them. The payment of interest raises problems which will not be analysed here. 12 We now tum to an analysis of the timing of the whole process. In a world free from uncertainty and populated by fully rational agents, the process just described would take place instantaneously. To begin with, no firm would ask for bank credit if not the very moment a payment is due. As Keynes writes: 'For there is no necessity to hold idle cash to bridge over intervals, if it can be obtained without difficulty at the moment when it is actually required. '13 Similarly, wage earners would spend their money incomes as soon as they get them, since idle balances would be both costly and totally useless to them. For the same reasons, firms would repay the banks as soon as they get back their initial outlays. As 1.B. Say once wrote: 'Lorsque Ie demier producteur a termine un produit, son plus grand desir est de Ie vendre, pour que la valeur de ce produit ne cMme pas entre ses mains. Mais il n'est pas moins empresse de se d6faire de l'argent que lui procure sa vente, pour que la valeur de I' argent ne chfime pas non plus. '14 The logical, even if disappointing, outcome of a similar situation would be that money, while still being a necessary element in a monetary economy, would no longer be an observable variable. A picture of such an economy taken at any moment of time would give the paradoxical image of a monetary economy in which money would be totally absent. 15 144 Money as Purchasing Power and Stock of Wealth In order for money to be an observable variable, not only money has to be the regular intermediary of exchanges but a second condition has to be met, namely that agents constantly keep a fraction of their money incomes in the form of liquid balances. In any economy in which money is an observable variable, there must be at least one agent who is in debt to the banks. In a monetary economy, pending debts are no longer contradictory to equilibrium. They are in fact part and parcel of the economic mechanism. To the extent that wage earners, instead of spending the whole of their incomes either on commodities or on securities, keep their savings in bank deposits, the firms are unable to repay their bank debts. If the previous level of output has to be preserved, the firms have to get new loans from the banks, which means that the 'stock' of money is increased. 4 MONEY IN KEYNESIAN THEORY It is a well known fact that in the General Theory Keynes defined money as a stock of wealth. If the stock of money is assumed to be constant, then in equilibrium firms must be able to repay, at the end of the production period, the whole of the loans granted by the banks in order to make production possible. Firms have therefore to sell securities, the value of which has to be equal to the monetary savings of households. For this to be possible, the rate of interest that firms offer on the securities they issue must be high enough to balance the liquidity preference of savers. Any increase in liquidity preference may bring about an increase in interest rates and a decline in investment and employment. The definition of money as a stock of wealth is therefore a necessary step to explaining demand failures and fluctuations in employment. However, if we leave the General Theory aside and consider the rest of Keynes's writings, a clear notion of money as purchasing power emerges. It may be interesting to recollect how Keynes, starting with his early monetary writings, openly declared his dissatisfaction with the state of monetary theory in Great Britain. In 1911, on reviewing Irving Fisher's Purchasing Power of Money, he wrote: The state of monetary theory and the literature of it in England present a remarkable contrast to their development in the United States. For nearly twenty years, ... no substantial contribution to this subject has been published in England. It is hardly an exaggeration to say that monetary theory, in its most accurate form, has become in England a matter of oral tradition. 16 His negative judgement on monetary literature in England was reasserted years later on reviewing Bendixen's book. I? Fifteen years later, when writing the Treatise, Keynes could quote but one British writer, namely F.W. Crick, while he Augusto Graziani 145 drew repeated attention to the continental literature, by mentioning more than once authors like Wicksell, von Mises, and L.A. Hahn l8 A first key point is firmly made by Keynes, namely that money is not and cannot be a commodity money. Even metal coins are token money, not different in substance from paper money. The Indian rupee, he writes, 'is a note printed on silver' .19 Similarly, in 1913, on reviewing a book on monetary theory, Keynes wrote in an even more explicit fashion: 'Not only has it been held that only instrinsic-value money is "sound", but an appeal to the history of currency has often been supposed to show that instrinsic-value money is the ancient and primitive ideal, from which only the wicked have fallen away ... Such a history is quite mythical.'20 Keynes was therefore fully persuaded of the fact that money is always in the nature of credit, and that money is created by banks the very moment they grant loans, without any previous collection of deposits being necessary or even possible: Practical bankers ... have drawn the conclusion that for the banking system as a whole the initiative lies with the depositors, and that the banks can lend no more than their depositors have previously entrusted to them. But economists cannot accept this as the commonsense which it pretends to be. 21 Similarly, in his writings concerning ancient monies, Keynes held the view that in the past, the circulation of money had never relied on the intrinsic value of coins, but rather on the confidence placed in the issuing agent. This is why he expresses the view that sanctuaries were the first banks and priests were the first bankers. 22 Having seen clearly that money is necessarily in the nature of credit, Keynes is able to give a lucid picture of the single phases of the circulation of money, starting with the initial creation of money and ending up with its final destruction. The First Phase: the Creation of Money The 'first link', as Keynes names it, takes place when banks grant credit to firms, who are thus enabled to meet the costs of production. The source of credit creation thus lies in a decision taken by a bank. When discussing how an increase in the value of output is financed, Keynes writes: 'In order that producers may be able, as well as willing, to produce at a higher cost of production ... they must be able to get command over an appropriate quantity of money ... Thus the first link in the causal sequence is the behavior of the banking system. '23 Again in 1937, on dealing with the same problem, he wrote: 'the finance required ... is mainly supplied by specialists, in pll'l'ticular by the banks'. 24 Keynes himself, a few lines later, put forward a slightly modified statement, by emphasizing the fact that in principle a firm wanting additional finance may make recourse to two possible sources, in that it may ask for a bank loan or it may issue securities on the financial market. 25 If Keynes insisted on the fact that the financial market, and not only banks, can be a source of finance, this was due to 146 Money as Purchasing Power and Stock of Wealth the fact that he did not want to reject the model of the General Theory, where he had placed the financial market at the centre of the picture, while pushing the banks in the background. The Second Phase: Financing Production Finance granted to firms is used for covering current costs of production, that is, total costs of producing both consumption goods and capital goods. On this point Keynes is absolutely clear. Even if sometimes, when discussing the general problems of liquidity, he speaks of financing investment, when his attention is centred on the problem his statements are unquestionable: 'Production of consumption goods requires the prior provision of funds just as much as does the production of capital goods. 26 As already mentioned, it is a common belief that while bank credit is required in order to carry out a new investment, consumption is automatically financed by consumers' expenditures. Similar ideas, widespread as they may be, lack any reasonable foundation. The Final Phase: Repaying Bank Debts Firms may get back their initial outlays both by selling commodities or by issuing securities on the financial market. This is once again a point which finds it hard to gain acceptance in current literature. Common statements tend to go in the opposite direction, in that they emphasize that only savings, in contrast to consumption, can be a source of final finance to firms. Once again, Keynes's statements are instead absolutely clear: consumption is just as effective in liquidating the short-term finance as saving is.27 Savings has no special efficacy, as compared with consumption, in releasing cash and restoring liquidity.28 Money which is spent on prior consumption flows into the same pool of available funds as money which is saved. 29 Keynes sees clearly that the role of the final phase of the monetary circuit is to enable firms to repay their debts to the banks. He consequently assigns to the financial market the role that it in fact has, which is not that of financing investment but of bringing back to firms the money savings of households, and thereby preventing them from remaining into idle balances. When the entrepreneur decides to invest, he has to be satisfied on two points; firstly, that he can obtain sufficient short-term finance during the period of producing investment; and secondly, that he can eventually fund his short-term obligations by a long-term issue on satisfactory conditions. 3o Augusto Graziani 147 It is unfortunate that in discussing problems of finance, Keynes moves to a context of increasing investment and aggregate demand, as though in stationary conditions the same problems were absent. This procedure is at the root of the widespread belief that a financial problem only emerges when the economy moves from a lower to a higher level of investment and income. 31 When dealing with a stationary economy Keynes almost eliminates the problem by defining the supply of finance as a 'revolving fund which can be used over and over again' .32 The trouble with the Keynesian revolving fund is that it suggests the idea of an automatic mechanism, supplying a constant amount of finance without any negotiation with the banks ever being needed. As Robertson rightly remarked, the fact of its amount being constant does not deprive the revolving fund of its being by nature a flow of liquidity supplied by the banks. 33 In stationary conditions, the amount of finance needed by firms is constant something which by no means implies that firms become financially independent from the banks. CONCLUDING REMARKS It emerges clearly from the preceding discussion that, while considering money fundamentally as a store of wealth, Keynes made an extensive use of the notion of money as purchasing power. It would hardly be incorrect to say that in his writings before the General Theory, the notion of money as an intermediary of exchanges prevails, while the model of the General Theory is clearly based on the notion of money as a stock of wealth. A more delicate problem of interpretation arises when we come to the writings following the General Theory, and in particular to the articles of 1937-9. It seems as though, after the General Theory, Keynes had in mind a sort of complete model, in which the money stock no longer comes from outside, but banks are included among the agents, and money is regularly created according to the financial needs of production. Implicitly, Keynes was considering money as an endogenous variable. 34 Still it would be equally hard to deny that Keynes never made clear statements concerning the limitations of the model used in his main book. In fact, he insisted on saying that any financial need of firms might be satisfied either by creation of money on the part of the banks, or through an increase in the velocity of circulation brought about by a placement of new securities on the financial market and by a consequent increase in the rate of interest (Keynes's insistence on there being two potential sources of liquidity was the source of a debate with Robertson).35 It should be recognized, however, that the attempts made by Keynes to make use of both notions of money reflect a contradiction which is intrinsic to the nature of money. If defined as purchasing power, money becomes a nonobservable-and therefore hardly tractable-variable; but the analysis of how 148 Money as Purchasing Power and Stock of Wealth purchasing power is created, distributed and utilized is fundamental for understanding the workirig of a market economy. In spite of having centred his main book on money as stock of wealth, Keynes was fully aware of the analytical relevance of money as purchasing power. One year after the General Theory, he made similar ideas explicit by writing: The control of finance is, indeed, a potent, though sometimes dangerous, method for regulating the rate of investment (though much more potent when used as curb than as a stimulus). Yet this is only another way of expressing the power of the banks through their control over the supply of money - i.e. of liquidity.36 Appendix: A Note on the Ambiguous Meaning of the Term Finance, the Role of Banks, Money and Saving The term finance is used in current economic literature in a number of different meanings which it would be advisable to keep rigorously separate. The role of finance is given sometimes to savings and at other times to money. Money and savings are also considered jointly, when households savings are thought to be at the origin of bank deposits and bank loans are believed to be financing investment. Since similar misunderstandings have also arisen recently, it seems appropriate to add a few considerations on this point. Three different uses of the word finance appear in the three following phrases: 'financing the economy', 'financing the exchanges' , 'financing investment'. When the phrase 'financing the economy' is used, what is meant is supplying the economy with an amount of money which satisfies the demand for cash balances coming from the market. In this case, money is defined as a stock, its amount at any moment of time being equal to the difference between the amounts of money previously created and destroyed. A second meaning appears in the phrase 'financing the exchanges', or 'financing aggregate demand.' Here the word financing is used in connection with the fact that, in a monetary economy, any demand for commodities has to be supported by a corresponding amount of money. Money is seen here as an intermediary of exchanges, the role of which is performed when it passes from the cash balances of an agent to the cash balances of the next. Once money has performed its role of financing an exchange, it may be kept in the idle balances of some agent, or, if the agent has a debt pending with a bank, it may as well be used for repaying the debt, and therefore destroyed. This means that money which is used for financing the exchanges need not be in the nature of an observable variable. Of course, the reverse is also true: the money stock in existence at any moment of time does not necessarily exhaust the amount of money which has been used to finance exchanges over the preceding period. Augusto Graziani 149 Finally, the same word finance is used in a third meaning in the phrase 'financing investment'. What is really at stake in this case is the formation of a supply of saving high enough to balance investment. The equilibrium between saving and investment belongs to the real sector and has no direct connection with the supply of money. Since this point is often the source of confusion, a few more considerations concerning the relationships between money and savings should be added. Let us start with a case in which savings and investment are equal. In real terms, the equality means that, if a certain portion of total output has been produced in the form of non-consumable goods, and if the community decides to abstain from consuming the same portion of output, the commodity market is in equilibrium. All this has to do only with the real sector, both on the supply and on the demand side. Let us now turn to money flows. Two cases can be envisaged: 1. 2. Money savings are totally spent on the financial market, where savers buy securities issued by firms. (Since households' savings are not entrusted to banks, this is not a very important case for the present analysis. We mention it in order to get a complete picture of the situation.) It is reasonable to assume that firms use the proceeds they get from the sale of commodities and from the placement of securities for repaying their debt to the banks. This means that an amount of money equal to the initial loan granted to firms gets destroyed, and the level of the money stock goes back to what it used to be before the production cycle was started. In this case, the banks are not at all concerned by money flows related to savings or to investment. They grant an initial loan to firms, and they finally get their money back. The conclusion may be drawn that if all savings are 'invested' (Le. spent on the financial market), which is the most perfect form of full equilibrium, banking activity has nothing to do with the savings-investment equilibrium. (Hayek would have said that in this case banks behave as they should since, while financing aggregate demand, they do not increase the money stock.). The opposite case is, from our viewpoint, the really relevant one. It occurs when savings, which are still assumed to be equal to investment, instead of being spent on the financial market are kept in liquid balances, Le. in bank deposits. In this case, banks do get hold of households savings and it can legitimately be asked what they do with them. If money savings are paid into bank deposits, firms are unable to repay their debt to the banks. There is obviously nothing banks can do in order to help firms repay their debts. All what the banks can do is to postpone repayment of old debts until a further date (or even indefinitely) and in the meantime grant new loans to firms as required, in order prevent a decline in the level of activity. For the situation of the firms to be fully balanced, the total amount of loans (and consequently the total money stock) has to increase by an amount equal to the ISO Money as Purchasing Power and Stock of Wealth increase in deposits, which we have assumed to be equal to current savings, as well as to investment. The correct interpretation of this case is a matter of debate. It goes without saying that Hayek would blame the banks for having increased the money stock; although he might make an exception and admit that in the special case we are now envisaging banks might be forgiven, since they have reacted correctly to an increase in the demand for idle balances on the part of the public. The current literature would definitely praise the banks for having performed their typical role of financing investment by means of savings. The crucial question still remains to be answered: who has really financed the firms? Is it the banks or is it the savers (or depositors)? Edwin Cannan, in a famous article of 1921 on the nature of bank deposits, explained that, in our case, the true origin of finance lies in the real savings supplied by depositors: 'It is the action of depositors to deposit which enables the bank to lend', the amount of deposits depending on the 'comparative desirability of direct individual investment and indirect investment through the medium of banks' .37 Banks, therefore in Cannan's view, act as intermediaries between savers and investors. In contrast to this, a more rigorous interpretation, such as the one supplied by Erich Schneider, stresses the fact that banks do not finance firms by means of deposits ('deposits are the liabilities of the banks and a bank cannot lend its Jiabilities!').38 Banks, while financing the economy and supplying as much money as is required by firms, do not act as intermediaries between savers and investors. Banks finance firms by means of newly created money, the proof lying in the fact that a new loan increases the total money stock while an act of saving does not. In the two cases so far envisaged, savings have been assumed to be equal to investment. Let us now consider a case in which investment is higher than savings and the economy is out of equilibrium. Again, this is not really an interesting case for the present discussion: there being no spontaneous saving flowing into bank deposits, no question really arises as to the nature of banking activity. As we all know, an excess of investment over saving might produce an increase in income until saving is again equal to investment; but, since the case we want to examine is not a case in which saving and investment are equal, we assume that the multiplier is not working. If real income is stable, an excess of investment implies an excess demand on the goods market and a consequent increase in prices. Here the banks come in and a decision has to be taken: I. As a first possibility, the banks may decide to accommodate the requests of the firms who are asking for higher loans. Provided wages rise less than prices and provided the propensity to save out of profits is higher, inflation comes to a halt as soon as profits have risen to a level high enough for savings to be again equal to investment (this is the case Hayek would con- Augusto Graziani 2. 151 sider as typical of bad bank behavior). If this happens, the banks may be said to have increased the nominal money stock, they have surely financed inflation, they have made possible an increase in profits, but they have certainly not done all this by means of a previous collection of deposits. Banks may supply more money; they cannot supply more saving. As a second possibility, banks may refuse to grant higher loans as price~ go up. As a consequence, firms will be forced to reduce the level of activity. Whether this would bring the economy back to equilibrium (as Hayek would have it), or whether it would be the starting point of an unemployment crisis (as Keynes would think) is something which can be debated. What is sure is that even in this case what has been cut short is the supply of money not the supply of saving. As Keynes said: 'The investment market can become congested through a shortage of cash. It can never become congested through a shortage of saving. '39 The implicit conclusion of the above discussion is that there is no conceivable case in which banks can act as intermediaries between savers and investors. Banks do control the supply of money but they have no direct control over the supply of saving. The commonplace according to which banks, by selling deposits, would be collecting savings and by granting loans would be financing investment is but a myth. Notes I. 2. 3. 4. 5. 6. 7. 8. Patinkin (1965), ch. V, and Supplementary Note D. Tsiang's is one of the rare cases in point. In his view, 'a vast bulk of the total supply of money does not lie in the portfolios of different economic units, but is found rather on the wing in various channels of circulation', Tsiang (1989), p. 20. However, Tsiang does not explain how any amount of money could be in existence at any moment of time without belonging to the cash balances of any single agent. The majority of authors consider credit under the viewpoint of management; credit is then an input, limited in amount, appearing in the production function; Vickers (1989). A general presentation of the so-called theory of the monetary circuit can be found in Graziani (1989). Schmitt (1984) is the strongest supporter of the idea that cash balances are wealth and not money, and that money proper only exists in the fleeting moment in which it passes from the cash balances of one agent to those of the next. Robertson writes: 'in those days, Mr. Keynes was so taken up with the fact that people sometimes acquire money in order to hold it, that he had apparently all but entirely forgotten the more familiar fact that they often acquire it in order to use it. In later articles Mr. Keynes has regained memory of this simple truth.', Robertson (1966), p. 161. Patinkin (1965), ch. V. Kaldor and Trevithick (1981) See Minsky (1980), pp. 507-9. Kregel suggests that the idea of financial innovation is already present in chapter 21 of the General Theory. 152 9. 10. II. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. Money as Purchasing Power and Stock o/Wealth A case in which the distinction between saving and liquidity is not always sufficiently clearly drawn is the treatment by Asimakopulos (1988). pp. 156-8. Hayek (1933). pp. 153 ff.• 190. The same conception of the role of banks to collect savings and finance investment is to be found in Tsiang (1989). pp. 110. 126. This point is dealt with more extensively in the Appendix. Graziani (1989). pp. 17-18. The problem is dealt with in Graziani (1989). pp. 17-18. where the point is made that firms can only pay interest to banks in kind. Other suggestions have been put forward in the literature. for instance by Parguez (1987) who sees the possibility of firms paying interest charges only in a dynamic context in which bank loans are constantly expanding. Keynes (1936). ch.15. I. p. 196. Say (1826). I. p. 183. A similar danger was clearly seen by Wicksell. In his model. Wicksell avoided the danger of the disappearance of money by imagining that firms cannot repay their debt before the end of the production period. In the meantime. the cash balances of savers are kept in bank deposits yielding a rate of interest which. in equilibrium. is equal to the rate of gross profit earned by entrepreneurs (Wicksell. 1898. ch. IX. section B). A similar point is raised by Hahn (1988). Keynes (1911). p. 393 (Collected Writings. vol. XI. p. 375) Keynes (1914). p. 418 (C. W.• vol. XI. p. 401). Bendixen was a follower of Knapp's state theory of money. Keynes (1930). I. ch. 12. Keynes (1913). ch. 3. Schumpeter liked the definition and. years later. also defined coins as 'notes printed on silver' (Schumpeter. 1931-2. repr. in Schumpeter. 1985. p.146). Keynes (I914b). p. 421 (C. W.• vol. XI. p. 406). In a similar mood. Hawtrey writes: 'Economists have at times been inclined to teach that this usage [of precious metals) ... approximates to a moral principle. as if the use of metallic currency were somehow essential to honest dealing' (Hawtrey. 1947; 1st edn. 1927). p. 1. Keynes (1930). I. ch. 2 (i). p. 25. Keynes (1920). pp. 223 ff.; the same remark in Keynes (1930). I. ch. 1 (vi). The interrelationships existing between money and religious institutions in the ancient world are analysed by Heinsohn and Steiger (1983; 1989). Keynes (1930). I. p. 182. Keynes (1937b). p. 666 (C. W.• vol. XIV. p. 219). Keynes (I937a). p. 246 (C. W.• vol. XIV. p. 208). Keynes (1939). p. 573 (C.w.. vol. XIV. p. 282). Similar statement in Keynes (1937b). p. 667 and (1939). p. 573 (c. W.• vol. XIV. pp. 221. 283). Keynes (1937b). pp. 667-8 (C. W.• vol. XIV. p. 221). On this point Keynes was possibly influenced by Hawtrey. whose description of the process of creation and circulation of money is even more explicit (Hawtrey. 1947. p. 13; 1st edn. 1927). Keynes (1938). p. 321 (C. W.• vol. XIV. p. 223). Keynes (1939). p. 573 (C. W.• vol. XIV. p. 283). Keynes (1937b). p. 664 (C. W.• vol. XIV. p. 217-18). The conviction that a problem of finance emerges only in a dynamic context is shared by a wide number of authors: Davidson (1978). p. 164 ff.; Minsky (1975). p. 75; Weintraub (1982). p. 365 ff.; Chick (1983). pp. 198-200. Keynes (1937a). p. 247 (C. W.• vol. XIV. p. 209). Robertson (1966). p. 164. Moore has repeatedly stressed the fact that money is in the nature of an endogenous variable at least in the first part of the General Theory (Moore. 1984; 1988). Augusto Graziani 35. 36. 37. 38. 39. 153 Robertson (1938a; 1938b), Keynes (1938). Keynes (I 937a), p. 248 (C. W., vol. XIV, p. 21O-11). Cannan (l921), pp. 34,35. Schneider (1962) ch. 2, p. 58. A point constantly emphasized by the supporters of the 'loans make deposits' view is that depositors never lose the free disposal of their deposits. Cannan of course, held, the opposite view: 'the fundamental fact that the banker is able to lend X, Yand Z more than his own capital because A, Band Care allowing him to temporarily use some of theirs', Cannan (1921), p. 32. A brilliant presentation of the 'loans make deposits' view in the French literature is due to Thomas (1977), p. 127 and (1989), p. 122. Keynes (l937b), p. 669 (C. W., vol. XIV, p. 222). References Asimakopulos, A. (1988), Investment, Employment, and Income Distribution (Cambridge: Polity Press). Cannan, E. (1921), 'The Meaning of Bank Deposits', Economica, 1(1), 28-36. Chick, V. (1983), Macroeconomics after Keynes (Oxford: Philip Allan). Davidson, P. (1978), 'Money and General Equilibrium', Economie Appliquee. Davidson, P. (1978), Money and the Real World, 1st edn, 1972 (New York: Macmillan). Graziani, A. (1989), The Theory of the Monetary Circuit, Thames Papers in Political Economy. Hahn, F. (1988), 'On Monetary Theory', Economic Journal, 98, December, 957-973 Hawtrey, R.G. (1947) The Gold Standard in Theory and Practice, 1st edn, 1927 (London: Longmans, Green). Hayek, F.A. (1933), Monetary Theory and the Trade Cycle (New York: Harcourt Brace). Heinsohn, G. and O. Steiger (1983), 'Private Property, Debts, and Interest, or: The Origins of Money and the Rise and Fall of Monetary Economies', Studi Economici, I. Heinsohn, G. and O. Steiger (1989) 'The Veil of Barter: The Solution of the Task of Obtaining Representations of an Economy in which Money is Essential', in Kregel (1989). Kaldor, N. (1982), The Scourge of Monetarism (Oxford: Oxford University Press). Kaldor, N. and 1. Trevithick (1981), 'A Keynesian Perspective on Money', Lloyds Bank Review. Keynes, I.M. (1911), 'Review of I. Fisher, "The Purchasing Power of Money .. ', Economic Journal, September, Collected Writings XI, 1983, pp. 375-81. Keynes, I.M. (1913), Indian Currency and Finance (London: Mamcillan) (CW, 1,1971). Keynes, I.M. (1914a), 'Review of: F. Bendixen, "Geld und Kapital"', Economic Journal, September (Collected Writings, XI, 1983, pp. 400-3). Keynes, I.M. (l914b), 'Review of A.M. Innes, "What is Money'" (New York, 1913), Economic Journal, September, Collected Writings XI, 1983, p. 404. Keynes, I.M. (1920), 'The Origins Of Money', Collected Writings, XXVIII (London: Macmillan, 1982) Keynes, I.M. (1930), A Treatise on Money, (London: Macmillan) (CW, 5, 6,1971). Keynes, I.M. (1936), The General Theory of Employment, Interest, and Money (London: Macmillan) (CW 7, 1973). Keynes, I.M. (1937a), 'Alternative Theories of the Rate of Interest', Economic Journal (CW, 14, 1973). Keynes, J.M. (l937b), 'The Ex-Ante Theory of the Rate of Interest', Economic Journal (CW 14,1973). Keynes, I.M. (1938), 'Comment' (A reply to D.H. Robertson 1938), Economic Journal (CW 14,1973). 154 Money as Purchasing Power and Stock of WeaLth Keynes, J.M. (1939), 'The Process of Capital Formation', Economic Journal (CW 14, 1973). Kregel, J.A. (ed.) (1989), Inflation and Income Distribution in Capitalist Crisis: Essays in Memory of S. Weintraub (London: Macmillan). Metzler, A.H. (1988), Keynes' Monetary Theory: A Different Interpretation (Cambridge: Cambridge University Press). Minsky, H.P. (1975), John Magnard Keynes (New York, Columbia University Press). Minsky, H.P. (1980) 'Capitalist Financial Processes and the Instability of Capitalism', Journal of Economic Issues 14(2), 505-524. Moore, B. (1984), 'Keynes and the Endogeneity of the Money Stock', Studi Economici, 22. Moore B. (1988), Horizontalists and Verticalists (Cambridge: Cambridge University Press). Parguez, A. (1987), 'Introduction' to La monnaie, les rentiers et La crise, Economie et Societes, S6rie Monnaie et production, n. 4 (Paris: Ismea). Patinkin, D. (1965), Money, Interest, and Prices, 1st edn, 1956 (New York, Harper & Row). Patinkin. D. (1976), Keynes' Monetary Thought: A Story of its Development (Duke University Press). Robertson, D.H. (1938a), 'Mr. Keynes and Finance', Economic Journal. June. Robertson, D.H. (I 938b), 'Reply' (to Keynes 1938), Economic Journal, September. Robertson, D.H. (1966), 'Mr. Keynes and the Rate of Interest', in Essays in Money and Interest (London: Collins). Say, J.B. (1826) Traite d'economie politique, 5th edn (Paris). Schmitt, B. (1984), Inflation. chOmage et malformations du capital (Paris: Economica). Schneider, E. (1962), Money, Income. and Employmelll (London: Allen & Unwin). Schumpeter, J.A. (1931-2), 'Weltkrise und Finanzpolitik', in Der Deutsche Volkswirt (repr. in J.A. Schumpeter, Aufsatze zur Wirtschaftspolitik, edited by W. Stolper and W. SeVl, Tiibingen, J.C.B. Mohr, 1985). Stiglitz, J.E. and A. Weiss (1988), 'Banks as Social Accountants and Screening Devices for the Allocation of Credit', National Bureau of Economic Research, Working Paper, no. 2710, September. Thomas, J.O. (1977), Inflation et nouvel ordre monetaire (Paris: Presses Universitaires de France). Thomas, J.O. (1989), Politique monetaire et autodestruction du capital (Paris: Economica) Tsiang, S.-C. (1989), Financial ConstrailllS and the Theory of Money (London: Academic Press). Vickers, D. (1989), 'The Money Capital Constraint and the Decisions of the Firm', in Kregel (1989). Weintraub, S. (1982), 'Money Demand Motives: A Reconsideration', Economie Appliquee, June. Wicksell, K. (1898), Geldzins und Giiterpreise (Jena: Fischer Verlag) (English translation, Interest and Prices, London: Macmillan, 1936). 6 Beyond Scarcity: A Reappraisal of the Theory of the Monetary Circuit* Alain Parguez 1 INTRODUCTION: THE SCARCITY PROBLEM Hayek's last words on Keynes's economics prove that for him Keynes could not be redeemed of his sin against the first commandment of the economist - to believe in the scarcity principle. 1 Believing in the principle is to know that economics as a science exists because productive resources, society's capital fund, are (and always will be) scarce relative to society's desire or lust for consumption. Rational behaviour, market, competition, price mechanism are just the set of institutions allocating this scarce capital in the most efficient way. The sine qua non of an increase in the accumulation of such a scarce capital is saving. Saving is that share of its real income that society does not want to spend for its current consumption: as soon as current saving is greater than depreciation, the available capital fund is increased. Hayekian saving is pure thriftiness assessing society's ability to postpone over time the satisfaction of its desires. No economist since the days of Adam Smith has therefore dared to reject the doctrine of the sanctity of parsimony in being the ultimate constraint on investment. 2 According to Hayek's interpretation, Keynes is the very first economist having dared to present a whole world outlook approach contradicting the First Postulate of Classical Economics, the Scarcity Principle.3 In Keynes's brave new world, thriftiness is no longer the sine qua non of accumulation. Rather, it is the major obstacle to investment through its negative impact on society's real income. Capital is still scarce but scarcity is a voluntary, desired scarcity which is forced upon society as a whole by those economic agents being so thrifty that they are afraid of spending to increase the rate of capital formation. Policy-makers should stop praising and rewarding thriftiness; they have to fight entrepreneurs and 'holders of liquid resources' having a high propensity to thriftiness. Hayek understood that. Central to the scarcity debate, was a debate on the nature of money in a market economy. The First Postulate implies the Classical theory of money - the three major requirements of which were emphasized by Hayek as crucial for understanding the revolution Keynes had initiated against orthodoxy: 155 156 1. 2. 3. Beyond Scarcity The prerequisite for a smooth coordination of plans through the market is the neutrality of money which holds as long as there is no creation of money through credit contracts between banks and enterprises. Assuming such an economy without pure credit, the quantity of money perfectly fits the requirements of the market. Money is a pure exchange money that must be governed by the Scarcity Principle.4 A Wicksellian credit system always evolves from the market economy because society strives to escape from the yoke of scarcity. Money cannot spontaneously be neutral. Its quantity cannot match the requirements of a smooth coordination, since it is created by banks at the request of enterprises (and government) to finance their capital expenditures. The Scarcity Principle is superseded by credit money, and investment is no longer constrained by thriftiness! Those 'false' investment expenditures initiate cumulative inflationary booms whose ultimate results are deep depressions associated with involuntary unemployment and with a fall in the stock of available capital. 5 Non-neutral credit money is the scourge of the market. Money is to be neutralized by imposing a set of rules and institutions that will force the Scarcity Principle upon society. This first postulate must thus permeate all the decisions of policy-makers. 6 Hayek rightly emphasizes Keynes's paramount loathing for the political economy of austerity implied by the third proposition. He is still right in the explanation of Keynes's critique of austerity: Keynes does not support the second proposition of the Classical theory of money because he cannot endorse the first proposition! Money can never be a neutral exchange money in Keynes's model of the market economy. This essential non-neutrality of money remains Hayek's ultimate proof of Keynes's supposed misunderstanding of the fundamental principles of economics. 7 Contemporary conventional wisdom is addressing economic policy problems in the spirit of Hayek's original judgement passed on the economics of Keynes. Keynesian policies of full employment are to be disregarded because they can trigger an unsustainable inflation. The Scarcity Principle is to be forced upon reluctant societies that must be taught that scarcity is their fate in saecula saeculorum. This economics of scarcity has three major aspects, all of them dealing with thriftiness as the prerequisite for accumulation. 1. The lack of competitive capital in the context of an integrated world economy is the result of a lack of saving. A country's relative competitiveness depends on its relative rate of domestic saving, saving reflecting the nation's inclination for thriftiness. Thriftiness is again the sole source of finance for a non-inflationary market-led accumulation. 8 Alain Parguez 2. 3. 157 Enterprises demand for capital goods is to be enhanced by a decline in the share of labour in the distribution of aggregate income. A falling share of labour is the source of an increase in enterprises' or business saving that would lead to more desired capital expenditures. This postulated theory of profit requires an inverse relation between profits and current expenditures out of wages and salaries, which is a logical consequence of the doctrine of thriftiness. 9 Government deficits are to be eliminated because they divert the so-called national saving from productive expenditures. Policy-makers have to raise national saving through greater thriftiness while endeavouring to cancel all deficits. Keynes and Kalecki had already shown the fallacy of this policy of generalized thriftiness. The fact that their warnings have been forgotten is the best proof of the counter-revolutionary content of the contemporary mainstream vision. It could not have been so successful, had it not been deeply rooted in mainstream neo-Classical models. All of them that were (are and will be) constructed, fail to integrate money within the framework of a comprehensive theory of accumulation. Money is still either a simple neutral quasi-barter money the existence of which is puzzling, or it is the major disequilibrium factor initiating rigidities in a 'neo-Keynesian' eqUilibrium framework. As long as mainstream economics still deals with accumulation, thriftiness rules supreme as the sole source of finance consistent with non-inflationary growth. The circulation approach to monetary theory directly addresses the legitimacy of the austerity view of a thriftiness-led world economy. Questioning the role of thriftiness requires a reappraisal of its theoretical foundation, the Hayekian theory of money and credit. With respect to the circulation approach, what must be demonstrated is that both Hayek and today's neo-Classical advocates of austerity are incapable of integrating money within their framework because they miss the major characteristics of a fully developed capitalist economy. 2 THE CORE OF THE CIRCULAnON APPROACH TO MONETARY THEORY The Principle of Essentiality of Money Central to the Circulation Approach is the Essentiality of Money. Money is the sine qua non of a capitalist economy, its existence condition. Four major characteristics of money explain its essentiality: 1. Money exists to match the requirements of production and therefore of accumulation. Through their credit contracts with banks, enterprises (and government) can raise the liquid resources they need to carry out their plans. 158 2. 3. 4. Beyond Scarcity In every state of the economy, the amount of liquid resources initiated by credit contracts is thus identical to the amount of newly created money. Money is essential according to this characteristic as a pure credit money. Production cum accumulation plans are carried out to match the requirements of the credit contracts. Enterprises carry out their current production plans which are based on future receipts in money out of which they can repay their initial short-term debt. They also carry out their investment plans because they bet on the future flow of profits out of which they could refund their initial investment. Money is the sole real unit of account within the economy ruled by (1) and (2). All values are to be expressed in money terms. Those values are assigned to all the elements of the production sphere - capital goods, consumption goods and labour. For each element of this production set, its value accounts for its meaning, from the point of view of the system as a whole. Capital goods are to be assigned values in money reflecting the market's expectation of the future flow of the income they must yield to their owners, firms having invested in their acquisition liquid resources initiated by credit contracts. Consumption goods enter the market with a predetermined set of values reflecting the profit constraint. This constraint embodies the amount of households' expenditures which is required by firms just to capture the profits needed to sustain their own plans. In such an essential money economy, the value of labour must also be expressed in money terms. In each state, the value is bestowed on labour through the payment to labour of an aggregate money-wage bill that is to be spent in the acquisition of the supply of consumption goods to bestow on firms the required profits. 10 Central to the production sets are firms whose values are expressed in money terms. The capital value or asset prices of a firm reflects the market's ability and desire to bet on this firm's future flow of net revenue or profit out of which it could meet its financial commitments. A firm's required flow of profit is that flow of profit which, according to the firm's managers bets on the market view, should adjust the firm's capital value to some required level. Essentiality of Money and the Monetary Circuit (l) and (2) explain why each state of the pure capitalist economy is characterized by a Monetary Circuit. Such a Monetary Circuit perfectly fits the major characteristics of the Monetary Production economy Keynes had in mind in the Treatise, in the first draft of the General Theory, and in the General Theory itself, albeit it lacked a model of this Monetary Circuit. II This Monetary Circuit is a perfectly ordered sequence of stages explaining the determination of the present state of the economy by firms' bets on the future responding to a data set embodying both the past results of the economy and the Alain Parguez 159 constraints firms have to integrate within their bets. The present set of bets on the future will also deeply affect the future states by generating their available equipment. The Monetary Circuit is thus an open system relative to time. The First Stage It is indeed shaped by firms' bets on the future whose stake is the profit flow that should adjust their capital values to their required levels. Initial value of a firm Vx according to (4) reflects the judgement banks have just passed on its effective ability to earn a flow of net income. Past performances relative to the firm's commitments are data that could influence banks' new set of bets on the future of firm x. For the same firm there is a level of the capital value it should reach because this would convince the banks of the perfect ability of the firm's management to meet their commitments whatever the increase in their current demand for a liquid fund. Banks would thus consider this required capital value V~ as the required collateral for their loans; firms are thus endeavouring to reach their target capital values that should protect them from any credit rationing or increase in the cost of credit. For any firm x, Vx and V~ are thus crucial data embodying the financial constraint stemming from ·banks impersonating the market for real assets. Responding to this set of data, firms have to carryon two kinds of bets on the future,I2 bets on the short-run and bets on the long-run. They have to bet on the short-run profits that should be both the outcome of their current sales and the proof banks need to support the rise in the effective capital value. Betting on those profit flows requires a rule of the game embodying the financial constraint. Such a rule is reflected by the required rate of return firms have to charge on their production costs, just to be sure that banks will thus be convinced that they do endeavour to reach the target capital values. In the most abstract model of the capitalist economy, production costs are reflecting the wage bill and the required rate of return is thus charged on expenditures. The rt being part of the financial constraint are also part of the initial data set, like technology embodied into equipment inherited from the past and the money-wage which is also a component of this data set, whatever the process determining its level. (For any firm x) technology is reflected in the average labour productivity, a x; all firms can now fully determine their current production plans. For any firm x, the game requires it to bet on the maximum amount of sales it can earn, dx, which must allow a profit whose ratio to the wage bill must be the required rate of return. The first stage of the circuit is thus illustrated for all x by the following set of equations: Yi=dx Yi= (1 + ri) Wx Yi= piQx (1) (2) (3) 160 Beyond Scarcity Qx=axNx (4) Wx = wNx (5) P;= ~ (l + r!J x (6) For a given data set including w, r;, ax, firm x knows what the money value of its real supply, must be to match the demand constraint, the stake of its bets (1); it also knows what the wage bill must be to allow for the required amount of profits (2); it therefore knows at once its required quantity if labour Nx to carry out the required amount of real supply Qx (4,5). Another way of expressing the same result is to emphasize that for any firm both the real supply and the required price are simultaneously determined in this first stage of the Monetary Circuit. As shown by (6), there is only one money price which is consistent with the data set including the financial constraint. The cost of labour w/a being given, prices are entirely determined by the required rates of return embodying the financial constraint. No bank would endorse the bets of x if x did not strive to grant the banks in the future a higher level of capital value that should reflect an increase in the banks' collateral. Any firm thus has to bet on a long-run increase in its ability to catch a profit flow. This bet on the far future is embodied into the firm's demand for newly produced equipment goods that must be integrated into its available future equipment. The x firm wants to acquire Ix of the available supply of equipment goods because it bets that by doing so it will increase the value of its equipment and thus its own value by an amount 11 Vx equal to Ix that should grant the banks the required increase in their future collateral. By investing Ix of liquid resources in this acquisition, the firm bets on the ability of banks to bet on the increase in the future flow of profits given the constraint on the production process, the required rate of return on labour expenditures. All firms are required to have an investment account, both consumption goods producers and equipment producers. Investment is thus crucial in the current demand producers are betting on to determine their own supply. Producers of equipment goods are betting directly on investment demand, while consumption goods producers bet on consumption out of income paid by producers of equipment goods. Investing firms shape their demand for investment by reacting to the same data set, explaining their current production plans. In this essential money economy, they are not constrained by inherited savings transmitted from the past. The major constraint is embodied into the generalized betting system: firms must be bold enough to bet on increased profits as an outcome of their investment and banks must be ready to legitimate those bets. r;, P; The Second Stage Banks now play the leading role by providing firms with the quantity of money or liquid circulating capital they need to carry out their bets. Alain Parguez 161 Banks form a credit network whose individual members can contract with firms on behalf of the network as a whole. Any firm can raise the liquid fund it needs by asking for credit from any member bank. As soon as the bank bestows its credit on the whole set of the firm's bets, it is betting on the success of those bets. Endorsing those bets, the bank - on behalf of the whole network - initiates a credit contract generating commitments or debts for both the bank and the firm: (a) (b) The bank is fully committed to generating the required liquid fund the firm has to spend to carry out its bets. This fund is thus created in the form of an amount of deposits reflecting an increase in the bank's liabilities. Those deposits can be freely transferred to households as wages or to producers of equipment goods as receipts. Newly created deposits thus reflect the amount of money created by a specific set of credit contracts. The firm is fully committed to generating an income out of which it could replenish its balance and thus pay back the bank's initial credit. Proof of a genuine commitment is given by the strict observance of bank-imposed rules of the game sustaining capital values, including the payment of interest on credit and the embodied profit norm reflected by the required rate of return on labour expenditures. This last commitment is the firm's debt to the bank, which becomes payable as soon as the firm acquires an income resulting from the success of its initial bets. This principle explains why individual credit contracts cannot always insist that current investment must be paid back or refunded entirely by current net income generated by current sales. The contract arranges that a share of new investment should be refunded by current profit, the other share being refunded later by a share of those future increased profits, whose expectation is the support of the rise in capital value which is the stake of current investment. The Third Stage Repayment of their debts is the outcome of firms' effective ability to generate an income by carrying out their bets, which is the third stage of the Monetary Circuit. The x firms have to spend a share of their newly created liquid fund in order to acquire the required amount of labour. Aggregate payments of wages create households' own income, reflecting their purchasing power on the current available supply of consumption goods. Another share of the liquid fund must be invested in the acquisition of the available supply of equipment goods generated by those producers who have bet on this investment. Receipts or gross income of the set of consumption goods producers are generated by households' expenditures equal to the excess of aggregate wage bill Wx - their own wage bill W2 plus the wage bill paid by equipment goods producers Wlx - over households' saving reflecting their propensity to thriftiness. Equipment goods producers earn as gross income the amount of aggregate investment I, their own acquisition of equipment 162 Beyond Scarcity goods I, plus acquisition of equipment goods by consumption goods producers, 12 , The Fourth Stage Each set of firms can now pay back its debt to banks by an amount equal to its gross income. The ratio of its current net income or profit to its investment expenditures reflects that share of investment which is paid back or refunded by the current profit flow, the effective rate of internal finance. This fourth stage of the Monetary Circuit is thus crucial for the determination of the new data set for the next state of the economy. Some components of the data set can change independently of banks' appraisals of firms' performances: the wage rate can be lowered by an increase in unemployment resulting from initial production plans and the rate of labour productivity can rise as new investment begins to be embodied into available equipment, which could reflect a change in technology. Banks play the leading role in changing their judgement over time on the effective ability of firms to generate enough net income to foster their expected rise in capital values, which was the stake of past and current bets of the future. Passing this judgement banks compare effective profits with the levels that are required - given their own ability to bet on the future - to convince them on confirming their initial bets of a firm's future. Either banks maintain their initial vision and the outcome of current investment is an equal increase in the set of capital values, or, in the case of a profit failure, banks can impose a rise in the rates of return on labour as the prerequisite for accepting the rise in capital values. The profit failure could be so bad that banks would now bet on a diminishing ability of firms to earn a profit and the whole set of capital values will fall. The Adjustment Process over Time The circulation approach is not consistent with the conventional distinction between static and dynamic analysis. An essential money economy is so constrained by the process of shaping the future by bets on what it should be, the present state being the mere outcome of those bets, that the economy moves through time displaying four major characteristics. (I) Within a circuit process, firms cannot adjust their effective results to their profit targets. For each set of firms, the proof of their ability to earn a net revenue is crucial. Significant results are illustrated by their effective profit flow and the resulting rate of return on labour expenditures. Assuming a zero household's propensity to thriftiness, 1T2, 1T" 1T, r2' r, (being respectively profits earned by consumption goods producers, equipment goods producers and aggregate profits, and rates of Alain Parguez 163 return for the corresponding set of firms), the outcome of the ongoing circuit process is: 1T1 = WI = II + 12 - 1T = II + 12 = 1 1T2 (7) W2 (8) (9) Assuming for the sake of the argument that if (9) holds, the distribution of aggregate profit matches the distribution of investment expenditures (10) 12 = 1T2 and thus (II) From (7) and (11) we derive the set of effective rates of return, the ratio of profit to labour expenditures for each set of firms: WI r2= - W2 II rl = - WI wNI = --=n (12) II = -=m 12 (13) wN2 Effective rates of return reflect the real structure (m, 11) evolving from initial bets: investment goods producers earn a rate of return equal to the ratio of their own investment to the investment of consumption goods producers; consumption goods producers enjoy a rate of return equal to the ratio of employment in equipment goods firms, N I , relative to their own employment of labour, N2 • There is no reason why (7)-(12) should always automatically meet the required rates. Let us assume that consumption goods firms have bet on too high a level of NI to reach their profit norm, their required rate of return ri in this model embodying their whole profit requirements. Their overoptimistic bet on equipment goods firms own bets has led them to employ too high a quantity of labour. The resulting distribution of labour imposes a rate of return which is inferior to the required rate. Consumption goods producers are thus in a state of profit failure. Equipment goods producers could be in similar if they bet on an overpessimistic vision of consumption goods producers, leading then to invest much less than they have effectively spent. Equipment goods firms would thus spend less than their own investment than they should in order to meet their profit target. The resulting distribution in investment would allow them a rate of return inferior to its required level. Nothing can compensate firms for their lack of profit. No price adjustment can help them since effective prices are uniquely and entirely determined by Beyond Scarcity 164 (7)-(13). Effective prices PI and P2 for equipment goods and consumption goods firms correspond to the amount of their gross income that allows them their effective rates of return; gross income reflecting effective sales which are equal to the effective money value of available real supply PI and P 2 levels are such that PI QI = Y I = 1=(1 + rl) WI (14) P2 Q2 = Y2 =C =(1 + r2) W2 (15) QI =a,N I (16) with Q2 = a2 N2 and so =(1 + rl) ~ ; with r. =m a. P = (1 + r2) ~ ; with r2 =n a2 PI (17) Prices are the mere reflection of the structure of initial bets for a given level of the wage rate and labour productivity. In a state of profit failure, effective prices evolving from the circuit process are inferior to their required levels. The circuit process proves that the sine qua non for the success of a strategy of imposing required prices is the ability for each set of firms to impose its required rate of return, through the control of the real structure of the economy (m, n). (II) Assuming an unchanged data set through time and an unchanged ability to bet on the future for both firms and banks, the economy is moving towards a state of profit consistency. The assumption holds as long as banks do not react negatively to profit failures by imposing a fall in the capital values of investing firms. Assuming an unchanged set of required rates of return, each set of firms will have to learn from its profit failures and its own bets will be increasingly consistent with those of the other set. The economy will converge through time in a state of profit consistency, corresponding to the circuit process whose outcome is the real structure (m, n) which matches the profit constraint such that r2 = n = ri r. = m = r( (18) In such a state, effective sales adjust effective rates perfectly to their required levels included in initial bets: Alain Parguez 165 (19) Convergence could be described by a set of dynamic equations but in this chapter the dynamic process over time will be addressed in an informal way. Emphasis is to be put on the economic constraint underlying the adjustment process: firms never stop in their fight against the profit failure because sooner or laler these could lead to a fall in their own capital values; the sole avenue towards profit consistency is consistency of bets on the future. In the simple model we are now discussing, this informal dynamic displays the crucial part consumption goods producers have to play. Assuming that they maintain over time an unchanged amount of inveslment in line with their bets on longrun future profits the banks are endorsing, equipment goods producers will soon learn what their own investment must be to match their profit constraint. They will be taught the game they have to play, betting on II = rj 12 (20) YI = 12 (I + rj) (21) WI =/2 (22) The game (20)-(21) leads them to adjust their own wage bill to what consumption goods producers spend as investment. Such a game perfectly matches the profit constraint of equipment goods producers. From (21) and (22) they derive their required amount of real supply whose distribution between themselves (for a share v) and consumption goods producers (1 - v) matches the profit norm; for this unique level of QI: VQIP I II V 12 1- v ----= -=m=r'j'=(1 - v) QIP I (23) Consumption goods producers will learn shortly what game to play equipment goods producers. Being taught the (21) rule of this game, they will know that their own game requires a unique amount of wage bill and thus of employment and real supply, the wage rate and labour productivity being both included in the data set: (24) 166 Beyond Scarcity (25) Their game requires that the higher is It the lower will be employment and real supply for a given amount of employment in equipment goods firms. In this model (21) shows that for consumption goods firms the secret of their success is a perfect consistency of their short-run bets (whose stake is their required employment) with their long-run bets (whose outcome is their unchanged investment through time). This game consistency requires (26) So drastic is the constraint of profit consistency that we can address a monetary production economy as if it were in a perfect state of profit consistency as described in (20) - (26). As long as we deal with an unchanged data set over time and with banks which do not depreciate capital values or impose new constraints just to respond to profit failures. (III) Any change in the data set which has an impact on the terms of the profit constraint evicts the economy from the past state of profit consistency and triggers a dynamic process of adjustment, evolving over time through a sequence of circuit processes. This sequence is to be addressed in the same informal way, emphasis lying on firms striving to reach a new state of profit consistency. Dealing with our simple economy and abstracting from changes in the money wage rate, changes in the required rates of return are crucial in determining changes in firms' capacity to spend in line with profit constraint. In both cases, past levels of employment, real supply and price in both sets of firms resulting from equations (19)-(26) would not correspond to a state of profit consistency. Both length of time and magnitude of adjustment depend on the banks' ability and desire to foster firms' capital values in the course of this adjustment of bets to the new profit targets. Here lies the prerequisite for firms' ability to maintain a degree of investment resulting from their bets on future profits, corresponding to the new data set. An unchanged level of investment over time would support a fast and smooth adjustment. (IV) Both aggregate demand and its distribution between investment and consumption are thus adjusted over time to the requirements of the profit constraint. This last characteristic is crucial for the understanding and putting in perspective the effective part firms are playing in the essential money economy: (a) Interpreting in our simple model rates of return as the rates of mark-up dealt with by Kalecki and Post-Keynesian theory, firms enjoy the power of imposing their mark-up. Alain Parguez (b) (c) (d) 3 167 This power fully integrates the absolute constraint demand imposed on firms' income. It exists because firms are committed to generating the real structure (m, n) of the economy allowing them to earn their required rates of return, as shown by equation (18). By controlling this structure, firms generate the amount of sales which adjust effective prices to pre-determined. required or normal prices. As already shown by Graziani (1990). in dealing with the circuit process which is the outcome of adjustment, we always find a price system, as in equation (19), which is both the exact set of prices matching the profit norm and the set of prices perfectly 'clearing the market'. Submitting to this profit constraint. firms cannot be neo-Classical and thus ruled by the marginal productivity principle. They must rather be the megafirms, mahager-ruled, with which Post-Keynesian economics is concerned. Their fixing-price power does not reflect an absence of competition a very high Kalecki-like degree of monopoly. It is the result of the absolute necessity of endeavouring to foster firms' own values over time that all managers bear in mind as their paramount requirement. This stems from the existence of a banking network whose sole collateral is the capital values of firms. It is therefore the mark of a true capitalist market economy where credit money is the sole source of liquid funds. There is therefore no reason why mark-up rates should remain unchanged over time. As the profit constraint becomes more stringent for managers, it triggers a rise in mark-up rates. Even assuming an unchanged view by mane agers relative to the future. we cannot dismiss the change in the banks' view of what firms' rates of return should be to support some level of investment. FIRMS' THRIFTINESS As mentioned above in the introduction, it is easy to derive from the Classical doctrine of saving the enhancement of firms' propensity to thriftiness. Dealing with our simple economy, the Classical doctrine would address profits as the available saving fund being the sole 'normal' source of capital for investment expenditures. The prerequisite for any increase in investment would thus be the rise in the amount of profits firms can exact from the circuit. If this view were true, firms would always be capable of generating more profits without increasing their expenditures; the sine qua non for a rise in investment would be the imposed fall in the share of labour in the distribution of aggregate income; there would be a rigid inverse relation between the wage bill and aggregate profits for any given level of aggregate income. Here lies the core of the Classical view: the less firms spend, the more profits they could earn and invest for the future! This conclusion is the apology for firms' thriftiness. but it is not supported by the characteristics of the capitalist market economy that have been derived from the model of the monetary circuit. Beyond Scarcity 168 The Kalecki Principle The monetary circuit leads to an explanation of profits whose principle is the following: in the circuit process, firms gain as profits what they have already spent as investment to acquire newly produced equipment goods. In our simple economy, the principle is always true for firms as a whole, as shown by equation (9): aggregate profits are created in the third stage of the monetary circuit when each set of firms has completed the sale of its current real supply. Their amount is always perfectly equal to the effective value of the real supply of equipment goods, as shown by equations (7) and (8); aggregate profits are thus perfectly equal to aggregate investment expenditures. By assumption, in the simple model if the principle is true for firms as whole, the principle is also true for each set of firms - consumption goods firms and equipment goods firms as shown by equations (10) and (11). The principle can be addressed as the Kalecki Principle because it is the core of Kalecki's own macrotheory of profits. 13 It has three major direct implications. The Crucial Role of Consumption Emphasizing investment as the sine qua non for the generation of profits, the principle does not contradict the leading role of consumption in the capitalist monetary economy. Investment being initiated by bets on the future course of demand, aggregate investment is the outcome of the ways of betting on future consumption, either directly or indirectly. Investment from consumption goods firms reflects their direct bets on consumption in the future; investment from equipment goods firms reflects their direct bets on future investment from consumption goods firms and thus their indirect bets on consumption in the distant future. The principle is thus cast in stone: Firms recoup as profits the amount of their bets placed on future consumption. The higher is the profit of equipment goods firms relative to the profit of consumption goods firms, the greater the number of firms as a whole which are indirectly speculating on consumption. In Keynes's terms in chapter 16 of the General Theory, this distribution factor of aggregate profits, d, is the reflection of the 'degree of roundaboutness' of the production structure. The greater is d, the more firms as a whole endeavour to postpone until the distant future the sanction of effective consumption as the ultimate endorsement of investment. In our model, d is equal to the rate of retum in equipment goods since d= 1T. =I. =m=r. "'1T'272 (27) Alain Parguez 169 Since adjustment imposes effective rates of return equal to their required levels, we can address any state of profit consistency as making true (28) Thus appears the ultimate formulation of the Kalecki Principle in the simple model: From a given amount of aggregate investment expenditures, consumption goods firms gain a profit whose amount is inversely related to the required rate of return in equipment goods firms. The Solution of the Riddle to the Generation of Profits Profits should not be the savings fund tapped to pay for investment. Since they are generated by investment expenditures, profits cannot exist before investment as an available liquid fund to be substituted for credit-money. The principle thus emphasizes that credit-money creation is the sole source of initial finance for investment in the capitalist monetary economy. The principle formulates a major aspect of the monetary circuit: Firms as a whole receive as profits the amount of money all of them have individually borrowed from banks as credit to carry out their bets on the futur~ embodied in the acquisition of equipment goods. In our model, each set of firms as a whole receives as profits the amount of money created by banks to finance the acquisition of equipment by all their members. The finance principle disposes of any riddle relative to the generation of profits that would b(' the outcome of the postulate that money creation initiated by banks' credit is equal to the aggregate wage bill. Such a postulate would mean that investment expenditures are financed by profits generated by the circulation of this amount of wage money. Dealing with our model, the postulate should explain how each set of firms can earn the amount of profit consistent with the required rate of return. Assuming for a while that the postulate is true, money creation would be equal to the wage bill paid by the two sets of firms because they bet on their required rates of return in money terms. Consumption goods firms would still earn a profit equal to the wage biII paid by equipment goods firms and they would now spend this profit to acquire equipment goods. Equipment goods producers have spent their wage bill because they bet on an aggregate amount of sales whose excess relative to the wage bill would be the amount of profit corresponding to the required rate of return. In the terms of our model, YT, Y. being respectively the required value of the supply of equipment goods and their effective value: Beyond Scarcity 170 rr = (1 + rj) WI (29) YI = 'TT2 =12 (30) and since YI=W I 'TTl =0 rl =0 (31) Equipment goods firms earn a gross income which is equal to their sales to consumption goods firms. They are therefore just able to pay back their wage bill; they cannot earn an effective profit in money terms; their effective rate of return is thus zero. They are left with an unsold real supply nobody could buy because there would be no more money in circulation that could finance this acquisition. The unsold supply of equipment goods would thus be deprived of value and being without value could not be the support of a positive rate of return. The system of equations in (31) can be generalized to any kind of production structure. It shows why the wage bill postulate is inconsistent with the essential money economy: In an essential money economy, all firms (and therefore all sets of firms) have to earn a positive rate of return to foster their value over time. In an economy ruled by the wage bill postulate there is always a set of firms which cannot earn a profit and which cannot meet the profit constraint imposed on the value of its members. The wage bill postulate is therefore an aspect of the first proposition of the Classical theory. Rejecting the first proposition dealing with profits as a pre-existing capital fund, the theory of the monetary circuit leads to the rejection of the wage bill postulate. Profits and Saving: the Equality of Saving to Investment Explaining the shortage of investment relative to the requirements of full employment as some fatal outcome of the shortage of profits, the second proposition of the Classical theory of profits expresses the Classical doctrine of saving when there is no saving in households. Aggregate profits are the sole saving fund to which investment is adjusted. This vision of a lack of saving is the missing link connecting saving with investment. Saving exists in the circuit process as both real saving and pure financial saving. Real saving accounts for the increase in society's real wealth generated by the circuit process. This wealth is the value of the stock of equipment providing real output in the future that should attract demand and thus gen- Alain Parguez 171 erate revenue. It is identical to the aggregate capital value of the whole set of firms owning and using this equipment. Real saving is thus reflecting the increase in this aggregate value of firms which is the value of that share of existing output which has been embodied in equipment for the future. Real saving is thus identical to aggregate investment: a lack of real saving is a true lack of investment, as already shown by Wray (1988; 1989). Financial saving accounts for the excess of aggregate income over the amount of consumption expenditures; it is that share of aggregate income that must be spent by providing permanent or final funding for investment initiated by banks' credit. Aggregate income accounts for the sum of wages and profits. In our model, wages and consumption being equal, financial saving is identical to aggregate profits. By virtue of the Kalecki Principle, financial saving is thus equal to aggregate investment expenditures. The sheer impossibility of a shortage of saving in the funding of investment is obvious: firms as a whole generate an amount of financial saving exactly matching the debts initiated by investment expenditures whose outcome is that amount of saving. From the theory of the monetary circuit, stems the first version of the Keynesian equality connecting saving to investment, S*, S, V*, V*, Y being respectively real saving, financial saving, aggregate capital value of firms included in the data set, its required target level, aggregate income reflecting the aggregate value of output. The circuit process is characterized in our model by the following system emphasizing the two aspects of saving. Real saving Financial saving S==Y-C==Y-W S*== Ll V == V* - V V* - V == I (32) (33) Y == W +1T == l( + Y2 (35) (36) S*== I (34) S==1T (37) (38) =I (39) S = I = S* (34) and (37) are mere identities expressing how S* and S are to be accounted for. Since (38) is the Kalecki Principle, 39 is not an identity. Albeit being true in any circuit process, it reflects the outcome of the whole sequence of stages which is the circuit process itself. S is created as the mere financial counterpart of I and thus as the mirror image of S*. Equality 39 cannot be connected with any kind of equilibrium conditions that would impose the adjustment of S to I through a sequence of circuit process. 1T Profits and Distribution: The Principle of Independence of Aggregate Profits from the Share of Profits The third and most crucial statement of the classical theory of profits is that being able to control the share of profits, firms as a whole can control the level of aggre- Beyond Scarcity 172 gate profits. This statement is still the hardcore of contemporary political economy of austerity whose principle is to reinforce firms's power of determining the share of profits. The second part of the proposition is false: aggregate profits being predetermined by investment expenditures, their level cannot be altered by a change in the share of profits. The first part of the proposition is perfectly true: firms can impose the share of profits they want. To prove the existence of such a power in the context of the monetary capitalist economy, let h be the share of profits and r the average rate of return: 1T r=W h=~= Y (40) rW W(l + r) = r 1 +r (41) (42) By adjusting their effective rates of return to their required levels, firms are adjusting the average rate of return to its required level in line with profit constraint: 1j = rt (43) r2 - - r* _w. _1 2 W2 (44) "'I = r2 W2 r = ~(l+1j)=r* 1+r2 (45) (46) Putting together the Kalecki Principle of (38), the expression of h in terms of r in (41) and (46) we put forward the role firms can play in distribution. Let us assume an unchanged required rate in equipment goods firms. Let us now assume that to maintain their investment, consumption goods firms impose a rise in their rate of return. They succeed in imposing a rise in the average rate of return which determines a rise in the share of profits. Investment being unchanged for consumption goods firms and thus for equipment goods firms, Alain Parguez 173 aggregate investment is constant relative to the past, and aggregate profits are thus constant relative to the past. The outcome of firms' games is a fall in the level of aggregate income reflecting the fall in the wage bill paid by consumption goods firms. Aggregate income has to fall until its level has reached its required level y* determined entirely by aggregate investment and the 'distribution multiplier' equal to the inverse of the share of profit. The characteristic of the circuit process meeting the profit constraint is thus Y= y* 1 =-1T h 1T=1 1 Y=-I h (47) (48) (49) For a given I, Y is determined by h, the distribution factor. Equation (48) emphasizes the principle of independence of the level of profits from the distribution factor. This principle is the best proof of the strong hierarchical relation connecting wage earners to firms. In the course of adjustment, firms impose the price of consumption goods in line with their distribution target and thus, for a given money wage rate, they fix the real wage rate in terms of consumption goods. Firms can impose both the real wage rate and the amount of employment. Firms' power is so binding that it contradicts the very possibility of a labour market in the neo-Classical sense, since wage earners cannot escape from unemployment by accepting a fall in their real wage rate. The fall in the real wage rate is the counterpart of the increase in unemployment which is rooted in the relation of the required rates of return with the amount of investment they support. The relation illustrates the blatant involuntary aspect of unemployment relative to the role wage earners can play in the circuit process. 4 THE IMPACT OF FIRMS' OWN THRIFTINESS According to the theory of the monetary circuit firms derive binding power from the very fabric of the capitalist monetary economy. Exercising the power which is bestowed on them, firms can follow two paths, the spendthrift path or the thrifty path. Following the first path, firms never gamble on bigger profits without first betting on more investment expenditures. They never try to increase the share of profits in order to raise the profit level. The more spendthrift they are, the more they are led by expansionist animal spirits and the less they are lured into the quest for larger rates of return. 174 Beyond Scarcity Following the thrifty way, finns are increasingly animated by their loathing to spend in respect the distant future because they are more and more afraid of losses of capital values. They contrive ceaselessly to increase the rates of return to foster their own values. There are three degrees of addiction in the thrifty way, each displaying a greater loathing for betting on the future. The First Degree of Firms' Thriftiness Managers impose a prerequisite for any rise in the level of investment relative to the former state of profit consistency, the increase in their rate of return. Let us assume that the new norm is reached by the rise in the rate of return in consumption goods firms, the required rate of return being unaltered in equipment goods firms. The game, according to rule (27), leads managers in consumption goods firms to an increase of their investment expenditures relative to an unchanged wage bill. Equipment goods firms' managers will accordingly increase their own investment expenditures (rule 20) just to maintain their rate of return. Aggregate employment has increased because of the rise in employment in equipment goods firms relative to the frozen level of employment in consumption goods firms. The new distribution of employment fits the new profit norm in consumption goods firms. In this game, managers do not play explicitly against the Kalecki Principle. Firms as a whole indeed impose a rise in the share of profits, which is consistent with a rise of their aggregate profits resulting from the rise in investment. Playing this game, firms enjoy both an increase in their average rate of return and in the absolute level of their profits because of the increase in investment ex pendi tures. The Second Degree of Firms' Thriftiness Managers now want an increase in their average rate of return in order to maintain the same level of investment characterizing the past state of profit consistency. They do not as yet long for an increase in the absolute level of their profits. The game requires a rise in the rate of return in consumption goods firms relative to a frozen level of investment expenditures. Consumption goods firms are thus committed to slashing their wage bill and together with their level of employment until the ratio of their investment relative to their wage bill is equal to their new required rate. While employment falls in consumption goods firms, its level is frozen in equipment goods firms playing with an unchanged level of investment expenditures and the same profit norm. The new distribution of employment can then fit the new required rate of return in consumption goods firms. Thereby, firms as a whole impose the rise in the Alain Parguez 175 share of profit, then providing for an increase of profits relative to the wage bill. Managers are not yet playing explicitly against the Kalecki Principle. The Third Degree of Firms' Thriftiness Managers now impose a rise in the average rate of return because they desire an increase in the absolute level of their profits, and have a growing dislike for an increase in their investment expenditures. This is a bad variant of the second game whose players are now trying to cheat the principle of independence and thus the Kalecki Principle itself. Imposing a rise in the share of profits is the sole reliable source of profits for players contriving to get rid of the very core of the monetary capitalist economy. The Outcome of the Thrifty Games: Inflation and a Fall in the Rate of Utilization of Equipment All the three games lead to inflation by pushing up the price of consumption goods. The first game at least compensates wage earners for their loss of purchasing power by allowing them winnings accounted for by the creation of jobs. The other games preclude the very possibility of any compensation by inflicting on wage earners a loss of jobs, thereby worsening the impact of the loss of purchasing power. All the games have an impact on the rate of equipment utilization in consumption goods firms - namely, the ratio of effective output to the maximum technologically feasible level, corresponding to the existing stock of equipment inherited from the past. Playing games two and three, managers are pledged to slash employment in consumption goods industries. Under constraint of the technological relation (16) where labour productivity is a given parameter embodying the existing technology, they are committed to slashing their effective output relative to its past level that matched the previous state of the profit norm. Their games result in a fall in the rate of equipment utilization relative to its past level, because of the fall of effective output relative to an increased maximum level expressing the rise in the stock of equipment accounting for past investment. Playing game one, managers are pledged to increase the stock of equipment in consumption goods industries while they freeze their employment. Assuming an unchanged technology, the maximum feasible level of output of consumption goods is to rise pari passu with equipment, while its effective level is frozen. Game one will also lead to a fall in the rate of the utilization of equipment. The Ultimate Outcome of the Thrifty Games: the Fall in Investment Let us start from a state of profit consistency with a given profit norm, managers not yet playing a thrifty game. 176 Beyond Scarcity The profit norm supports some stable level of investment and sustains over time an unchanged rate of unemployment and a constant rate of utilization of equipment. Firms' performance fosters their capital values which, over time, can increase because of investment, albeit there is a rate of equipment utilization not equal to 100 per cent. Now, managers begin playing thrifty games. The rate of equipment utilization is falling relative to its previous level and this faU must be regarded as permanent since it is the result of a new long-run strategy. The fall in the rate of utilization accounts for the increased share of equipment which is turned into pure waste and thus deprived of any value. The faU in the rate of utilization therefore accounts for the suddenly increased inability and unwillingness of managers to foster their past bets, embodied in equipment. The drop in the rate of utilization proves that managers no longer want to spend enough to create sufficient value to legitimate their past bets on the future. According to the fourth law of monetary circuit theory, this failure will be reflected in a new data set by a fall in the money value of consumption goods firms. Their values must faU relative to the level of the stake of the thrifty game. This planned or target value was the initial value inherited from the past state, plus the amount of investment expenditures consistent with the new game. Let V~, fi, V ~ be aggregate initial values of consumption goods firms inherited from the previous state of profit consistency, the amount of investment carried out when the new game starts, and the resulting aggregate capital values in the next data set. The fourth law leads to 6 V2 accounting for an increase in wealth: (50) (51) with (52) Whatever the case prevailing in (52), there will be a fall in the amount of investment in consumption goods firms. Managers are victimized by the very laws of the monetary circuit they contrived to forget by following the thrifty way. Being convinced of the necessity of following this path, they cannot understand the reason for their losses, and miss the feedback effect of their games. They are led to the sole interpretation fitting their addiction to thriftiness: there has been a permanent reduction of their ability to generate wealth from their investment. They will thus bet on a Alain Parguez 177 sharply reduced target capital value whose excess beyond the effective level accounts for new investment; so significant will be the fall in the target capital value that its excess over the effective level of capital value must be inferior to what it was before the aftermath of the thrifty games. The fall in investment in comsumption goods firms is thus the fatal outcome of the fall in the rate of utilization. I~ be respectively effective and target aggregate Let V~, V?, V~I, vL capital values in consumption goods firms when firms began playing thrifty games, then target capital values and effective capital values in the circuit process after the beginning of the game. Investment in circuit process 0 and in circuit process 1 reflects the aftermath of the games: ti, (53) (54) (55) and so (56) The more they have disregarded the laws of the monetary circuit, the more firms must be wounded by the resulting fall in the rate of utilization and the bigger wiII be the fall in investment in consumption goods firms. The most dramatic fall in investment is to be observed in the aftermath of game three whose players wanted the impossible for their goal - a rise in the level of their profits as a reward for their absolute thriftiness. No rise in the investment of equipment goods firms can compensate the fall in investment of consumption goods firms. In the immediate aftermath of the games, equipment goods firms enjoy a 'windfall rise' in their rate of return for having frozen their investment, according to rule (22). This 'windfall gain' cannot protect their capital values against the failure of consumption goods firms. As shown above while discussing the Kalecki Principle, bets on the future course of consumption are the ultimate source of value for the whole equipment structure, whatever might be its degree of roundaboutness. Direct bets on consumption are the source of value bestowed on equipment embodied in consumption goods firms, and direct bets on the value of this equipment (and thus indirect bets on consumption) are the source of value bestowed on that part of equipment embodied in equipment goods firms. In the aftermath of the games, the collapse of capital values will thus spread all over the equipment structure and equipment goods producers will cut their own investment, aware of their diminishing ability to generate wealth. Beyond Scarcity 178 All the thrifty games are therefore leading to a fall in the aggregate amount of investment. In games two and three, investment falls below the level that was sustained before managers began playing with thriftiness. In game one, the fall in investment of consumption goods firms according to (56) should wipe out any increase in investment relative to its past pre-game level. Equipment goods having slashed their investment relative to its pre-game level, in game one also aggregate investment has to fall below its past level in line with the pre-game profit norm. Rule (56) therefore leads to (58) where j, fl, /1 are respectively aggregate investment before firms played one of the thrifty games, the amount of investment that was the stake of the game, and the level of investment which is the outcome of the thrifty game: (57) (58) Rule (58) is the ultimate consequence of the Kalecki Principle, emphasizing the dramatic feedback effect of firms' thriftiness on the future course of the economy. Every time firms impose a higher share of profits because of their increased propensity to thriftiness, there will be a fall of aggregate investment expenditures relative to the level that matched the past degree of firms' propensity to thriftiness. Without trying to address a full explanation of the level of investment, from this principle we can derive the (60) relation emphasizing the negative impact of the rise in the share of profit over the level of investment. Such a relation embodies the whole dynamic sequence which is the short-run aftermath of the change in the share of profits (50)-(58). I = I (h) (59) <0 (60) I~ The Impact of the Feedback Effects on Firms' Propensity to Thriftiness and the Role of Banks Such an induced fall in investment triggers a cumulative process of deflation whose first stage is both the increase in unemployment and the fall in aggregate profits determining a new fall in capital values all over the economy. This process cannot end before banks succeed in stabilizing the capital values of firms, by overcoming the impact of the cumulative fall of profits on their own bets on the Alain Parguez 179 future. Banks could then convince firms to bet on a new stable level of investment and the effective fall in profits would thus be stopped. Without greater insistence on the long-run dynamic, it is crystal clear that stabilization would take effect at a very depressed level of capital values, investment and profits relative to the past state of profit consistency concurring with the former degree of firms' propensity to thriftiness. As long as they are pledged to foster their firms' own capital values, managers should thus take care of the feedback effect of an increased inclination to thrift. The feedback effect should therefore preclude managers from becoming addicted to thrifty games. It would stabilize the share of profit and thus the average rate of return, and, more specifically, the rate of return in the consumption goods firms. The prerequisite for the instability of this rate of return is the fact that managers can ignore the possibility of the feedback effect. The sine qua "lOn of such blithe ignorance is that banks themselves succeed in preventing the feedback effect from being engineered by firms' growing propensities to thriftiness. What is then at stake is the banks' own increased propensities to thriftiness. Banks are now so loath to endorse bets on the future that they support firms' desires for an increased rate of return. They can even impose this rise as the prerequisite for an endorsement of firms' new bets on the future. Banks are therefore pledged to supersede the impact on capital values of the fall in the rate of equipment utilization. Banks are thus committed to determine capital values as being the mere reflection of firms' power to impose the required average rate of return. By playing thrifty games themselves, banks can tum the monetary capitalist economy into a pure Keynesian rentier economy, whose dynamic short-run process is the complete opposite of the normal 'non-rentier' regime, as shown by the following system where V and 2 are respectively firms' aggregate value and the average rate of utilization. Rentier regime Normal case 2 = 2 (h) (61) 2;, < 0 V= V (h) 2 = 2 (h) (63) 2;, <0 (62) V= V(h) V;,: 0 Feedback effect No feedback effect (64) In both cases, a more stringent profit norm (the rise in the share of profit reflecting the rise in the average rate of return) leads to a fall in Z. In the renlier regime, the fall in the rate of utilization induces no further losses of capital values. Nothing 180 Beyond Scarcity could deter managers from imposing any rise they want in the average rate of return and the share of pront, and there would be no further obstacle to the rise in unemployment adjusting the effective rate of return in consumption goods industries to its required level. Rule (64) displays the possibility of a much worse aspect of the rentier regime when the rise in the share of profit creates windfall gains in capital values. Banks are now betting on the improved quality of management which should be reflected in the rise in the rate of return. Investment is no longer the prerequisite for the increase in capital values. Nothing could now deter managers from lowering their investment expenditures and thus slashing their own profits. The theory of the monetary circuit emphasizes the cumulati ve effect of thriftiness: (a) (b) (c) (d) (e) Superseding the feedback effect of firms' own thriftiness banks suppress the major obstacle to the thriftiness of managers; To succeed in their strategy, banks are led to determine capital values by emphasizing firms' power to increase their rate of return; They are thus led to suppress any short-run link between capital values and effective 'real' profits; Cumulative thriftiness thus increases the discrepancy between 'real' or 'normal' capital values, reflecting the ability to generate 'real' profits from investment expenditures and 'effective' capital values, which are the pure outcome of banks' striving to cheat Kalecki's Principle; As the gap grows, the whole set of capital values becomes more and more 'fragile' relative to the state of the real economy afflicted by increased unemployment, an increased fall in the rate of utilization and a progressive drop in profits. Such a growing fragility requires more and more intervention from the whole credit network whose strategy must be perfectly coordinated, with the full commitment of all policy-makers to support the thriftiness of firms relative to wage earners. Without further insistence on the long-run course of the rentier economy, we can assume, as Keynes and many others did,I4 that it could not escape from the liquidation crisis of writing off 'fictitious' gains in wealth. 5 BANKS' OWN THRIFTINESS AS THE EXPLANATION OF THE RATE OF INTEREST Lending Banks and the Central Bank in the Credit Network The existence of a perfectly consistence set of commercial or rather lending or credit banks is the institutional foundation of the monetary capitalist economy. A consistent network of credit consists of a set of iending banks with the following characteristics: Alain Parguez I. 2. 3. 4. 5. 181 The credit granted by a bank must be instantaneously endorsed by the whole set of banks; The counterpart of any credit granted by a bank to a particular firm must be money for the whole set of banks and this can be used by any other firm to write off its debt initiated by the credit of another bank; All member banks are pledged to the same judgement of firms' ability to generate a profit; there must be in the data set of a circuit process a unique set of capital values; According to the fourth principle of the core of the monetary capitalist economy, the ability of a lending bank to meet its requirements must be reflected in its own capital value accounting for the bank's success in catching a flow of income; Short-run failures by firms to meet their profit target should not lead to an automatic fall in the value of banks. This is the prerequisite for preventing a dramatic increase in banks' thriftiness expressed by their growing dislike for granting credit. As long as we are not in a perfectly self-regulated credit economy, the consistency of the credit network requires a regulation agency, namely, a central bank which is a branch of the apparatus of government. The normal mission of the central bank is to foster lending banks' own values by sanctioning their own endorsement of firms' bets. Carrying out this fifth principle, the central bank provides the lending banks with perfectly liquid assets whose value in money terms is independent of the fluctuations of firms' rates of return. Lending banks acquire those assets, the so-called reserves, through a loan they have to write off at the end of the circuit process. The central bank imposes a ratio b of these reserves to the amount of credit granted by the lending banks. The central bank also charges a rate of interests i' on its loan of reserves. Through this policy of fixing both b and i', the central bank determines the minimum income lending banks have to earn to maintain themselves in the network. Carrying out principles (1)-(3) embodying the homogeneity and consistency of lending banks' credit, the central bank is committed to imposing the harmonization of their profit norm which is the stake of their credits. The fourth consistency rule thus explains the equalization of the required rates of return banks have to reach. For the sake of the fifth rule, the central bank imposes a minimum rate of return on banks as the prerequisite for its endorsement of their bets and thus for the endorsement of their own values. The Rousseas Principle In any circuit process, for lending banks as a whole their net revenue is equal to the excess of their gross income over their operating costs. In our model, their sole source of gross income is the amount of interest income firms have to pay. Beyond Scarcity 182 Let us assume that their operating costs are equal to interest payments to the central bank. The ratio of their profits to their operating costs must be the required rate of return whose excess over the minimum rate accounts for banks managers own desire for profits. In any circuit process, the amount R of interest paid by firms must tally with the consistency condition, where i, F, rB, rBCB , kB are respectively the rate of interest charged on credits, aggregate credits, the rate of interest charged on credits, the rate of return lending banks require, the minimum rate of return fixed by the central bank, and the factor reflecting bank managers own desire for profits: R= iF (65) R = (1 + rB)i'bF (66) =(1 + rB)i'b (67) i rB = (1 + kB)rC: (68) The theory of the monetary circuit makes full sense of the principle adamantly emphasized by Rousseas (1986): Neither the existence nor the level of the rate of interest depend upon a specific demand for money as an asset relative to its supply: both the existence and the level of the rate of interest are the outcome of the management of the banking network as the specific industry providing with credit the production industries; the rate of interest is determined like any other price by the application of a mark-up rate to the unit operating costs; the rate of mark-up is reflecting the profit norm of the credit 'industry'. What is, of course, specific to the banking 'industry' is the permanent identity of the effective rate of interest, given by (67) with its required level generating the required rate of return whose level is reflected by equation (68). Relative to its customers, the credit industry enjoys the situation of a perfect moriopoly which is never constrained by demand in the determination of its price. The whole (65)-(68) system reflects this absence of demand as a binding factor, which is the foundation of the Rousseas Principle; this non-existence of the demand constraint is embodied in the two following rules explaining the fluctuations in the rate of interest l. 2. For a given kB the rate of interest is entirely determined by three factors controlled by the central bank, the minimum rate of return and the two parameters expressing the monetary policy, the rate of interest on reserves and the reserves ratio. For a given level of the factors controlled by the central bank, the rate of interest reflects the specific desire of bank managers for profits, their own propensity to thriftiness. Alain Parguez 183 The central bank plays a crucial part in the fluctuations of the rate of interest. The central bank's power is not bound by any constraint other than its own agenda. Keynes had already emphasized that full employment is spontaneously ignored by central bankers when shaping their agenda. Being perfectly exogenous relative to the requirements of full employment, the rate of interest is not the outcome of a pure whimsical power. Central bankers are spontaneously led to believe that the consistency of the credit network requires the perfect stability of the value of money, reflected by its purchasing power in terms of goods. Stabilizing money prices of goods should protect wealth-holders against losses of purchasing power; money values of firms would thus rise; this increase in real wealth could support a sound increase in investment. In this process lie the roots of central bankers' absolute pledge to impose price stability by imposing a rise in the rate of interest. The central bank is animated by its unmitigated faith in thriftiness as the avenue leading to a zero inflation regime. Forcing a rise in the rate of interest, the central bank expresses its increased desire for a persuasive increased thriftiness whose outcome should be the fall in expenditure, leading to the required unchanged level of money prices. The central bank's own thriftiness thus sustains the rentier economy described by (63) and (64), where the generation of wealth is not dependent on investment expenditures. The thriftiness of the central bank leads to the major consequences of the Rousseas Principle: When the rate of interest has become exogenous relative to the requirements of full employment, it is endogenous relative to the central bank's target of zero inflation. As long as the rentier regime is supported by the central bank, increases in the rate of interest should not lead to an automatic fall in capital values. These two propositions are rooted in the relation between the rate of interest and the firms' own rates of return. The central bank is betting on the cumulative resulting fall in employment that should induce a fall in the money wage relative to labour productivity, which would write off all the factors of inflation. Without insisting on this relation connecting the fluctuations of the money-wage rate with the growth in unemployment, we must emphasize the channel through which the fluctuations in the rate of interest are transmitted to the data set controlling the circuit process. The Transmission Channel: the Impact of the Rise in the Rate of Interest over Firms' Required Rates of Return The transmission channel that the theory of the monetary circuit emphasizes can be derived from the impact of interest payments over firms' retained profits. In our simple economy, let us assume that both lending banks and the central bank do not spend their net income in acquiring equipment goods. There is thus no feedback effect of interest payments on firms' gross income. As already shown Beyond Scarcity 184 by Graziani (1990), interest payments are a net levy on aggregate profits which are now distributed between firms and banks. 1T, 1Te, 1TB being aggregate profits, firms' own profits, banks' own profits including lending banks' net income and the central bank's income, the Graziani Principle implies that 1T= (69) 1 1T = 1Te + 1TB (70) 1TB=R (71) 1Te= I-R (72) Aggregate profits being still determined by the Kalecki Principle (69), the higher are interest payments, the lower are aggregate firms' own profits (72). Interest payments reflect the amount of initial credit firms cannot pay back from their sales-generated gross income. For the sake of the consistency of the credit network, banks are pledged to finance this deficit instantaneously by recycling their profit on the so-called market for financial assets, the stock exchange. In the monetary capitalist economy, firms' capital values are reflected by the market value of the stock of shares issued by firms' managers. This stock of financial titles to real wealth is held partly by firms themselves and partly by banks (including the central bank). In each circuit process, banks recycle their interest income in the acquisition of shares issued by investing firms as the financial counterpart of their investment. Through this recycling process, firms can write off their whole initial credit and banks accumulate net real wealth, which supports their own capital values. The recycling process is perfectly consistent with the investment-saving relation stemming from the Kalecki Principle. S*, S being as before real saving and financial saving, a v, aVe, a VB accounting for aggregate increase in wealth, firms' increase in their own wealth and banks' accumulation of net wealth, S's/=AV (73) S==aVe+AVB (74) == 1Te + 1TB S == 1T 1T = 1 (76) S=I=S* (77) (75) Investment still accounts for real increase in wealth or real saving (73). Financial saving accounts for the increase in the value of the stock of shares held respectively by firms and banks (74). The source of this increase in financial wealth is the value of the shares newly issued by firms' managers having to Alain Parguez 185 recycle aggregate profits. Banks spend their who.le profit in acquiring a part o.f the new shares. The remaining part is retained by firms as if they were spending their o.wn pro.fit in buying shares o.n the market, so. as to. write o.ff their debt. Financial saving is thus stilI identical to. aggregate profits (76) who.se amo.unt is fixed by investment expenditures (76). The Kalecki Principle thus impo.ses anew the strict equality o.f financial saving to. investment as real saving. Fro.m this impact o.f interest payments o.n pro.fits stem two. paramo.unt co.nsequences: I No. free fund fo.r investment can be generated o.n the financial market. 15 The proo.f o.f this pro.po.sitio.n is included in the (73)-(77) system. Shares can be so.ld because there are already available profits needing to. be spent. Tho.se profits canno.t be generated by the market: they are created by credit-financed investment. II Any rise in the rate o.f interest leads to. an increase in firms' required rates o.f return who.se o.utco.me is a cumulative thriftiness. Each set o.f firms is pledged to. generate an amo.unt o.f profits at least equal to. the interest bill. Let us address the ultimate case when profits are just equal to. their minimum. The Kalecki Principle still ho.lds and minimum pro.fits are equal to. the amo.unt o.f investment expenditures. Let 1T'j'and rj'be, respectively, the minimum amo.unt o.f profits and the co.rrespo.nding minimum rate o.f return: (78) 7T'j' = Ij fo.r j = 1,2 (79) and so. m . m 7Tj I J Wj l-i r· = - = - - (80) Equatio.n (80) gives the required level o.f the rate o.f return when firms canno.t retain a share o.f their profits. This minimum rate o.f return is the same fo.r all sets o.f firms. To. meet their pro.fit no.rm, firms are pledged to. retain a share o.f their pro.fits who.se ratio. to. their wage bill acco.unts fo.r their specific ability to. fo.ster their capital values. Fo.r each set o.f firms, the required rate o.f return is therefo.re the sum o.f two. co.mpo.nents, the minimum rate and the required o.wn rate o.f return rEi' The set of Beyond Scarcity 186 rates of return is therefore entirely determined by the own rates and the rate of interest, the outcome of dynamic adjustment being always >I< rj == >I< rEJ i +-1-. -I (81) j == 1,2 (82) for both consumption goods firms and equipment goods firms. The rj reflects the real structure of the economy (n, m), and the transmission mechanism of the rise in the rate of interest is rooted in the determination of the rj according to (81) and (82). Let us assume a rise in the level of the rate of interest, whatever its origin, the spontaneous increase in lending banks' propensity to thriftiness (embodied in the k8 factor in (67», the central bank induced rise in their minimum own rate of return (the rCf term in (67», or the autonomous increase in the central bank's propensity to impose thriftiness relative to it prices target (the factor costs in (67». The rise in the rate of interest has triggered a dynamic process of adjustment whose major phases embody the following sequence: First Stage All managers are pledged to raise their rates of return, the rj, in order to protect their own rates of return, the rEj. They are thus lured into playing the second game of thriftiness, imposing a rise in their rates of return without increasing their acquisition of equipment goods. In consumption goods firms, managers have to slash employment while they increase prices, until n, the distribution factor of employment is equal to their new required rate of return. Managers of equipment goods firms, according to their profit norm in (22), have to raise their investment expenditures relative to the unchanged level of investment in consumption goods, until the distribution factor of investment expenditures m is equal to their new required rate. Real supply of equipment goods being frozen, the required rate of return is reached through the price rise of l:quipment goods that imposes the necessary increase in the share of equipment goods firms in the distribution of the real supply. Profits of equipment goods firms are raised but this is a pure profit inflation that banks have to finance for the sake of consistency. The outcome of the first stage is thus: (a) (b) Inflation - the price level of both consumption goods and equipment goods having been raised; An increase in unemployment; Alain Parguez (c) 187 A fall in the share of consumption goods firms in the distribution of aggregate profits, reflecting the increased degree of roundaboutness of the production structure induced by the change in the distribution of new equipment goods. Second Stage Let us assume that the central bank wants to delay as long as possible the impact of the feedback on capital values. The central bank responds instantaneously to the effective increase in prices by imposing a new rise in the rate of interest. Even assuming an unchanged propensity of lending banks to thriftiness, the central bank's pledge to thriftiness imposes on firms' managers a new rise in their required rates of return, to maintain their own rate. There will be more inflation, more unemployment and the degree of roundaboutness will be increased again. The pace of the thriftiness race accelerates when both lending banks and firms strive to maintain their 'real' own rates by increasing their money own rates to compensate for inflation. This 'real own rates' strategy is imposed by the central bank as the prerequisite for its support of capital values. Third Stage The cumulated growth in unemployment must lead to a collapse of wage earners bargaining power. The money-wage rate starts falling relative to labour productivity. This drop in the cost of labour is accelerated by the fall in investment expenditures. The drop in investment starts either when the central bank is not able to compensate any further for the feedback effect, or when it bets on the necessity of a dramatic fall in capital values to write off inflation. The ultimate outcome of the thriftiness race must be the increase in unemployment that would impose such a dramatic fall in the labour cost that it would fully write off inflation, including central bank induced inflation. Without addressing this long-run adjustment any further, it is obvious that it would lead the economy to a new state of stable profit consistency, displaying, relative to the former condition a much higher rate of unemployment, a higher share of profit in line with the required increase in the rates of return, a lower level of investment and a lowered amount of aggregate income. The theory of the monetary circuit thus fully endorses Keynes's conclusion relative to the impact of the central bank's pledge to maintain price stability by rising the rate of interest. This commitment of the central bank is indeed the scourge of the monetary capitalist economy. It enslaves the whole of society to thriftiness. 6 THE IMPACT OF HOUSEHOLDS' THRIFTINESS This accounts for the share of their income that households do not want to spend in the acquisition of consumption goods. In our simple model introducing the Beyond Scarcity 188 thriftiness of households it is assumed that wage earners do not wish to spend their whole income in the market for consumption goods. Their thriftiness accounts for their desire to substitute titles for wealth that could generate purchasing power in the future for current consumption expenditures. Here lies Hayekian saving in its purest embodiment. abstinence! One of the major contentions of the theory of the monetary circuit is that Hayekian abstinence leads to increased scarcity. This Keynesian contention is anchored in the dramatic impact households' thriftiness has on firms. lending banks and even central banks' own propensity to thriftiness. The Generalized Kalecki Principle Wage earners' saving is a pure non-compensated leakage in the process of generating gross income accruing to consumption goods firms. Wage earners' saving is thus a net loss of profits for those firms relative to the case when there was no leakage. Consumption goods industries gain a profit which is now the excess of the wage bill in equipment goods industries over aggregate wage earners' saving. SH = W. -SH (83) 1Tt =1- W. (84) 1T =I-SH (85) 1TZ as always and so Equation (85) is the generalized Kalecki Principle. Firms as a whole receive as aggregate profits the excess of their investment expenditure over wage earners' saving. For a given amount of investment expenditures and a given wage bill. the higher is wage earners' propensity to thriftiness. the greater is their saving. and the lower must be the amount of aggregate profits firms as a whole generate by their investment expenditures. From the generalized principle stem three major consequences illustrating the dire effect of wage earners' thriftiness. 1. For a given investment. a given wage bill. the rise in the wage earners propensity to save determines a fall in firms' retained profits relative to the unchanged interest bill: Alain Parguez 189 (86) (87) (88) 2. 3. There is not the least reason why the rise in s, the propensity to save, should induce a fall in the interest bill. The whole impact of the rise in s is thus sustained by firms' retained profits, as shown by (86). For a given investment and a given wage bill and interest bill, the fall in consumption goods' firms aggregate profits accounts for the whole impact of the rise in s. An increased wage earners's propensity to thriftiness is thus reflected by a rise in the ratio of profits generated by equipment goods firms to the profits generated by consumption goods firms. The further s rises, the less profits generated by consumption can sustain the value of the whole equipment structure. Assuming a given amount of investment and a given wage bill in both sets of firms, the rise in s cannot affect the effective rate of return in equipment goods industries whose aggregate profit is unaltered. Since the rise in s induces a fall in consumption goods firms' profits, their effective rate of return will fall. The ratio of labour in equipment goods industries relative to consumption goods industries being as always n, the new effective rate of return in consumption goods firms is derived from the determination of profits: 1T2 = WI (1-s) -sW2 r2 = n (l - s) - s (89) (90) In Post Keynesian terms, as shown by Graziani (1990) and Seccareccia (1984, 1985), the rise in s lowers the effective rate of mark-up in consumption goods industries and thus their price level. For a given profit norm, the rise in wage earners' propensity to thriftiness is challenging the very capacity of firms to meet their profit target. This challenge cannot be ignored because there is not the least obstacle to the rise in wage earners's propensity to save which could arise from the circuit process. The Recycling of Households' Saving in the Circuit Process and the Generalized Keynesian Equality of Saving and Investment Wage earners' saving accounts for the gross deficit of firms as a whole which is that share of initial credit firms cannot pay back by their sales-generated gross income. Wage earners' saving cannot exist if it is not instantaneously invested into the acquisition of available debt or titles to wealth. Wage earners' saving is Beyond Scarcity 190 thus instantaneously spent in the acquisition of an equal amount of debt titles. In our model, the sole source of debt titles, directly or indirectly, is the gross deficit firms are pledged to finance by writing off their short-run debts. There is thus a required equality connecting wage earners' saving to the gross deficit it has initiated. This required equality holds in any circuit process. The gross deficit is reflected by the loss of profits induced by wage earners' saving, and the recycling process is described by the following relations: /-1T = SH (91) or S being aggregate financial saving S* == / = 1T + SH = 1TE + 1T8 + SH = S (92) Neither (91) nor (92) are mere identities because they are the outcome of a recycling process which is a sine qua non condition for the very existence of the credit system. Its crucial teaching is that no real increase in financial wealth can be generated without being invested in the refunding or the permanent financing of the newly produced equipment goods. As long as the economy is not a pure rentier economy, the value of the increase in financial wealth is thus equal to the increase in real and productive wealth S· which is the amount of investment expenditures. Investment generates an equal amount of financial wealth, aggregate financial saving, which is split between firms (retained profits), banks (banks' own profits) and wage earners. A 'lack of saving' cannot occur so that any lack of investment must be considered a shortage relative to the requirements of full employment. The recycling process cannot be hindered by the so-called households' preference for liquidity. Households do indeed have the choice between two investments of their savings, acquisition of direct claims on firms or acquisition of direct claims on banks. The first investment is operated through the financial market by acquiring shares newly issued by managers, channelling income directly to firms. This investment leads to an identical reduction of the gross deficit. There now remains an ultimate deficit which is equal to that share of their savings that households want to invest in claims on lending banks. Those claims are 'permanent' deposits whose money value is perfectly secure from the fluctuations in the capital values of firms; they are absolutely secure from the constant possibility of change in the future perspectives. As a consequence, they are secure from the fluctuations in the level of the rate of interest. Here lies that famous Keynesian preference for liquidity which can lead wage earners to invest their savings in long-run deposits. The more they fear the future fall in capital values, the higher is their preference for liquidity, and the greater the share of their savings which is not invested through the financial market. For the pure sake of consistency, lending banks are pledged to channel this saving into the refunding of the remaining deficit. They substitute a long-run claim on firms for their remaining deficit. Alain Parguez 191 The Two Sides of Lending Banks Activity in the Monetary Circuit and the Impact on their own Thriftiness The theory of the monetary circuit has emphasized the two roles lending banks have to play in the circuit process. In their credit role, they generate the liquid fund, the newly created quantity of money, invested into the achievement of firms' bets. Playing their recycling role, the so-called 'intermediary' role, they cannot create a fund that could finance a new round of expenditures. They merely increase their stock of firms' debts which is the reflection of the stock of liquid assets with which they provide wage earners, to quench their thirst for liquidity. The theory of the monetary circuit, as already shown by Graziani (1990), makes a clear-cut distinction between money initiated by credit responding to bets on the future, and the stock of liquid assets held by households in the form of claims on lending banks reflecting the long-run claims on firms banks hold as assets. The recycling role has an impact on the credit activity affecting both the interest bill and the constraint imposed on firms. The magnitude of this feedback effect depends on the propensity to thriftiness of both the lending banks and the central bank. The Generalized Rousseas Principle Being pledged to accumulate illiquid assets, lending banks are committed to fostering their own capital values by maintaining a ratio of their stock of firms' debts to their net wealth. Relative to the case when they were not required to accumulate a stock of firms' debts, they are now pledged to increase their own profits by imposing a higher rate of return on their costs. The central bank raises those costs by maintaining a ratio of reserves to the stock of illiquid assets. Let us assume that banks succeed in maintaining a zero rate of interest on wage earners' permanent deposits by putting forward the perfect liquidity of these assets. The source of lending banks' increased income is the interest bill firms have to pay on their debt. Let us assume that this supplementary interest bill is included in interest payments resulting from new credits. Firms are now required to pay an interest bill equal to iF = (l + rB) b i F (l + e) (93) e being the ratio of the stock of debts to new credits and thus i = (1 + rB) b i' (l + e) (94) with (95) with rs r gD>O (96) accounts for the rate of return agreeing with the case corresponding to the absence of illiquid assets, and rB ego) accounts for the rate of return that should 192 Beyond Scarcity generate enough profits to cope with the expected rate of increase in the stock of debts, thus allowing the maintenance of the required ratio of illiquid assets to net wealth. Equations (94)-(96) emphasize the generalization of the Rousseas Principle embodied in three propositions. 1. 2. 3. Households' saving does not have the least negative impact on the rate of interest by the virtue of a particular market mechanism. For a given lending banks' and central bank's propensity to thriftiness, households' saving is indeed pushing up the rate of interest imposed on new credits. Banks recycle their supplementary profits in the acquisition of more shares which firms' managers have to sell on the market. This result can explain why banks' managers strive to attract household savings. 16 The more they lure households into holding their wealth in liquid assets, the more they increase their claims on firms, and the more they can increase their net wealth. The more lending banks are thrifty, the more they raise the rate of interest when they bet on an increased preference for liquidity, leading to a rise in the rate of increase of their stock of illiquid assets. The generalized Rousseas Principle leads to a major characteristic of the monetary capitalist economy: any increase in the preference for liquidity determines, ceteris paribus, a rise in the rate of interest. The liquidity trap is much deeper than that which Keynes had in mind in the General Theory! The Imposed Ratio of Aggregate Profits to Investment To express their dislike of endorsing bets that could force them to hold too much illiquid assets, lending banks can persuade firms to meet a more stringent profit constraint. 17 This far more demanding strategy imposes on firms to generate enough aggregate profits to fund a required share q of their investment expenditures. This constraint of internal finance determines the amount of aggregate profits firms have to raise from the circuit process to meet the required internal financial target and pay the interest bill. For a given interest bill, this amount of required aggregate profits is expressed as ratio 1'to the amount of investment expenditures,fbeing entirely determined by the required ratio of internal finance. Aggregate profits being effectively determined according to the Kalecki Principle, the state of profit consistency now requires that fl=l-sW (97) from which we derive W=/ 1 - f s (98) Alain Parguez 193 with f=f(q) (99) !q>O (100) where (98) detennines the wage bill adjusting effective profits to their predetermined required level, according to the new profit norm expressed as an increasing function of the required rate of internal finance q, and taking account of the interest bill banks want to receive in any case. For any amount of investment, the wage bill and thus the amount of employment have to fall When banks become more thrifty (q is raised, which raisesj) and when households themselves become more thrifty by raising s. This adjustment imposes the amount of effective households savings matching the profit constraint for a given amount of investment. This adjustment leads also to the generalized form of the process of aggregate income determination. In our economy, the average required rate of return is now r*=~=sL ]1-f I-f (101) s which determines the required share of profits as a function of both s andf r*=~=sL ]1-f I-f (102) It has to be increased when both s andfare raised to cope with the induced rise in the required average rate of return. Adjustment will always lead to the amount of aggregate income tallying with the profit constraint and thus meeting the condition h *Y = ] - s(l- h*)Y (103) Y=]---- (104) h =h* (105) h * +s(l- h*) and thus substituting for the required share of profits its expression in (103). Y=] 1 sll-j(l-s) (106) This last expression is fonnally similar to a Keynesian multiplier but it is the outcome of an adjustment process which is rooted in the imposed fall in labour 194 Beyond Scarcity income inducing the required profits. It does mean that for a given amount of investment and required profits, firms as a whole can generate enough effective profits by imposing the required rise in the share of profits. This result does not vindicate the Classical doctrine since adjustment is carried out by the rise in unemployment. For a given I, the amount of income meeting the profit constraint is inversely related to bothfand s. Households' thriftiness is not the avenue leading them to wealth. Contrary to what policy-makers and their advisers think, by being more thrifty wage earners are pledged to suffer from more unemployment. The Anti-Hayekian Theorem The Impact of the Saving-induced Profit Norm on the Required Rates of Return The whole impact of wage earners' saving on sales, profits and deficit is sustained by consumption goods firms. We may thus assume that the required rate of return of equipment goods firms is left unaltered by the new profit norm. Consumption goods firms cannot ignore this norm which is imposed solely on them. Their required rates of return are the ratio of their required aggregate profits to their required wage bill in line with the equality of their effective profits to these required profits. Let (1 - v) be, as always, the ratio of consumption goods firms investment to aggregate investment. Their required profits are thus 1T;;::f(l-v)/ (107) From equation (97) given the required aggregate wage bill we derive the required wage bill in consumption goods industries (108) or W;;:: 1[1-s f __+1_]r. 1 (109) From (107) and (109) we derive the required rate of return of consumption goods industries which must be imposed as their effective rate by consumption goods firms • r2 = f(l- v) 1- f __I_ s 1 + Ii (110) while as shown before r2;:: n (1 - s) - s with as always (111) Alain Parguez W. N W2 N2 11=_1 =_1 195 (112) For a given rate of return in equipment goods industries and a given distribution of investment expenditures, the higher are both / and s, the more stringent is the profit norm, and the higher is the required rate of return in consumption goods industries, which accounts for the rise in the required average of return and in the share of profits. The Impact 0/ an Increased Household Propensity to Thriftiness on the pace 0/ the race to thriftiness Any rise in s at the same time increases the required rate of return in consumption goods industries and squeezes their effective rate corresponding to the former distribution of labour. There is not the least reason why equipment goods producers should increase their employment. Consumption goods firms are thus pledged to slashing their employment until the II factor has been raised sufficiently to adjust the effective rate of return to its required increased level. The list of the dire effects of the rise in the propensity to save is an impressive agenda: The rise in unemployment occurs while the price level of consumption goods is increased. It is thus perfectly untrue that households' thriftiness is an obstacle to inflation, which proves the fallacy of contemporary stalwart advocates of households' thriftiness as the prerequisite for a full stabilization of the price level. (b) As shown above, the rise in wage earners' savings will lead lending banks to increase their own required rate of return, pushing up the rate of interest and thus raising the/factor for a given ratio q (according to 99). The rise in / determines a new rise in the required rate of return in consumption goods firms, while the rise in the rate of interest also raises the required rate of return in equipment goods industries. For a given amount of investment, prices and unemployment will rise again. (c) The central bank is to join the race by imposing a new rise in the rate of interest to curb inflation. (a) The cumulative rise in unemployment will affect the amount of investment negatively either directly through the feedback effect of the fall in the rate of utilization or indirectly as the outcome of the rentier economy. Here lies the core of the Anti-Hayekian Theorem: The increase in abstinence relative to income leads to a cumulative increase in thriftiness whose outcome is a fall in investment. Thriftiness is thus increasing scarcity by preventing the generation of real wealth.IR 196 7 Beyond Scarcity CONCLUSION: BEYOND SCARCITY The Anti-Hayekian Theorem has been demonstrated in a simple model of the monetary capitalist economy. This model could easily be generalized by introducing income payments to rentier households holding financial assets, as in the Graziani model and by the integration of a government sector into our economy. In our model government was acting solely through the intervention of an 'autonomous' central bank having no commitment to full employment. It has been shown that the theory of the monetary circuit makes full sense of the Keynesian major vision: 'Classical Scarcity', Hayekian scarcity does not exist in the positive monetary capitalist economy. Scarcity does exist but it is the outcome of a persvasive social thriftiness reflecting the impact of self-imposed constraints upon the expenditures that anlicipate and thus create the future. Scarcity is a 'man-made artificial scarcity' whose major responsibility lies in firms' managers own thriftiness and in that of the credit network. Austerity-led policies are gUilty of creating scarcity by always sustaining thriftiness. Three perfect examples of the dramatic impact of contemporary wisdom in economic policy have been emphasized; policies of supporting firms in their search for a fall in the share of labour in distribution; policies of encouraging households to become more thrifty or forcing them to save more in search for the protection against the future fall in their income; and policies of the central bank pledged to maintain the stability of prices by raising the rate of interest. The theory of the monetary circuit leads, therefore, to the absolute dismissal of the new fiscal orthodoxy. The dislike of governments' deficits is a purely ideological vision of the economy which is inconsistent with the core of the circuit process. 19 The theory of the monetary circuit should thus be considered as a contribution to a synthetic approach to the 'After-Keynes' Economics that can lead to a political economy agenda vindicating the core of the revolution Kalecki, Keynes and many others in their time started against the austerity agenda. 2o Notes * I. 2. The author wishes to thank Ghislain Deleplace (Universite de Paris VIII), Thomas Ferguson (University of Massachusetts), Edward Nell (New School) and Mario Seccareccia (University of Ottawa) who worked hard on the first draft of this paper. Hayek. 1941: 374: 'Although the technocrats and other believers in the unbounded productive capacity do not yet appear to have realized it, what he [Keynes] has given us is really the economics of abundance for which they have been clamouring so long. rather. he has given us a system of economics which is based on the assumption that no real scarcity exists'. Hayek. 1941: 393: ' ... ultimately. therefore. iUs the rate of saving which sets the limit to the amount of investment that can be successfully carried through.' Alain Parguez 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 197 Hayek, 1941: 394: Throwing away the principle leads to the Keynesian assumption that 'the only scarcity with which we need concern ourselves is the artificial scarcity created by the determination of people not to sell their services and product below certain arbitrarily fixed prices'. This condition requires that money should just be a pure intermediate between exchanges initiated by the decisions of firms and income-earners. Whatever the real rate of expansion of the economy could be, the quantity of money should be absolutely independent of any demand for money. Any quantity of money can fit the requirements of exchange within the production structure. Hayek has never deviated from' this crucial postulate! (see, for instance, Hayek, 1941: 360. Hayek's view of unemployment is very different from the contemporary wisdom in political economy. Unemployment is imposed upon wage earners as the outcome of the collapse of a credit-financed production structure, initiated by the banking system in its search for earnings. Unemployment can be worsened by the rigidity of money wages but this rigidity is not the initial factor. Contrary to Keynes, Hayek emphasizes that unemployment will be automatically wiped out as a new production structure evolves from the normal market mechanism channeling savings both individual and corporate into the new investments. This view is addressed by scant explicit statements in the whole theoretical work of Hayek, most of which are in Hayek (1939). It explains why Hayekian economics leads to the rejection of the so-called banks' unbounded power to create money from credit. This could prove that Keynes's economics is turning into a permanent state of what is an anormal 'non-natural' state according to conventional economics - economics having evolved from the writings of the great economists of the past (see, for instance, on this point Hayek, 1941: 356. Keynes's economics in Hayeks' interpretation would thus lack the very reference to an equilibrium state. On the new view of thriftiness in the context of a world economy, see Parguez (1990). Hayek explicitly includes firms' saving or corporate saving in its definition of the available saving fund (Hayek, 1939, 1939: 159-160). Firms' savings are initiated by retained profits of corporations available for investment just like individual savings. There is no doubt in exegetical terms that Hayek relies on the Ricardian inverse relation when he deals with the determination of the level of profits, albeit this problem is rarely explicitly addressed in Hayek's works. Contemporary political economy is, in some way, much more explicit in its dealing with the postulate. Which, as already shown by Graziani (1987, 1990), explicitly leads to the rejection of the labour value theory. On this point, we must also refer to the pioneer work of Heinsohn and Steiger (1988) which addressed the fundamental link connecting capitalism to the existence of a credit system, creating funds as the counterpart of debts whose collateral is the value bestowed on the property of the debtor. On this point, see the outstanding exegetical paper of Graziani (1987) on Keynes's Finance Motive. See, for instance, Kalecki, 1966: 14-15. Albeit Kalecki does not emphasize the distinction between the two stages of the financing process, there is no doubt that profits are the outcome of already realized expenditures in the whole Kalecki model. See on this point Graziani (1990). The problem of firms' thriftiness has also been addressed in a very extensive and formal way in Bhaduri and Marglin (1990). Their conclusions seem to differ with the analysis we are presenting, because in Bhaduri and Marglin's model there is the possibility of a positive impact of the rise in the share of profit over the amount of investment. 198 14. 15. 16. 17. 18. 19. 20. . Beyond Scarcity On this point, see Hotson (1987) for an extensive discussion of the literature. On this crucial'proposition, see also Terzi (1989). This active role of banks in the search for savings has been forcefully emphasized by Chick, 1992: 193-205. Our formulation of the financial constraint is indeed integrating a crucial aspect of Post Keynesian theory addressing, like Wood (1975), the ratio of sustainable external financing to investment and thus determining the required amount of profits. This theorem had already been forcefully emphasized in the French literature by Jean de Largentaye in a paper first published in 1944 and published again under the reference, Largentaye (1988). Jean de Largentaye, the translator of Keynes's General Theory, introduced Keynes's economics in France. On this major point see, for instance, Parker Foster (1990). Nell (1988) and Wray (1990,1991). For the initial propositions of a New Keynesian Agenda, see Lavoie and Seccareccia (1989), Nell (1988) and Parguez (1990). Reference Aasland, A. (1990), 'The Scarcity of Money: A Reality, a Myth or Both', Economies et Sociites, Monnaie et Production, 7 (6), 37-50. Asimakopulos, A. (1990), 'Investment, Finance, Saving and Profit: A Kaleckian Approach to the Dynamic Circuit', Economies et Sociitis, Monnaie et Production, 6 (2), pp. 35-49. Barrere, A. (1990), 'Signification generale du circuit: une interpretation', Economies et Societes, Monnaie et Production, 6 (2), 9-34. Bhaduri, A. and S. Marglin (1990), 'Unemployment and the Real Wage: The Economic Base for Contesting Political Ideologies', Cambridge Journal of Economics, 14, 375-93. Chick, V. (1992) On Money, Method and Keynes, Selected Essays, ed. by Aristis, P. and S. C. Dow (London: Macmillan; New York: St Martins Press). Davidson, P. (1986), 'Finance, Funding, Saving and Investment', Journal of Post Keynesian Economics, 9 (I), 101-110. Fitzgibbons, A. (1988), Keynes's Vision: A New Political Economy (Oxford: Clarendon Press). Graziani, A. (1987), 'Keynes's Finance Motive', Economies et Sociites, Monnaie et Production, 4 (9), 23-42. Graziani, A. (1990), 'The Theory of the Monetary Circuit', Economies et Sociites, Monnaie et Production, 7 (6), 7-36. Hayek, F. (1939), Profits, Interest and Investment (London: Routledge). Heinsohn, G. and O. Steiger (1987), 'Private Ownership and the Foundations of Monetary Theory', Economies et Socieris, Monnaie et Production, 4 (9), 229-43. Heinsohn, O. and O. Steiger (1988), 'Monetary Theory and the Historiography of Money, or Debts, Interest and Technical Progress in Economies with and without Private Ownership', Economies et Socieres, Monnaie et Production, 5 (9), 139-53. Hotson, J. (1987), 'The Keynes Revolution and the aborted Fisher-Simons Revolution', Economies et Socieres, Monnaie et Production, (9), 170-85. Kalecki, M. (1966), Studies in the Theory of Business Cycles (Oxford: Basil Blackwell). Kregel, J. (l986a), 'Shylock and Hamlet or Are There Bulls and Bears in the Circuit', Economies el Sacieres. Monnaie et Production, 3 (8,/9), 11-22. Kregel, J. (l986b), 'A Note on Finance', Journal of Post Keynesian Economics, 9 (I), 91-100. Alain Parguez 199 Largentaye, J. de (1988), 'L'ccueil de I'cconomie monctaire', Economies et Sociitlis, Monnaie et Production, 5 (9), 11-19. Lavoie, M. (1984), 'Un modele post-keynesien d'cconomie monetaire de production fonde sur la theorie du circuit', Economies et Societes. Monnaie et Production, I (4),233-58. Lavoie, M. (1985), 'Monnaie et Production: une synthese de la theorie du circuit', Economies et Societes, Monnaie et Production, 7 (6),123-34. Lavoie, M. and M. Seccareccia (1989), 'Les idees revolutionnaires de Keynes en politique economique et Ie dedin du capitalisme rentier', in 'L'austerite: theories et politiques', ed. by A. Parguez, Economie appliquee, 1,47-70. Nell, E. (1988), Prosperity and Public Spending (London: Unwin Hyman). Parguez, A. (1984), 'La dynamique de la monnaie', Economies et Societlis, Monnaie et Production, I (4), 83-118. Parguez, A. (1986), 'Au coeur du circuit ou quelques reponses aux enigmes du circuit', Economies et Sociites, Monnaie et Production, 3 (8/9), 24--40. Parguez, A. (1989), 'Hayek et Keynes face a l'austerite', in O. Dostaler and D. Ethier (eds), Friedrich Hayek: Philosoph ie, Economie et Politique (Paris: Economica), 43-60. Parguez, A. (I 989b), 'Money and Financial Money Capital within a Keynesian Framework', in A. Barrere (ed.), Money, Credit and Prices in Keynesian Perspective (London: Macmillan), 3-15. Parguez, A. (1990), 'Keynesianism and Austerity', Economies et Sociitlis, Monnaie et Production, 7 (6), 107-122. Parker Foster, O. (1986), 'The Endogeneity of Money and Keynes's General Theory', Journal of Economic Issues, 20 (4), 953-68. Parker Foster, G. (1987), 'Financing Investment', Journal of Economic Issues, 21 (I). 101-112. Parker Foster, G. (1990), 'Keynes and Kalecki on Saving and Profit: Some Implications', Journal of Economic Issues, 24 (2»,415-422. Ranson, B. (1983), 'The Unrecognized Revolution in the Theory of Capital Formation', Journal of Economic Issues, 17 (4). Rousseas, S. (1986), Post-Keynesian Monetary Economics (New York: Sharpe). Seccareccia, M. (1984), 'The Fundamental Macroeconomic Link between Investment Activity, the Structure of Employment and Price Changes'. Economies et Sociites, Monnaie et Production, 18 (4),165-219. Seccareccia, M. (1985), 'The Role of Saving and Financial Acquisition'. Economies et Sociitlis, Monnaie et Production, 2 (8), 253-72. Terzi, A. (1989), 'Three Questions on Finance: Preliminary Remarks', Economie Appliquee, 2, 115-24. Wood, A. (1975), A Theory of Profits (Cambridge: Cambridge University Press). Wray, L.R. (1988), 'Profit Expectations and the Investment Saving Relation', Journal of Post Keynesian Economics, II (1),131-47. Wray, L.R. (1989), 'A Keynesian Presentation of the Relation between the Government Deficits, Investment, Saving and Growth', Journal of Economic Issues. 23 (4), 977-1002. Wray, L.R. (1990) Money and Credit in Capitalist Economies: The Endogenous Money Approach (Aldershot: Edward Elgar). Wray, L.R. (1991), 'Saving. Profit and Speculation in Capitalist Economies', Journal of Economics Issues, 25 (4). 7 Payment Systems and Dynamics in a Monetary Economy Jean Cartelier The purpose of this chapter is to elaborate some implications of a very simple idea about money. Money fundamentally allows economic agents to carry out either totally (in equilibrium) or partially (in disequilibrium) their desired transactions. A monetary economy differs from the barter economy described by orthodox theory, not only because it includes a commonly accepted means of payment but also because the acceptance of this means of payment is not subject to the establishment of an equilibrium. In other words, money makes it possible to deal with the actual disequilibrium positions of an economy (and not the fictitious ones described in tatonnement models). Let us assume that the market process drives the economy toward an equilibrium at time T. The actual time-path of the economy is made of some disequilibrium positions (before 1) and of one equilibrium position (at 1). The question is now: is the final equilibrium dependent on the actual time-path? Most economists, using common sense, would answer: yes. However, most economic models are built as if the answer were: no! They rule out path-effects and implicitly follow the tdtonnement paradigm. This disparity between economic theory and common knowledge about the market is the result of contradictions between some fundamental assumptions of economic theory (e.g. general eqUilibrium theory) and what is required to conceive of effective disequilibria (whatever the special assumptions may be: perfect or imperfect competition, exchange or production economy, etc.). The study of non-tatonnement models, i.e. the only ones dealing with effective transactions out of equilibrium, clearly shows a double drawback; either - - they treat a barter economy with a resulting contradiction between the market efficiency and barter (F. Fisher, 1983) or between market efficiency and the principle of voluntary exchange (d' Autume, 1985); or, they consider a monetary economy, but because of the general properties of Arrow-Debreu models (Ostroy, 1987), there is no possible justification for the detention of money. 200 Jean Cartelier 201 It is likely that this double drawback cannot be overcome without amending some basic assumptions of economic theory. A simple consideration of organization of transactions in disequilibrium suggests two assumptions that should be added to ordinary models: 1. 2. a nominal unit of account (say, the dollar), and a principle of settlement for the balances of payment (or a lender of last resort). These additional assumptions are detailed in the first section of this chapter. The concept of the payment system is put forth. Moreover, the twin characteristics of an actual disequilibrium in a monetary economy I are detailed: - a market disequilibrium (a non-zero aggregated excess demand) a monetary or individual disequilibrium (an excess or a deficit between receipts and expenditures). The settlement or the postponement of these balances is at the heart ofa monetary economy's dynamics and plays a decisive role in determining the stationary dynamic equilibrium of the economy. Several different payment systems· can be imagined, each with its own rules. Several examples are presented in the second section. It is shown that the form of social wealth differs according to the accepted rules of the payment system; a simple merchant economy and a capitalist economy are described by very different payment matrices. The third section deals with a monetary economy with a wage relationship. Some monetary foundations for the Keynes-Kalecki approach are suggested. A simple model is presented which bears some resemblances with IS-LM. But, because of the payment system, this model exhibits very specific properties. The concept of dynamic equilibrium, related to the payment system, is opposed to that of static equilibrium. Comparison shows that the existence and the stability of dynamic and static equilibria are subject to very different conditions. Moreover, the locus of stationary dynamic equilibria appears to be a generalization of the static one. 1 THE DRAWBACKS OF BARTER AND THE CONCEPT OF THE PAYMENT SYSTEM The Rebuttal of Barter and the Organization of Transactions The reason for considering the barter model as irrelevant for the study of a decentralized economy is that barter violates the assumption of market efficiency. It is 202 Payment Systems and Dynamics in a Monetary Economy easy to show that. even in equilibrium. barter does not necessarily allow all feasible transactions to take place (Starr, 1976). This result a fortiori holds in disequilibrium (Fisher, 1983). Accordingly, we have to assume that goods do not buy goods, as Clower (1967) pointed out. This assumption is, at the same time, a necessary condition for the existence of markets in the ordinary sense of the word. If goods could buy goods, there would be n(n - 1)/2 trading-posts and each commodity would be traded in (n - 1) different places according to the commodity used as its equivalent. Although they would supply and demand the same commodity, agents would not meet if they went to different trading-posts (absence of a double coincidence of wants). Such a situation is exemplified by the following excess demand matrix (prices are all equal to I, commodities are in the rows and agents in the columns): z= [~o ~1 ~I ~ 1 0 -1 0 1 -1 0 1 No transaction can be carried out although all desired transactions are mutually compatible (equilibrium) since there is at most one agent at each trading-post. Assuming that goods do not buy goods, we can imagine a unique market for each commodity where everyone has to go in order to realize a transaction in that commodity. The impossibility of barter proves that the mere assumptions of the existence of agents and of commodities are not sufficient conditions for a theory of market economy. Something more than the auctioneer is needed; he is relevant only in the case of perfect competition. A special institution is needed regardless of the type of competition. The purpose of such an institution is to allow the realization of desired transactions. A minimal assumption seems to be that of the existence of a unit of account ($) in which prices are expressed. This assumption is minimal because it is merely the rebuttal of barter's resulting definition of prices as a ratio between commodities. This assumption will be completed where it is insufficient to account for the existence of efficient transactions in disequilibrium. Concerning the organization of transactions, the main assumption here is the same as that in the non-tatonnement models: desired transactions may be carried out independently of their mutual compatibility. All feasible transactions, that is respecting the principle of voluntary exchange, are realizable dependent only on the rules of the payment system. The question of the payment system is thus central: the actual disequilibrium position of the economy cannot be assessed unless the organization of payment is known. Jean Cartelier 203 Before making this organization explicit it is necessary to add some brief considerations about the concept of time adopted in this chapter. Although continuous time is easier to deal with from a mathematical point of view, discrete time is more in accordance with the view adopted here. Continuous time would imply the idea of an infinite velocity of circulation, a very inappropriate starting point for the study of transactions. Thus, discrete time is adopted for the description of market functioning. This market period is not decomposable in subperiods. 2 There are no sequential processes inside the period. Purchase and sale orders are decided at the beginning of the period and are irrevocabLe. Whether they are realized at the beginning or at the end of the market is insignificant. They cannot be modified during the market. In any case, agents have to close their accounts only at the end of the market. In this context, it seems convenient to assume circulation velocity for the means of payment equal to unity, since every agent has to pay for his or her purchases, only at the end of the market. The above assumptions emphasize an essential feature of a monetary economy: the transactions carried out by an agent in different markets are independent of each other. This point deserves elaboration. Clearly, desired transactions are not independent because they result from a unique and. rational decision (maximization of profit or utility under constraints) as do the orders at the beginning of markets, which are similarly non-independent for this reason. But this interdependence does not hold as far as the realization of the transactions is concerned. This is a very important difference between a barter and a monetary economy. In barter, the 'sale' of com is not independent of the 'purchase' of barley, because barley is the 'price' of com. In a monetary economy, it is quite possible that an agent succeeds in selling (or purchasing) com but is unable to buy (or sell) barley because of a shortage. Purchases and sales, for a given agent, are independent because they depend on conditions which are specific to the various markets. 3 Two necessary and sufficient conditions for a desired transaction to be effective are the coexistence of purchases and saLes orders (voLuntary exchange) backed by an appropriate amount of means of payment (active orders in the terminology of Arrow and Hahn, 1971). Markets are assumed to be efficient: they are closed as soon as there are no longer any active orders of the opposite sign to be carried out. The dynamic process of a monetary economy is conceived of as a sequence of markets where desired transactions are partially or totally carried out according to the rules of a payment system. The Concept of the Payment System: Circulation and Balances Is the unit of account a sufficient basis for a system of payments? How can agents perform purchases and sales? 204 Payment Systems and Dynamics in a Monetary Economy Since desired transactions respect budgetary constraints (desired sales are assumed to equal desired purchases), each agent could accept the IOUs (expressed in dollars) of other agents as a means of payment. Reciprocity would guarantee the ability of each agent to carry out his or her purchases. Such a system of payment would be minimal since it implies only the unit of account. But, obviously, a necessary condition for such a system to be feasible is that prices must be at equilibrium. Equilibrium only ensures the perfect clearing between the different IOUs. In an economy where transactions are not allowed until equilibrium is reached, the system of payment could be minimal in this sense. In contrast to barter, there would never be any hindrance to the realization of desired transactions. IOUs are related to agents and not to markets (or trading-posts) so that there is never any lack in the double coincidence of wants. In the case of the matrix Z above, the mutual acceptance of IOUs suffices to overcome the drawbacks of barter. In a tatonnement economy, a minimal system of payment is thus perfectly appropriate. But when transactions are allowed out of equilibrium the story is quite different. No seller can be sure that the purchaser will be able, at the end of the market, to offset his or her IOUs with those received from other agents. It is quite possible that the purchaser will have, at the end of the market, more liabilities than assets (his or her sales being more constrained than his or her purchases), in which case, the seller runs the risk of not being paid. Consequently, he has reason to refuse the would-be purchaser's IOUs. But the same applies to all parties and thus no transaction is realized. It is worth noting that the problems encountered here are not those of barter. In a barter economy some transactions could take place (in every case of double coincidence of wants) whereas, in our simple monetary economy, no sale (and no purchase) can be carried out (unless people accept the risk of non-payment). An economy with a unit of account differs from the barter kind in that disequilibrium there takes a very specific form. In both economies, there is a market disequilibrium expressed by a non-zero excess demand. But, in a barter economy, no agent is allowed to 'buy' more than he or she 'sells'. In non-tatonnement models, for example, the exchanges (and not only the notional transactions) are constrained by resources so that no individual disequilibrium between payments is conceivable. This is in contrast to an economy with a unit of account. If all possible transactions were realized (through the mutual acceptance of IOUs) there would be, in addition to market disequilibria, some individual disequilibria expressed by negative or positive balances between 10Us. These non-zero balances are specifically responsible for the failure of a minimal payment system. The mere existence of a unit of account, responsible for the latter form of disequilibrium, does not suffice to make a minimal payment system feasible. The rules by which individual actions are coordinated are not independent of ques- Jean Cartelier 205 tions of payment: an IOU payment system, suitable for a tdtonnement economy, is not appropriate to a non-tdtonnement model. Again, the fact that transactions are allowed out of eqUilibrium is crucial. Can this difficulty be surmounted through the introduction of agents specialized in the trading of IOUs? Let us see briefly why the answer is negative. Let Zd be the matrix of desired transactions (at unit prices): 0 01 +1 -1 o +2 -1 0 o 0 +1-1 . -I -1 0 +1 Zd= [ The non-acceptance of IOUs, due to the risk of non-payment, prevents any transaction from being carried out. The economy is in autarchy, each agent suffering from utility losses. If, for extrinsic reasons, some transactions were possible, agents would be better off. Consequently, they are ready to pay for this option. There is now room for specialized agents who accept the risk of non-payment against a remuneration. Let us· call these private agents trading in IOUs banks. Would-be purchasers have to exchange IOUs with a bank (swap). The bank's IOU, being less risky than an individual one (diversification of risk), may be accepted as a means of payment and purchases could take place. In this way, the intervention of the banks ensures the market efficiency. In the case of matrix Zd (if agent 4 buys commodity 4 from agent 2) markets would close with the following matrix Zm of non-realized transactions: However, bank intervention does not really solve the problem raised by individual disequilibria. There is no reason for a unique bank. Banks, here, are nothing other than partial aggregations of individual transactions. If, on the whole, the customers of a bank experience an excess of purchases over sales, the bank would be in a deficit and could not meet its liabilities to other banks. The problem of individual disequilibria is still pending at a higher level. In the case above, if two banks A and B act respectively for even and odd agents, bank A would be in debt of 1 dollar to bank B.4 Bank A would be unable to pay this debt. Consequently, its IOUs would not be accepted as a means of payment and we are back to the original problem. 206 Payment Systems and Dynamics in a Monetary Economy With or without banks. the partial realization of transactions in disequilibrium generally prevents the clearing between IOUs. thus making a minimaL system of payment unpracticable. The settLement of non-zero balances (the specific form of disequiLibrium in a monetary economy) appears to be a crucial problem for any payment system. The Notion of a System of Payment: Non-zero Balances and Finance Let us now consider the closure of the markets in which some agents are unable to meet their deficits. Such a situation contradicts the fundamental principle of equivalence in exchange, because a deficit means that the amount of payments does not equal the value of the commodities bought in the market. s The sanction of this contradiction is bankruptcy. A bankrupt is deprived of any access to the market in the future and his or her commodities are sold according to a special procedure, which differs from the ordinary market. Nothing ensures that creditors will obtain the exact value of their claims. In brief, it is as if the realized purchases and sales, which led to this situation, were not validated. As far as I know, the dynamics of an economy experiencing bankruptcies has not yet been formalized. Neither is this question treated herein but it is useful to keep it in mind and to be aware of the narrow limits in which a monetary economy evolves. In order to rule out bankruptcies it is necessary to specify how deficits can be settled or postponed to future periods. Let us begin with the question of settlement. 6 Suppose that there exists an official procedure for the settlement of deficits (e.g. through gold coins of a certain type). In this case, if deficit agents are able to pay the liabilities resulting from the working of the markets, the period is closed. Agents leave the markets and, maybe, consume their allocations. As a result of the disequilibrium, their wealth after the market is not equal to that which they expected before it. But it should be noted that agents no longer have reciprocal obligations. 7 Economic time is over. In an economy with no durable commodities. a balance settlement would make successive markets intrinsically independent. 8 No store of vaLue wouLd link successive periods and a true dynamics is not conceivable. Balance settlements are to be interpreted as restoring the principle of equivalence in exchange in case of disequilibrium. A deficit agent has acquired more commodities in the market than allowed by this equivalence. Balance settlement is the means of filling this gap: the deficit agent uses part of his or her final allocation (his or her wealth) to discharge his or her debt. Let us call this condition the monetary constraint. This constraint works according to the rules of the payment system (see section 2 below). The settLement of all balances amounts to validating the disequilibrium of sales and purchases realized in the markets and to making the monetary constraint effective. Jean Cartelier 207 Before dealing with the diverse modes of settlements we have to consider whether it is possible to avoid this monetary constraint. Since the absolute value of aggregate deficits is identical to that of aggregate excesses, it is always possible, if agents accept it, to postpone payment of the debt to the next period. Excess and deficit agents decide, if possible, the conditions of this operation. But the main problem is determining the conditions of such an agreement between debtors and creditors. This coordination problem is quite different from that raised by ordinary circulation operations (purchases and sales) because deficits and excesses are unexpected and not desired. They are neither the solution of a maximization procedure nor are they rational. Two different cases must be considered according to the opportunities open to agents: postponement may be an alternative to the payment of the balances or may be the only way to avoid bankruptcy. In the first case, the agents have the choice. They compare the present balance with an amount of dollars available in the next period. 9 There is then a need for a coordination of individual actions, which takes the form of a very special market. The specificity of this market is that it can never be in equilibrium or in disequilibrium. Whatever its outcome may be (depending on interest rates or the prices of the private IOUs to be paid in the next period), there will be no balances remaining even if the desired transactions on this market are not realized. The effective transactions in it are the result of an effective disequilibrium in the ordinary markets. The 'market' for the postponement of balances is logically posterior to the ordinary markets. JO Its outcome is self-validated in the following sense: if agent i does not succeed in postponing her deficit she has to pay it immediately (which is always assumed possible). No individual disequilibrium, no non-zero balance arises as a consequence of the working of this 'market'. The function of this 'market' is to elucidate the conditions required for closure of the (ordinary) markets. Between the two extremes (either the postponement or the settlement of all balances) all intermediary cases are possible. The monetary constraint is allowed to take any value between zero and unity. Let us call finance the operations by which balances, rather than being paid in cash, II are transferred to the next period. The finance 'market' works under special conditions. Its very existence is the result of a disequilibrium in the economy, and its outcome (the degree of monetary constraint) represents a bet on the future. There is no objective basis for the evaluation of the ability of agents to meet their obligations in future periods: the awareness of disequilibrium may generate perverse expectations and mimetic phenomena. 12 A social (or conventional) evaluation of the future is necessary for the smooth working of the finance 'market'. In disequilibrium, the first role to be played by a monetary institution is to establish 'the' rate of interest as a basis for a decentralized determination of the different individual rates (spreads) or prices of IOUs. This function cannot be compared with that of the auctioneer in the 208 Payment Systems and Dynamics in a Monetary Economy ordinary markets. The auctioneer is required only on the basis of a special assumption (perfect competition), whereas the determination of the rate of interest is the general result of a disequilibrium in an economy with a unit of account (regardless of the kind of competition). Moreover, finance (and the rate of interest) is relative to agents and not to markets. This seems a good reason to speak of a monetary institution completely distinct from the orthodox auctioneer. In the present approach. the rate of interest is necessarily exogenous because the quantity of means of payment is endogenous. This is in sharp contrast to orthodox theory in which money is exogenous and the rate of interest determined in ordinary markets as other economic magnitudes. 13 In the second case (certain agents cannot settle their deficits), the viability of the economy is at stake. A partial accord between agents may not be enough to avoid bankruptcies. The only suitable outcome for the finance market is the postponement of all balances (no monetary constraint). But such a result is never guaranteed. It may happen that, for a given rate of interest, creditors refuse the transfer of their IOUs. This provides the monetary institution with a second role: the lender of last resort which stands in for reluctant creditors. Generally speaking, the monetary institution is the name given to all the tasks linked to the realization of transactions out of equilibrium. A monetary institution is defined by the following three functions: 1. 2. 3. establishing the unit of account (a necessary and sufficient role in a tatonnement economy); setting the rate of interest (a necessary and sufficient role, jointly with the definition of the unit of account, in a non-tatonnement economy where viability is satisfied); lender of last resort (a necessary role in general). These generalities are meant to aid the understanding of the following analysis and are summed up in Table 7.1. 2. PA YMENT SYSTEMS AND BALANCE SETTLEMENT One essential point has not yet been dealt with: the different ways of settling balances and their consequences on the wealth of agents. This, however, requires further details about the concept of the payment system. The settlement question is closely related to the definition of the unit of account and to the conditions necessary for obtaining some units of account (mintage). 14 As soon as prices are expressed in money account (dollars), the realization of a transaction requires that the buyer transfers to the seller the amount of dollars corresponding to the price. Replacing this transfer by a promise of transfer at a 209 Jean Cartelier Table 7.1 Disequilibrium Equilibrium Complete realization of desired transactions not possible Complete realization of desired transactions possible Barter Market efficiency not warranted Market efficiency not warranted Unit of account Non-sufficient condition for the realization of transactions (autarchy) (may be worse than barter) Necessary and sufficient condition for the realization of transactions Market efficiency warranted No balances, no finance Unit of account Rate of interest and lender of last resort Sufficient conditions for the mutual acceptance of IOUs. Market efficiency warranted. Non-zero balances Exogenous rate of interest: finance If non-viability: lender of last resort later time is possible but, as explained above, only when the monetary constraint is zero (at eqUilibrium or when all the balances are transferred to the next period). In general, some transfers of dollars are required. It is thus necessary to explain the way by which dollars are introduced into the economy. The definition of the unit of account is more or less explicitly related to the 'content' of a dollar which, in tum, provides an idea of how agents may go about acquiring a given amount of dollars. Schematically, the conventions or rules of the game which 'institutionalize' money solve two distinct, but related, problems: 1. 2. How can an agent acquire units of account (if not from a payment by other agents)? How can these units of account be transferred from one agent to another? The first question is that of the mintage, the second one that of circulation proper. Any payment system may be derived from answers to these questions. 210 Payment Systems and Dynamics in a Monetary Economy The Metallic System Let us consider the metallic system as an elementary but essential example (since it is an implicit reference for most theories of money). This system can be summarized by the following rules: IS (a) (b) (c) (d) There exists a commodity called 'gold' which is known as such by all agents; The unit of account (the dollar) is defined by some determined weight of gold, which is to say that gold has a legal price in dollars (if $1 is defined by 0.05 g of gold, the legal price of gold is $20 per g); The transfer of gold coins (of a predetermined type) is the only way of transferring units of account from one agent to another; Minting is free of charge at the Mint and it is possible to melt down coins. 16 Rule (c) answers the question of circulation and rule (d) that of mintage. These four rules form a payment system which, in tum, constitutes the economic space in which economic relations between agents are possible. The outcome of the circulation is self-paid balances and a recomposition of wealth. This point deserves further comment. If b is the vector of individual gold endowments and cr the legal price of gold, the vector of the maximum number of coins which can be obtained from the Mint is crb (in $). If the velocity of circulation is taken equal to unity, the maximum vector expenditures in the market is also =(e 11+e)w=mw (4) where m =(Ell + E) is the mark-up. As a result, price does not depend on time.25 Desired level of activity (for employment) is then: (5) Jean Cartelier 221 For these particular (rational) values expected profit is: n U) *e = (m -1)wQ *(1) -[I + i(l_I)]Du_1) (6) The demand for accumulation 26 is assumed to be independent of the current state of the market. A(t)* =A* (7) A * denotes the long-term expectations of entrepreneurs. The independence of long-term anticipations is true to the Keynesian view, according to which the market is a useful guide for correcting short-term expectations but not for longterm ones. 27 The desire for accumulation does influence the current state of the economy, but the reverse is not true. For the sake of simplicity, accumulation is also assumed to be independent from the rate of interest. This difference with IS-LM is not an essential one. More generally, no rational intertemporal arbitrage is assumed. Entrepreneurs' aims are satisfied if the accumulation takes place at the desired level without any intervention on the part of the lender of last resort. Recourse to the lender of last resort means thatthe viability of the economy is at stake and that entrepreneurs run the risk of bankruptcy. Their future activity then depends on an arbitrary decision, and not on the normal course of the market. Entrepreneurs desire to repay, at the end of the period, all the funds they have borrowed, i.e. the value of current transactions A*p* + wN(t) * + [l + i(l_I)]D(I)' with receipts from the sale of the commodity or the temporary sale of bonds. Desired sale of bonds is thus equal to he expected balance between current receipts and outlays: (8) If this amount of bonds were effectively sold, entrepreneurs would respect the monetary constraint. If not, the lender of last resort would enter the stage. Let us now consider wage-earners' behaviour. At each period, wage earners expect to get the following receipts: - wages equal to wfi., where fi. is the total quantity of labour they are ready to supply at the current wage, repayment of bonds purchased during the preceding period with interest: U + 1(1-1)]8(1-1), liquid assets at the bank L(I_I) (necessarily equal to E(I_I) interest collected by the bank on the preceding period's loans of the [~,_I) EO_I)] which is assumed to be spend by the bank as wages. 222 Payment Systems and Dynamics in a Monetary Economy On the whole, the wage-earners' expected receipts of R;,) are: R~t) = w!:!... + [1 + !It-I) ]B(I_I) =[1 + !It-I) ]E(I_I) (9) The demand for the commodity is: (10) Ct,) = bRtt) / p' whereas the demand for liquidity (issued by the bank) is: 4.t) = fR(:) - gi(l) (11) Because of the budgetary constraint, the demand for bonds is: B(:) = (1 - b - f)Rtt) + gi(t) (12) Let us take a global view of the desired transactions of this economy at the beginning of period t (in Table 7.4). In order to determine the outcome of the market in disequilibrium it is necessary to recall some of the assumptions made above. Table 7.4 Expenditures WQ(I~ = w N;.,) A"p" [I + i(I_I)]B(t_I) [1 + i(t_I)lE (I-I) Entrepreneurs Wage Commodity Bond Liquidity Receipts " • e Q(lp Ap -tr(t) Total circulation Expenditures Wage earners Wage Commodity Bond Liquidity Receipts , wN + i(l_I)E(I_I) Total Lender of Last Resort Expenditures Total circulation Income Wage Receipts '1'(1) i(l_I)E(I_I) Jean Cartelier 223 Entrepreneurs hold the initiative for circulation because they have exclusive access to the means of circulation. The wage-earners' realized transactions are subject to the transactions desired by entrepreneurs. This asymmetry can be seen on the labour market where the demand for labour N* determines the effective level of employment (it is assumed that if N* > l:!.. the demand for labour will be satisfied by recourse to external workers or to individuals taking this opportunity to enter the market). In the market for the commodity. when excess demand is positive. entrepreneurs are assumed to fill the gap by a negative stockpiling. A positive excess demand leads to an unexpected diminution of stock (the reverse holds if excess demand is negative). The report of a disequilibrium (N* i' l:!.) affects the wage-earners' behaviour. For the sake of simplicity. it is assumed that the parameters band f remain fixed and that they apply to the unexpected effective receipts. The actual demand for the commodity is thus: C(I) = bR(t/p* (13) whereas the actual demand for liquidity is: (14) L(t) = fR(t) - gi(t) Budgetary constraint implies the following demand for bonds: B(t) = (I - b - f) R(I) + gi(t) (15) This demand is always satisfied because bonds are issued on tap, which raises the following problem: the subscription for bonds may be greater than B~,), which implies a negative loan on the part of the lender of last resort (and a negative interest ending in the payment of a negative wage). In real economies this result is harmless because such a negative loan simply means commercial banks become less indebted to the central bank. But. in the present model. we have assumed. for the sake of tractability. that all debts must be repaid in one period (at each period financial stocks are equal to flows). It follows that it is impossible to rule out the difficulty mentioned above; this is the price paid for the compromise between realism and tractability. Realized transactions are reported in the accounts in Table 7.5. Let us briefly comment on Table 7.5. and notably its differences with Table 7.4 which immediately precedes it. The wage-earners' effective resources are R(t) which differ from Re,) uniquely for the unexpected level of employment. N - l:!.* is the involontary unemployment. Its consequence is to be seen in a proportional fall in the purchases of the commodity and in a change in financial behaviour. This. in tum. affects the situation of the entrepreneurs. They succeed in fulfilling their desired level of activity but not necessarily that of accumulation. although they spend A*P* (total effective circulation is equal to the desired one). 224 Payment Systems and Dynamics in a Monetary Economy Table 7.5 Entrepreneurs Expenditures Receipts [1 + i(t_I)]B(I_I) [I + i(t_I)]E (I_I) Wage Commodity Bond Liquidity /R(r)-gi(t) 'I'~t) Total circulation 'I'~t) wNit) A'p' A'p'+bR(I) ( l-b-j)R·(t)+gi(t) Wage earners Expenditures Receipts bR(t) ( I-b-j)R(t)+gi(t) L(I)=/R(t)-gi(l) Wage Commodity Bond Liquidity L(I_I)=E(I_I) R(I) Total R(t) wtv;t)+i(t-I)E(I_I) [1 + i(/_I)]B(I_I) Lender of last resort Expenditures 'I'~/) i(l_I)E(I_I) E(I) Receipts Total circulation Income Wage Loan '1'(/) i(/_l~(I_I) L(I) Effective accumulation now depends on the difference between current production Q(t) and current demand from wage earners. It may be superior, equal or inferior to desired accumulation A *. It can be seen below that the non-realization of desired accumulation amounts to a difference between expected and realized profitability (because p* is invariant). The financial position of entrepreneurs depends also on the behaviour of wage earners (specifically on their financial behaviour). A positive demand for liquidity means an intervention on the part of the lender of last resort since such a demand prevents entrepreneurs from covering their expenditures through the sale of the commodity and bonds alone. This demand appears as a specific and independent factor of disequilibrium (reflecting the influence of the rate of interest, as well as the other factors common to all markets). Figure 7.1 shows the realized circulation of our economy (flow-of-funds). Entrepreneurs consider the outcome of the market from two points of view: that of desired accumulation and that of their financial position. Jean Cartelier Transactions on commodity Transactions on bonds Bank finances 225 6. Assets goods [P*Q(t) L-b~t) 6. Liabilities 1ft} E(t} Figure 7.1 Flow of funds The realized accumulation is equal to the current production Q* (I) minus what is bought by wage earners [bR(/)]lp*. This quantity may be different from the desired one A*. We denote the unexpected accumulation by U. U(I) =Qtl) -[bR(I)]lp*-A* (16) Unexpected accumulation, evaluated at p*, is equal to the opposite of the unexpected profit (difference between realized and expected profit). Realized profit is the excess of sales over costs: (17) Equation (17) is simply the expression of the Kalecki Principle when there is wage-earners'saving. The difference between realized and expected profit, using (6) is: n(1) -n;:i =~~)p* + bR(1) - C4~)p* =-P*U(/) (18) 226 Payment Systems and Dynamics in a Monetary Economy The financial position desired may be described by a zero level of intervention of the lender of last resort (the bank). As it is clear from above, this is possible only if wage earners decide to hold all their saving as bonds. As soon as wage earners accumulate liquid assets (at the bank), entrepreneurs experience a deficit: their receipts from the sale of commodities and bonds are inferior to their expenditures. This deficit cannot be covered except by a loan from the lender of last resort. This loan E(I) is thus identical to L(t), allowing us to express the financial position of entrepreneurs by: E(t) = L(r) = fwR(t) - gi(t) (19) = fw0.~) + /[1 + i(I_I)]D(t_I) - gi(t) V(t) "* 0 and E(I) "* 0 are two independent expressions of disequilibrium. Entrepreneurs and the payment system will react to these disequilibria in the next period. Entrepreneurs will modify their short-term expectations according to the signs V(I) [or TI(t)- TI~,)1 and £(1). An unexpected positive accumulation reveals maladjustments between longrun expectations and the current level of activity. The latter will be adapted by a change of opposite sign. An unexpected intervention by the lender of last resort is a signal that the economy runs the risk of non-viability. Reproducing the economy depends on the willingness of the lender of last resort. Entrepreneurs will reduce their activity if E(ll > 0 and will augment it in the opposite case. Alternatively, the past intervention of the lender of last resort may make the payment system more restrictive, obliging entrepreneurs to reduce their activity. In any case, a positive £(Il will be associated, ceteris paribus, with a reduction in activity during the next period. To sum up, short-term expectations change according to: (20) with I) and '"Y > o. As a result, the evaluation of production is: 0..1+1) = 0..,) -aV(I) - {3E(t/p* (21) a =8 (mw)E and (3 =1 (mw)' Using the above expressions for gives: 0..1+1) V(tl and E(t) and rearranging the different terms = [1-a {1- (bl m)} - (3(f I m)]0.·r) + {(1 + i(t_I)/p}(ab - (3/)D(t-l) +aA + {3gi(t/p* (21a) Jean Cartelier 227 The total debt is detennined by the wage-earners' saving: D(I+I) = (1- b)w~~+1) + (1- b)(1 + i(t)D(t) (22) Let us define a new variable: I'{I+I) = {(I + i(t)lp}(cxb - f3f)D(t) (23) (which measures the dynamic effect of the existence of the debt on consumption). Replacing Q*(I+ I) with (2Ia) allows us to rewrite (22) as: D(I+I) =(1- b)w[l-cx {I- (blm)} - f3(f I m)]~:) (22a) + (I - b )(1 + i(t) )D(I) + (1- b )wl'{l) + (1- b )w[cxA + {3giu/ p*] Equations (21 a), (22a) and (23) form a linear dynamic system which, in matrix notation, becomes: Y(I+ I) (24) = HY(I) + h where: [l-cx{I-(blm)}-{3(flm)] 0 I H= ( (l-b)w[l-cx{1-(blm)}-{3(flm)] (l-b)(l+i(l) o {(1+i(I)lp*}(cxb-{3f) cxA + (3gi(1) I P * h= [ (1- b)W[~ + f3gi(t)] ] (l-b)w 0 1 Comments and Results Putting aside for the moment the problem of stability conditions, let us consider the stationary solution of this model. This solution is in fact the fonnal expression for dynamic equilibrium. It depends on the reaction coefficients of the entrepreneurs and, possibly, on the tightness of the credit rationing (according to the interpretation adopted for f3). Equilibrium here appears as the actual result of an effective market process (if stable) and not as the condition of mutual compatibility of desired transactions. Dynamic equilibrium is preferred over static equilibrium; this has been justified above by the fact that transactions are realized out of equilibrium thanks to the payment system. It is only when equilibrium is the prerequisite for the realization of transactions (tatonnement economies) that the two concepts are equi valent. The stationary solution (dynamic equilibrium) is, for a rate of interest constant over time (i(I) = i for all t): 228 Payment Systems and Dynamics in a Monetary Economy y** = [I - H]-Ih (25) This gives: Q** = [aA + J3gilp*][1 - (I - b) (I + i)]1 [at 1- blm) - (1 - b) (1 + i)] + J3(flm)] (26) with i ~ [ab(m - 1) + J3j]lam(l - b) D** = [aA + J3gilp*][(I-b)w]1 [at 1 - blm) - (1- b) (1 + i)} + J3(f7m)] F** = {(l + i)lp) (ab -13f) D** The steady-state ratio of the debt to the value of production is: 8** = D**lp*Q** = (l - b)lm[1 - (l - b) (1 + i)] (27) Let us compare these results to those of a static equilibrium analysis. Such an analysis would start with two equilibrium conditions V(I) = 0 (expected accumulation is perfectly achieved) and L(I) = E(I) = 0 (the lender of last resort does not intervene). Two relations between Q(I) and D(I_I) may be deduced from these conditions. From the first one we have: Q(I) = [b (1 + i)/w(m - b)] D(I_I) + Aplwl(m - b) (28) and from the second follows: Q(I) = -[(1 + i)flfw] D(I_I) + gilfw (29) As equilibrium Q(tl depends on DCt-1) in both cases, it would be interesting to determine the stationary value of the debt. The entrepreneurs' debt is the result of the wage-earners' saving: D(I) = (I - b)wQ(t) + (1 - b) (l + i) D(I_I) (30) For D* = DO-I) = D(I) all t, one must have: D* = (l - b)wQ/[1 - (l - b) (1 + i)] (31) Substituting D* for D(t-I) in (28) and (29) gives the steady-state values of Q satisfying the first equilibrium condition (denoted by Qu-) and the second (QE-) respectively. We get: Qu- = m [1 - (l - b) (1 + i)]AI[b(m - QE-=g [l-(1-b) (l + i)] ilfw \) -m(1- b)i] (32) (33) A steady-state full eqUilibrium of the economy means equality between Quand QE-' This condition holds only if a determinate relation exists between the desired level of accumulation (A *) and the rate of interest (0. Making Qu- and QE- equal gives the following condition: Jean Cartelier 229 ~UO ~£>O DU0 Figure 7.2 gm (1 - b)i 2 - gb (m - 1) i + fp A* = ° (34) Condition (34) is fulfilled for real values of the rate of interest only if the desired accumulation is not too high. More precisely, the following inequality is required: (35) Figure 7.2 is drawn for: m = 1,25; b = 0,75; w = 1;/= 0,1; g = 10 and A = 2.5 (> 2.25). If A * is greater than this value. the (static) steady-state equilibrium does not exist. On a diagram drawn on the plane (i, Q) the two curves Qu- and QE- do not intersect if A* > [gb 2(m - 1)2 [4fwm\1 - b)]. In this case, no static equilibrium is found. However, a dynamic full equilibrium Q** > exists for all values of i in the range of 0:::; i:::; b/(1 - b) and i:l- [ab(m - 1) + 13j]/am(l- b) as equation (26) proves. If A * is such that the above condition is strictly fulfilled, a pair of steady-state full static equilibria exist. 28 Figure 7.3 is drawn for the same parameters values as above, except that here A is equal to unity (condition (35) holds). ° * 230 Payment Systems and Dynamics in a Monetary Economy ~UO ~E>O IiU>O L-.JE 0, the equilibrium exists and is unique. Moreover, this model allows us to clarify an important point. In Keynes's view, there is no reason for long-term expectations to be in harmony with the rate of interest. The former are governed by animal spirits whereas the rate of interest equilibrates the preference for liquidity with the existing money stock. Here the same is true, mutatis mutandis. The monetary authority determines the Jean Cartelier 231 rate of interest without any knowledge about the desire for accumulation and, in general, the rate of interest is not a root of (34). Unless it can be proved that a rule exists for correcting the rate of interest in reaction to the observed disequilibria, such that one of the static equilibria may be reached, we must conclude that market coordination has failed. But, in contrast with IS-LM, this failure is manifest not through an inadequate equilibrium, but rather by an absence of equilibrium. As it is well known, IS-LM strictly respects the static general equilibrium method. No disequilibrium position can be imagined and actual economic magnitudes are determined by the simultaneous equilibrium of all markets (except the labour market). The model presented above is very different. Because the payment system allows for effective disequilibria, the static equilibrium method is no longer relevant. Actual economic magnitudes are determined according to the rules of the payment system. The means of circulation are endogenous. 3o The quantity of the means of payment is closely related to the degree to which entrepreneurs realize their desired transactions (with or without credit rationing). The monetary authority determines the rate of interest (and not the quantity of money) which, in tum, governs the mode of closure of the disequilibrium (postponement of balances settlement with the possibility of intervention on the part of the lender of last resort). As a result, the actual position of the economy may be somewhere out of the curves. Even if the rate of interest were pegged at an equilibrium level (if one exists), nothing guarantees that this equilibrium would be acceptable for wage earnersY The desired level of employment l:i. (or 12) may lie anywhere above or below one of the static equilibria. But wage earners cannot directly determine the level of activity. They cannot carry out their desired transactions on their own account since they have no access to the means of payment.32 Their market power is conditioned by the realization of the entrepreneurs' transactions. The influence wage earners exert on economic activity is only indirect and measured by the degree to which entrepreneurs realize their expectations. Let us now compare the static analysis of the model with the stationary solution. It is not surprising that the two static equilibria, if they exist, are also possible dynamic equilibria of the model. In Figure 7.3, the Q** curve (drawn for a =13 = I) intersects the Qu- and Qe- curves at static equilibrium points. It is easy to see that substituting A* (given by (34) for any value of i) in Q** (known by (26» and Qu(known by (32» allows one to get the value, gi[1 - (I - b) (1 + i)]/fw, which is equal to Qe*. In this very special case, the economy is in full equilibrium. Not only do entrepreneurs realize their desired transactions exactly but, as in general equilibrium theory, money is not held: in an economy where an asset with a positive interest is available, there is no room in equilibrium for money. There is no rational reason to hold money in a stationary equilibrium (no uncertainty is conceivable in this case) if it is possible to transfer purchasing power from one period to another through a store of value with a positive rate of return. 232 Payment Systems and Dynamics in a Monetary Economy But confonnity to general equilibrium analysis is confined to the two special cases above. In general, condition (34) is not fulfilled. The economy is in an effective disequilibrium position. An adjustment process takes place through time according to equation (24). This dynamic evolution is irreversible since transactions are effective. The 'history' of the economy depends on the reaction coefficients a and {3. As a result,. the process, if stable, will converge toward a position which is path-dependent and which depends on a and {3 as shown by (26). This dynamic equilibrium Q** differs greatly from the static one: it exists and is unique for all values of i in the range of 0 ~ i < bl(} - b) (except for i:F [ab(m - I) + (3f]lam(l - b); (b) entrepreneurs do not realize all their desired transactions and, in general, the lender of last resort has to intervene to ensure the viability of the economy; (c) consequently, money is present at equilibrium, although there exists an asset yielding a positive return, wage earners prefer to hold money (liability of the lender of last resort). (a) In general, the dynamic equilibrium (defined as the stationary solution of the adjustment process) in an economy with a payment system differs radically from the static equilibrium, if it exists, of the same economy in which transactions are realized only at equilibrium. As a matter of act, static equilibria appear at best to be special cases of a model of effective dynamics. These special cases concern: - either an absence of reaction of the entrepreneurs and of the payment system to the monetary aspect of disequilibrium ({3 = 0); if so, (26) gives the first condition of static equilibrium (32): Q(~=O)** - = m[1 - (I - b) (l + i)] A/[b(m - I) - m(1 - b)i] = Qu- (36) or an absence of reaction of entrepreneurs to non-desired accumulation (a = 0); here (26) is equivalent to the second condition of static equilibrium (33): Q(a=O)** =g[l- (I - b) (I + i)]iifw = Q(E)- (37) Besides these special cases, the dynamic equilibrium positions of the economy cannot be analysed by the static equilibrium method. Because it fails to take into account market functioning in disequilibrium, and because it ignores the role of the payment system, traditional orthodox theory cannot deal with the most common situations of an ordinary economy and is limited to special cases. The specificity of dynamic equilibrium is thus to be found in the view adopted about money and credit: it is only in an economy with a system of payment (lender of last resort included) that transactions in disequilibrium are conceivable; they 233 Jean Cartelier lead to path-dependent steady-state situations unknown to traditional static equilibrium theory. Let us now consider the stability of the above model in order to check that these dynamic equilibrium properties are not cancelled out by too restrive stability conditions. Asymptotic stability is ensured if eigenvalues of H, be they real or complex, are inferior to unity in absolute value. It will be convenient here to use the condition according to which, if all elements of H are positive, the leading principal minors of the matrix (I - H) must be positive. 33 To ensure that all elements of H are positive (~ 0) we shall assume: (a) (b) (c) a [1 - (blm)] + f3(jlm) ~ I which amounts to saying that the reaction coefficients are not too high a/f3;;;' fib which means that adjustment to an unexpected accumulation (real effect) must not be too low in relation to the reaction to an unexpected debt to the bank (finance effect) i < bl(I - b) The last assumption is not very restrictive because, for i = bl(l - b) Q** Now, the necessary and sufficient conditions for stability are: = O. I - [I - a{ 1 - (blm)} - f3(jlm)] > 0 which is guaranteed by assumption (a) 1 - [I - a { 1 - (b 1m)) - f3(jlm)] 0 I. 2. -(1- b)w{1 - [1 - all - >0 (blm)) - f3(jlm)]) b(l + i) - i which is the result of assumptions (a) and (c) 3. [TJ-HJ>O This condition is equivalent to: [a{l - (blm)} + (3(jlm)] {b(l + i) - i) > [( 1 - b) (l + i)lm](ab - (3f) or to: m{ 1-[1 - b) (l + i)] > b - ({3la)f which is ensured by assumptions (a), (b) and (c) for i < [ab(m-I) + (3f]lam(l - b) (which amounts to Q** > 0 according to (26)]. For plausible values of the parameters, it is only for strong reaction coefficients that the economy is unstable (not a very surprising result).34 On the whole, stability conditions do not seem to impair the conclusions above. If stability conditions are satisfied, the model exhibits four transitory regimes of effective disequilibrium. They are summarized in Figure 7.4. Payment Systems and Dynamics in a Monetary Economy 234 U < 0 Recovery E<0 U >0 C · . E > 0 rlStS U < 0 Expansion E >0 U > 0 Recession E<0 0.0764 5.6 32.4 50 Figure 7.4 Two regimes are characterized by real and finance effects working in the same direction. In traditional business cycle analysis these regimes correspond to recovery and crisis (increase or fall o( activity induced by real and financial factors). Two other regimes result from the opposition between real and financial factors. Supposing that the real effect is stronger, situations of expansion are described by the extreme areas between QE- and Qu- (above the higher and below the lower static equilibrium points). Accordingly, situations of recession correspond to the inner area between QE- and Qu-. Notes I. 2. 3. See Benetti and CarteHer (1990). Schmitt (1984). This conception differs radically from the orthodox one which can be summed up by the following argument: either transactions take place simultaneously and money is not necessary; or transactions are sequential and a sine qua non condition for a means of payment is a store of value (Cass and Shell, 1980). Jean Cartelier 4. 5. 235 This argument misses an essential point: the separation between the sales and the purchases is not only a question of time but also one of market coordination. Note that, even in disequilibrium, the algebraic sum of balances is zero. For each agent, desired sales are equal to desired purchases. If some of these sales are not realized, the agent suffers from a deficit equal to the value of the nonrealized sales. But, because of Walras's Law, there must be some equivalent nonrealized purchases on other markets and, as a consequence, some agents experience an equivalent excess of sales over purchases. In the usual diagram for a two-commodity economy, the individual would be out of his or her budgetary constraint: = x initial endownment x* =desired allocation x- = actual allocation Commodity 2 Commodity 1 L-...l---L....I.---". . . 6. 7. 8. 9. 10. II. 12. 13. 14. 15. 16. 17. 18. In a barter economy, x- necessarily lies within the budgetary constraint. See Benetti and Cartelier (1980) and Aglietta and Orl~an (1982). For the sake of simplicity, we rule out any desired intertemporal transactions in order .to concentrate upon the consequence of effective disequilibrium alone. A new market period would arise from a decision of the auctioneer (non-tatonnement with consumption allowed only at equilibrium) or from new endownment from Heaven. The arbitrage is restricted to two terms: liquidity or future payment. As in Keynes's General Theory, commodities or real capital are not concerned. This simply means that agents are aware of their situation resulting from the working of the ordinary markets when they choose between the payment and the postponement of balances. As desired intertemporal transactions are excluded (for simplicity) finance denotes only the financial transactions which take place as a consequence of an actual disequilibrium in ordinary markets. In the real world it is often difficult to make a clear distinction between ordinary financial transactions and finance in the sense it is used here. See Orl~an (1987). Keynes's General theory included. The endogeneity of money has been emphasized by Kaldor (1982), Moore (1979, 1988) and others. As it is used here, this term refers specifically to the definition given to the French term monnayage in Benetti and Cartelier (1980). It does not imply an exact reference to the history of metallic money even if it is in some sense a conceptual generalization of it. Cleai'ly, what follows is only a 'stylized' description common to several historical systems but not true to anyone in particular. A theoretical view very close to the present approach has proven to be very fruitful in economic history (BoyerXambeu, Deleplace and Gillard, 1994; Deleplace, 1989). In most historical systems melting down of gold coins was forbidden and some seignorage was due. See Tooke (1844). Consumption credit is not taken into consideration. 236 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. Payment Systems and Dynamics in a Monetary Economy Marxian theory of the value of labor power is a striking (but not coherent) combination of both views. In Benetti and Cartelier (1980) the French expression soumission monetaire is used in this sense. Kalecki (1943). As a consequence, such a theory allows us to account for the conventional character of most economic phenomena (self-fulfilling prophecies included). The analysis in terms of dynamic equilibrium does not necessarily require the technique of nonlinear dynamics (very useful for sophisticated models). It can be applied to textbook models. Equation (2) describes how individual entrepreneurs perceive demand: a change in the individual price is perceived as changing the entrepreneur's market share; this elasticity is not that of the aggregate demand. This explains why E is different from -I, which is the price elasticity of the aggregate demand adopted below. This is the obvious consequence of the invariance of wand of E. Such an hypothesis is arbitrary but does not imply that markets do not work satisfactorily. The capacity effect of accumulation is neglected in the present version of the model. As a consequence, accumulation and entrepreneurs' consumption are not distinguished. This treatment usual in IS-LM macroeconomics, is highly questionable. Keynes (1936) chapter 12. If condition (35) is fulfilled as an equality, then the equilibrium is unique. Introducing the rate of interest into the determination of A'" would not alter this result. The reader can see that replacing A '" by A - e i(l) where A is exogenous, does not change the main result. See Kaldor (1982), Moore (1979; 1988). Let us assume that the monetary authority applied an efficient rule to find one of the two eqUilibrium rates. It is entirely possible that this rule leads to the 'low' equilibrium even if the 'high' is the only one which ensures full employment. Consumption credit would not Significantly change the situation. At best, wage earners could choose between incomes available in different periods but, the Kalecki Principle would still apply: incomes of different periods still be directly determined by entrepreneurs. Gandolfo (1985), p. 138. For the values of the parameters used in the text and for a = /3, the coefficients should be more than 2 to generate instability. References Aglietta, M. and A. Orl~an (1982), La violence de la monnaie (Paris: Presses Universitaires de France). Arrow, K. and F. Hahn (1971), General Competitive Analysis (San Francisco: Holden-Day). d' Autume, A. (1985), Monnaie. croissance et desequilibre (Paris: Economica). Benetti, C. and J. Cartelier (1980), Marchands. salariat et capitalistes (Paris: Maspero). Benetti, C. and J. Cartelier (1990), 'Monnaie et formation des grandeurs economiques', in J. Cartelier (ed.), La/ormation des grandeurs economiques (Paris: Presses Universitares de France). Boyer-Xambeu, M.-T., G. Deleplace and L. Gillard (1994), Private Money and Public Currencies. The 16th Century Challenge (Armonk, NY: M.E. Sharpe). Jean Cartelier 237 Cass, D. and K. Shell (1980), 'In Defense of a Basic Approach', in 1. Kareken and N. Wallace (eds), Models of Monetary Economies (Minneapolis: Federal Reserve Bank of Minneapolis). Clower, R.W. (1967), 'A Reconsideration of the Microfoundations of Monetary Theory', Western Economic Journal, 6, pp. 1-9. Deleplace, G. (1989), 'Compte, paiement et change', IOF, Universite d'OrIeans, mimeo. Fisher, F. (1983), Disequilibrium Foundations of Equilibrium Economics (Cambridge: Cambridge University Press). Gandolfo, G. (1985), Economic Dynamics: Methods and Models (North Holland). Kaldor, N. (1982), The Scourge of Monetarism (Oxford: Oxford University Press). Kalecki, M. (1943), 'The determinants of profits', repro in M. Kalecki, Selected Essays on the Dynamics of the Capitalist Economy (Cambridge: Cambridge University Press, 1971). Keynes, I.M. (1936), General Theory of Employment, Interest and Money, repro in I.M. Keynes, Collected Writings, ed. by D. Moggridge (vol. VII) (London: Macmillan, 1973). Moore, B. (1979),'The Endogenous Money Stock', Journal of Post-Keynesian Economics, 2, pp. 49-70. Moore, B. (1988), Horizontalists and Verticalists: The Macroeconomics of Credit Money (Cambridge: Cambridge University Press). Orlean, A. (1987), 'Anticipations et conventions en situation d'incertitude', Cahiers d'Economie Politique, 13. Os troy, I.M. (1987), 'Money and General Equilibrium Theory', in 1. Eatwell, M. Milgate and P. Newman (eds), General Equilibrium, The New Pal grave (New York: Norton). Schmitt, B. (l984),lnjlation, chOmage et malformations du capital (Paris: Economica). Sraffa, P. (1960), Production of Commodities by Means of Commodities (Cambridge: Cambridge University Press). Starr, R.M. (1976), 'Decentralized Nonmonetary Trade', Econometrica, 44, pp. 1087-9. Tooke, T.L. (1844), An Inquiry into the Currency Principle (London: Longman, Brown, Green and Longmans). Part III The Two Traditions: Commentary and History Introduction to Part III Part III contains eight chapters arranged in three, which aim at preparing the field for parts IV-VI, where a possible convergence between the Post Keynesian and the Circulation Approaches on specific points is explored. Section A tries to specify the circulation approach by appealing to streams of thought which have deeply influenced it or have illuminated features which are central to it. Section B stresses the difficulties faced by the two variants of mainstream economics, the General Equilibrium Theory and the Keynesian Theory, in their analysis of money. It also suggests implications of that critique for an heterodox approach. Section C presents two ways of comparing the two approaches, through the relation between money and finance and through the analysis of investment decisions. A. SEARCHING FOR THE ROOTS OF CIRCULATION THEORY Nell in chapter 8 starts from the observation that CA writers assume that the supply of medium of exchange is equal in value to the aggregate gross product. He shows that this is impossibly expensive and implies irrational behavior on the part of some agents. Alternative accounts of the circuit are presented: an incomplete circuit for a medieval economy, then complete circuits for a mercantilist system, and for early and then modern capitalism. Complete circuits are shown to depend on the balance between the wage bill in the capital goods sector and the capital requirements in consumption goods, together with multiplier-like-respending within the capital goods sector. Balanced expansion, in turn, depends on the Golden Rule, which in a two-sector model can be shown to imply, and be implied by, the wage-bill/capital requirements relation. Deleplace in chapter 9 raises the question: does circulation need a monetary standard? The concept of a monetary standard has disappeared from modern theories of money, as a consequence of the definition of the value of money by its purchasing power over a basket of commodities. The role of that concept (as distinct from the standard of value) is emphasized in Ricardo's theory. The comparison between a model of metallic money regime and a model of fiat money regime reaches the conclusion that a 'debt standard' might be understood in the same way as a gold standard. Two consequences are derived: the general price level must be ruled out as a measure of money depreciation, and monetary stability requires discount lending by the central bank at a low and constant rate. Mehrling in chapter 10 reconstructs the ideas of the 'Banking School', developing them in modern form. He shows that the approach rested on a set of conceptual 241 242 Introduction distinctions and a style of argument that combined to form an analysis shared by its prominent members. This begins with the contrast between a circulation resting on a commodity-based money, e.g. gold, and one based on bills of exchange. The circulation of a developed economy, however, employs both money and credit, but the distinction is fundamental, since credit is adapted to the needs of trade. A distinction between lending capital and lending credit helps to clarify the 'Law of Reflux', which states that an overissue of a convertible currency will automatically be corrected. Criticisms of this idea are largely based on misunderstandings, especially of its relation to the 'Real Bills Doctrine'. The chapter concludes by showing how this approach can be adapted to the analysis of contemporary institutions: sophisticated forms of asset and liability management play much the same role that the banking school attributed to its classic channels of reflux. As in the earlier case, these reduce the ability of the central bank to control the quantity of credit. The banking school held that money takes care of itself provided credit is kept sound - a challenge that still faces central banking. Schmitt and Greppi in chapter 11 offer a survey of circular flow analysis in the German language. Following the main argument of Schmitt's chapter 4, - i.e. that the very essence of this analysis is the general validity of Say's law, in the context of actual monetary economies - chapter II analyzes the insights of several past or contemporary authors: Peter and the definition of exchange, Reichardt and accounting rules, Krell and the predominance of identities over eqUilibria, Fohl and the identity between aggregate saving and investment. The issues of the existence of macroeconomic profits and of the time-dimension of circular flows are also explored. B. CONFRONTING THE MAINSTREAM Benetti in chapter 12 considers the problem of the integration of money into the General Equilibrium theory of value. He concentrates on the widely accepted proposition by Hahn (1965), according to which the set of equilibria of an economy with fiat money contains a non-monetary eqUilibrium (with a zero-price for money). Benetti provides a proof that Hahn's proposition is faulty: in an exchange economy with fiat money, there does not exist a non-monetary equilibrium. For that purpose, three types of exchange economies are distinguished, and a critique of Hahn's demonstration shows that the monetary type is heterogeneous to the Arrow-Debreu type. This result reinforces the need for an approach in which money is an integral part of the description of an economy. Dymski in chapter 13 starts from the fact that New Keynesians and Post Keynesians have developed different but seemingly complementary approaches to instability in financial markets and its effects. The Walrasian model is taken as a benchmark: money, credit and financial intermediation are all redundant. Different ways of relaxing the assumptions then permit the development of money and credit markets. Exogenous uncertainty is distinguished from Keynesian uncertainty; both lead to roles for money and credit, but the latter generates a far more significant Introduction 243 need. Once money and credit exist, there may lead be market arrangements in which activities are finance-constrained. Environments can be classified according to whether such a constraint is combined with certainty or uncertainty. The former can be called Schumpeterian, the latter Post Keynesian. Asymmetric information can be introduced, but when there is true uncertainty it is difficult to calculate risk. Three micro-problems are then explored: the degree of competition in money and credit markets, the attitudes towards risk of financial agents, and the closure rule under which financial agents operate. Micro-macro interaction is then explored. C. COMPARING POST KEYNESIAN AND CIRCULAnON APPROACHES Seccareccia in chapter 14 challenges the view, expressed by Weintraub and Moore, among others, that the function of bank credit is chiefly to provide 'working capital', i.e. funds to pay wages. On the contrary, bank credit also supports the purchases of capital goods. Were this not the case, the payments system would not be sufficient to enable the full realization of money profits. This problem is central to understanding 'monetary reflux', the circuit of payments enabling production and concluding with the cancelling of debts. This point is developed and explored in two-sector models, investment and consumption goods, with a pure credit system, and classical savings behavior. It is first extended to continuous borrowing and lending, and then to further cases. Arena's chapter 15 addresses the critique made to the Circuit Theory, according to which it is incomplete, as compared with the Post Keynesian Theory, because it does not consider the financial determinants of investment. Focusing on the respective importance of industrial and financial factors in investment decisions, Arena compares Kalecki's and Keynes' views, and then examines two synthetic approaches, offered by Steindl and Minsky. He shows that these syntheses cannot involve the circuit approach, which, because of its logical structure, rules out the influence of financial markets on investment. Then the Circuit Theory is less incomplete than different, and may be considered as more Kaleckian than Post Keynesian. This conclusion suggests that both approaches may be more complementary than analogous. A. Searching for the Roots of Circulation Theory 8 The Circuit of Money in a Production Economy Edward J. Nell Explaining the role of money in the economy by means of supply and demand has often seemed inadequate, even awkward. Money is evidently an institution - not just another good whose excess demand equation is to be added to those of the other goods, to form a monetized general equilibrium where before a barter system prevailed. A single excess demand equation, derived from preferences, endowments and technology, hardly seems sufficient to express all the economically relevant differences between a generalized barter system and a monetary economy. Even worse, the equation, when added, appears to create serious problems of interpretation, even of consistency, suggesting that the approach is not adequate to the task.· In addition, there are problems forming the equation. Money performs many functions - unit of account, medium of exchange, store of value, means of settlement, at least. Should there be supply and demand functions for cash? Can they be aggregated? This last is particularly significant in regard to the store of value and medium of circulation functions, which appear on the face of it to be alternative uses of money, and therefore not independent,2 But if the supply and demand approach is set aside as inadequate, what can take its place? There is a long tradition in the early discussions of the quantity of money - not the Quantity Theory - in which money is understood as an institution grounded in both the market and the state, and designed so that when proper policies are in place, its quantity will adapt to the needs of the circulation (Marcuzzo and Rosselli, Ricardo, Wicksell). Aspects of this tradition have been revived in the French Circulation approach. 1 THE CIRCUIT OF MONEY AS AN INSTITUTION If money is to be considered as something more than another good, expressed in another excess demand equation, it must be shown how a monetary system fits together with and affects market relationships. What difference does it make that transactions are carried out in money? An answer requires showing how money circulates against goods. It will emerge that this is not seriously, and further, that the most common representation, familiar from every textbook, is seriously defective. It will help to start with a definition: A monetary system is an institutional arrangement in which a specially designated medium of circulation, which also 245 246 The Circuit of Money in a Production Economy serves as unit of account, and is juridically acceptable in settlement of debts, exchanges against the goods produced for market. (Indeed, the prospect of such exchange will be the motive animating the production of goods.) Such a pattern of exchange can be called a circuit. A circuit will be called complete if, in normal conditions, all the goods produced for market can be exchanged for the designated medium, and it will be considered closed if when such exchanges have taken place, funds and goods respectively are positioned to begin a new round of production and exchange, without outside assistance or interference. This definition puts monetary arrangements on a level with the market itself indeed, it claims that the monetary system is inseparable from the market in capitalism. To model the market it is necessary first to understand the rules of the game - property, wage labor, contracts, competition, rational choice, and so on. The monetary system requires a further spelling out of these rules. Prices are to be stated in money of account, while transactions will be conducted and obligations settled in a medium of exchange. A full monetary system means that all transactions must be conducted in money; to be consistent with the market, the completion and closing of the circuit must be consistent with market forces, and all monetary transactions must be motivated by market incentives. Competitive conditions and market motivation, however, do not necessarily imply that circuits will always close or be completed successfully, within a normal time period. It will emerge that breakdowns - crises - are always possible. Nor does completing and closing the circuit imply full employment, or even full utilization of existing capacity. A complete, closed circuit at less than full employment may be analogous to 'underemployment equilibrium'. But it must be possible for the circuit to be completed and closed in accordance with the dictates of tht; market, and without reliance on accidental conditions or exogenous assistance. The circuit must be workable in principle. If monetary institutions called for an unworkable circuit, over time such institutions would be amended or replaced. 3 2 THE EXCHANGE CIRCUIT Consider a medieval county fair.4 The county may be assumed to consist of a set of feudal manors surrounding a town, populated by merchants and craftsmen. Individual choice is not at issue; people are born into definite roles and behave accordingly - yet they are nevertheless motivated by prospects of gain. On the manors peasants grow crops, owing rent for their land to the lords. From their crops they retain their subsistence and market a surplus equal in value to their rental obligations. (They may barter subsistence goods and services among themselves, with bailiffs keeping records and arranging clearings.) The surplus will be marketed to merchants whose working capital is a supply of silver. Wi~h the proceeds of their sale the peasants pay their rents. Edward J. Nell 247 The lords may then enter the market with the silver. Merchants, having traded their silver for goods, now keep a portion of the goods for themselves, and advance the rest to the craftsmen - or, better, obtain an advance of silver from the lords, and, in tum, advance this to the craftsmen, who then buy the materials and wage goods they prefer. The craftsmen next work up the materials into fine goods and luxuries, adding value to the goods equal to the support of the merchants. (The peasants produce a surplus over their subsistence to support the lords, the craftsmen produce a surplus over theirs to support the merchants.)5 At this stage the craftsmen repay the advances from the merchants by delivering the finished goods to them; the merchants then tum over these goods to the lords, receiving the rest of the silver, thus completing the transaction - and the circuit - by returning all the silver to the merchants. All the goods brought to market have exchanged for money, the peasants' rents have been paid in money, merchants have been supported, craftsmen have bought their materials and wage goods for money, and the lords have obtained a better selection of goods, and goods of finer workmanship than they could have had individually from the surplus of their own estates (Nell, 1992, chs 12, 13, esp. pp. 253-4 and n. 58). A simple diagram illustrates this circuit (Figure 8.1). On the main circuit money flows from merchants to peasants to lords back to merchants. In a subsidiary circuit, money and goods circulate between merchants and craftsmen. Goods flow from peasants to merchants through craftsmen to lords, being partly consumed along the way by merchants and craftsmen, and worked on further by craftsmen. The most striking fact about this diagram is its great simplicity: like the famous textbook diagram, money flows one way, goods the opposite, and the two flows are equal - the whole amount of money swaps for the whole amount of goods. The value of silver = the value of the peasant surplus = the value of rents owed = the value of the lords' consumption = merchant income + craftsmen wages + raw materials. The supply of silver exchanges for each of these in successive stages of the circuit. Goods are valued by their direct plus indirect labor content, and the same holds for silver. It is produced by a mining sector and may be coined or stamped by some authority, at the mine or elsewhere. But its value is measured by its weight and purity, and depends on its labor content, not on the stamp of authority. Under normal conditions the labor value of silver will equal the labor value of the goods for which silver exchanges. When the goods brought to market exceed the normal level, however, prices will fall and the value of money in circulation will rise. Since more can be had for silver in circulation than it cost to produce, the output of silver will rise. The increased silver in circulation will raise prices and lower money, until normal prices are re-established. (As the mines get deeper and the veins exhausted, the labor content of silver at the margin may rise, raising silver permanently.) So the supply of silver will adjust through arbitrage to the demands of the circulation, and this pattern of adjustment is stable (though, as we shall see, this is an especially simple case).6 l Peasants $ $ 2. $ rnNT Lords Fig.8.1 RAW GOODS 1. $$$ SALE PROCESS $$$ 4. I The exchange circuit J Merchants FINISHED GOODS $/G 3. $/G Craftsmen ~ 00 Edward J. Nell 249 This is an exchange model which allows for a limited kind of production, devoted to improving the use-value of the already existing goods being marketed. The market exchanges have no effect on the initial supplies of goods. This is very similar to the Walrasian system, in which goods are marketed by self-subsistent agents, whose trade improves their utility, but has no bearing on their ability to support and maintain themselves (Koopmans, 1958; Rizvi, 1992). So money exchanges only against those goods brought to market; the larger part of the goods produced remain in the countryside, where they are consumed directly by producers, or bartered, or circulated by makeshift coins of wood or leather - really tokens of indebtedness, markers or counters, rather than genuine coins. The system is closed, but it is not complete. Hoarding and Instability Suppose, however, that in addition to serving as medium of circulation, silver were hoarded as a store of value; that is, a portion of household, guild or manorial wealth is held in the form of precious metal. Coins, however, are easy to lose and, moreover, being small, portable and readily concealed, can easily be stolen. Hence it is dangerous to store wealth in the form of coins. But precious metal can be molded into candelabra and other decorative and/or useful artefacts, neither portable nor easily concealed. When needed, these can easily be melted down and minted into coins. When more than the normal quantity of goods are brought to market, prices will fall, and the value of money will rise. But the rise in the value of money will raise the value of hoarded silver, also. Here we must consider two effects: first, the effect on circulation of a higher value of silver/money, and second, the effect of rising value. At first glance, it might seem that the higher value of hoarded silver would lead to an increase in current spending by the wealthy - a real balance effect. This would be a mistake. As Hicks pointed out in reviewing Patinkin (Hicks, 1967), a higher value of hoarded money will first lead to a portfolio adjustment, a rearrangement of asset holdings, which will affect asset prices, but not put any additional money into circulation. Thus holders of silver might spend on (or trade plate or bullion for) paintings and objets d'art, or they might purchase additional land and buildings, or further rights to monopolise or positions in the market. None of these is currently produced; such spending therefore does not put money in circulation; it simply redistributes existing assets, changing asset values in the process. 7 However, this is not the end of the matter. There is also ongoing expenditure on long-term projects. These have been planned and are under construction, or are being improved, e.g. merchants building counting-houses. As a result of the highter value of money, the same effects can be achieved for less money. Hence money allocated to these projects can be added to hoards, or could be spent otherwise. In particular, the projects could be expanded. Will the elasticity of expendi- 250 The Circuit of Money in a Production Economy ture on continuing projects with respect to the value of money be greater or less than unity? If it is unity, the project will be expanded to absorb the originally allocated funds; if it is less, then funds will be saved; only if greater than unity will additional funds be thrown into circulation. Thus the effect of a higher value of money on current spending is ambiguous, but it is not likely to lead to an increase. Indeed the more likely effect, particularly in the short run, is that the lower price level will lead to less money being put into circulation than before. But this will be destabilizing since it will put further downward pressure on prices. The effect of the rising value of money is not ambiguous at all, however. As prices fall and money rises, the prudent course of action is to withhold money, since over time its value is increasing. Such withholding, practiced by everyone, reduces the quantity of money in circulation, so that prices fall further, and the value of money continues to rise. The expectations are self-fulfilling and the effect is destabilizing. A fall in the value of money, rising prices will similarly tend to be destabilizing, although the impact of lower silver values may lead to saving to build up precautionary hoards. Taken together, then, when a significant stock of hoarded silver exists, changes in the value of money are no longer stabilizing. To put the point sharply, when money serves both as a medium of circulation and as a store of value, changes in value tend to be destabilizing. A very simple measure deals with this, however. By introducing the mint, and the concept of money as legal tender, acceptable in taxes and settlement of debts, the link between the value of money in circulation and the value of hoarded gold or silver can be broken, thereby eliminating the connection essential to the instability. This will be explored in detail below. 3 INTERDEPENDENCE AND PRODUCTION Consider now what happens when the craftsmen of the town learn to manufacture, not luxuries, but, say, iron plows and leather harness. Let us assume that these are superior to existing means of production, with the result that they raise the productivity of agriculture (Nell, 1992, ch. 13). Agriculture depends on industry for tools, but industry depends for wage goods and raw materials on agriculture. For reasons similar to those mentioned earlier, pressures will develop to bring the rates of surplus value of the two sectors into equality.8 Let us further suppose that with growing interdependence lords seek to manage their holdings more efficiently, and decide to market their entire produce, paying their peasants wages, instead of collecting rents. Guild-masters likewise handle their output, and pay their craftsmen wages. All goods will therefore come to market, to exchange against money. Equally important, production is now governed in the light of monetary incentives. Edward J. Nell 251 All goods now come to market, but not all on the same terms. Before, all the goods were on the same footing, so to speak, but now some are replacement means of production or subsistence, while others are surplus, and the former will not be needed until the latter have been sold. Some transactions are now contingent on others. This is not the only difference. In the exchange circuit the value of the goods in the circuit is always the same, and equal to the value of the money. But when all goods are brought to market this is no longer so. At the outset we have the means of production and subsistence, which presumeably exchange for money. Production then takes place, resulting in gross output - replacement means of production and subsistence plus the surplus - a larger value of goods, which must now exchange for money. In Marx's notation, (M for money, C for commodities, with primes indicating expanded quantities), the circuit is: M - C - P - >C' - M' C - P - > C' represents the production of the surplus. It is explained by the theory of surplus value, or in conventional economics by the theory of factor productivity. Whether either of these theories is adequate is not the issue; they nevertheless provide an account explaining how it is possible for C become C' must exchange for a larger sum of money. M'. How does M become M'? There is no theory to explain that - particularly if M is a sum of silver or gold, (Nell, 1967; 1986). This is problem the theory of the circuit must solve. Treating Production as Exchange Evidently there are differences between circulating only the surplus in exchange, and circulating the goods gross output of an interdependent production system through monetary transactions. Nevertheless, these have generally been ignored and the latter case has been treated as if its pattern of circulation were the same as that of the former. The idea is simple. and owes its origin to Marx.9 There is no initial sum of money M which is followed later by a larger sum M'; there is only M'. The gross product swaps in a single exchange for an equivalent sum of money - just as the surplus swaps for the silver in the exchange circuit. There is an annual market; all goods are brought to market at one time and exchanged against an equal value of money. Merchants now have goods, producers money. Incomes are paid and spent, and the goods are bought back again. The surplus is consumed, means of production and subsistence are employed productively, and a new gross output is produced, ready to be exchanged for money again. This approach faces a number of problems. First, as Marx noted, the money capital advanced is the working capital; the only money 'thrown into the system', as he puts it, is that advanced by capitalists. This will always be less than the gross output, at least under conditions in which the net product is entirely consumed. Capitalists will seek to minimize their costs; they will therefore advance 252 The Circuit 0/ Money in a Production Economy less money than they draw out from the sale of output. This is the basis of the problem. Second, when productivity rises in the exchange circuit, the quantity of goods brought to market at a given time will be larger, and so the quantity of money must increase, since the money must exchange with the goods at the marketing date. But when productivity rises in an interdependent production system, the quantity of goods may not increase. Productivity may increase because goods are being processed/aster. The same quantity of goods may go to market as before; the only difference is that they will go earlier, or more frequently, but there will be no increase in the quantity being marketed at any given time. To say that over a period of time, the quantity of goods on the market has risen is to obscure the fact that at any given occasion of marketing the same quantity comes to market. So there is no need to increase the quantity of money; productivity changes may have different effects in the two systems. Third, to base circulation on M' is expensive. The total quantity of money needed must equal the value of the gross output. In the case of circulation by a metallic currency, that is a large sum to keep tied up unproductively. In the case where circulation is funded by bank loans, it will require substantial interest charges, which will eat into profits (and which must be accounted for in the circuit.) If a less expensive system could be devised, it would be profitable to replace the exchange circuit format. Fourth, because of interdependence, not all goods can be marketed at the same time. No one will be in the market for means of production until net output has been sold. Production takes time, and during this time wages must be paid and household consumption goods purchased. There is a sequence of transactions, which follows the sequence of activities in production and distribution. This did not appear in the case of the medieval county fair because manorial production was no governed by monetary transactions. But now the sequence has to be considered, which means that total output cannot swap at one time for the total money supply. A Modern Treatment of Production as Exchange There is another, modem, version of this account of monetary circulation, in which the surplus, or part of it, is invested while banks provide the funds for circulation, yet the basic pattern remains the same. 10 This is the approach favored by the French Circulation writers as described by Deleplace (see Poulon, 1982, ch. 2). Banks advance working capital to firms who pay wages to households, who, in tum, spend the wages for consumption, returning the funds to firms who repay the banks. Banks make investment loans to firms, who buy investment goods. Output equals C + I; banks make loans equal to C + I, and these funds are spent buying the output. At the end of the period banks are supposed to be repaid; if there is a failure to complete the circuit, as will often happen, financial instruments will be issued, enabling the circuit to begin again. Over time financial Edward 1. Nell 253 obligations can be expected to build up, until they become sufficiently burdensome to cause a crisis, and are wiped away in a restructuring. I I This can be illustrated with a simple diagram (Figure 8.2). Banks advance working capital and investment funds to entrepreneurs in each sector. Wages are paid to households who spend them on consume goods, returning the entire wage bill to the firms of the consumer sector. Entrepreneurs in both the consumer goods and the capital goods sectors spend on investment, buying goods from the capital goods sector, returning the entire advances for investment to the firms of the capital goods sector. The funds thus return to the firms and can be repaid to the banks, provided that certain conditions are met. For when part or all of the net product is invested, and banks advance gross investment funds, which together with working capital, add up to gross output, even if the funds are promptly spent, it does not follow that the banks can be repaid. Consider: while, in the aggregate, the funds return to the aggregate borrowers, namely the firms, these are divided into two kinds, operating in different sectors. Firms in both sectors will have borrowed investment funds, but only those in capital goods earn them back in sales, and both will have borrowed wage funds, but only consumer goods firms get them back in sales. Will the borrowing and earning from sales balance overall? This will depend on an equilibrium condition, which itself reflects the pattern of interdependence. The equilibrium condition, of course, is that the wage bill of the capital goods sector should equal the investment expenditure of the consumer goods sector. This can easily be seen from the social accounts: the capital goods sector spends on wages and capital goods, but earns from the sale of capital goods; the consumer goods sector spends on wages and capital goods, but eams from the sale of consumer goods for wages. The net sales of capital goods are those it sells to the consumer goods sector for investment. The net sales of the consumer goods sector are the goods it sells to the workers of the capital goods sector. For equilibrium these two must be equal. 12 But the French authors do not explain how this comes about. Indeed, the strength of the French approach is precisely that it recognizes that the sectors are interdependent. Yet it retains the idea that the total money supply equals the value of total output. This results in a pattern of circulation that appears at times to run counter to economic common sense. If the total wage bill and funds for total investment are both advanced, then the capital goods sector will find itself awash in unnecessary cash. For when the capital goods sector sells investment goods to the consumer sector, it will receive funds equal to its wage bill. Yet it has just borrowed its wage bill! Why does it borrow this money (incurring expenses and interest) when it can earn it? The problems can be seen by re-examining the diagram. Unless it is expected that its sales to the consumer goods will cover its wage bill, the capital goods sector would not borrow working capital from the banks; but if such sales would cover the sector's wage expenses, why should its firms borrow? Would they not be better off asking for advance payments on orders? (Nothing is specified in the BANKS ~t ,\ I~ I II I - Fig. 8.2 4. II 21 II .J 7 \, \, .," II -I II II pay We 2w. 2w. Spe ndI sell C-goods spends W +"1 pay WI F==:::~ spend Ie CONSUMER GOODS 21 CAPITAL GOODS .J 7 Banks advance loans equal to gross output repay V\(: +WI sell Ie 4. tJ.-.,-....--- ., .. repay Ie+II 1oanWe+WI 1. 1. 11 1 ..I ""------..J \ r:------~.I.I -----~" "~~----, 3. HOUSEHOLDS 0 I ~ Edward J. Nell 255 diagram about timing or priorities, e.g. that wages have to be paid before the goods are sold.) When the capital goods sector sells to consumer goods, why do they repay the loan of their wage capital at once? If they kept the proceeds, they could use it to circulate the remaining investment goods among themselves, working out a clearing system for transactions within the sector. Brokers and dealers would see the opportunity to take business away from the banks by providing a clearing service that would be cheaper than taking out loans. They would then not need to borrow investment funds from the banks, saving themselves expenses and interest. Similarly, the consumer goods sector is assumed to borrow its entire wage bill from the banks. Yet the sector sells its own products to its own workers. Would it not be advantageous to find a way to reduce the borrowing costs, by advancing claim checks to workers and clearing them among themselves? A small amount of money would suffice for this, an amount which could continuously circulate. Even if no clearing system were devized, the whole wage bill would not have to be borrowed. A fraction of the whole wage bill would do, since it would be advanced, then spent, returning to the firms, who would pay the funds out again. Suppose the production period is two months; firms might borrow one week's wages for that period, paying it out eight times, receiving it back each week as workers spend it on their household consumption. They certainly need not borrow the full eight-week wage bill for eight weeks. A further problem concerns repayment: when firms repay their loans, they must surely pay interest? Where do the funds for this come from? The only money in the system is that advanced by banks. How do the firms earn the money with which to pay interest?'3 This blends into a more general difficulty. Firms are assumed to borrow their entire working capital plus their prospective investment, and to repay both at the conclusion of the circuit. Ignoring interest for the moment, this implies that they borrow the entire gross product, and repay it. How do they make a profit? Under these conditions, only the banks could earn anything - namely, interest - and that will only be possible given an answer to the previous objection. Enough has now been said to indicate that monetary circulation in a surplusproducing system should not be treated as an exchange circuit. Interdependence implies sequencing in exchange, and competitive pressures will lead to economizing on the use of money. Let us now examine the simplest case of a closed, complete circuit in a production economy - the 'no-surplus' case. 4 THE TWO-SECTOR 'NO-SURPLUS' PRODUCTION ECONOMY Consider a highly simplified economy consisiting of two sectors. One sector produces a composite good which serves as means of production; the other produces a composite consumption good. Assume that the means of production are fully The Circuit of Money in a Production Economy 256 used up each round of production, and that the consumption good supports labor for a definite period of time. The coefficients of production then are: ab AB (Only a single technique for each industry will be considered, namely the average technique for each industry currently in operation. Changing techniques requires scrapping and rebuilding plant and equipment, a process which cannot be studied until we understand circulation in given conditions. 'Switches of techniques' will not be considered here.) The lower case letters (a, b) stand, respectively, for the means of production and the number of labor units required to produce one unit of means of production. The upper case letters (A, B) stand for the means of production and the number of workers required for the production of one unit of the composite consumer good. The column a, A shows the use of means of production, while the column h, B shows the labor requirements in the two sectors. But the phrase 'labor requirements' has to be understood carefully. What (h, B) show are the number of workers needed. So far, nothing has been said about how long or how had they work. So we must defined (h" B') as the number of workers, working for a unit period of time. Before defining this period, however, let us examine the question of the labor value of the two 'commodities', or sectoral outputs. The labor value of an output is the sum of the labor directly and indirectly embodied in it, i.e. the direct labor producing the output plus the direct labor in the inputs, and in the inputs to the inputs, etc. In a two-sector economy, however, it is simplified so that labor values are given: AI=aA,+b' (1) (2) where AI and Ae are the labor values of unit outputs of industrial and consumer goods, respectively. The first column of the RHS then shows the indirect labor inputs and the second the direct labor inputs. According to the special labor theory of value, goods exchange in proportion to the direct and indirect labor time embodied in them. Hence: VAe = v =b'/[ah' + B' (1 - a)] (3) = (labour/capital goods)/(labour/consumer goods) = consumer goods/capital goods. The equations are: Price Equations PI = aP I + wb'P2 (4a) P2 = API + wB'P 2 (4b) Edward J. Nell 257 where a b' AB* is the matrix of coefficients of machinery and labor inputs per unit output, with (bo, BO) representing the required number of workers, working for a unit time period; w is the amount of the consumption good which supports one worker for the unit time period; PI is the price of machinery in accounting money and P 2 is the price of the consumer good in accounting money. Dividing both sides of both equations by P2: P/P2 =a(P I IP2) + wbo 1= A(PlP2) + wRo. Let v =P I IP2 and solve both equations for w; then v = PlP2 =b'/[ab' + B'(1 - a)] (3') The units of accounting money cancel out. Quantity Equations XI =aX I +AX2 (Sa) where XI is the amount of machinery produced. X2 = wb'XI + WB'X2 (5b) where X2 is the amount of the consumption good produced. Or rewriting XI =aX I +AX2 X/w = bOXI + B'X2 X21w is the amount of the consumption good produced, measured as a multiple of the subsistence wage. It shows how many workers can be supported for the unit period. Dividing both sides of both by X21w: wX I /X2 = a[wXI /X2] + wA 1 =b'[wXlX2] + wB'. Let q =wX I/X2; then q=aq+ wA (Sa') 1 = b'q + wB*, where (5b') q == XI/[X/w] = wX I /X2, which is the ratio of machinery to the consumption dgood, expressed a multiple of the subsistence wage. Then, eliminating w, we get: q =AI[Ab' + .8*(1 - a)] (6) 258 The Circuit of Money in a Production Economy giving us an expression for q which is symmetrical to that for v, the barter exchange ratio. Finally, eliminating v and q, respectively, we have from either set of equations: w::: [1 - a]/[Abo + If(l - a)] (7) Next we put the quantities and prices together to form the complete system. Multiply the value equations by the quantities (q,1) and the quantity equations by the values (v,l) qv::: qav + wbOq::: vq::: vaq + wAv (8) I::: Av + wBo::: 1 = bOq + wBo (9) It clearly follows that the condition for balance is that Av = b°q; the wage bill in the sector producing means of production (Department I) must equal the replacement cost of the means of production in the consumption good sector (Department II). We are here supposing that this economy produces just enough each round to replace the means of production used up in producing, and to support the workers during the time taken by production. Remember we have redefined the coefficients, adding asterisks, to indicate that (b*, BO) now give the number of workers required for the time it takes to produce exact (direct and indirect) replacement, without surplus. At the end of the production round each sector has its own output but needs the output of the other. The consumption goods sector must trade its surplus beyond its own requirements for the means of production for the next round, while the means of production sector must trade its surplus for the consumer goods it needs. Looking at the allocation of output another way, capital goods output goes to replace its own means of production and to replace the consumer goods means of production, while consumer goods output is divided between supporting its own workers, and those in the capital goods sector. Thus we have two sets of equations, expressed here as one for relative values and one for relative quantities. The value equations show the price ratio at which goods must be exchanged between the sectors in order for reproduction to take place; while the quantity equations show how the sizes of the sectors are scaled to each other, dividing the labor force between them. The quantity system defines the initial capacity of the consumer goods sector, and this will be an important basic unit, since its output defines a period of time during which the labor force can be supported. That is, this capacity, operated during a time, must produce an output which, at the normal subsistence level, will support the labor required both in it and in the other sector for that period. This defines the unit period, and the unit size. The coefficients (a, A) and (bo, BO) are on a different footing. The first group (a, A) shows the means of production required to produce one unit of output, regardless of whether this unit of output is produced in a day, a week, or a year. The second shows the labor time required to produce exactly as much as is used up in production. Edward J. Nell 259 The first set is independent of time; the second set has two dimensions. On the one hand, there is a number of workers needed to operate the process at unit level (think of places along the assembly line); on the other hand, there is the time it takes to run the process (how fast the assembly line moves). These different relationships to time are .the basis of Marx's distinction between 'constant capital', which merely transfers its value to the product and 'variable capital', which is capable of creating surplus value. The key to the creation of surplus value lies in the fact that while labor is supported for a given period of time by the subsistence wage, it can work more or less intensively (or for longer or shorter hours per day) during that time. The Circuit of Money in the No-Surplus Economy The key to the pattern of circulation will prove to be the condition for 'balance' between the sectors, namely Av = b·q = wb·[X t/X2 ] This states that the wage bill of the capital goods sector, which depends on the wage rate, the unit labor requirements in that sector, and the size of the sector relative to the consumer goods sector, equals the value of the replacement capital required in the consumer goods sector. The actors in the drama of circulation are: merchants who bring money to the market and specialize in buying and selling; producers who sell their output for money, which they use to buy inputs and hire labor, and workers who receive money wages, which they spend on consumer goods while working. Producers are divided into capital goods and consumer goods firms. Circulation proceeds through the following steps: 1. 2. 3. At the outset, production is complete, and merchants have silver coins on hand equal in value to the wage bill of the capital goods sector. Merchants buy the net output of the consumer sector, leaving that sector in possession of consumer goods equal in value to its own wage bill. The consumer sector now buys its replacement means of production from the capital goods sector, with the funds which equal the value of that sector's wage bill. This is the crucial step, drawing on the balancing condition. A secondary circulation now takes place within the capital goods sector. The capital goods sold to the consumer goods sector were produced by a subsector of the whole capital goods sector; call this subsector IB, and let the rest of the capital goods firms, who sell only to other capital goods firms, be subsector IA. The value of the replacement goods needed by IB, the subsector that sells to the consumer good firms, will equal the wage bill of the firms that produce them within sub-sector IA. However, the same subdivision can be made between those firms who sell only to other firms in lA, and those who 260 4. The Circuit of Money in a Production Economy sell to IB; call these subsectors lAA and lAB, respectively. Then the wage bill of subsector lAA must equal the value of replacement goods produced by lAB. Repeating the procedure, we can see that the wage bill of subsector IAAA must equal the value of replacement goods produced in lAAB, and so on. As the wage bill successively circulates in the capital goods sector, it will, in the limit, enable the unexchanged amount of produced goods to be reduced below any arbitrary level - until we finally reach a set of producers who employ only their own goods as means of production. 14 This 'final subsector' produces its own means of production, as well as the means of production for the next-to-last subsector, the proceeds from the sale of which provide its wage fund. But the goods produced in the final subsector may nevertheless circulate against money. Consider a case where the both products and processes are the same except for aesthetic differences, which appeal to different producers. Thus RED produces red machines using WHITE; WHITE in turn produces with BLUE, while BLUE will only employ RED equipment. All outputs sell for the same price. Either a promissory note or a sum of money starting with, say, RED, will first exchange for WHITE, then BLUE, and finally RED. Again, the actual transactions could be cleared, but in principle, the exchanges are against money. None of the transactions between the subsectors of the capital goods sector actually have to be carried out in sequence. Indeed, to do so would take an unconscionable amount of time. Instead, being desired and anticipated, they can be performed with promissory notes, which can be cleared in a central clearing house. Since the balances cancel in the aggregate, all transactions can be settled, so that all firms receive their wage fund in money. As the capital goods sector completes its 'secondary' circuit, production will begin again, since output has been sold and the wage bill is on hand. As wages are paid, they will be spent, and the funds will return to merchants, as they dispose of their stocks of consumer goods. Upon the receipt of the first funds, some merchants, perhaps specialists taking advantage of the 'short circuit' in the consumer sector, will buy additional consumer goods. These funds will provide the wages for consumer goods producers to pay their workers; as the workers receive their wages, they will spend them buying consumer goods, returning the funds to the merchants, who, in turn, will stock up with additional consume goods, providing firms in the consumer sector another round of wage payments, and so on, until the revolving fund has completed the delivery of the stock of wage goods to the workers, over the period required for a full round of production. Thus a small fund, by revolving frequently, accomplishes a large set of transactions. 15 At this point, when both sectors have paid out their wage funds, and the workers have spent their wages on consumer goods, all funds will have returned to the merchants, both sectors will have completed production, the consumer Edward J. Nell 261 goods sector will need means of production, and both sectors will need to earn money to pay wages. The circuit can be illustrated by a diagram (Figure 8.3), in which the contrast to the exchange circuit is clear. It can also clearly be seen that there are three different patterns of exchange of money for goods. There is first a straightforward swap, of a sum of money for a set of goods, when the merchants buy the consumer goods' surplus, and consumer goods in turn buys, all at once, its replacement capital goods. This exchange takes place at one moment because the consumer goods sector must have its entire means of production on hand in order to begin the next round of production. There then follows a sequential pattern of exchange, as the wage funds make their way through the capital goods sector, enabling each subsector to exchange its output, destined for use in that sector, for its wage bill, concluding with a sequential exchange of means of production in the 'final subsector'. Finally, in the exchange between consumer goods firms and their own workers, we find a 'rotation' in which a small sum of money continuously circulates between workers, merchants and firms, enabling wages to be paid and spent for goods, as production takes place. Money exchanges against goods in three distinct patterns, which must be coordinated for the circuit to be completed. 16 These exchanges can all be made easier and speeded up through the use of bills of exchange and a clearing system. Less money will have to be moved around, and protected, and many transactions can be completed without waiting for shipments of money. (But once bills of exchange come into use, 'bills of accommodation' will not be far behind.) 5 THE CIRCUIT IN AN ECONOMY WITH A SURPLUS The 'no-surplus' economy is interesting only as a hypothetical starting point that makes it easy to see how the circulation works. Because of its simplicity, the principles stand out. But the basic question raised earlier has been avoided; namely, how can a given amount of money exchange first for a set of inputs of a given size, and then later against a larger set of outputs? This must now be confronted. The Emergence of a Surplus Starting from the subsistence-level system above, consider a speed-up resulting from employers cajoling or coercing workers into working faster or harder. Inputs are processed into outputs in less time, but the amount of subsistence available to support workers during the production time remains the same. The actual conditions of work in the two sectors will very likely be quite different, but the speedup will have to be the same in both. Suppose Department I speeded up more than Department II. Then means of production would be completed before consumer goods, who would neither need II merchants $$$ ( goods Figure 8.3 consumers goods workers - goods = bq econdary cicuit " $$$ $=bq I1 capital goods workers - consumer goods $bq wages II {,oil- consumer mfg. , II d ;> capital mfg. I ~ N 0N Edward J. Nell 263 them yet, nor have products to trade for replacements. The completed means of production will sit unsold until the products of the consumer goods sector are completed and ready to market. So the workers in Department I must either be laid off or kept idle. If the latter, then the speed-up (beyond that of Department II) achieved nothing; but if they are laid off then worker resistance to speed-ups will be encouraged. Similarly, if Department I lagged behind Department II, when producing in the latter is finished, it cannot begin again, because means of production are not available, so workers will have to be kept idle or laid off. Furthermore, the faster producing sector will find it advantageous or encourage the fastest-producing firms in the slower sector by offering premium prices. Thus the faster sector will be slowed down while the slower sector will be encouraged to speed-up. Hence the pace of work will tend to the same rate in both sectors, and will become institutionalized in formal and informal work norms. (This conclusion clearly generalizes to any number of independent sectors, allowing for variations permitted by inventory policy. In general, the slowest rate of production sets the pace, a familiar result from critical path theory (Nell, 1991). Let t be the number of complete turnovers made within the period defined by the subsistence level system. In the subsistence level system itself, t = 1. Suppose there is a speed-up, so that t> 1. The production in both Departments will be completed before the wage goods produced last period are exhausted. But the means of production for a complete production run are now available, and the newly produced consumer goods are sufficient to support a full period of production at t = 1. There is therefore a surplus of consumer goods - and only of consumer goods. Both goods are available, so trade can take place and production can begin again. What happens to the surplus? That, of course, depends on who appropriates it, and how. This requires some specification of the class system. This is not the problem at issue, so I will present only the merest sketch of what might be called a 'mercantilist economy'. Each sector is operated by guilds, the masters of which compete in the display of opulence, and exploit their journeymen and apprentices. These latter are forced to work from fear of the power of the masters and by lure of preferment - the possibility of eventually rising to the status of masters themselves. The masters control the sale of products and determine the disposition of revenues, paying journeymen and apprentices only the socially defined subsistence wage, appropriating the surplus for themselves. This can be thought of as coming about through the degeneration of a system of once democratic guilds, which have fallen under the control of the masters, who have reduced wages to subsistence, while using their power to speed up production. But the surplus is not used to expand or transform the production system; the institutions of production - guilds, manors cannot be bought and sold. It is spent essentially on consumption, which in tum is a means to political power and prestige, preferment at court, for example. Under such conditions capital would not be fully developed. Workers would be hired for pay, at a social subsistence level, and the surplus value, in proportion to direct labor, would be appropriated by owners of the means of production. But The Circuit of Money in a Production Economy 264 there would as yet be no investment or mobility of capital, which would be bound by mercantilist restrictions and impediments, licensing, legal monopolies and letters patent. There would, however, be no need to assume uniform organic compositions of capital and in what follows I shall, in fact, assume that the compositions in the two sectors are different. So let us examine the system when a surplus exists and its value is distributed in proportion to the direct labor employed. The equations for value and quantities, respectively, will be: = tav + (1 + s)wb* t = tAv + (1 + s)wB*, = taq + twA t = (1 + c)(b*q + wB*) tv tq (10) (11 ) (1 + s) represents the rate of surplus value, while (1 + c) is the ratio by which the output of the consumer goods sector exceeds the amount needed to support workers during production. Solving both systems, we see that: (1 + s) = t(1 - a)/[Ab* + B*(I - a)]w (12) and (1 + c) = t(l - a)/[Ab* + B*(l - a)]w so that: (I + s) = (1 + c) = t (13) Further: v and = b*/[Ab* + B*(l - a)] q = Atl[(1 + c) [Ab* + B*(I - am (14) = A/[Ab* + B*(1 - a)] (These results easily generalize to many sectors; Nell, 1964; 1992, chapter 4). Comparing this with the earlier equations for the no surplus economy, we see that the speed-up has produced a surplus consisting of the consumer good, while leaving both the exchange ratio and the relative size of the sectors unaltered. Moreover - the central point of Marx's theory of value - exchange at that ratio will bring about both the reallocation of goods necessary for reproduction and the distribution of the surplus, with surplus value being received in proportion to direct labor time. Thus, for example, suppose that after a speed-up exchange takes place at the exchange ratio, v. The consumer goods sector will retain its full period wage fund, wB', and swap the rest for its required means of production, Ar. But this provides the means of production sector with its full period wage fund wb*q. Moreover, having swapped Av of its output, it still has on hand the reminder, qav, just the amount required for the next round. But neither sector will need their full wage fund, for the speed-up enables them to complete a round of production in lit of the unit period. Hence they will only Edward J. Nell 265 need lit of their wage funds for production. They can consequently devote the rest, 1 - lit, to paying surplus value. The rate of surplus value will therefore be: s = [(1 - lit) wages]/[(lIt) - wages] =(t - 1) (15) and this rate will be realized in the exchange which accomplishes the redeployment of commodities required to make the next round of production possible. Notice the importance of exploitation in reaching this result. Because of pressure to produce a surplus, workers will be forced to work as hard and fast as they can, subject to the constraint that no sector can gain by working faster than the slowest basic workers in the slowest basic sector. Thus the pace of work is set by the maximum speed obtainable from the slowest workers, and competitive pressures force all to adjust to this speed. By contrast, suppose there were no exploitation and workers chose their own pace; suppose, in short, that the guilds were democratically controlled. There would then be no systematic pressure to maximize the pace of work, nor would there be pressures to adjust all production to a common work-speed, for each group of workers would be free to pick its own position on the trade-off between the pace of work and the enjoyment of leisure. The market will tend to favor the exploitative system. The Circulation of Surplus Value The emergence of surplus through a speed-up implies that a smaller wage fund will be needed each new, shorter period of production, in each sector. But in each period, a full complement of replacements will be needed in each sector. The consumer sector must therefore purchase goods equal to A veach period, hence must sell goods equal to bq, which is what the merchants are willing to buy, and able to fund. However, such a sale of goods equal to bq provides the capital goods sector not only with their wage bill, but also with their surplus value, since this stands in proportion to their new, smaller wage bill: 1+s = bql(l/t)bq, where s is the rate of surplus value, and lit is the ratio of the new period to the old. The same holds for the consumer goods sector, as we saw. Capital goods thus earns $bq from the sale of replacements A v to the consumer firms; this sum is then redistributed through the sector in the secondary circulation, enabling each subsector both to employ workers for money wages and to payout surplus value. As in the no-surplus case, these funds will be spent on consumer goods, returning the money to the merchants. Merchants, in turn, finding their stocks moving, will now buy additional consumer goods, enabling firms in the consumer sector to employ labor and payout surplus value. As before, these funds will circulate until both wage goods and the goods corresponding to the surplus have been sold, at which point (if consumption out of surplus value takes place at the same pace as consumption of wage goods) production should be complete. This system is still mercantilist, not yet capitalist, since restrictions prevent the sale of enterprises, although labor is exploited and employers or masters are able The Circuit of Money in a Production Economy 266 to appropriate the surplus. These restrictions channel the forces of competition so that the appropriated surplus will brought into proportion with direct labor. rather than total advances. i.e. capital. A tendency to uniformity in the rate of surplus value. rather than a rate of profit, will be established. With the speed-up of production, timing in the pace of consumption becomes crucial to the smooth working of the economy. A new set of consumers is brought into prominence - the recipients of surplus value, presumably the guild masters, the gentleman farmers, merchants. etc. If these new consumers spend at the same rate as workers. then if funds are paid out regularly, they will rotate so that the circulation will be completed at the same moment that production is finished. But if the recipients of surplus value spend more slowly. or if they save or hoard funds. the circuit will not be closed at the point when the round of production is finished. The merchants will not have their funds back; moreover. they will still have stocks of unsold goods. and a commercial crisis will ensue. As Mandeville and others saw, the obvious remedy is to encourage the regular and complete consumption of the surplus earnings; saving and hoarding. prudence and parsimony, should be discouraged. Profligacy may be a private vice. but it is a public virtue. 17 Merchant Earnings But surely the merchants have to earn something? Merchants speed up the process of circulation; they help to make sales of output and purchases of inputs quicker and more convenient. Less time and effort is wasted in storage and bringing goods to market. This represents a saving on wage costs through a quickening of the pace of circulation. that is a rise in t. Merchants, in the no-surplus economy. would be represented by a subsistence level of consumption. As turnover speeds up. therefore. they will obtain surplus value in proportion to their subsistence funds, exactly as in the producing sector. We can write out the equations for this very easily. Merchants' labor time can be included in b· and B" the coefficients giving the labor required for each sector. In the no-surplus case, merchants buy at the price representing the value, and sell to workers at a higher price, thus obtaining the whole wage fund back for less than the whole amount of consumer goods. keeping back enough for their own consumption. (If they kept back more. other guilds, prodded by disaffected workers, would help to set up rival merchants, undercutting them.) With the speed-up, funds and goods will tum over faster, more sales will be made in the same time, and merchants will earn surplus value. This: 1m =(1Ts -1TB)(qb· + B·w) (16) where 1m is merchants' labor measured in consumer goods and (I +s) 1m = (1Ts - 1TB)t (qb· + B·w) (17) Here 1Ts and 1TB are the merchants' selling price and buying price, respectively. and must average to the value of consumer goods expressed in money. Under Edward J. Nell 267 these conditions the rate of surplus value to merchants will be governed by t, the turnover speed, so will equal the rate elsewhere in the economy. A uniform rate of surplus value will prevail throughout the economy. Very shortly, we will see that the same rate of surplus value, governed by t, will also be established in the production of gold coins - mining and minting. Of course, as mercantilist barriers fall, and a rate of profit begins to form, the economy will cease to be characterised by a uniform rate of surplus value. Additional Remarks In practice a metallic currency has always consisted of a number of metals, even if only one served as the official standard. Most frequently, the combination included gold for large transactions and taxes, silver for commerce, and copper for small change and daily wages (Copper will often be valued, not by weight, but as a token.). This means that after producers of large or expensive goods sell their output they will have to convert at least part of their receipts into copper. Merchants selling consumer goods will receive copper which they will have to convert to silver. Silver receipts, in tum, will have to be converted to gold for taxes and large transactions. First gold or silvermiths and then banks will find it profitable to provide this service, and whoever does so will have to have on hand stocks of the various metals and/or appropriate coins. Such holdings do not circulate themselves and they are held as an asset, since earnings arise from the service of making change. In the same way, businesses might hold money in order to avoid paying the service charges for conversion. Such holdings can be correctly described as 'transactions demands' for money as an asset, and they are clearly separate from money which is temporarily on hand, pending expenditure. 6 THE MONEY SUPPLY FOR THE PRODUCTION CIRCUIT Precious metals are produced by mining and prepared for circulation by minting. The quantity of money required will be equal in value to the replacement requirements of the consumer sector, which in tum equals its net output - the wage bill of the capital sector plus the surplus of both sectors. Mining and minting must be arranged so as to regularly replenish the money in circulation, since a certain percentage of coins will normally be lost and misplaced each year and others will become too worn to be acceptable. Since precious metal is produced it will have a value determined by the labor needed, directly and indirectly, in its production. The total quantity required, M, can therefore be calculated from the equation: MP$=Av (18) where p$ is the value of money, as determined by the equation for mining and minting. The annual output, then, must be such as to keep this at the correct level (assuming a stationary economy). The Circuit of Money in a Production Economy 268 This is not a supply and demand equation in the usual sense. It is not derived from summing over the individual equations of all those in the market, nor does it depend on the solution to any specific maximizing problems - although it is assumed that agents will seize on profitable opportunities for arbitrage. In the history of monetary thought, the Cost of Production and the Quantity Theories have been seen as alternative explanations of the value of money (Wicksell, 1895, chs 4,5; 1967, vol. 2, pp. 141-33.) This formulation reconciles them. The quantity of money in circulation determines its value in the short run; the difference between this value and the 'cost of production' value, will, in the long run, lead to arbitrage which will adjust the quantity so that the two values are equal. This is a theory of the adjustment of the quantity of money; it is NOT a theory of inflation, where that refers to a general rise of prices over a long period of time. Such sustained inflation requires an explanation of the persistent pressures driving up prices. In particular, it cannot be inferred from this approach that general price increases are normally caused by a rise in the quantity of money in circulation. This may sometimes be the case, but the mere availability of money, for example, does nor explain the motivation to use it in ways that will drive up prices of goods, when it could be hoarded, loaned or expended in the asset market. In metallic currency systems, hoarded money is always available in large amounts; the cause of an inflation will not be such money, it will be whatever draws it - or paper issued against it - into circulation. Some examples: Bills of exchange will speed up transactions and reduce costs; they will therefore come into general use. But once they do, bills of accommodation follow - credit drawn on the general reputation of a house, rather than against specific goods. They will develop because they are useful - and people will pay for them. As long as they are not issued in excess, no harm is done, and commercial activity will be further assisted. However, only the prospect of deterioration against metal stands in the way of an excessive issue - a weak defense at best when the pressures to expand are strong. Even worse, when paper falls against metal, Gresham's Law comes into play - metal will be hoarded, rather than brought back into circulation. Hence a fall of paper will encourage faster circulation of paper, intensifying the inflation. Cost of the Monetary System Gold is supplied by gold mining, which uses labor and means of production and produces enough gold each period to replace the coins worn out or lost during the year. Let m be the means of production required to mine and mint gold, and n be the labor required. s = mAl + n = [mb*/(1-a)] + n (19) Hence slAc = [mb* + n(I-A))ILb* + B*(1-a)] (20) Edward 1. Nell 269 = labor value of money expressed in terms of consumer goods, which is to say, labor time. (The reciprocal of this will be the gold price of consumer goods, 'IT.) Ae/As = (labor/consumer goods)/(Jabor/money) (21) = (money/consumer goods) = [Ab* + B*(l - a»)/[mb* + n(l - a») The labor value-money exchange equations, multiplied by the quantities are: tqv = tavq + (l + s)wb*q (22) where: (l + s) = [t(1 - a»)/[Ab* + B*(l - a»)w =I t = tav + (l + s)B*w IPs (23) = tmv + (l + s)nw (24) where v = b*/[Ab* + B*(l - a»). Clearly the q cancels out in the first equation. The unit size of the gold equation is taken as the amount of gold needed each period to replace worn-out and damaged coins in the no-surplus system. With the speed-up of production generating a surplus, circulation also speeds up and coins are worn out, damaged and lost relatively more frequently. Workers in mining and minting are a fixed labor force, required to man the mines and the mint, whether much or little output is required. Hence the output of the gold mining and minting sector depends on the speed of production in the other sectors, or equivalently, on the size of the surplus. The output of gold adapts to the need for it in circulation. As a direct consequence the rate of surplus value will be the same in gold production as in the other spheres, since it is brought into being by the same speed-up of activity. Solving for Ps, the consumer goods price of gold, by substituting and rearranging: Ps = [mb* + n(l - a»)/[Ab* + B*(l - a») (25) = the labor value of money. We can assume that gold mining and minting is controlled by the crown, and that the surplus value in mining and minting is appropriated by the crown as seigniorage - the charge for minting new coins or for exchanging old coins for new. These funds will then be spent on consumer goods, supporting the crown's retainers, judges, public officials, etc. Introducing new coins into the circulation takes place during the normal course of exchange. Coins will be lost or damaged at some average rate which will be proportional to the circulation, and production of new ones will be adjusted to this. Using stock on hand the mint will buy (mv) worth of means of production The Circuit of Money in a Production Economy 270 t times during the period, paying wages to n labourers to use these means to mine, refine and mint coins. The capacity output of the consumer goods sector must cover the needs of the workers in the mines and mint: 1 = qb' + (B' + n) w. (26) The quantity equations now read, q = qa + (A + m)w t So, (from equation 5a') = (1 + c)[qb' + B' + n)w] (from equation 5b') (l + c) = 1(1 - a)/[w(A + mW + (1 - a)(B' + n)w] (27) (28) (29) = 1(1 - a)/[[Ab' + B'(1 - a)]w + [mb* + n(1 - a)]w]. Substituting for w using equation (7), (1 + c) = t[Ab' + B'(1 - a)]/[[Ab' + B'(l - a)] + [mb' + n (l - a)]. Now divide by Ab' + B'(l - a), and substitute (1 + c) = 11(1 + Ps) where t = (1 + s) from equation (27). So (l + Ps) (1 + c) = (1 + s), and or P s = (s - c)/(1 + c) c = (s - Ps)/(1 + Ps) (30) Now multiply the quantity equations by prices, and then look back to the price equations multiplied by quantities. From the first equation of each set we have an expression (for qv); hence, subtracting we obtain: (A + m)v= b'q (new balancing condition). (31) A word on dimensions. How can Ps be subtracted from the pure number s? Surely P s seems to have the dimensions of labor time per unit output of gold, whereas s is a pure ratio, so dimensionless. But the discrepancy is only apparent. Labor time translates directly into consumer goods, because we have assumed that consumer goods support labor for a definite time, which defines the period of circulation. Since Ab' + B'(1 - a) is labor time per unit of consumer goods the dimensions cancel. The same is true for mb' + n(1 - a), which has the dimensions labor time per unit of gold. The unit amount of gold is the amount which buys the initial capacity output of consumer goods, which defines the unit amount of labor time. Therefore the unit of gold and the unit of labour are interchangeable measures, so the dimensions cancel, and Ps, like s, is a pure number. It shows by how much the physical productivity of labour has to be reduced below the value ratio to pay the costs of keeping the circulating medium intact. Hence, this also provides us with a measure of the saving that would be had by shifting from a metallic currency, with 'real' value, to a paper one, with only a nominal value. If the nominal value of the paper equals the real value of the gold that it replaces, then prices will be unchanged in nominal terms, but c, consump•. Edward 1. Nell 271 tion per head, will rise in proportion to the difference between the real resources required to produce the paper and those required to produce the gold. Hoarding and Regime Instability Once a surplus exists, the formation of hoards can be expected. 18 Part of wealth will be held in the form of money, for precautionary and other reasons, and money held as an asset will be kept for safety in the form of bullion or plate, or objects d'art. (In an economy with a metallic currency, the latter are doubly significant as stores of wealth: the workmanship will increase in value with inflation, partly because embodied labor will rise to keep pace with current labor, and partly because the scarcity value of the artistic work will rise to reflect the general price level. But the value of the precious metal will serve as a hedge against deflation.) As we saw earlier, hoarding creates instability. A rise in the value of money, if reflected in the value of hoards, will divert new gold and silver into hoards and will encourage the melting down and withdrawal of currency. Both effects will further depress prices, adding to the rise in the value of money. A fall in the value of money will encourage dishoarding, and further fuel the inflation. The instability in question does not arise from particular patterns of expectation, nor does it depend on any special assumptions about risk aversion or any other patterns of preferences. It does not depend on the kind of maximizing practiced, or on the degree of perfection of markets. It is not tied to any particular model and does not depend on whether solutions are unique or multiple. 19 It is a kind of meta-instability which will be present in the institutional arrangements governing monetary circulation, so long as there is a motive for gain, what ever the other expectations and preferences, or the exact form of maximizing objectives or degree of market perfection. This form of instability can be called 'regime instability', to distinguish it from the more commonly discussed market instability. Market instability, by contrast, depends on the precise assumptions made about the individual agents operating in the market - their sources of information, their degree of mobility, their exact motivation (the form of the function they maximize), and the position from which the market began. From this data, first, supply and demand functions are derived - usually for 'representative agents' - to determine the equilibrium, and then reaction functions are defined to describe the response to deviations from the equilibrium. None of this is relevant to regime instability, which instead, refers to the rules of the system, in this case the monetary regime. Here the question is simply one of arbitrage. If the value of money in circulation rises or falls, opportunities for arbitrage will be created - and anyone might take advantage of them at any stage of the circuit. No assumptions need be made about 'representative agents', nor does a full maximizing model have to be defined. All that has to be assumed is that agents, of whatever class or occupation, will take advantage of simple opportunities for gain. The question is, if such opportunities arise, will 272 The Circuit of Money in a Production Economy the arbitrage tend to drive the system back to the proper position, that is, the position considered correct by the authorities who make the rules. Circulation involves a sequence of transactions in a number of different markets; some may, others may not, be in equilibrium. Whether or not they are in equilibrium, the question for a monetary system is, will its normal working create incentives to provide the correct amount of money, and adjust an incorrect amount? Market stability or instability does not imply a corresponding condition for the regime. Regime instability does not imply any particular market instability, although clearly there must be some kinds of disorder in markets. Such instability is not to be confused with the commercial problems that may result from an 'insufficient rapidity of circulation', to use the old-fashioned phrase. This last is a manifestation of an effective demand failure, and so is on a different level. Effective demand does depend on expectations, etc; it is the outcome of a particular pattern of behavior, whereas the instability under discussion exists for all patterns of behavior - though it may well come to the fore more readily in some. Nevertheless, it could easily interact with effective demand problems and the consequences could be indistinguishable in practice. A failure to spend rapidly enough, on the part of the recipients of surplus value Mandeville's problem - will lead to a shortage of demand and a fall in prices; therefore, to a rise in the value of money, stimulating a tendency to increase speculative hoarding, which further reduces the circulation - and so on. The Role of the Mint To prevent such instability it will be useful to break the link between the value of money in circulation and the value of bullion or plate as an asset. If bullion or plate could simply be melted down and stamped by private individuals, and if coins can be melted into bullion and plate, then the value of hoards will reflect that of circulating money. So much metal will yield so many coins. But if melting and minting is the prerogative of the state, a partial barrier is interposed: conditions must be accepted (e.g. prohibitions or limitations on melting), seignorage must be paid, and an official price for precious metal has to be accepted. Delays will occur. A certain weight of metal will no longer translate directly into command over goods. Moreover, coins with the official stamp will be accepted at face value, in settlement of taxes and debts, irrespective of weight and fineness. The value of money in circulation is not the value of the metal in the market, considered as an asset. As a result when prices fall and money rises, bullion and plate will not necessarily rise, or will not rise by as much. This will be especially true for short-run variations. Hence it reduces the temptation to withdraw money and melt it down, i.e. to hoard for speculative reasons. On the other hand, it also makes it more difficult to take advantage of lower prices, since it will take longer to convert hoards into coin. Hence, as well as blocking instability, whatever stabilizing influences may exist may be reduced. When additional money is needed for circulation, it will no Edward J. Nell 273 longer reliably be provided from hoards or from the mines. It will have to come from somewhere else - from the creation of paper capable of circulating. However, the establishment of a mint presents a new set of problems. Faced with a shortage of revenue in relation to need, a government may resort to debasement of the currency, (alternatively: overissue of inconvertible paper money) as a short-term measure to cover its deficit. This will set off a monetary inflation. But such an inflation is a self-stimulating cumulative process. For the rise in prices - fall in the value of money - will lead holders of money balances, which are being accumulated with an eye to a specific future expenditures, to spend these balances sooner, i.e. to move their purchases ahead in time, so as to avoid the loss due to the falling value of money. But this puts additional upward pressure on prices. Thus a given debasement, setting off a monetary inflation, will cause a further loss in the value of money. Hence whatever the initial shortfall, it will end up being larger. To avoid (or at least control) these problems rules are required prohibiting, or limiting, debasement and overissue of paper. 7 THE CAPITALIST PRODUCTION CIRCUIT: PROFITS AND GROWTH The regime of surplus value is at best a partial form of capitalism. As industry expands the barriers preventing the formation and movement of capital, and especially the buying and selling of enterprises, will be swept away and in the new system the market will gravitate around prices corresponding to the rate of profits. Profits, in turn, will underwrite investment. This requires rethinking the pattern of circulation. Equations First, to show the characteristic tradeoff in capitalism between wages and consumption on the one hand, and profits and growth on the other, the two-sector model must be adapted, showing changes in productivity, as before, but now exhibiting both a rate of profit and a rate of accumulation, while retaining the assumption of circulating capital. (Fixed capital will come later.) We begin with the same the formulae as before, except we now write labor as b, B, measuring labor in the equivalent consumption goods; as before, t represents the number of turnovers in a unit period made possible by the speed-up; given a 'no-surplus' system: PI =aP I + bP2 (32) P 2 =API (33) + BP2, then let there be a speed-up, leading to a more rapid turnover of capital goods, producing a surplus of consumer goods. However, if the composition of this surplus is to include any of the capital good, the sizes of the sectors will have to be changed; this is done by multiplying by XI and X2, keeping the total quantity of labor the same. 274 The Circuit of Money in a Production Economy [tP I > taPI + bP2]X2 [tP 2 > tAPI + BP2]X2 Value relations do not depend on quantities, however. Hence, where p = PI /P2 and R = 1 + r, tp = Rtap + wb (34) t = RtAp + wB, where (35) w is the wage rate, meaning the competitively determined ratio of payments to labor to the cost of living, a pure ratio, either in value terms, or expressed as consumer goods paid to consumer goods needed. When this rate is unity, or 100%, labor's earnings just cover the cost of living; when it is, say, 110% labor earns 10% above the cost of living. Then, solving for w in terms of R, we have w = t(1 - Ra)1 [B + R(Ab - aB] = t(1 - Ra)/D, dwldR = -tAbl[B + R(Ab - aB]2 = -tAbID 2, (36) (37) which shows that profit and wage rates are inversely related, and p =bl[B + R(Ab - aB)] =bID, so that, (38) > or < 0 according to the sign of aB - Ab, that is, according to the relative capitaVlabor ratios of the sectors. The quantity relations start from the outputs in each sector in relation to their use as input. Initially, we have, dp/dR = b(aB - Ab)1D2 XI =aXI +AX2 (39) tX2 > bX I + BX2· (40) By varying XI and X2, a positive surplus of the capital good can be created. These two multiply the price equations, establishing the sector sizes. But profits will be invested, and wages consumed. Hence the uses of output must reffect this. Accordingly, the output of the capital good will be divided between replacement of its use in its own production, expansion in proportion to profits, and use in the production of consumption goods in proportion to consumer demand. Consumer goods output will be divided between replacement of basic wages in capital goods, expansion in proportion to the investment of profits, and providing wage good input in proportion to consumer demand in consumption goods. XI = GaX I + cAX2 (41) tXI = GbX I + cBX2, (42) so that letting q = XI IX2, and dividing by X2, q=Gaq+cA (43) t= Gbq + cB, (44) Edward J. Nell 275 from which we derive, c = t(1 - Ga)1 [B + G(Ab - aB)] = t(1 - Ga)ID' (45) where D' differs from D having G in place of R. Clearly G and c are inversely related, in the same relationship that holds between Rand w. And we also see that q =tA/[B + G(Ab - aB)] =tAID', (46) dqldG == tA(Ab - aB)1D2 > or < 0 according to the sign of Ab - aB, the ratio of the capital-labor ratios of the sectors. When G = 0, c will be at its maximum level, and the whole net produce will consist of the consumer good; that is the point from which we start when there is a speed-up. To form a net product which includes capital goods, there will have to be an adjustment of the relative sizes of the sectors. The equations for the use of the outputs show a relationship between growth and consumption which is dual to that between profits and wages. The relative quantities of the sectoral outputs are related to the growth rate in the same way prices are related to the profit rate. The Golden Rule Finally, note that since p =bID and q =tAID'. if D = D', that is, if R = G, pq = value of capital goods per worker == tAblD 2• But dwldR = -tAbID2. Hence, dwldR=-pq, (47) which could also be written, -KIN. That is, the slope of the wage-profit tradeoff equals the value of capital per worker when the Golden Rule holds. This means that Ndw, the increase in payments to labor exactly equals -KdR. the reduction in payments to capital, or vice versa. The implication for circulation is that a change in distribution causes no net change in financing requirements; any change can be managed by shifting funds from financing capital payments to financing payments to labor or the reverse. It also follows that the elasticity of a point on the wage-profit tradeoff equals the ratio of the share of capital to the share of labor (Nell, 1970). Look again at the equations for prices and quantities, above. When r == g, D = D'; hence, dividing prices by quantities, plq = bltA, (48) the ratio of prices to quantities is invariant to the distribution. This implies that dplp = dqlq, or that (Plq)dqldp == 1. The relative quantity of capital goods to consumer goods moves uniformly with the relative price. Cross-multiplying, we obtain, 276 The Circuit of Money in a Production Economy tAp = bq, (49) the sectoral balancing condition! This holds for all distributions between wages and profits, so long as the Golden Rule is satisfied. This conclusion can be reached by another route. Multiply the price equations by the quantities and the quantity equation by the prices, and cancel. The result is, RtAp =Gbq, or since we shall assume R =G, tAp = bq. (50) Given the speed-up, then as long as all and only profits are invested, whatever the distribution, prices and quantities will stand in a fixed, i.e. invariant, relationship. To put it another way, the value of the relative output of capital goods, evaluated at the relative price of consumer goods, is a constant. And this is the central relationship in monetary circulation. Money At this point the analysis will continue on the assumption that money is based on gold and precious metals - a gold-silver-copper combination, for example. But money is not gold; rather, gold and other metals are suitable, because of their value, durability and malleability, to serve as money. Money is whatever serves the monetary functions. The primary monetary function is that of providing a medium of circulation, since it implies the others. Because production takes time and circulation mirrors production, a medium of circulation must also be a convenient store of value, and since circulation requires calculation and the ability to make change, the circulating medium must also be divisible into convenient accounting units. A medium of circulation, then, is whatever enables circulation to be accomplished; it must be the means of settlement, and the pattern of circulation must complete all transactions so that a new round of production can begin with the same price level and the same relative burden of debt. Credit may be issued during the circuit but must be extinguished by the end. The distinguishing mark of credit is that it earns net interest during circulation. Credit money - credit functioning as the medium of circulation - on the other hand, does not. Any interest earned will be offset by interest paid over the course of the circuit. But the medium of circulation need not consist of a metal. The advantage of a produced commodity is that the value of money is set by the cost of production; it does not have to be determined by a convention enforced by regulation. The disadvantage is that a commodity money system is both expensive and hostage to the fortunes of mining and exploration. Instead of a produced commodity the medium of circulation could be a token, endowed with value by convention and designated as acceptable means of settlement by government fiat. (Contracts will be considered legally settled when such Edward J. Nell 277 tokens are presented in payment.) Paper money is an example; but the tokens need not be material - bank deposits, for instance. What makes such tokens money is not merely the government's fiat, but the fact that by means of them the transactions of a circuit can be closed and completed. But token money does significantly enlarge the role of government. Under a precious metals regime the government sets and enforces standards, operates the mint and regulates the currency. It is also, of course, responsible for enforcing contracts and preserving public order in the conduct of trade. But with token money it must also regulate the issue, in order to preserve the value of currency, since it is purely conventional. This, of course, opens the door to temptation. To see the gain from introducing paper we must examine how a metallic system works when a rate of profit is earned and the economy is growing. Under these conditio~s the currency must expand in pace with accumulation. The Circuit of Profits First, continuing with the assumption that money is based on a gold-silvercopper combination, we shall explore how it may be earned by an export industry, which produces a good for sale abroad. The size of this sector is determined by the amount of additional money needed each period to finance accumulation and profits. It will be shown that this amount is $rbq = R(tmp + wn), where m is now redefined as the coefficient of capital goods in the export sector and n is the corresponding labor coefficient. Since the size of the sector is determined by the monetary requirement, we will fix it arbitrarily as unity, so that $ = 1, the per capita unit quantity of gold. (Later we will consider paper money, issued by banks in proportion to their reserves.) So we must rewrite our equations, adding the export sector. tp =R(tap + wb) q t =R(tAp + wB) t =G[aq + w(A + m)] (51) =G[bq + w(B + n)] (52) $ =R(tmp + wn) (53) Now multiplying the price equations by the quantities and the quantity equation by the prices, we obtain, t(A +m)p=bq (54) At the outset, the amount of gold will be taken as fixed, and any increments will have to be earned during the period. The stock of money will serve as working capital, that is, to pay wages. The amount required will be equal to the wage bill of the capital goods sector. The previous period's output of consumer goods is available, so as funds begin to circulate, production can commence, businesses will pay workers, who, in turn, will buy consumer goods, returning the 278 The Circuit of Money in a Production Economy funds to the businesses in the consumer goods sector. As these funds return to it during production, the consumer goods sector will order its replacements from the capital goods sector, in which the returning funds will circulate, and both sectors will be enabled to payout profits, which in tum will exchange for the net product. Part of this will be in the export sector, which will earn the funds needed to augment the working capital. Let's take this step by step. Starting from the beginning of production, each sector pays wages. The consumer goods sector can do so by issuing credits worth $wB to its workers against its inventory of consumer goods (last period's output); the capital goods sector pays wages with funds equal to $bq; consumer goods advances the export sector funds equal to $wn for its workers, who spend them on the current inventory of consumer goods. (b) At the end of the production period the inventory of consumer goods will be depleted and all credits and funds will have returned to the consumer goods sector. The credits will be cancelled, and consumer goods will have $bq on hand, plus a debt of $wn, plus interest, owing from exports. Capital goods will have its output on hand, and exports will be ready to sell. (c) The export sector now sells their output for gold worth $rbq, which they proceed to transfer to consumer goods, cancelling their debt of $wn, and paying interest of $rwll; consumer goods now owes the export sector Rtmp worth of capital goods. (This is the key to the solution of the so-called 'probleme du profit' (Parguez, 1987). (d) The consumer goods sector now has on hand funds of $Rbq, with which it purchases capital goods for replacement and new investment, both for itself and for the export sector, in an amount equal to $Gt(A + m)p. (e) This triggers a secondary circulation in the capital goods sector. Call the subsector of capital goods that sells only to capital goods, lA, and the subsector that sells to consumer goods, lB. (Firms that sell to both must be treated as divided into two artificially separate parts.) Then the wage bill of subsector IA will necessarily equal the gross profits (= gross investment) of subsector lB. As before, subsector lA, in tum, can be divided into two, lAA and lAB, and the wage bill of lAA will equal the gross investment of lAB, and so on. Since these transactions can be readily foreseen, they can be carried out simultaneously with bills of exchange which can then be cleared through a central exchange, leaving wages funds of $Rbq, exactly what is needed for next period's wages, in the pockets of the capital goods businesses. (f) The consumer goods sector now has its own gross investment on hand, and is carrying the inventory of its just completed production, which is G times larger than last period's. It delivers Rtmp to the export sector, cancelling its transactions debt there, and both sectors are ready to produce again. (The consumer goods sector will advance credits for wages to the export sector, and to its own workers; the capital goods sector has on hand an expanded wage fund.) The quantity of money thus required for an (a) Edward J. Nell 279 equilibrium circuit is $bq, and it must grow during that circuit by $rbq; hence dMIM =r =g. Export Adjustment When the rate of profit exactly finances the rate of growth, so that r =g, an equilibrium circuit exists - provided the export industry can earn the required increment of gold. An informal analysis shows that this is by no means guaranteed, although there are forces working toward such an end. The export sector sells its product abroad for gold or precious metals, or most likely for coins made of such metals. If the export sector earns more than $rbq each period, then the domestic supply of money will increase; under certain circumstances this will tend to cause prices to rise, thereby creating pressures for money wages to keep pace. If either p or w rise, and all the more if both rise, then the price of the export on world markets will rise. Initially, this may increase earning still more (and destabilizing activity by hoarding speculators may also exacerbate the inflation), but eventually the rise in price will ruin the market, especially since the absorption of gold may lower prices elsewhere. Earnings will then fall. If they fall below $rbq, then the domestic economy will not have enough gold, and prices will tend to fall, exacerbated by the actions of speculators. This will lower p and w, cheapening the export good and improving its competitiveness. This specie-flow mechanism, however, is suspiciously oversimplified. If the earnings imbalance is large enough and continues long enough, no doubt prices will adjust; but there are certainly other possibilities along the way. Suppose, for example, that the earnings are only a little greater. Instead of - or along with - a permanent rise of prices, activity levels may increase, perhaps from a temporary price surge, leading to higher employment and higher earnings for labor, thus increasing the requirements of circulation. As another possibility, the excess money may not, at first, enter circulation at all; it might go into hoards, raising bank reserves, and leading to lower lending rates of interest. These lower rates, in turn, could well facilitate both accumulation and lUXUry consumption, leading to higher growth rates and greater levels of activity. Again, the needs of circulation for money will be raised, eventually absorbing the excess. On the other hand, while there will normally be some room for expansion in activity levels, in a craft-based economy there cannot be much. But the possibility does help to validate the beliefs of the mercantilists that earning precious metals abroad promoted domestic prosperity. A shortage of funds would depress prices, and very likely depress them at least initially, relative to money wages, squeezing profits. It would also tend to raise interest rates and reduce activity levels, perhaps leading to adjustment, but at the price of prosperity. Suppose that activity has expanded as far as possible, interest rates have fallen, and now the excess funds do enter circulation, raising prices. But now suppose that there is an excess supply of labor, perhaps because of labor-saving invest- 280 The Circuit of Money in a Production Economy ment in the countryside, so that money wages do not keep pace with the increases in prices. Profits will therefore increase, and so will accumulation. In so far as this results simply from increased activity and employment, the requirements of circulation will be raised - but let us ignore this effect, since we have just considered it. Assume also that the rise in accumulation is irregular, so that r does not rise in price with g. A higher rate of accumulation implies a change in q. If aB > Ab, that is, if consumer goods is labor intensive, then q will rise with G, and the higher rate of accumulation will imply a greater need for circulating funds; but if the consumer goods sector is capital intensive, q will move inversely with G, and the rise in accumulation - apart from its effects in increasing the level of activity - will reduce the needs of circulation, making the situation worse. 8 SPECIE-FLOW AND BALANCE OF PAYMENTS If there is an insufficient influx of gold the currency will not expand to keep pace with the increase of production, and prices will tend to fall. If it is excessive, they will tend to rise. In each case the value of money will change inversely to prices. When prices are low, according to the traditional argument, money will be pulled in either because the low prices make goods attractive, or because the higher value of money makes it an attractive market for gold merchants to sell in. In other words, gold may flow in (or out) either on current or on capital account. Suppose gold flows in on current account, in excess of the amount required. Let us consider the effects and the relationships to a trading partner. Given capitalistic production, the rise in commodity prices would translate into a rise in profit (fall in the real wage). (In a relatively early stage of capitalism, the techniques of production may be considered fixed and difficult to change. Substitution effects would be unlikely.) An inflow of gold leads to a 'profit inflation'. But if profits are higher, then ownership of enterprises will be more desirable than ownership of fixed-interest financial assets. Hence bonds will be sold and enterprises bought. So bond prices willfall, not rise, and interest rates will rise, pari passu, with prices - as in fact they did during most of the formative years of capitalism, giving rise to what Keynes called 'Gibson's Paradox'. This also suggests that a gold inflow could be destabilizing. By driving up prices and profits, it leads to a shift in favor of real capital (or shares, claims to real capital) thereby raising interest rates and attracting a further inflow. The process could come to an end as labor resists the erosion of its real wage and demands a comparable rise in money wages, converting the profit inflation into a wage-price spiral. But labor may not be in a position to make such demands, e.g. there may be surplus labor. The high prices, however, are likely to cause problems for exports, and perhaps lead to a shift of spending in favor of imports. (The proportional loss of exports plus loss of sales due to substituting imports for domestic goods must be less than the proportional increase in spending brought about by the influx of gold. Otherwise domestic prices would fall back to their original level.) Hence the domestic country will simultaneously lose gold on Edward J. Nell 281 current account, and attract gold on capital account. Prices will stop rising when the loss on current account just balances the gain on capital account. Now suppose that the initial inflow of gold had come from the trading partner. As a consequence of the loss of circulating gold its prices will have fallen. It will therefore begin to earn on current account, but as its profits have also fallen its interest rates will decline, and it will begin to experience an outflow on capital account. Its prices will stop falling when its earnings on current account just balance its losses on capital. If there are only two countries they will arrive at this 'balanced disequilibrium' together. This suggests, then, that adjustment through current account and through capital account need not be complementary, but, as in this example, can work against each other. Of course, it would require a more complex model to work out what finally will happen; we would have to consider relative growth rates, etc. A similar conclusion might be reached by a different route. We need not accept the claim that an influx of gold directly bids up prices, while not affecting money wages - at least in the short-run - since output and employment are fixed (or only affects them with a lag). Suppose, instead, that we took the position that the influx of gold expands bank money. This would initially lower interest rates and expand investment. If the influx came on current account the lower interest rates might lead to a corresponding outflow on capital. (If the influx were on capital account, the decline in interest rates would presumeably bring it to a halt.) Again the two could conceivably reach a balance, supporting not higher prices but an expanded level of activity. Suppose, however, that the output of investment goods could not be expanded because of capacity constraints. Prices would be bid up relative to money wages and profits would increase. So interests rates have declined and prices have risen this is the case closest to the traditional discussion. The problem is that the decline in interest rates can only be temporary; if profits do increase, and growth takes place at a higher rate, interest rates will be pulled up as investors abandon fixed-interest bonds for higher earning ownership shares. The higher profits and growth rates will attract overseas investors and funds will flow in on capital account, while the higher prices will damage exports and encourage imports, leading to an outflow on current. We are back to the earlier case, with a temporary detour through lower interest rates. Adjustment, then, is possible, but by no means certain. The mercantilists were surely correct to emphasize the importance of export earnings in maintaining high activity and prosperity. But it is an expensive system; a significant level of real resources must be devoted to earning the funds necessary to underwrite the circulation of investment. Paper Money Instead of earning the required increment of funds by selling abroad, a financial services sector could be developed to generate the supply endogenously. The advantages would be threefold: first, there should be a saving of resources, in the form of 282 The Circuit of Money in a Production Economy lower labor and capital inputs into financial activities than in exports. Secondly, if the new system were properly designed, the ability to adjust funds to the requirements of circulation would be greater. Thirdly, investment transactions are typically large scale, and often require amassing funds in one place and spending them in another. For such purposes, coins made of precious metal are inconvenient, compared for example to paper notes or other written promises to pay. Let m' and n' be the coefficients for the banking sector, where the unit size is the scale that provides the services and funds necessary for a closed, complete circuit. If, as seems likely, both are less than m and n, the respective export coefficients, then banking is unambiguously a less expensive method of obtaining the circulating medium than exporting. But if one coefficient were less and the other greater, then the decision would depend on the level of the rate of profit. At a certain level there will be a 'switch' - which implies the possibility of 'reswitching', as well. The circulation of profit under a banking/paper money system will follow the pattern already examined. However, instead of earning the increment of new money from exports, the financial sector will earn profits that will increase its capital assets enough to justify an increment in the note issue of just the amount needed to service the increase in capital and employment. In a 'convertible' paper money system, paper is issued to supplant or augment the precious metals in circulation, a stock of these metals being held, in the form of bullion or plate, as a reserve against the paper. If paper is overissued, its price in circulation would fall against gold and silver, so that the paper would quickly be presented at the banks for conversion into metals. Arbitragers will be able to make a profit borrowing paper and presenting it for conversion. (Even if convertibility is suspended, banks will have to accept their own notes back in repayment of debts; so gradually excess notes will be withdrawn.) So long as the gold and silver exchanged for the paper does not circulate, but is held as an asset, e.g. as bullion, conversion removes money from circulation, reducing inflationary pressure, and raising the value of money. If paper were underissued, compared to the needs of circulation, prices would fall and paper rise, signalling to banks that it would be profitable to issue additional notes. The precious metals act as the standard for money, and as an asset, but the reserve metals do not and cannot circulate. (For if they did, an overissue could not be corrected by the reflux of paper, since the amount of money in circulation would not change, the metals simply replacing the paper.) This corresponds to the traditional account of the 'law of reflux'; but there is no reason to require that the reserves be precious metal. They could be securities, titles to land or claims of any kind; all that is necessary is that when the paper issue exceeds or falls short of the needs of circulation, it will fall or rise, and consequently swap against whatever assets are held as reserve. Banks will advance paper notes to business firms for the purpose of paying wages. So the capital goods sector, anticipating orders, will begin hiring and producing, and with their wages workers will purchase consumer goods. So the consumer goods sector will also hire and start production, activating the 'short Edward J. Nell 283 circuit'. Beginning a new round of production, however, requires the consumer goods sector to purchase its full complement of replacement goods plus its net investment, since these must be on hand from the beginning. They will do so with their sales receipts plus, if necessary, advances from the banks, in anticipation of further receipts. The purchase by the consumer goods sector sets off the secondary circulation within the capital goods sector, enabling it to realize profits and make its own investment purchases. Thus, as before, all goods will be circulated by money, where the amount required will equal $bq, the wage bill of the capital goods sector. However, this amount was borrowed from the banking system. (Even if the capital goods sector could borrow less for wages, the consumer sector would have to borrow the remainder, in order to pay for its replacement plus investment purchases.) Hence interest will have to be paid when the loan is settled. These payments must be deducted from the respective sector's earnings. One implication, of minor concern here, is that to maintain a uniform profit rate, prices will have to be adjusted to reflect each sector's borrowing requirements. But the major point is that as a result of these earnings the capital assets of the banking system increase. If the interest rate earned equals the growth rate of capital, then given fixed coefficients, employment and output, bank assets will expand pari passu with the system. If the banking system's assets initially supported an issue of notes sufficient for the circulation, they will now support an issue just sufficient for the new level of activity, following the expansion due to investment. 20 Notice that the expansion could take place in a way analogous to the way the export system worked. Banks, finding their assets and actual or expected deposits (and therefore reserves) increased, will need to expand their services. To do this they issue additional notes - justified by their earnings - to pay for their replacement and capital expansion, and for their wage advances. So far so good; then in analogy with the export earning model, they would pay new funds equal to $rbq = G[tmp'p + wn'] to the consumer goods sector. When the consumer sector concluded its transactions with the capital goods sector, these funds would reappear as deposits, as anticipated. Thus the 'probl~me du profit' is solved, and dMlM = r= g. But there is no reason to suppose that the banking sector is anywhere near as expensive as this. To earn gold, exports had to be worth what they sold for; the export sector therefore absorbed real resources. Banking and financial services will absorb resources, too; but the difference is that these services do not exchange for the new funds. The new money is created. This creation is justified by the earnings of the sector, which are added to reserves, but there is no reason that the real resources absorbed by the sector have to equal the value of the new money. Paper money, being cheaper, can therefore be expected to displace a metallic currency. Paper Instability A paper money system, however, is subject to its own particular form of instability. Competition forces banks into a race for customers; hence there is pressure to 284 The Circuit of Money in a Production Economy overissue. For anyone bank at any given time, this will be profitable; moreover, for any individual bank, an excessive issue of banknotes need not create any special problems. As loans are repaid, notes can be retired and reserves built up; curtailing new loans will reduce the loan portfolio to the right level. This will bring things back to the correct relationship without any fanfare, provided, of course, that no run develops. But matters are different for the system as a whole. Since the monetary circulation has to grow to keep pace with the expansion of output that results from accumulation, banks will tend to hold reserves in the form of earning assets, rather than metal. Competitive pressures will virtually require this - but it has important consequences for stability. If competition leads all banks to overissue, then paper notes will depreciate relative to metal reserves and hoards. Holders of paper will tend to redeem it for metal, reducing the paper in circulation; this tends to stabilize the issue. But when securities are held as reserves, the process is different. When competition leads to an overissue, prices will tend to rise. In an early capitalist system, output and employment will be inflexible. Initially, therefore, the overissue will lead to a profit inflation; paper depreciates, but profits are up, so interest will be pulled up, and bank earnings will rise. Hence with securities earning more, a higher issue is justified in terms of the value of reserves. Banks will tend to raise lending rates to keep pace, and the competitive pressure will continue to stimulate still further overissues. Profit inflation leads to higher capital goods prices. Eventually money wages will tend to rise, perhaps even catching up, in any event making the inflation general. At this point profits will no longer increase; the rate of profit may even begin to decline if wages catch up. But a general inflation means money is worth less, so money will be traded for securities and other assets. But the immediate effect of a tendency to shift from money into securities will be to support or even drive up the price of securities, so that bank assets will appear to be high, and may even seem to justify continuing to issue notes. But the inflation, meanWhile, will have eroded the earning power of banks; their old loans are now being repaid in depreciated currency. Banks will feel the pressure of lower earnings; moreover, with lower profits as wages rise, the securities they hold will be worth less in fundamental terms, even as the shift from away money is supporting their inflated prices. Having over-issued, and finding securities shaky, banks will now find themselves without adequate backing, at a time when the public is inclined to exchange notes for reserves; banks may be compelled to sell securities, e.g. for bullion, to redeem their notes, a move which will tend to drive down security prices. But with a drain of securities their earning position will be further undermined; moreover, as securities begin to decline, they will no longer have adequate resources to redeem their notes, which may well provoke a run. A likely result is that general over-issue, followed by loss of reserves, will continue until a collapse takes place. To prevent this regulations restricting note issue have been widely adopted. Edward J. Nell 9 285 THE MODERN ECONOMY So far the argument has been developed in the context of circulating capital. But mass production requires fixed capital, which brings with it a new set of problems for the analysis of circulation. Fixed capital turns over slowly - it lasts longer than the normal period of production; its production and sale must therefore be financed. This finance, however, should be considered a normal - inescapable part of the circulation. The Circulation of Fixed Capital The peculiarity of fixed capital goods is that they outlast the period of circulation, with the result that their producers cannot sell replacements to their customers at regular marketing times, nor will new investments take place regularly each period. Yet producers of these goods must somehow continue to produce if they are to stay in business and meet the demands that will eventually come due, and for this they need circulating capital - to cover wages and current inputs. This means that the producers of circulating goods must lend to producers of fixed capital goods, and the effects of this interest must be accounted for. Certain simplifying assumptions will be useful. First, circulating goods can all be assumed to move in one common period of time. Second, let us suppose that the industries that produce fixed capital goods employ circulating capital, while circulating capital industries use fixed capital; the use of fixed capital in fixed capital can be ignored for the moment. Third, all fixed capital goods of a given kind will be traded at the same time; a common period of depreciation will be assumed. Prices, wages and profits will be taken as given. During the course of time the durable goods are gradually used up, so that at the close of a period of circulation an industry employing fixed capital must set aside a certain part of its earnings to cover depreciation. Producers of fixed goods must acquire circulating capital in order to continue production, but they cannot sell their output as yet. Hence they must borrow the depreciation funds to cover their circulating capital needs. The amount of borrowing a fixed capital producer needs to undertake depends on the ratio of the write-off period to the period of circulation (i.e. the period in which fixed capital goods are produced.) Bearing in mind that the circulating capital for the first period after sales will be earned by the sale, the borrowing required, B, in a given industry, i, will be f3j = [(wp - cp)/cp]Cit where 'wp' is the write-off period, 'cp' is the circulation period, and C is the circulating capital. In the same way, the amount that producers of circulating goods in industry j, will be willing and able to lend, L, will depend on the same period, and on the level of depreciation per period, D: Lj = [(wp - cp)/cp]Dj 286 The Circuit of Money in a Production Economy To show that borrowing will equal lending in an economy in balance, consider a system with fixed capital but no surplus (or one in which surplus is devoted to paying wages). The circulating capital industries use fixed capital, while the fixed capital industries use circulating capital. If there is no surplus, then the circulating industry's demand for fixed capital must just equal the fixed industry's demand for circulating capital. All that is necessary is to adjust to a common write-off period, say, n periods of circulation. Then every nth period, the fixed capital producers will deliver the durable goods, and each of the following n periods, circulating capital producers will payoff lin of the price. Thus C = FIn. Now suppose that the system is expanding each period through the reinvestment of profits, so that r = g. Then the output of both fixed and circulating goods must grow each period at 1 + g; which implies that the circulating goods consumed by producers of fixed goods will expand at the rate (1 + g)q'CIP, while the fixed goods consumed by the producers of circulating goods must also expand at the same rate, to maintain balanced growth. But the transaction in which this balance is realized can only take place at the end of the write-off period, i.e. after n periods of production/circulation. Hence compound growth must be figured in each case: we require that (l + g)nq'C,tP = (l + g)nq'Fcp Here Cf is the matrix of circulating inputs in the fixed capital industries, and Fe is that of fixed inputs in the capital industries; q' is the vector of industry circulating sizes in balanced growth proportions, and P is the price vector corresponding to a uniform rate of profit. The output of fixed capital goods for eventual sale to users, who produce circulating capital goods, must grow over the write-off period at the same compound rate as that at which circulating capital goods output - and rest of the economy - is expanding. How can this be arranged? Fixed capital producers must borrow circulating capital at a growing rate in order to reproduce an expanding output. Fixed capital users must accumulate funds at an expanding rate in order to have the wherewithal on hand to make the eventual purchases of appropriately expanded capacity. Thus each period fixed capital users will invest their (growing) depreciation allowances in a sinking fund, which will earn compound interest. This sinking fund, in turn, provides loans of both current allowances and its own earnings, to producers of fixed capital goods, enabling them to continue production on an expanding scale. So long as this interest compounds at the rate i = g, and provided the depreciation allowances each period are correctly calculated, and grow from period to period at rate g, the sinking fund will be just the amount required to enable the users of fixed capital to purchase the expanded output of durable goods, thus providing them with appropriately expanded productive capacity. (How can users of fixed capital expand their output before the fixed capital is augmented? Surely fixed capital must constitute a capacity constraint, which will act as a barrier to the expansion of the circulating capital industries. This misun- Edward J. Nell 287 derstands the nature of fixed capital. Moreover, to the extent that this is so, in a growing economy, producers will overbuild at the outset. But fixed capital is not a rigid barrier; it is the nature of mass production that utilization is flexible, and that output from given plant and equipment can nearly always be expanded in the short run (Nell, 1992; Steindl, 1952).) Money, Capital and the Sinking Fund When producers of circulating goods complete a period of production they may be assumed to sell their output to traders for money. An appropriate portion of this money is put into a sinking fund to cover eventual replacement costs; the rest goes to pay income and to cover the new period's circulating charges. The sinking fund is then loaned (by a bank or broker) to fixed capital producers, who use the proceeds of the loan to purchase circulating goods, thus returning the funds to the traders. This puts them in position again to purchase the next period's production of circulating goods. Each period the size of the circulating industries expands at the rate I + g; each period a given fraction of the value of fixed capital is set aside for replacement and, together with the accumulated compound interest, is available for lending. The fixed capital called for at the time of replacement will not be the original amount, but an amount reflecting growth. This will be calculated by considering the amount of fixed capital used up each period - the portion written off - which can be considered to circulate. Hence each such portion will grow at the rate at which circulating capital grows, compounded over the write-off period. Depreciation allowances should be calculated so as to keep capital intact while growing at the prescribed rate. This if fixed capital is F and the allowance is A, where g is the rate of growth and i the rate of interest, capital at the end of the first period will be (F-A) + (1 + g)C + (l + g)A, where C is circulating capital (wages, materials, energy, etc.) That is, (I + g)A, rather than A must be set aside, because capital is growing. At the end of the second period capital will be (F - 2A) + (I + g)2C + (l + g)A + (l + g)(l + i)A, at the end of the third period (F - 3A) + (l + g)3C + (l + g)A + (l + g)(1 + i)2A, and so on. As the existing stock of fixed capital is worn down, the sinking fund is built up and compounds at a rate of interest which in equilibrium will be equal to the rate of growth. In this way fixed capital can expand in pace with circulating capital. The sinking fund will be loaned to producers of fixed capital, providing them with the circulating funds they need to expand their operations so as to produce and warehouse the increased set of fixed goods that will be needed at the replacement date. Each period they will borrow the new allowances plus the earnings from the sinking fund. Their accumulated borrowings will equal the value of the output of fixed capital goods; which is to say, that the sinking funds will be just the sum required to purchase the expanded output required for the growth-augmented replacement. 288 The Circuit of Money in a Production Economy (The loan in period 3, say, was that period's depreciation allowance, so equal to (1 + g)3 times the initial allowance, A. By the end of the write-off period, this will have compounded to (1 + i)n-3 (1 + g)3A = (1 + i)nA, since i = g So at the end of the write-off period all loans of depreciation allowances will equal the current period's depreciation charge, since the amounts borrowed have grown at rate g, while the amounts borrowed have compounded at rate i.) At the end of the write-off period circulating producers sell their current output, and, instead of putting the last period's depreciation allowances into the fund, they will be employed to make the first payment towards the replacements. At this point the loans will be called in; as we saw, the funds will be exactly the sums needed. As loans are called in, payments are made; the funds will be deposited, and the increased deposits will justify banks in expanding their note issues. Hence there will be no shortage of funds. The repayments of loans will recirculate as payments towards the purchase of capital goods, until all loans are settled and all durable goods paid off. Continuous Production and Circulation Underlying the discussion so far has been the assumption that production processes take the 'batch' form: they begin and end at definite moments of calendar time. It has been tacitly assumed that these moments are coordinated, so that all producers in an industry operate more or less together - as, for example, farmers do. Further, the producers of inputs and purchasers of outputs of any given industry coordinate their activities with the timing of that industry. In the history of the advanced world these assumptions would not be problematic prior to World War I, and perhaps even up to World War II. By the postwar era, however, batch processes have been replaced or are being phased out. Mass production replaces the batch system with continuous output. Batch production means that a plant begins processing a certain amount of raw material, running it through a number of successive stages. Each stage basically occupies the attention of the entire plant while it is running. At the end, the batch is complete, ready for marketing; then, and not before, another batch can begin. Continuous production means that as soon as one batch of raw material has begun to processed, another can be started up. Input goes in continously - or at regular but frequent intervals - and output emerges steadily. It is not necessary to wait for a process to be complete to begin work on another set of inputs. Continuous production implies continuous circulation; there will be no clear beginning and ending points on which all processes converge. It might seem, therefore, that a different approach is required. If production is continuous, for example, it will not be the case that producers in the capital goods sector all finish at the same time, ready to sell replacements to the consumer sector, and to one another in the secondary circulation. Edward J. Nell 289 But the fact that processes run continously need not imply that the pattern of circulation will work differently. Processes take time; production starts at the point where the raw material is acquired and set up for processing, and finishes when the good is turned over to be marketed. There is still a definite sequence to the operations, and this sequence defines the relation of the process to other processes. Thus, for example, each day work on a new set of auto bodies will begin, requiring as input a certain amount of appropriately alloyed steel. The output of this steel from the capital goods sector must be adjusted to this rate of input, with appropriate precautionary inventories. Refrigerator inputs of a different kind of steel may be required only weekly, and engine blocks in larger amounts, monthly, but only in the six months prior to the assembly of the new models. The various processes that make up the economy must still be coordinated. Even if there are no common starting up and 'completing points, outputs and inputs must be delivered in the right amounts, and in the required sequences, so that they are available at the points when they are needed in the continuous process. The case of fixed capital showed how mutually dependent processes with different turnover periods are adjusted to one another through credit. This can be applied to circulating processes with different turnover periods as well. Even under mass production the seasons, traditional holidays and social customs provide a framework that sets definite marketing dates towards which manufacturers aim. Think of summer vacation time, Christmas shopping, the New Year sales, spring fashions, the Fall Back-to-School sales. So, while under continuous production there need be no common starting and finishing points, these will often exist, nevertheless. In any case, there must still be coordination of inputs and outputs, and this requires that different processes be adapted to one another. Outputs designed to serve as inputs in other processes must emerge in a flow that provides them in the required amounts at the necessary times. Hence the timing and size of processes must be adjusted. Bank Deposits Mass production cannot function effectively with only paper money; fixed capital requires bank finance, and so do purchases of consumer durables. Carrying about large amounts of paper money is dangerous and inconvenient. But it is easy to shift from paper to bank deposits and the writing of checks, although this does call for careful monitoring to prevent fraud. Consumer purchases can be made easier with credit and debit cards. The circuit will work in the same way - banks create deposits by making loans, instead of issuing notes. Banks issue lines of credit to firms, to cover their wage bills. These will be drawn upon as firms adjust their output to demand. As credit is drawn down, the deposits of firms rise; as wages are paid, business deposits fall, and those of households rise. But the level of deposits - money in circulation - adjusts in pace with the level of employment, which in tum is governed by the level of demand. Bank deposit money is 'endogenous'. 290 The Circuit of Money in a Production Economy There is an important difference, however, between this and the earlier circuit. Commodity money enters circulation with a value derived from production. Paper notes acquire a derivative value from commodity money by being convertible. But bank deposits are purely nominal; they have no 'anchor' in a value derived from a production process. Such money therefore, has no 'natural' or normal value. If it is anchored to anything, bank deposit money is tied to the money wage. 21 But money wages are or can be market-driven, so there is nothing to prevent the market from reacting to the movement in the value of money, as opposed to the deviation from the normal level. As we saw earlier, if the market reacts to movements, rather than levels, instability will tend to result. Moreover, just as competition drove banks to over-issue notes, so it will tend to drive banks to over-extend loans, creating a new form of instability. As in the case of paper, the instability arises because, for each individual bank, overlending tends to be profitable in the short run; moreover, if a given bank restrains its lending, while its competitors over-lend, it will lose market share. But for the system as a whole, over-lending leads to inflation. 22 Inflation, in turn, drives up the prices of securities, and thus encourages further over-lending, together with arbitrage against reserves. The result is a steady pressure weakening the system. The corrective in this case is the imposition of reserve requirements, together with a system of bank inspection to ensure compliance. 23 The Central Bank More is needed, however. Even well-managed banks can run short in a crisis, and any bank can be ruined in a panic. The reason bank liabilities can serve as currency is that they are perfectly liquid; they are available on demand. But bank assets have to be marketed, so are not perfectly liquid, especially in a time of financial disorder. A lender of last resort is needed, to protect well-run establishments in crises, by providing them with the liquidity to meet their liabilities. Such an authority should also possess the powers to preside over liquidations, to prevent a disorderly collapse whenever overissuing gets out of hand, and to prevent closures from injuring innocent parties. This is necessary to prevent any general tendencies to withdraw funds permanently, for this would endanger the whole banking system. Just as the Mint establishes a price at which it will buy and sell metal, so as to tend to stabilize the currency, and also to redeem underweight coins, thus validating the coinage, so the Central Bank establishes a rate of interest that tends to stabilize the system of notes and bank deposits, while at the same time making loans and accepting deposits at that rate in its capacity as lender of last resort, 'validating' deposits. And just as the Mint must recognize that, although larger and carrying the stamp of authority, it is still one player among many in the market, so the Central Bank must face the limits to its powers. By virtue of size and authority, it can - within limits - fix the rate of interest. It is the 'price leader'. But market forces adjust the quantity of circulating medium to 'the needs of trade', to use the old-fashioned phrase, and the Bank cannot affect this. Of course, by restricting Edward J. Nell 291 credit and/or raising interest rates, it can alter the level of investment spending, thereby influencing the composition (Craft economies) or the level (Mass Production economies) of aggregate activity. Since the quantity of money in circulation adapts to activity, this indirectly modifies the money supply. But the Central Bank cannot directly set the quantity of money. As a matter of policy, to limit inflation, and to control the volume of speculative funds, the Central Bank's interest rate, of course, should be the 'Golden Rule' rate, i.e. the rate at which the proportional increase in money in circulation would equal the proportional growth of capital. This point will be developed in a moment. 10 MONEY AND THE GOLDEN RULE A stable value of money requires stability in the value of goods against which money exchanges-so that changes in distribution can be accomodated by switching finance - as well as stability in the relationship between money and goods. Whether the medium of circulation is metallic, paper or bank deposits, the stability of the monetary system depends on 'the Golden Rule', both to maintain the value of the standard of money, and to ensure a supply that grows at the proper rate. The Pattern of Circulation In the no-surplus case, the circuit began with merchant purchases; in the Mercantilist and early Capitalist economies, the circuit began with the issuance of wages. In each of these cases the output of both sectors was assumed to be used up within the period. But modem capitalism is driven by investment in durable capital goods. Here the French theorists surely have a point; the circuit cannot get under way unless there is financial backing for the proposed investment spending. Current production - whether for stock or to order - can be financed in the usual way, by drawing on lines of credit for working capital. These lines of credit will be repaid as firms receive (or borrow) depreciation funds. The sale of large-scale, long-lived capital goods requires long-term finance as well, which means that the capital goods will be paid off over time. As we have seen financing the sale of currently produced long-term investment goods calls for mobilizing depreciation funds. New investment demand in the consumer goods sector provides the starting point, purchasing the output of the capital goods sector. This output will have to expand at the going rate in order to keep pace with the growth of the consumer sector. The capital goods sector will therefore produce output for investment both in itself and in consumer goods. Since these are long-lived goods, their sale will have to be financed by drawing on the accumulated depreciation funds, which have been growing at the rate of interest. As the investment sector produces in response to the investment demand, it will payout its wages, which, when spent, then provide the profits of the consumer goods sector, enabling the consumer goods sector to service its debt. The consumer goods sector issues its own credit 292 The Circuit of Money in a Production Economy to pay its wages, on the 'short circuit', as before. In practices, in any modern economy, the actual circuit is likely to be a mix of advances from lines of credit for working capital, and advances for investment spending. The Value of Money In the Mercantilist economy, and in early Capitalism, the commodity which serves as money, and/or as the standard for money, enters circulation through the Mint with a value acquired during the process of production. For the circuit to be closed and complete a certain amount of the money commodity must be minted and put into circulation. Money will then circulate at a value governed by the mint price, which in turn reflects the cost of production of the money commodity. Adjustments will take place when the demands of circulation vary; money will be brought into or removed from the circulation according to the needs of trade. However, this does not just happen automatically. The rules have to be designed to ensure that market pressures will adjust the amount of money minted and put into circulation in the right way. Destabilizing reactions are always possible. The Mint price, the conditions of convertibility at the banks, and laws governing melting coin, clipping and other practices, and the rates at which different coins are acceptable in settlement of debts and for payment of taxes, all can be adjusted to help ensure that the money supply is kept adequate and that reactions to deviations tend to stabilize it. . In order to design and adjust these regulations and laws, it will be useful to know the cause when an apparent change in the value of money in circulation takes place. The change may have arisen because of a change in the quantity or value of goods on offer, due perhaps to productivity advances, distributional changes, or to accidents of the harvest. Or it may have originated on the monetary side - a rise in hoarding, a slowing down of production of precious metal. Or it may be due to the effect of productivity changes or distributional changes on the value of precious metal. When the quantity of goods changes, and relative prices remain constant, money prices will change, i.e. the value of money changes, so the amount of money in circulation must adapt to restore the price level. Similarly, if accidental forces cause the amount of money in circulation to rise or fall, goods and relative prices constant, its current value will fall or rise, and to correct this the quantity must be brought back to the proper level. In both of these cases the aim of policy will be to see that market forces maintain a constant price level in the long run. By contrast, when the quantities are unchanged, but relative prices change, a change in the price level may be required. And this mayor may not require a change in the supply of money in circulation, depending on the effect of the price changes on the cost of production of the money metal. But even if the quantity or money must change, if relative prices have changed, the objective of monetary policy might Edward 1. Nell 293 not be to restore the previous price level. For the value of money, as measured by its cost of production, may now be different (leaving aside special cases). To see what adjustment will be required it is necessary to know whether relative prices have changed, money the same, or whether the value of the money commodity has changed, other prices the same. In the first case, (the case Ricardo hoped would be the norm) the quantity will change, but the price level will not; in the second, both the price level and the quantity might change. Of course, all relative prices - goods and the money commodity - might change. It was in order to analyze and distinguish these cases properly that the search for an 'invariable measure of value' was undertaken. The Golden Rule provides the basis for a way of handling these difficulties since it defines a case in which both the price level and the money required for circulation are unaffected by a change in relative prices. As a result, it also provides a solution to the problem faced by a financial system when income distribution changes. 24 When distribution changes, relative prices change. However, these variations may be accommodated without requiring a change in either the amount of money or in the price level if r = g. Consider the implications of the Golden Rule: Suppose between one period and the next, the wage rate rises, so that the rate of profit falls. Suppose further that r = g continues to hold before and after this change; the relative sizes of industries are adjusted to the change in distribution. When r = g, the wage bill of the capital goods sector will equal the replacement demand of the consumer goods sector (plus the requirements of the banking sector). A further implication of r = g, derived in the previous section, is that the slope of the wage-rate/rate of profit tradeoff equation equals the (negative of the) labor to capital ratio, and the elasticity measures relative shares. That is, -r/w(/dr) = rKlwN (Nell, 1970). Hence, as we saw: -Kdr=Ndw Suppose the circuit is partly financed by lines of credit for working capital and partly by advances for the purchase of capital goods. The wage rate then rises. The funds needed to pay the increased wages bill are exactly those released by the fall in profits due to the change in the profit rate, given the adjustments in industry sizes. Finance can thus be shifted from capital spending to underwriting wage payments. There will be changes in relative prices, but there need be no change in the price level, nor, in pure banking systems, in the quantity of money. For if the available supply of money, circulating in the normal pattern (at the normal velocity), drawing on the available sources of credit, sufficed to circulate all goods at the initial rate of profit, then, when for example the rate of profit falls, the decline in the profit flow is exactly matched by the rise in the wage bill. The funding and credits have simply to be transferred, e.g. from financing investment to financing wage payments and consumer spending. Of course, there may be practical difficulties, but in principle, the new circulation can be accomplished with the monetary resources that managed the old. (In a metallic currency system, 294 The Circuit of Money in a Production Economy however, the relative price of the metal could be affected, if, as would normally be the case, the capital/labor ratio in the industry producing precious metals differs from the 'Standard Ratio'.) This shows that the financial system can readily adapt to changes in distribution - provided that there are market forces that tend to bring rand g into line with one another. This must now be explored. Stability in Monetary Growth It has traditionally been held that competitive pressures will tend, in the long run, to bring the rate of profit and the rate of interest into line with one another. Consider holding assets for income: a given sum invested can bring a (somewhat uncertain) return in profit, when placed in a venture, or h can bring a (safer) return in interest, when placed in bonds of suitable rating. In addition, in a growing economy, in which investors can choose between growth through investment in real assets, on the one hand, and growth through compound interest on bonds, on the other, arbitrage should tend to bring the rate of interest into line with the rate of growth. (Nell, 1992 Ch. 22) Investors can choose to hold shares in partnerships or to own companies, which will plow back their earnings and grow; or they can hold long term bonds, reinvesting the interest. A well-managed portfolio will have a selection of safe and risky income-yielding assets, and a similar potpourri of growth assets, the mix of income and growth itself being determined by the needs of the asset-holder. So it appears that competitive pressures exist, tending, in theory, at least, to bring the rate of profit, rate of interest and rate of growth into line with one another. When the rate of interest equals the rate of growth, the earnings of the banking system enable it to expand its monetary issue - or its credit lines - at the same rate as the economy is growing. Thus the money supply automatically expands to keep pace with the needs of trade. But does this arrangement provide a stable monetary regime? First, consider deposits and lines of credit for working capital. When i = g credit for working capital will expand at the same rate as the demand. When i < g credits will expand by less than the demand; alternatively, granting such credits will be riskier, since the capital to back them is lacking. Hence there will be a tendency for i to rise. The higher level of i together with the shortage of credits will tend to reduce activity and investment; hence there will be a tendency for g to fall. Conversely, if i > g then the ability to offer credits will grow faster than the demand for them. Competition will tend to bring interest rates down, while the lower rates and the easy availability of credit will tend to encourage expansion. The relationship therefore appears to be stable. On the other hand when i < g, g will tend to rise. This is, after all, the traditional relationship-when credit is cheaper, investment will expand. Wicksell (1898, 1967) argued that when the money rate of interest lay below the real rate of return, a cumulative inflation would take place. And when i > g, according to Edward J. Nell 295 Wicksell, a cumulative deflation would take place. His argument was set in the context of a Craft economy; adapting it to a Mass Production economy, j < g would imply an expansion of output, up to full employment, with inflation first accompanying the expansion then replacing it, while j > g would lead to contraction rather than deflation. In short, the relationship may be unstable - with the proviso, however, that the impact of interest rates on investment is not very strong, a finding confirmed by many studies. Next, consider the financing of fixed capital. As we have seen when i = g sinking funds expand at the rate necessary to purchase and install the expanded replacement capital. But when i < g, the sinking funds will not be growing at the required rate; there will be shortages of credit for working capital to finance the current operations of the fixed capital producers, and the funds will not be sufficient to purchase a full complement of replacements expanded at the rate g. Demand for credit to expand operations at the rate g will tend to drive up interest rates, while the shortages will curtail operations, tending to reduce the growth rate. Similarly, when i > g, the sinking funds will be growing faster than demand for working capital, tending to bring interest rates down, while the easy availability of credit will tend to encourage additional expansion. Again the relationship appears to be stable. But once again, other things being equal, a low rate of interest will encourage expansion, and a high rate discourage it. A discrepancy between i and g will tend to perpetuate itself through the stimulus provided to investment. Unfortunately, this is not all. The financial side gives rise to a more significant instability. Besides credit for working capital and finance for fixed capital, there are stocks, in particular, growth stocks. And here the relationships are problematical. For if a company has a certain number of shares outstanding, and retains its profits to re-invest them, the price of its shares must rise, since they are now claims to a larger enterprise. (Of course, this presumes what the market must judge, that the new investment is well-considered and likely to succeed). Since such companies retain their earnings, they will not distribute more than nominal dividends; hence their stocks and shares will be held for appreciation, reflecting the market's judgement of the success of the investments of such companies. If $ is the price of a share, then d$l$ will be the rate of share price appreciation, and if investments are well-planned, so that with a given technology and market conditions, new investments earn the same rate of profit as old, d$l$ will equal g, the rate of growth. Arbitrage should then ensure that, allowing for risk, d$l$ = i, the rate of interest. But at this point speculative activity enters the picture, creating a problem analogous to that which arose with hoarding in the case of a metallic currency. Suppose a flood of opportunities opens, raising g above i; this will lead the stock market at the margin to consider d$l$ > i. The initial quotes will reflect a small number of purchases based on expectations and the knowledge of well-placed traders. Reservation prices will rise. As information about the successful investments spreads through the market long-term bonds will be sold, and the funds will be 296 The Circuit of Money in a Production Economy shifted into growth stocks and shares. This will tend to depress bond prices, raising i, a move in the correct direction. But the shift of funds into stocks and shares will bid up prices, tending to RAISE d$l$ further, maintaining the inequality. (Whether the difference widens or narrows depends on the relative volume of bonds and growth stocks; but even if it narrows, it will never disappear.)25 At this point the effect of arbitrage with the bond market will have been to raise d$l$ above the rate of growth of the real capital stock. Hence companies considering investment must pause to reflect - stocks are risky, but they are rising. It might be worth delaying projects to take advantage of the rising stock market. As in the case of hoarding, this is a self-justifying move. If money is diverted from real investment, and from bonds, to investment in stocks and shares, d$l$ will continue to rise. The effect, of course, will also be to pull up interest rates. Both of these moves tend to justify additional issues of notes or credits by the banking system, adding fuel to the speculative fire. But as funds are diverted from real investment, g will fall. The problem began with g > i; market pressures will drive i up, and the diversion of funds will bring g down. But there is nothing to make the process stop once g and i are in balance! For the arbitrage is not between g and i directly; it is between i and d$I$, and the process has raised d$l$ above g. As long as funds are diverted into the stock market, d$l$ will continue to rise. Moreover, rapid price appreciation means that de$l$ is out of line with current real growth. At some point firms will not wish to further curtail their real investment, and portfolio managers will begin to pick up bond bargains, looking for a capital gain. At this point funds will begin to flow into the stock market at a reduced rate, and d$l$ will slow down. Considering that the stocks are overvalued and that bonds are a bargain, portfolio managers will begin to sell, and stock prices will come crashing down. But one again there is no obvious stopping point. Selling sets off a self-justifying downward spiral. Stabilizing a modem monetary system, then, requires stabilizing stock market booms and busts, or at least, insulating the real economy from them. Implications of the Golden Rule for the Monetary System On the one hand, then, the working of the monetary system depends on the Golden Rule, and, on the other, the working of the system also tends to enforce that Rule, provided that Government keeps the system's tendencies to instability under control. That is, changes from one point on the wage-profit tradeoff to another tend to be matched by correlative movements between equivalent points on the consumption-growth tradeoff. The matching is brought about by two kinds of market forces. On the one hand there are expenditure pressures. Changes in the normal wage tend to bring equivalent changes in consumer spending. Changes in normal investment spending tend to generate equivalent changes in normal profits. On the other hand, if g and r, or either and i, are out of alignment, arbitrage will tend to pull them back Edward J. Nell 297 together - barring outbreaks of speculation. On this interpretation the Golden Rule is not an equilibrium condition; it is an extension of the 'law of one price' to the earnings from capital. 26 The system depends on the Rule in four ways. First when r = g, the 'balancing condition' holds in a simple form. This is the foundation for the circuit. Second, if i =g then changes in the wage bill, due to a rise or fall in the real wage, will exactly equal the corresponding changes in payments to capital, due to the correlative change in the profit rate. So when distribution changes, funds that went to underwriting investment can be switched to financing payments to labor, or vice versa. Thirdly, when i = g, bank capital will expand at g, enabling the growth of justified bank liabilities to keep pace with the growth of the economy. Thus bank deposits and loans will grow at the same rate as the same rate as the demand for them. Fourthly, fixed capital producers can borrow from sinking funds to spend on circulating capital while continuing to produce and expand to keep pace with the rest of the economy. The Golden Rule is a condition relatingfiows of profit, interest and investment. It is not to be understood here as an optimality condition, nor as relating to 'bestpractice' techniques. The Golden Rule is the condition for the circuit to work, and it tends to be established and maintained by competitive pressures, provided that the system is not destabilized. The monetary system tends to pull i and g together through arbitrage. As we just saw if the two are not equal, funds will shift; in some cases this will pull them closer, but in other cases it may set off destabilizing speculation. Finally, it is worth remarking that there seems to be no tendency for i =g to prevail in the international economy, and there appear to be few curbs on destabilizing speculation either. . 9 THE SIGNIFICANCE OF CIRCULATION ANALYSIS The shortcomings of the supply and demand approach, treating money as an asset, have long been evident (Hicks, 1967; Moore, 1988), Until recently, there has been nothing to put in its place. Circulation analysis offers an alternative, which emphasizes that aspect of money most conspicuously neglected in the normal approach, namely, its role as a medium of circulation. It, too, relies on 'supply and demand', that is to say, on market forces, but in a context in which the institutional arrangements are paramount. The study of circulation reveals how expenditure takes place. The circulation approach relates the pattern of monetary expenditure to production. Competition and interdependence determine the form of the circuit. The form of the circuit can be seen most clearly when production is carried out by means of 'batch' processes: there will be a definite beginning and ending and a clearly defined sequence of activities. This provides a new way of treating time. Instead of the 298 The Circuit of Money in a Production Economy 'long run' and the 'short run', defined by whether or not the stock of capital can be varied, time is classified by the relation of activities to the circuit. There are, first, those that take place entirely within one circuit, next there are those variables that can be altered from one circuit to the next, and finally, there is the write-off period for fixed capital. Activities within a circuit, for example, consist of current spending, and current lending; whereas prices and money wages (in the modem economy) will only change from one circuit to the next. Embodied Technology and the capital stock change with the write-off period. Drawing on this makes it possible to describe the exact way a sum of money effectuates the exchange of goods, and the analysis defines the time required for circulation. There are three different patterns of exchange between money and goods in the circuit - a straightforward swap, a sequential process, and a rotation. Completion of the circuit requires that these be coordinated. Once the circuit is properly described, it is possible to explain the way the supply of money for circulation adapts to the need for it, without reference to the specific hypotheses concerning expectations, preferences or market behavior which appear to be required by the more conventional approaches. All that is necessary is that agents should be competitive in some broad sense, motivated by the desire for gain in some form, and generally aware of trends in the market. The argument depends only on the most general features of monetary institutions. But spelling out the working of different monetary systems reveals a new set of problems: these systems tend to generate instability. Metallic currencies run risks from hoarding; paper money systems are endangered by the tendency to overissue. Bank deposit systems tend to overlend. Maintaining monetary growth depends on an unstable stock market. In each case the instability can be seen to arise from a common source - the complex relationship between money's role as medium of circulation and its ability to function as an asset. In each case, as money develops and new problems of instability emerge, a new set of questions for practical policy analysis can be defined. To control or eliminate such instabilities there will have to be appropriate rules, and restrictions will have to be developed. These rules will have to be set and enforced by the state; monetary analysis becomes political economy. In addition, it is possible to develop expressions for the costs of the various monetary regimes. A metallic regime depends on the costs of equipment and labor in the mines; a banking system depends on the labor and resources required to provide financial services at the appropriate level. Together these two points, the need to control regime instability and to cover the cost of the monetary system, demonstrate how money is 'non-neutral'. It creates a different kind of potential instability, and a new category of costs. In addition, it can be misleading to examine the economy as if all behavior were settled simultaneously, as tends to be implied by supply and demand functions. In circulation activities take place in sequence. The actual economy must be studied as a monetary production system. Edward J. Nell 299 Notes 1. 2. 3. 4. 5. 6. Patinldn has characterised this as the 'invalid dichotomy'. The goods and factor equations determine relative prices and quantities, apparently leaving room for the money supply and demand equations to determine the general price level. But the money equations cannot be incorporated in this way; according to Walras's Law, if n - 1 equations are in equilibrium, so must be the nth. Yet this is not so; if all goods and factor markets are in equilibrium, the money market could still be out of balance, since the goods excess demand equations will be homogeneous of degree zero, while the monetary excess demand equations homogeneous of degree one (Patinkin, 1965). A more recent approach seems to explain money it terms of transaction costs. In Wicksell's account of the Quantity Theory it is clear that money is 'held' for the purpose of spending it (Wicksell, 1898: 39-41). But how to distinguish such balances from those that are held as a store of wealth, i.e. not to be spent, or spent only on other assets? Wicksell argues that the distinction cannot be made. (ibid., 1967: 21). Yet 'idle balances' have a completely different economic role. Hicks distinguishes money's role as a 'running asset' - more or less working capital - from its role as a portfolio asset and notes that there may be difficulties in reconciling them. A stronger case will be argued here, namely that the relationship between these two roles systematically creates instability in the monetary system, a pattern of instability which can be traced through metallic, paper and bank deposit systems, and which requires state intervention to control and contain it. It seems that Cartelier and perhaps others of the French School would deny this, holding that a dysfunctional circuit creates demands for financial assets. However these assets, in turn, eventually fail. Surely after experience, there would be learning and reform. Fairs were held several times a year, notably once each season, as the manorial economy had different products to market each time. Three kinds of pressure will operate to bring the town and countryside rates of surplus value (surplus over subsistence) into equality. First, factor mobility: in spite of ties to the land, labor will be mobile enough to establish a rough equality in levels of subsistence. Thus craft workers and peasants will tend to have comparable standards of living, adjusted for investment in acquiring skills. The same applies to merchants and lords. If merchants earn high rewards, lords can set up merchant operations in towns; less readily but occasionally, merchants can become landed. Second, monetary mobility: if merchants are not earning enough, they can move to other towns, or take their funds to other fairs. The shortage of funds will raise prices. Third, the level of rents and feudal dues can be adjudicated in court; prices will be set by town aldermen, and can also be appealed in court. The standard is the 'just' price, which allows each station the right or normal standard of living. It can be shown that this implies that prices equal direct and indirect embodied labor, which in turn implies equal rate of surplus value in different sectors. Suppose some of the merchants at the fair are from abroad; they sell their wares and depart with silver, reducing the supply for the country. Consequently, prices at the next fair must fall and the value of silver rise. It will therefore be worthwhile for merchants from abroad to travel there to make purchases - thus restoring the supply of silver. These are the circumstances that Hume had in mind, in arguing that 'the same causes ... must for ever, in all neighboring nations, preserve money nearly proportionable to the art and industry of each nation. All water, wherever it communicates, remains always at a level ... it is impossible to heap up money, more than any fluid, beyond its proper level .. .' (Hume, 1898, vol. I: 330-41). 300 7. 8. 9. 10. The Circuit of Money in a Production Economy Hicks argued that changes in the value of money would not be symmetrical in their impact on current spending. A rise in the value of money, as seen, will chiefly affect portfolio composition, with little or no impact on current spending. But a fall in the value of money (a general price rise) may require additional saving, to build money assets to the level required for transactions and precaution (Hicks, 1967). It should be remembered, however, that the burden of debt is similarly affected, but provides an opposite influence to that of the value of real balances, so even these admittedly tentative conclusions need to be considered cautiously. Competitive pressures could not form a profit rate in these conditions, however, because the means of production were not available to be bought and sold. Land was still held in feudal tenure. It was not alienable, therefore no market for it could develop. Hence rents could not be capitalized. Similarly, trades and crafts were organized by guilds - but guilds could not be bought and sold. There was therefore no market for firms. Production was still organized largely through the households. (In fact ways were found around these obstacles, but they were costly.) It may be objected that these pressures are weak and imperfect. In that case rates of surplus value will only weakly and imperfectly tend towards uniformity. The labor theory of value and its corollary, the rate of surplus value, provide an organizing principle explaining relative prices and distribution in certain pre-capitalist conditions. Some writers (e.g. Samuelson, 1941; Morishima and Catephores, 1978) have contended that the labor theory of value should never be interpreted as a description of actual historical conditions. Then in Western Europe, prior to the emergence of markets for means of production, what was the organizing principle governing relative prices? If it is argued that there was no general principle, why does the price data exhibit regularity and order? Why did the commentators, from Aristotle to Petty, believe that prices reflected labor values? Marx posed the problem, 'how the capitalist manages always to withdraw more money from circulation than he throws into it. The question at issue ... is not the formation of surplus value ... [A) sum of values employed would not be capital if it did not enrich itself by [creating) surplus value ... The question ... is not where the surplus value comes from but whence the money comes into which it is turned' (Marx, 1967, II, p. 330). Capitalists as a class advance the total money capital; it is then converted into productive capital, 'which then transforms itself within the process of production into commodities worth [more than the original capital) ... [so that] ... there are in circulation not only commodities equal to the money-value originally advanced, but also a newly produced surplus-value ... This ... surplus-value is thrown into circulation in the form of commodities ... But such an operation does not by any means furnish the additional money for the circulation of this additional commodity-value' (ibid., p. 331). He offered several solutions to the problem. The gold-mining sector could produce the additional funds; but then the entire surplus would end up in the hands of gold producers. He abandons this idea, and toys with the notion that 'the problem itself.. does not exist. All other conditions being given, such as velocity ... a definite sum of money is required in order to circulate commodities worth [a definite sum) ... independently of how much ... of this value falls to the share of the direct producers' (i.e. of how much it surplus value.) In other words, M' exchanges for C, the proposal discussed in the text. Later Marx notes that money is thrown into circulation by capitalists, who advance only their money-capital, equal to the means of production and wages, therefore to M, not M'. At this point he tries to set out a sequence of intersectoral transfers which will circulate surplus-value. Methodologically, this is the correct approach, although his account is defective (Nell, 1986). A peculiar exchange circuit appears in Blanchard and Fischer (1989: 178-9). The government issues money and provides it to 'individuals' as a transfer payment! Individuals deposit their endowments of goods with firms in exchange for bonds Edward J. Nell II. 12. 13. 14. (the firms are supposed to operate a 'storage technology', in which goods deposited with them increase over time); individuals buy goods from firms with money. Firms deposit the money with banks, who use it to redeem the bonds. This is not a plausible account of circulation. The government's behavior is unbelieveable; the banks do not seem to be necessary - or even convenient; the firms magically produce goods which individuals want without labor; and finally, taking the system on its own logic, neither money nor bonds appear to be necessary, since the firms could issue claim checks to the individuals (as interest on their bonds) which could be used to buy the goods. A modern variant of the exchange circuit sets the rate of growth of the money stock equal to the rate of growth of output (Foley, 1983, 1986; Semmler and Franke, 1989). The French account of the circuit depends on this condition being fulfilled, through some form of adjustment, yet many authors take the very strong position that markets provide wholly inadequate coordination (Cartelier, Deleplace). How then is this condition established? It will be argued below that a correct account of the circuit explains this. The payment of interest also presents a problem in Wicksell's theory of circulation (Wicksell, 1936, chapter 9). He proposes that interest expands the money supply in the same proportion that productivity expands the supply of goods - thus providing M' to purchase c', in Marx's notation. But he requires banks to accept costly deposits for which, in the context of his scheme, they have no possible use, and he fails to explain how they find the money with which to pay the interest (Nell, 1967). The secondary circulation consists of a potentially infinite set of transactions, in which. purchases of replacement capital goods provide revenue which subdivides into wages and gross profits. Let WI be the initial wage bill, which, spent by the merchants, is next received by the capital goods sector in the form of replacement spending by the consumer goods sector. These receipts will be divided into wages and gross profit, the latter, (1 - wb) WI' being spent on replacement. The wage bill for the first stage will be, W/c wb(l - wb)WI) of the second stage, Wile wb(wb(1 - wb) WI = wb 2(1- wb)WI and so on. This converges to [11(1 - wb))[l- wb)WI = WI. Hence the proper wage funds are distributed to all producers in the sector; as they pay them out to workers they will be spent on consumer goods, so the fund will return to the merchants. In the same way, the total gross profit of the sector can be calculated by summing the infinite series. The result is: [1 - wb)/wb)WI to which must be added the profit, in the form of its own investible surplus, in the 'final' subsector. The number of times this wages fund turns over, in order to facilitate the sale of that portion of the output of the consumer goods sector destined for its own workers, will be its 'velocity of circulation', using that term in the sense favored by both James and John Stuart Mill. 'The essential point is not how often the same money changes hands in a given time, but how often it changes hands in order to perform a given amount of traffic' (Mill, Bk3, ch. 8, S.3). Wicksell complains that this definition requires that we already know the value of money, so that velocity cannot become an independent factor in determining the value of money. But on the approach taken here, the value of (metallic) money is known, and the object of the analysis of circulation is to determine the required quantity - and to show that mint regulations are sufficient to ensure that the monetary regime will be stable. In the literature examining monetary exchange these patterns are described, but usually as if all exchanges took place in the same manner. Thus the French authors suppose that all money swaps for all goods in one exchange; Wicksell describes sequential transactions, as does Marx; while Keynes refers to a 'revolving fund' that ensures the circulation of goods. (Marx appears at different times to adopt all three, but never in combination). = 15. 16. 301 = 302 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. The Circuit of Money in a Production Economy Of course, saving for investment meant a move to a different system, one in which capital formed the basis of earnings, and competition engendered innovation, thus leading to the transformation of social life. In this case private virtues once again coincided with public - until the Keynesian era. Galiani believed that in the mid-eighteenth century, in Naples, hoards amounted to three to four times the amount of money in circulation. Others have argued for even higher ratios (Braudel, 1980, vol. I). Blanchard and Fisher (1989, chapter 5) survey the assumptions that give rise to instability in contemporary models. The earned interest will have to be invested in safe securities, such as government bonds, in order to be considered part of the capital - non-borrowed reserves - of the banking system. There will therefore have to be a regular flow of newly supplied such securities. This is the point of Keynes' remarks in Chapter 17: 236--9, where he argues that the standard of value must be whatever wages are most stable in. Does overissue of paper, or overlending, (leading to inflation) drive down interest rates? Real andlor nominal? If the interest rate were equal to the profit rate, would this prevent overissue andlor overlending? These questions have been extensively debated, but the discussion cannot be surveyed here. However, the key lies in the relation of the rate of interest to the rate of growth. The Golden Rule requires that these be equal; then the growth of money in circulation will just keep pace with the growth of output. And the growth of capital funds will just equal the growth of fixed capital requirements. But this equality is easily destablized, although there are also market forces tending to establish it. Recently, a 'New Monetary Economics' has developed, arguing that a competitive payments system could dispense with money altogether, making payments in the form of electronically transferred claims that will be settled in securities (Black, 1970; Fama, 1980; Greenfield and Yeager, 1983; Hall, 1982). This cannot be taken altogether seriously, since it assumes fully developed markets of all kinds, continuous eqUilibrium, and no uncertainty. Under these conditions all assets are equally liquid. In a more realistic economy there will be differences, and Gresham's Law will ensure that only the least reliable assets circulate, which will prove inconvenient if not unworkable. This may account for Ricardo's implausible argument that the gold mining industry had an 'average' capitaVlabor ratio, so that gold would not change in value with changes in distribution. As interest rates and stock prices rise, the value of bank portfolios 'Will rise. It is true that high interest rates imply lower values for bonds, but once the rise in the stock market begins to attract investment funds from companies the increase in the value of stock will outweigh the loss in bond values. In any case earnings from interest will be high, and demand for credit - for speculation on the market - will be strong. Hence banks will expand their offerings of credit. If i = r, the 'valuation ratio', v, must be unity - and, if both also g, following Kahn's formula literally, it will be undetermined (Kahn, 1972, Ch. 10). But there is a conceptual difficulty in Kahn's approach, which illuminates our arguments. Let N = number of shares, s = the retention ratio (corporate saving), r = profit rate, p = price of shares in money, i = rate of interest, g = growth rate of capital, K = capital stock expressed in Standard value (or labor value, anyway, in real terms) Kahn proceeds from v = pNIK, to dv + vdKlK = v(dNIN + dplp), after manipulating. In steady growth, dv 0, so vg v(dNIN+dplp), i.e. g dKlK = dNIN + dplp. Kahn then sets forth two definitional equations: = = = = Edward J. Nell 303 Investment is financed partly from retained earnings and partly from selling shares g :::: sr + pdN/K :::: sr + (pN/K)dN/N :::: sr + vdN/N, implying dN/N :::: (g - sr)/v. The return of shares, i, equals the dividend yield plus appreciation i :::: [J - s)rKJ/pN + dp/p = lIv(l - s)r + dp/p. Combining these, i =(ilv)(1 - s)r + g - dN/N =[(1 - s)rJ/v + g - (g - sr)/v, so that =(r-g)/v, andv= (r-g)/(i-g). i-g Hence if i:::: r, v:::: I, and if i:::: r:::: g, V is undetermined. But neither of the definitional equations makes sense dimensionally. Both i and g are pure numbers, as are dp/p and sr. But pdN/K has the dimensions 'money/ Standard commodity' (or other measure of real value.) These are also the dimensions of v, which therefore cannot equal an expression made up of dimensionless variables. Similarly the first term of the second equation, [(1 - s)rKJ/pN, has the dimensions 'Standard commodity/money', so cannot be added to dp/p, which is dimensionless. dN/N is likewise dimensionless, so cannot be equated to (g - sr)/v, which is not. Two changes must be made to correct these problems. The second term of the first definitional equation must be multiplied by I/v, to express the amount of real capital equipment the money raised by the new shares, pdN, will purchase. It then follows that g :::: sr + dN/N. From the basic identity, Kv pN, we can derive dKJK :::: pdN/Kv + dp/p dN/N + dp/p. Hence dp/p :::: sr. Correcting the second definitional equation requires rewriting its first term, to show the amount of share value the commodity dividend represents: = = i:::: [(1 - s)rKvl/pN + dp/p. Then cancelling and substituting we have :::: (1-s)r+sr= r. The Kahn formula cannot be used to object to our approach. References Black, F. (1970). 'Banking and Interest Rates in a World without Money', Journal of Banking Research, I, pp. 8-20. Blanchard, O. and S. Fischer (1989), Lectures on Macroeconomics (Cambridge, MA: MIT Press). Braudel, F. (1980), The Wheels of Commerce, vol. 2 of Civilization and Capitalism. 15th18th century (Paris: Presses Universitaires de France). Fama, E. (1980), 'Banking in the Theory of Finance', Journal of Monetary Economics, 6, pp.39-57. Foley, D. (1983), 'Money and Effective Demand in Marx's Scheme of Expanded Reproduction', in P. Desai (ed.), Marxism, Central Planning, and the Soviet Economy: Essays in Honor of Alexander Erlich (Cambridge, MA: MIT Press). Foley D. (1986), Money, Accumulation and Crisis (New York: Harwood Academic). Goodwin, R. (1974), Elementary Economics from the Higher Standpoint (Cambridge: Cambridge University Press). Greenfield, R. and L. Yeager (1983), 'A Laissezjaire Approach to Monetary Stability', Journal of Money, Credit and Banking, IS, pp. 302-15. Hall, R. (1982), 'Explorations in the Gold Standard and Related Policies for Stabilizing the Dollar', in Hall (ed.), Inflation (Chicago: University of Chicago Press). Hicks, J. (1967), Critical Essays on Monetary Theory (Oxford: Oxford University Press) 304 The Circuit of Money in a Production Economy Hume, D. (1752) 'Of Money' in Political Discourses (Edinburgh). Kahn, R. (1972) Selected Essays on Employment and Growth (Cambridge: Cambridge University Press). Koopmans, T.C. (1958), Three Essays on the State of Economic Science (New York: McGraw-Hill). Marcuzzo, C. and A. Rosselli (1991) Ricardo and the Gold Standard (London: Macmillan). Marx, K. (1967), Capital, vol. II (New York: International Publishers). Mill, J.S. (1848), Principles of Political Economy (New York: Augustus M.H. Kelley (1987». Moore, B. (1988), Horizontalists and Verticalists: The Macroeconomics of Credit Money (New York: Cambridge University Press). Morishima, M. and G. Catephores (1978), Value, Exploitation and Growth: Marx in the Light of Modern Economic Theory (London, McGraw-Hill). Nell, E.I. (1964), Models of Behaviour with Special Reference to Certain Economic Theories, PhD dissertation. Nell, E.J. (1967), 'Wicksell's Theory of Circulation', Journal of Political Economy, 75 (4), pp. 386-94. Nell, E.J. (1970), 'A Note on Cambridge Controversies in Capital Theory', Journal of Economic Literature, 8, pp. 41-4. Nell, E.I. (1986), 'On Monetary Circulation and the Rate of Exploitation', Thames Papers in Political Economy, Summer. Nell, E.J. (1992). Transformational Growth and Effective Demand: Economics after the Capital Critique (New York: New York University Press). Parguez, A. (1987), 'La monnaie, les rentiers et la crise', Economie et Societes, Series Monnaie et Production, 21 (9). Patinkin, D. (1965), Money, Interest and Prices, 2nd edn (New York: Harper & Row). Poulon. F. (1982), Economie generale (Paris: Dunot). Rizvi, S.A.T. (1992), Annual Supplement to Cambridge Journal of Economics. Samuelson. P. (1971) 'Understanding the Marxian Notion of Exploitation; A Summary of the So-called Transformation Problem Between Marxian Value and Competitive Prices'. Journal of Economic Literature. June, 9 (2), 394-431. Semmler, W. and R. Franke (1989). 'Debt Financing of Firms. Stability and Cycles in a Dynamical Macroeconomic Growth Model'. in W. Semmler (ed.), Financial Dynamics and Business Cycles (Armonk. New York: M.E. Sharpe). Steindl. J. (1952), Maturity and Stagnation in American Capitalism (New York: Monthly Review Press). Wicksell. K. (l898).lnterest and Prices (New York: Augustus M. Kelly (1965». Wicksell. K. (1936) Interest and Price. Trans R.F. Kahn (London: Macmillan). Wicksell, K. (1967), Lectures in Political Economy, vol. 2: Money (New York: Augustus M. Kelley). 9 Does Circulation Need a Monetary Standard?* Ghislain Deleplace Loosely speaking, the notion of a standard in economic literature usually carries the idea of something invariable, which helps to stabilize the economic system in on~ way or another. Three different meanings of this notion may be distinguished: 1. 2. 3. The standard of money, in which the monetary unit is defined, through the official fixation within that unit of the price of one unit of the standard. As an example, during more than two centuries England was on a gold standard, because the pound was defined as 123.24 grains of standard gold, which corresponded to the official price of £3 17 10 112 an ounce of standard gold, invented in 1717 by Isaac Newton, then 'Master of the Mint' . The standard of international monetary relations, in which all the monetary units of a given zone are directly or indirectly defined, and which may be obtained and used at a fixed price in each unit in order to settle international balances. As an example, most advanced economies have formed a gold standard system in various forms during the century up until 1971, when President Nixon 'closed the gold window'. The standard of value, in which relative prices may be measured in such a way as to neutralize the effects on measurement of the changes affecting the numeraire. The search for this standard is associated with a distinct strand of thought, originating with Smith and Ricardo, and has found an analytical expression in the concept of 'standard commodity' invented by Sraffa. These three standards have to be considered separately, although some links obviously exist between them. Moreover, their importance in economic literature is variable: the monetary debates in England at the beginning of the nineteenth century insisted on the (gold) standard as a domestic monetary norm,' attention was shifted at the end of the nineteenth century to the standard as an international exchange mechanism, and this is still the prevalent use of the word today, even after the adoption of an alleged 'fiat standard'; the standard considered as an analytical device which 'may give transparency to a system and render visible what was hidden' I is a specific issue which had completely disappeared during more than a century and is found uninteresting by mainstream economics. 305 306 Does Circulation Need a Monetary Standard? The analysis of circulation on which the Post Keynesian and circulation approaches concentrate may give more importance to the standard of money than to points (2) and (3) above, and I shall focus on it in this chapter. Nevertheless, the two other meanings cannot be completely ignored: one claim of this chapter is that the shift of emphasis from the standard of money to the international standard is a consequence of a theoretical change in the definition of the value of money, which altered the meaning of a standard relative to it; besides, the double role played by gold in Ricardo, as a standard of money and a standard of value, obscures the issue, and another claim is that both concepts are completely alien to each other and do not need each other. It remains that the focus here is on a concept of monetary standard which has practically disappeared from economic literature, whether orthodox or heterodox. It is suggested that such a disdain is not justified, particularly in two approaches which rely on the articulation between money creation and money circulation. A possible way of defining what could be a monetary standard in the monetary regime of today is proposed; it is based on the notion of standard debt. Two practical conclusions are derived, one negative and one positive: first, the conditions of monetary stability cannot be correctly analyzed as long as the value of money is defined as the reciprocal of the general price level; secondly, a necessary (but not sufficient) condition for monetary stability is discount lending by the central bank at a rate which is constant and as low as possible. The chapter is divided into five parts. Part 1 stresses the fact that the monetary standard has disappeared from modern economic literature. Part 2 returns to the gold standard, starting from Ricardo and presenting a simple model. Part 3 shows that the monetary standard and the standard of value should not be confused, and that a change in the definition of the value of money has been responsible for an unfortunate shift in the conception of the standard. Part 4 suggests that a debt standard may play the same role in today's fiat money regime as gold in the old days. Finally, part 5 draws some conclusions for the Post Keynesian and circulation approaches. 1 THE LOST STANDARD It is a common view that, since 1971, the world economy has moved from an (already modified) gold standard to a 'fiat standard'.2 What does that mean? As a first approximation, a gold standard regime is one in which the monetary unit is legally tied to gold. To what is it tied in a fiat standard? To government fiat money? But this would mean nothing except to say that the monetary unit is tied to money. The fiat standard must not be confused with an 'abstract' standard, as opposed to a commodity standard. Under the Bretton Woods agreement, money in most adva"ced countries was tied to the dollar, i.e. to a foreign monetary unit; whatever the link between the dollar and gold was, those monies could be said to be on a dollar standard, in the same way as they were on a gold standard in the nineteenth century. A so-called fiat standard, in which the monetary unit is legally Ghislain Deleplace 307 tied to nothing, means only that the monetary regime has no standard at all. There can only be a standard of money if money is tied to something other than itself. That a fiat money has no standard was clearly seen by Keynes in his Treatise on Money, where he makes the distinction between 'commodity money' and 'managed money' on the one hand, which 'are related to an objective standard of value', and 'fiat money' on the other hand, which 'has no fixed value in terms of an objective standard'3; the objective standard could be a commodity (e.g. gold standard) or foreign money ('exchange standard').4 The relation referred to by Keynes between a managed money and an objective standard seems surprising when confronted with the type of standard he advocates in the subsequent chapters. Among the different types of price levels he selects the 'consumption standard', 'for the position of the ideal objective standard in terms of which the monetary unit should be stabilised';s later, he answers in the negative to the question: 'is there such a thing as an "objective mean variation of general prices"?', thus criticizing Jevons and Edgeworth. 6 A basket of consumption goods may provide an objective standard of the value of money, but it takes the form of an arbitrary index number; nevertheless, it looks satisfactory to Keynes at a theoretical level because the consumption standard fits the requirement of 'corresponding par excellence to what we mean by the purchasing power of money', 7 and at a policy level because it provides 'a clear understanding of the social consequences of changes in the purchasing power of money'.8 Keynes's consumption standard, designed by the policy-maker to stabilize the value of managed money, is in sharp contrast with what is found in the traditional description of a commodity money regime, where the 'objectivity' of the standard has a different meaning: the choice of gold is inherited from history, constrained by social conventions and embedded in legal rules, and it plays the role of a regulator of money, imposed on monetary policy. As Robertson (1928) puts it, when he opposes two roles of gold in an 'arbitrary standard' and a 'commodity standard': 'It is open to such a country to maintain what is in effect an arbitrary standard, and to make the value of gold conform to the value of her money instead of making the value of her money conform to the value of gold'.9 Nevertheless, today the conception of a standard as a policy guide seems to be generally adopted, both by those who favor an active monetary policy and those who object to it. The understanding of a standard as the legal anchor of a self-regulating mechanism has remained only in the field of international economics, although with some confusion due to the abuse of the expression 'fiat standard' in reference to floating exchange rates. In what follows, I shall call monetary standard anything that is endowed with three properties: t. 2. 3. it is not money, i.e. cannot be used as a means of payment (no legal tender); the monetary unit is legally defined in terms of it; some specific rules exist which, combined with behaviors depending on selfinterest, aim at stabilizing the value of money in circulation in terms of it. 308 Does Circulation Need a Monetary Standard? Understood in that way, it seems that this notion of a monetary standard has disappeared from the pure theory of money with the vanishing of gold. Mason (1963), who already observed that it had left the lofty spheres of monetary theory to be confined to the applied field of international economics, traced this phenomenon back to the end of the nineteenth century and attributed it to two factors: the development of the banking system and of government fiat money, which historically disconnected money increasingly from the accepted standard, gold; and the adoption of a 'subjective theory of value', inconsistent with the previous reliance on the 'objective' value of gold. Mason proposed to rehabilitate the concept 'by conceiving a monetary standard as the criterion, or referent, of monetary policy': 'whatever is, in fact, used as the criterion of monetary policy constitutes the monetary standard'. \0 This proposal looks very similar to that of Keynes and seems to confirm that a standard of money can only be conceived as a policy J guide, and that no monetary standard with the above definition is still meaningful in modern monetary regimes, if it ever was. Maybe economists should stop thinking about concepts which no longer correspond to reality - maybe the concept of a monetary standard is one of them. But before doing so, let us check that all the lights in the house are correctly switched off. Which suggests the question: why were they switched on in the first place? 2 GOLD AS A MONETARY STANDARD One way to start is to raise the simple question: what was a monetary standard when gold was that standard? If Mason is right, this question should not be addressed to post-I 870s economists who buried the issue. Ricardo seems to be a good candidate (note that the return of England to the gold standard in 1925 was considered by Keynes as the final application of Ricardo's Ingot Plan).11 What did Ricardo mean by gold being the monetary standard? I shall argue that, for Ricardo, the monetary stability of an economy which is on a gold standard depends on its capacity to hold the market price of gold constant, the way to achieve that goal being independent of the circumstances which determine the real value of gold as a commodity. I shall then present a model of a gold standard regime, in which the devalorization and the depreciation of money are distinguished. Ricardo on the Gold Standard The question of the standard is treated by Ricardo during the 'bullion controversy' over monetary stability; for him the role of the standard is to regulate the value of money: 'During the late discussions on the bullion question, it was most justly contended, that a currency, to be perfect, should be absolutely invariable in value ... The only use of a standard is to regulate the quantity, and by the quantity the value of the currency'P After Mason (1963), but in a more convincing manner, Marcuzzo and Ghislain Deleplace 309 Rosselli (1991) have argued that for Ricardo gold was not a proxy of the modem consumption standard, precisely because the use of gold Was designed to separate the effects of monetary and of real causes on the general price level. 13 The 'bullion controversy' concerned the respective influences of two factors on the decline in the domestic and foreign value of money: the state of the balance with the rest of the world, and the state of money. The 'anti-bullionists' stressed the first one and the 'bullionists' (including Ricardo) the second one. Moreover, two strands of argument coexist in Ricardo. The main one deals with the determination of the domestic value of money, because for him it is the domestic factor which is dominant; here gold acts as a monetary norm, which stabilizes the domestic value of money. The second one deals with the determination of exchange rate movements; here gold is the cheapest exportable commodity, which is exported when a foreign balance has to be paid and which therefore stabilizes the foreign value of money. The two lines of reasoning may be separated, and, in accordance with the purpose of the present chapter, only the first one is evoked here. 14 The 'absolute invariability' in the value of money, which, according to the quotation above, would make it 'perfect', is out of reach for Ricardo. It cannot be achieved without a standard,15 and no commodity is free of variations in its value. Nevertheless, a satisfactory monetary regime may be obtained if the commodity with the least variability in value is selected as the monetary standard, and if monetary rules are adopted which prevent the occurrence of any variation in the market price of the standard that has a monetary origin. If both conditions are fulfilled, the domestic value of money, measured by the reciprocal of the market price of the standard, will vary only in response to those small real changes which affect the standard. 16 Two interpretations of this position can be distinguished. The first one is to consider that the accurate selection of the standard is outside the field of monetary analysis, which is confined to the specification of the rules which will make money conform to the standard. Ricardo's selection of gold is then explained by historical observations, later replaced by an improvement in the theory of value (see Section 3 below), and the monetary rules are understood as a set of constraints placed on the issuing policy of the Bank of England. Analytically, the value of the standard is exogenously given, and the purpose of Ricardo's Ingot Plan is to design rules which do not affect its money price. Keynes's endorsement of this plan departs from this interpretation. It relies on the assumptions that, first, another more accurate standard may be chosen, and, secondly, that the rules offer only a framework for money management using the standard as a policy guide. This latter aspect appears to contrast with Ricardo's mistrust in the managing abilities of the directors of the Bank of England, and eviscerates the self-regulating properties which Ricardo attributes to his plan. But the former one introduces the idea that the problem is not to neutralize the influence of money on the price of the standard (leaving the real causes of variation operative), but to design money so as to stabilize the price of the standard, 310 Does Circulation Need a Monetary Standard? whatever the circumstances. This idea is not logically connected to the adoption of Keynes's consumption standard or to his subsequent defense of a monetary policy. According to such an approach, Ricardo's view amounts to considering the market price of bullion as endogenously influenced by the monetary regime itself.17 Ricardo gives two signals in that direction. One is that, after having started chapter xxvn of the Principles with the affirmation that the value of gold money is determined by the cost of production of gold as a commodity, he turns his theory upside down and adopts another principle (the Quantity Theory), which breaks that link between the value of money and the conditions of production of gold. The second signal is that Ricardo's proposals in the Ingot Plan tend to disconnect supply and demand on the market for gold itself from the influence of production: the removal of gold from domestic circulation and its concentration in the Bank of England's vaults, together with the regulation of the issl>e of bank money, aim at providing a market adjustment without relying on an increase in the production of gold. 18 Continuing that line of reasoning, the crucial point is to understand precisely in what sense an increase in the market price of gold bullion measures the current depreciation of money; the description of a monetary economy in a gold standard regime may be of some help. A Model of a Gold Standard Regime: Devalorization and Depreciation of Money I shall here consider a pure gold standard regime, in which money is only composed of full-bodied gold coins. The extension of the model to light coins and to bank notes, either convertible or inconvertible, will be evoked later. Such a monetary economy contains a mint, which converts gold bullion into gold money (coins), and a gold market, where bullion is sold for coins. Money creation occurs in two stages: first, a physical unit of measure is defined for gold of a given (standard) quality, and the official price in terms of a unit of account is fixed for it, at which the mint purchases gold of that quality without limit; secondly, the content of the coin in standard gold is decided, and the price at which the coin will legally circulate is fixed. There are then two official prices of the physical unit of the standard gold in terms of the unit of account: one is the price of uncoined gold, the other is the price of coined gold, i.e. money. Fixing the second one defines the official value of the monetary unit as a weight of standard gold. This is summarized in the following model I: (1) (2) with SG as the weight of uncoined standard gold purchased £1 by the mint; T as the official price of uncoined standard gold at the mint; VM as the official value of Ghislain Deleplace 311 the pound as a weight of coined standard gold; and PM as the price at which one unit of coined standard gold must legally be taken in circulation. Usually, PM> T, the difference being the cost of minting (seigniorage).19 Then VM < SG, which means that the transformation of gold into money entails a loss of standard, which I shall call a legal de valorization of money, because £ I of gold loses value (here, plainly weight) when coined in money. If 8 is that rate of legal devalorization of money, we have: hence: (3) In circulation, i.e. in the market for gold bullion, standard gold is exchanged at a price P 6, and the current value of money is: vA: = IIP6 (4) The current de valorization of money 8* may be defined as: 8* = (SG - V A:)IvA: hence: s* = (P ri - T) IT (5) There is an obvious relation between 8 and 8*, which is derived from (3) and (5): s - s* = (PM - P ri)IT The level of s* measures the percentage loss of weight borne by gold when used in the form of money in circulation. SG being given, any decline in V A: is reflected by an increase in 8*. A depreciation of money, defined as a decline in the current value of money in terms of the standard, is revealed by an increase in the market price of gold bullion P ri; the official price of uncoined standard gold T being given, a depreciation of money means an increased current devalorization of it. Is there any limit to that depreciation? The working of the market for gold bullion provides the answer. Many factors may influence P ri: cost of production, demand for industrial use, etc. But whatever they are, no seller will agree to be paid in the gold market below the price at which he or she may sell to the mint (1); if c is the percentage cost of melting coins (including the cost of fraud if it is forbidden), no buyer will accept to pay more than PM (I + c) to obtain one unit of gold bullion. Therefore: (7) Does Circulation Need a Monetary Standard? 312 or, put another way: (8) These inequalities only exhibit price relations, and it would seem that they must be complemented by a quantity side, which in a monetary economy would concern money flows. For example, thanks to the Quantity Theory of money, the enforcement of the inequalities would be explained by the relation between a change in P 6 and the quantity of money. Any increase of P 6 above PM (1 + c) would lead to melting down coins, hence reducing the quantity of money and provoking a decline in P 6. Reciprocally, any fall of P 6 below T would induce people to bring their gold to the mint, hence increasing the quantity of money and therefore P 6. But nothing like that is necessary. Provided the mint is ready to absorb any quantity of gold, paying for it with the coins it makes, T will be an operative lower limit for P 6; the only consequence may be the lengthening of the time necessary to make the coins, hence an increase in s. Provided there are enough coins in circulation to be melted down, PM (1 + c) will be an operative upper limit for P 6; the only consequence may be an additional cost (included in c) to find them. This means that the range of variation of P 6 may not be fixed (depending on the changes in s and c). But this is not a reason to consider that only a process of adjustment to equilibrium-such as the one contained in the Quantity Theory of money - may stabilize P ci. There is a simpler way to achieve the same result, and it is relying on individual arbitrage and institutional arrangements which lower s and c - hence the range of variation of P 6. Although it is clear that the quantitative approach goes through Ricardo's monetary theory, one could not understand many of his monetary proposals if the 'rules of the game' approach had not been also present. This may appear more clearly by stressing five conclusions, which clarify the understanding of a gold standard regime. 1. Given the official price of uncoined gold, the current devalorization of money may only increase - therefore money may only depreciate - up to the limit set by the official price of coined gold plus the melting cost. Any diminution of the melting cost reduces then the possible range of depreciation of money. If c =0, inequalities (7) and (8) become: PM~pri~T (7') SG ~ V': ~ VM (8') Hence P ri cannot exceed PM, and the value of money cannot fall below its legal value. Hence the restrictions on the melting and the exportation of coins must be lifted; a still better measure would be to adopt a money which could be converted into its standard at no cost. Ghislain Deleplace 2. 313 In the case of no melting cost, equation (6) leads to: (6') s'" s*'" 0 There can still be a current devalorization of money (hence an increase in it, i.e. a depreciation of money), but it is limited by the legal one. The lower the cost of creating money with the standard, the smaller the possible range of depreciation; the ideal would be a money created at no cost. 3. The opposition of Ricardo to the use of metallic money and his plan for a public bank issuing paper money are not only rooted in the fact either that a precious metal is expensive to produce or to obtain by foreign trade, or that the physical use of coins in circulation leads to a degradation of the metal. The cost which has to be considered entails the transformation of the standard into money (cost of minting) or of money into the standard (cost of melting). Both costs are to be avoided because they impede the stability of the monetary regime by increasing the range in which the current value of money may vary. 4. Inequalities (7) and (8) concern the value of a money defined by equation (2), i.e. the weight VM of standard gold in full-bodied coin; in that case, the value of the monetary unit (the pound) and the value of the standard means of payment (the sovereign) are in a fixed ratio. When other means of payment exist side by side with the full-bodied coin (or drive it out of circulation), any increase in the market price of bullion above PM (1 + c) cannot be attributed to a depreciation of the monetary unit in terms of the standard, but to a depreciation of the means of payment used in the bullion market, in terms of the monetary unit. This is the case for coins underweighed because of wear and tear, or for undervalued Bank of England notes, either when a 'run' for convertibility obliges the Bank to purchase gold in the market to replenish its reserves, or in a situation of inconvertibility. 20 5. The limit set by a gold standard regime on the depreciation of the monetary unit is independent of any particular cause of depreciation (whether domestic or foreign) and of any particular policy designed to prevent or 'correct it; it is merely the result of the monetary rules embedded in that regime. Besides, no assumption is made about the quantity of money issued and the relation between that quantity and the value of the monetary unit; the quantitativist view of the value of money is superimposed by Ricardo on the analysis of the monetary rules, and the conclusions drawn by him as to what the behavior of the Bank of England should be concern the regulation of its notes, not the value of the monetary unit. This does not mean that the working of a gold standard regime can be reduced to the above model because foreign exchange relations have been left aside; in particular, it remains to be understood what rules and institutional arrangements may prevent an external shock from drying up the reserves of the Bank of England in case of convertibility.21 A last remark must be added which will be of consequence below. Ricardo is able to concentrate on how to make the cos~ of 'money as low as possible because he assumes that, in the conditions of his time, they are officially constant. In other words, he does not have to argue against their increase because there is no legal 314 Does Circulation Need a Monetary Standard? threat in that direction: the legal value of money has not changed in England since Isaac Newton's reform of 1717 (and it would not change until 1931). But one should remember that if the cost of melting is nil, so that the range of depreciation of money is limited by the legal devalorization of it (the cost of minting), a necessary (but not sufficient) condition for monetary stability is the constancy of the legal devalorization of money. This was the main question raised by monetary debates in France in the 1570s, which, at that time also, concerned the instability of the monetary regime; and the conclusion was that the ruler should stop changing the official value of money. 22 At the time of Ricardo, this conclusion was well understood, and debates could switch to the relation between sand s*, and to the best ways to make the constant level of s as small as possible. A striking feature of the above model is that only the market price of gold in monetary unit P~ shows up, never its value in a commodity num~raire; therefore the model exhibits no link whatsoever between the value of money and the value of gold as a commodity, and hence between the underlying theory of money and value theory. This raises the question of the relation between the concept of a monetary standard and two other concepts: the standard of value used to measure relative prices in the Ricardian tradition, and the tabular standard used to measure the value of money in mainstream tradition. 3 MONETARY STANDARD, STANDARD OF VALUE AND TABULAR STANDARD The Monetary Standard and the Standard of Value The concept of a standard seems to provide a link between Ricardo's theory of value and his theory of money, because for Ricardo gold is not only the monetary standard but also the invariable standard of value. This view has been strengthened by Marcuzzo and Rosselli (1991), who advocate that 'as the standard 'of currency and the standard of value, gold serves two distinct purposes', although 'the function of gold as a monetary standard is independent of its function as a measure of value'. 23 Two questions should be raised in order to clarify this point: what sort of invariability is required by the function of the standard of money and the function of the standard of value? How is this invariability ensured in both cases? To specify the nature of the invariability, the first task is to give a definite meaning to Ricardo's invariable measure of value. A broad agreement exists that Sraffa's 'standard commodity' is the correct expression of what Ricardo was looking for, although invariability is restricted to changes in distribution only (excluding changes in production methods). The criterion of that invariability is the equality between the value-ratio of net product to means of production in the industry producing the standard, and the same value-ratio 'in all the successive layers of the industry's aggregate means of production without limit',24 Ghislain Deleplace 315 Ricardo's analysis of the monetary standard aims at finding a commodity, the market price of which would be fairly constant in respect to any circumstance (except a 'run' on the convertibility of banknotes). Obviously, the two meanings of invariability do not overlap. The problem of the standard of value concerns the relative prices of the commodities (in their relation to distribution); the problem of the monetary standard concerns the monetary price of one commodity (in any circumstance). The same contrast appears in the method of ensuring invariability. In the case of the standard of value, the condition of 'recurrence' of the ratio implies that the standard commodity is a composite one, built in a way which reflects the conditions of production of the basic commodities in the system. In the case of the standard of money, the constancy of its market price is obtained, as seen above, via adequate monetary rules combined with rational individual behavior in the market for gold, irrespective of its production. This monetary approach to the standard of money contrasts with the real analysis of the standard of value, based on conditions of production and distribution. This contrast between both standards may be stressed in another way. The 'recurrence' condition of existence of a commodity invariant in value is very restrictive, and inconsistent with what is expected from a monetary standard, i.e. invariability in any circumstance. Then the quest for a monetary standard invariant in real value seems impossible. But we do not need that; all we need is a commodity invariant in money price, i.e. a commodity whose market price may stay constant even if there is a change in conditions of production and distribution (which affect its real value, i.e. its relative price in terms of any other commodity). As far as domestic influences only are considered, this task will be achieved if monetary rules are designed which, in most circumstances (i.e. apart 'runs'), will constrain this market price between narrow boundaries; this is the purpose of the Ingot Plan. Ricardo will still be interested later in the relation between absolute value and exchange value, but this does not concern the problem of the monetary standard, which he has solved in a monetary way. The concept of a standard of value is completely distinct from the concept of a monetary standard, and is useless in Ricardo's theory of money; contrary to what Ricardo assumed, no standard can provide a link between the theory of value and the theory of money. The Tabular Standard: A Change in the Definition of the Value of Money The emphasis put by Jevons (1875) on the 'tabular standard'25 and its popUlarization by Fisher (1911) changed the conception of the value of money. First used as a standard of deferred payments, complementary to gold used for all other transactions, it progressively replaced gold in the literature to measure the value of money. The tabular standard is constituted by a basket of commodities, which is considered as representative of consumption, and the value of the monetary unit is the value of the basket which it may purchase; then the value of money is given by the reciprocal of the general price level, computed on this standard basket. 316 Does Circulation Need a Monetary Standard? One may observe that the success of this tabular standard and its substitution for gold to measure the value of money occurred while the gold standard regime was still operating, indeed at its climax, with its generalization to most advanced countries in the 1870s. This suggests that this shift was not rooted in historical phenomena, but in a theoretical change in the conception of the value of money. As a consequence, although gold continued to be the domestic monetary standard (the official definition of the pound had not changed), this notion disappeared from the theory of the value of money; and, in a parallel movement, gold was confined to the pivotal role in the international exchange mechanism. The tabular standard did not replace gold as the monetary standard, but it replaced the monetary standard in the pure theory of money, thus shifting gold from that theory to the applied field of international economics. This theoretical change has two consequences. First, the standard used in the theory of money is no longer independent of the theory of value. For Ricardo, the current value of money being defined as the purchasing power of the monetary unit over the monetary standard, gold, its measure was a weight of gold, as was the official definition of the pound, and not the value of a weight of gold. The definition of the pound has not changed in Jevons's and Fisher's time, but the current value of money is now defined as the purchasing power of the monetary unit over a basket of commodities, not over the monetary standard. Thus its measure cannot be assessed without computing the value of the aggregate basket. While for Ricardo a change in the relative price of gold in terms of other commodities did not affect the value of money (but only the purchasing power of money over those commodities), a change in the relative prices of the commodities composing the basket affects the value of money. Hence the long-lasting controversiesabout how to construct an index-number which measures the 'true' value of money. This shift of emphasis even affects the analysis of the gold standard itself, made by authors like Hawtrey who carefully approach it through the definition of the monetary unit: The foundation of the gold standard is the tying of the value of the monetary unit to the value of gold by the fixing of the price of gold. Inasmuch as gold is a commodity with a world market, it has a world value, and therefore the gold standard gives a world value to the monetary unit itself. 26 The 'value of the monetl!.ry unit' is its 'wealth-value, that is to say, its value in goods and services' ,27 measured by the reciprocal of the general price level index. Being legally defined as a weight of gold, the monetary unit has a value in terms of goods and services which is by construction equal to the value of this weight of gold in terms of goods and services, determined on the markets at world level. A change in the relative price of gold means ipso facto a change in the value of money. One may see that the definition of the value of money as a purchasing Ghislain Deleplace 317 power over all the commodities makes it dependent on value theory, which determines the system of relative prices. A corollary is that it is no longer possible to distinguish in the changes in the general price level between those which have their origin on the side of money from those originating on the side of the commodities; all of them are by definition attributed to a change in the value of money, although one knows that some of them will be provoked by a change in the relative prices of the commodities, while others will be purely monetary. This means that even if the standard is conceived as a policy guide for money management, the use of that standard may be misleading to design such a policy. Money Creation and Money Circulation A second consequence of this definition is its inconsistency with the analysis of money creation. When the value of money in a gold standard regime is defined in the monetary standard itself, it is the same commodity, gold, which is coined at an official price (thus defining the monetary unit), and which is traded for money in the market. Under its two forms, in coin and in bullion, gold connects money creation and money circulation. When, in the same regime, the value of money in circulation is defined in a tabular standard, this connection is broken. A basket of commodities measures the value of money in the markets, but this basket is not 'minted', i.e. transformed into money at an official price; the creation of money continues through the minting of gold, or the issuing of banknotes convertible into gold. The analysis of money creation and of money circulation fall apart, the former in the institutional description of administered prices, the latter in the market theory of relative prices. The following model IT sums up such an analysis: SG= liT (I) VM IIPM = (2) s= (PM- nIT (3) PM(I+c)~pci~T (7) V:=IIPi. (9) with SG as the weight of standard uncoined gold purchased £ 1 by the mint; T as the official price of standard uncoined gold at the mint; VM as the official value of the pound in weight of standard gold; PM as the price at which one unit of coined standard gold must legally be taken in circulation; s as the cost of minting (seignorage); c as the cost of melting; P ci as the market price of gold bullion; V: as the current value of money; Pi. as the general price level. Equations (I), (2) and (3) are the same as in model I, but equation (9) replaces equation (4) to define the current value of money. Hence inequalities (8) con- 318 Does Circulation Need a Monetary Standard? straining the movements of V J cannot be derived from inequalities (7). No builtin monetary rules will stabilize the current value of money, which can only be influenced by a monetary policy controlling the quantity of money, provided the Quantity Theory of money is accepted. The disappearance of the monetary standard and the emergence of the tabular standard in the literature have paved the way for the present state of monetary theory in mainstream economics28 , which is split into two separate levels. At one level, the question of the existence of money (the positivity of its price in terms of the general equilibrium numeraire) is raised, money being defined by its functions of medium of exchange and store of value; this question is solved through an analysis of the demand for money. At another level, the rationale of the demand for money is posed by its being an argument in the household utility function, and the question of its value (the general price level in money terms) is raised and solved thanks to the assumption of an exogeneity of the supply of money. The first level is satisfied if a microeconomic rationale for the use of money is demonstrated, but the arbitrary character of the assumptions which have to be made in order to prove that money exists does not permit any determination of the quantity of money required, hence the need to approach the question of its supply. At the second level, the existence of money is simply assumed; the exogeneity of its supply is then consistent with the ignorance of where it comes from. 29 The Post Keynesian and circulation approaches could also be accused of discarding the need to prove the existence of money. But they are not in the same position as general equilibrium theory: they do not assume that there is a nonmonetary model of the economic system which gives the essential laws of its functioning. Hence they do not have to prove that money can be introduced as an afterthought into this model, in order to represent a monetary economy. This does not mean that microeconomic rationale should not be considered in the Post Keynesian or circulation approaches, but that the use of money in interindividual relations is not merely a matter of choice by agents: they have to comply with this rule when they are in a monetary economy. The central problem of monetary theory is therefore raised in another way than in the general equilibrium tradition: it consists in stating the conditions under which public rules and private behaviors are consistent with one another. This task is not easier than proving the existence of money and dissatisfaction about the results is probably as great. By not being constrained by a real theory of value, the Post Keynesian and circulation approaches are better able to link the creation and the circulation of money. It seems that the Post Keynesian attack on monetarist and standard Keynesian exogeneity of the money supply has led this approach to focus on the endogeneity of the quantity of money, and to neglect the analysis of the conditions of money creation. But some Post Keynesians like Moore 30 acknowledge the importance of this analysis and concur with the circulation approach which has always taken it into account. 3l Ghislain Deleplace 4 319 WHAT COULD BE THE MONETARY STANDARD OF TODAY? Monetary Regime, Monetary Standard and Standard Money A monetary economy may be defined as an economy in which all transactions are measured in a common unit of account and executed by a means of payment. More precisely, three features characterize such an economy: 1. 2. 3. a common unit of account (monetary unit), in which the amounts figuring in private contracts are written; a procedure allowing the transfer of those amounts, i.e. a means of payment (money); a procedure allowing private agents to have access to money, i.e. an issuing principle. The unicity of the unit of account is a definitional feature of a monetary economy, in contrast with a barter economy in which each transaction is measured in its own unit. This point has been acknowledged by a long heterodox tradition,32 in which the measure of prices is part of the theory of money, while the general equilibrium approach relies on a numeraire to perform that function and starts the analysis of money with its function of medium of exchange. The monetary unit and the means of payment are two distinct features. This was clear when two different names were used in practice to designate them (e.g. the pound and the sovereign), but it is no less true today: a dollar written on a price-list is not the same thing as a dollar printed on a Federal Reserve banknote. This distinction permits several means of payment to circulate side by side, provided their values are all expressed in the same monetary unit. A monetary regime may be defined as a particular set of rules concerning the monetary unit, the means of payment and the issuing principle. It has a monetary standard if two conditions are fulfilled: the monetary unit is defined in terms of that standard, and anyone may bring it to a special institution to obtain a means of payment with legal tender, which is issued on demand, in unlimited amounts and at a legal price; this is then called the standard money. The standard may be anything: gold or silver in a metallic standard regime (gold and silver in a bimetallic one), a foreign currency (e.g. the dollar) in an exchange standard regime. The existence of a monetary standard does not depend on its material form, but on the legal access to standard money which it provides for its owner. The monetary standard is then a concept of its own: first, it is different from the monetary unit, the latter being defined in the former; secondly, it is different from money because it has no legal tender: one cannot pay in bullion even in an economy with full-bodied metallic coins or in dollars in a non-US domestic economy (like everything, it can be used as a means of payment on the basis of a private agreement, but it is not money because this use is not fixed by law). 320 Does Circulation Need a Monetary Standard? When the monetary unit is not legally defined in something, as is the case since the collapse of the Bretton Woods system, there is no monetary standard; but it does not mean that the monetary unit is just a name, which may be fixed arbitrarily by law. It is a common feature of all monetary regimes, having a monetary standard or not, that the enforcement of the monetary unit cannot just be ensured by the apparatus of the state. For private agents, the meaning of the monetary unit appears only in transactions, when they use a means of payment to execute transactions written in that unit. Without an official relation between the monetary unit and the means of payment, no unicity of the unit of account could exist: each sovereign or each one-dollar banknote could be used to settle a different amount of the monetary unit in different transactions. Therefore, the existence of a monetary regime requires the legal declaration of the relation between the monetary unit and the means of payment, and it is not just a matter of providing common information to private agents. The only relation in which the declaration of the monetary unit may be enforced by the state is when private agents have to obtain money from a public authority, in compliance with the issuing principle; this is why, even when gold is 'sold' to the mint by someone for coins, this 'sale' has a public character as distinct from private transactions. A corollary is that an analysis of a monetary economy based on the concept of a common unit of account needs to link money circulation (the relation between the monetary unit and the means of payment in private transactions) and money creation (the relation between the monetary unit and the means of payment in the public issuing of the latter). The issuing principle is, within the same operation, the legal definition of the monetary unit and the money of legal tender. In a regime with a monetary standard, this definition is made in terms of this standard: the monetary unit is defined as a definite quantity of the standard, and the money. of legal tender, called standard money, is defined as containing a definite quantity of the standard. As a consequence, there is an official value of the currency in terms of the monetary unit. There can be no monetary standard without a standard money. For example, in England after Liverpool's Act of 1816, the pound was defined 123:24 grains of standard gold, and the sovereign legally contained 123:24 grains of standard gold, thus bearing a value of £1. This correspondence is not only assessed in proclamations, but may be checked at any time at the mint, when an ounce of standard gold is monetized in sovereigns on demand and without limit at the fixed price of £3 17 10 112 (which denotes £1 as 123:24 grains). The nominal inscription of £1 on the sovereign is only a matter of information about the legal value of the coin in circulation, but the proof of the gold standard de jure does not appear there: it lies in the monetization of gold in the shape of the sovereign at the mint price of £3 17 10 112. Another example is provided in a bimetallic regime, as in France in the middle of the nineteenth century. The monetary unit, the French franc, is defined in two ways: either as 0.29 grams of pure gold or as 4.5 grams of pure silver - implying Ghislain Deleplace 321 a monetary ratio of 15.5, as famous in Europe at that time as the English mint price of gold. The corresponding standard monies are, on one hand, the 20F coin, containing 6.45 grams of gold 90011000 fine and, on the other hand, the IF coin, containing 5 grams of silver 90011000 fine. This relation between one single monetary unit and two standard monies, assessing its definition in two distinct but strictly connected standards, may be checked at any time at the mint, where a kilogram of pure gold is monetized in 20F coins at the fixed price of 3434.44F and the kilogram of pure silver in IF coins at the fixed price of 218.89 F.33 A third example is the non-US countries belonging to the gold exchange standard system. The domestic monetary unit is defined in terms of the dollar, which acts as the standard, and any amount of dollars can be exchanged at the central bank for the domestic currency at a fixed exchange rate. 34 I have noted that there may be several means of payment in a monetary economy with a single monetary unit. This means that the issuing principle of each one must either include an official relation between the money and the monetary unit (legal tender), or include a rule of fixed convertibility against the money of legal tender on demand and without limit; both rules must be endorsed by the institution which issues the particular money. In a monetary economy, the official relation between the monetary unit and money ensured by the issuing principle of the money of legal tender may not always be respected in private transactions (in circulation): it is not possible to put a policeman or a judge behind every pair of traders. Money of legal tender may therefore have a 'voluntary' value in the monetary unit as distinct from the legal one. This discrepancy will take different forms according to the regime involved (a 'voluntary' currency value of the coin; a bullion market price different from the mint price, a 'black market' price of the domestic banknote in terms of the dollar, etc.), and this devalorization of money is the sign that the monetary link is dis functioning. Standard Debt in a Fiat Money Regime The theoretical foundation given above to a concept of monetary standard consistent with different regimes of a monetary economy may allow us to look for such a standard in a regime with fiat money. Three questions must be asked: I. 2. 3. What is the issuing principle in such a regime, i.e. against what is money of legal tender issued on demand, in unlimited amounts and at a legal price in the monetary unit? How may the value of money be assessed in such a regime? What are the respective criteria of a legal devalorization of money (a loss of standard provoked by its monetization), of a current devalorization of money (a loss of standard provoked by its use in money form), and of a depreciation of money (an increase in its current devalorization). Does Circulation Need a Monetary Standard? 322 If these questions are answered, results may be reached as to how to provide monetary stability assumed from past metallic standards in such a regime. A fiat money regime may be defined as one in which an inconvertible money issued by a central bank is legal tender. The creation of this base money results from the monetization by a central bank of various sorts of debt owned by banks. If money creation is endogenous, this monetization occurs on demand, in unlimited amounts and at a fixed price through discount lending, which provides borrowed reserves for the banks. Central bank money might then qualify for the role of standard money, and the type of debt eligible to the 'discount window' of the central bank might apply for the role of monetary standard; in the same way as a given quality (,standard') of gold is assumed in a metallic regime, a given type of such debt may be called standard debt. Model I of a gold standard regime may then be transposed to a debt standard regime, and this gives model III: SD= liT (10) (11) with SD as the amount of unmonetized standard debt purchased £1 by the central bank; T as the official price of one unit of unmonetized standard debt at the central bank; VM as the official value of the pound in amount of standard debt; PM as the price at which one unit of monetized standard debt (central bank money) must legally be taken in circulation. If r is the official discount rate, T = 1/( I + r); the monetary unit being defined as one unit of central bank money, PM = l. Hence: (12) (13) As with the gold standard, the legal devalorization of money is the loss of standard entailed by its transformation into money, at a rate s = (SD - VM)/VM; then: s=r (14) This result is quite trivial: the percentage cost of money creation is equal to the official discount rate. In circulation, i.e. in the money market, standard debt is exchanged at a price P and the current value of money is: Z, V~ = I/PZ (15) The market price of one unit of standard debt PZbeing 11(1 + i*), with i* the rate of interest in the money market, the current value of money is given by: V~ = I + i* (16) Ghislain Deleplace 323 Equation (16) might lead to a trivial conclusion: a depreciation of money (i.e. a decline in V:) is reflected in a falling rate of interest in the money market. But one should refrain from going so fast, for two reasons. First, the Keynesian flavor of the interest rate measuring liquidity preference is misplaced: equation (16) concerns money as a means of circulation endowed with general purchasing power (opposed to assets deprived of that quality), not money as a store of value (compared to other less liquid financial assets). JS Secondly, the reciprocal of P ri in (15) does not measure in itself the current value of money in circulation, just as the reciprocal of P ri in (4) of model I does not measure it in itself in a gold standard regime. If it were the case, there would be no reason to discard the reciprocal of the general price level pi as an appropriate measure of the value of money in equation (9) of model II. In fact, Pri and P ri may play that role (and not pi) because debt or gold act as a monetary standard, and this property is assessed by the existence of a legal price T of unmonetized standard debt or standard gold; IT in the gold standard regime. A depreciation of hence the use of s* = (P ri money cannot simply be computed on the basis of any market price, but defined as an increase in the current devalorization of money, i.e. in the percentage loss borne by the standard used in the form of money in circulation. In the case of a debt standard regime, this current devalorization of money is s* = (SD - V :)/V:; it follows from (10) and (15) that: n s* = (P ri - nIT (17) which is analogous to equation (5) in a gold standard regime. Another expression of s*, derived from (12) and (16), is: s* = [(1 + r)/(1 + i*)]-1 (18) Is there any limit to the depreciation of money? In the money market, no seller of standard debt will agree to be paid below the price for which he may discount it at the central bank (n; no buyer will agree to pay more than PM to obtain one unit of standard debt (by comparison with the gold standard regime, this means that the cost of 'melting' one unit of central bank money into standard debt is nil, because central bank money is itself standard debt par excellence). Therefore: (19) hence: (20) or: (21) 324 Does Circulation Need a Monetary Standard? Using inequalities (21) in equation (18) gives: r;;;' s*;;;' 0 (22) Five conclusions may be derived from this analysis: 1. There is no fundamental difference between a fiat money (debt standard) regime and a metallic money (gold standard) regime. In both cases, the stability of the regime depends on the combination of monetary rules and of microeconomic arbitrage between the market of the standard and the issuing institution. 2. A fiat money regime with a positive official discount rate is characterized by a legal devalorization of money, the magnitude of which is equal to the discount rate. This phenomenon is analogous to the devalorization of money produced by a positive cost of minting (seignorage) in a metallic money regime. 3. The criterion of monetary stability in a fiat money regime is not the constancy of the general price level, but the constancy of the rate of current devalorization of money. Any increase in the difference between the current price and the official price of standard debt (equation (17», i.e. any increase in the difference between the official discount rate and the interest rate in the money market (equation (18», means a depreciation of money. The comparison between the change in the rate of current devalorization of money and the change in the general price level allows us to distinguish between monetary and real causes of inflation. 4. The increase in the rate of current devalorization of money, which reflects its depreciation, is limited by the official discount rate at the central bank. Hence a very simple monetary rule may enhance the stability of such a regime: the official discount rate should be kept constant, and as low as possible. 5. The limit set by a debt standard regime to the depreciation of the monetary unit is independent of any particular cause of depreciation, and of any particular policy designed to prevent or to correct it; it is merely the consequence of the monetary rules embedded in that regime. Besides, no assumption is made about the quantity of money issued, and about the relation between that quantity and the value of the monetary unit. The meaning of depreciation associated with equation (18) may look strange, not only because it repudiates the usual reference to the general price level, but also because the meaning of a current devalorization of money seems unclear in the case of fiat money. However, a common interpretation of the two regimes depicted in models I (gold standard) and III (debt standard) is possible. Money creation by monetization of a standard entails a cost equal to the legal loss borne by the standard when it becomes money (seigniorage); but, in circulation, standard money may regain part of that loss, and the actual level of devalorization is measured by the percentage difference between the market price and the official price of the standard. When the market price of the standard is at its minimum (Pi; = Tor P jj = T, hence i* = r), money bears no devalorization in circulation; Ghislain Deleplace 325 private agents regain in the use of standard money the loss of standard which they had to accept in order to convert the standard into money. A money which depreciates is one where the current devalorization increases, i.e. which imposes on private agents a growing loss of standard (that is of value) in circulation; this loss, however, cannot overtake the legal one imposed by the public issuer of money. As in the above analysis of the gold standard regime, the international aspects of a debt standard regime have been left aside; the purpose of the model is simply to attract attention to the benefits of looking at fiat money from the starting point of the monetary standard. Put in a historical perspective, the recognition that the constancy of the official discount rate is a necessary (but not sufficient) condition for monetary stability is close to the statement, common in the French monetary debates of the 1570s, that the 'prince should not manipulate his money. The condition that the discount rate should be as low as possible corresponds to what Ricardo looked for in the English monetary debates of the early 1800s. But the prerequisite of both conditions is the rehabilitation of the concept of the monetary standard, which has so long obsessed economists before disappearing from the pure theory of money. 5 CONCLUSION: THE RELEVANCE OF THE MONETARY STANDARD FOR THE POST KEYNESIAN AND CIRCULATION APPROACHES The above suggestion of a monetary standard conceived as a debt standard seems to fit together with some central aspects of Post Keynesian and circulation approaches. First, it is consistent with the Post Keynesian statement that, in modern economies with a fully developed banking system, the rate of interest is the independent variable, fixed by the monetary authorities, and the quantity of money is endogenous. It is suggested here that, not only the rate of interest at which the central bank monetizes the debt owned by the banks is exogenously determined, but that it must be so, and at a constant level, to fulfil a necessary condition of monetary stability. Secondly, the role of the central bank as a lender of last resort is reaffirmed, in a way advocated mostly by the circulation approach: this role is not only of the fireman's type, which bursts into action when a crisis occurs; it is part of the monetary regime, in that the behavior of the central bank in its discount activity contributes to the stability of that regime. This view may be extended by the addition of a convertibility rule of other means of payment (such as bank money) in terms of the central bank standard money; various payment systems could be designed. 36 The concepts of monetary standard and of standard money provide the analytical link between money creation and money circulation. They may apply equally to a pure metallic regime, with coins made of gold and/or silver, to a regime where gold and/or silver coins circulate side by side with convertible bank money, to one where inconvertible government fiat money officially linked to a foreign currency circulates side by side with bank money convertible within it, 326 Does Circulation Need a Monetary Standard? and to one where the same monies have lost any official link whatsoever with a commodity or a foreign currency standard. This shows that, when the analysis of money is built on the basis of the relation between the unit of account and the means of payment, which reflects the link between money creation and money circulation, the construction of a unified theory of a monetary economy, in which various regimes are encompassed, is possible. This construction may be one of the tasks assigned to the Post Keynesian and circulation approaches, if they find the way to work together. Notes * Since its first version in October 1992, the present chapter has benefited in various discussions and correspondence from the comments of Carlo Benetti, Marie-TMr~se Boyer-Xambeu, Suzanne de Brunhoff, Jean Cartetier, Bernard Courbis, Lucien Gillard, Augusto Graziani, Jan A. Kregel, Christophe Lavialle, Maria Cristina Marcuzzo, Edward J. Nell, Annalisa Rosselli, and Bernard Schmitt. Many thanks to all of them. 1. 2. 3. 4. 5. 6. 7. S. 9. 10. 11. 12. Sraffa (1960), p. 23. 'Since then (1971), every major country has adopted an inconvertible paper or fiat standard, not as an emergency measure expected to be temporary, but as a system intended to be permanent. Such a worldwide fiat monetary system has no historical preCedent' (Friedman, 1990, p. S7). 'Commodity money and managed money are alike in that they are related to an objective standard of value. Managed money and fiat money are alike in that they are representative or paper money, having relatively little or no intrinsic value apart from the law or practice of the State... Moreover, managed money is, in a sense, the most generalised form of money - which may be considered to degenerate into commodity money on the one side when the managing authority holds against it 100 per cent of the objective standard, so that it is in effect a warehouse warrant, and into fiat money on the other side when it loses its objective standard' (Keynes, 1930, part I, p. 7). 'I define an exchange standard as a managed representative money the objective standard of which is the legal-tender money of some other country' (Ibid., p. 16). Ibid., p. 56. The question is ibid., p. 71, and the answer p. 77: 'Thus our quaesitum, namely a measure of the "intrinsic value" of money, has no separate existence, but is merely one of the currency index numbers over again.' Ibid., p. 47. Ibid., p. 55. Robertson (192S), p. SO. Mason (1963), p. SI. '(By virtue of the Gold Standard Act 1925) pre-war commodity money, in the shape of the sovereign, has not been restored; Ricardo's proposal of a hundred's years previously has been adopted; and sterling is established by law as a pure managed money ... namely a representative money managed so as to conform to an objective standard' (Keynes, 1930, part I, p. IS). The concept of a managed money explains this surprising filiation between Ricardo and Keynes, who adds: 'If Ricardo had had his way with his ingot proposals, commodity money would never have been restored, and a pure managed money would have come into force in England in ISI9' (Ibid., p. 14). Ricardo (lS16), pp. 5S-9. Ghislain Deleplace 13. 14. 15. 16. 17. 18. 19. 20. 21. 327 'Where has it been disputed that there are not other causes besides the depreciation of money which may account for a rise in the prices of commodities? The point for which I contend is, that when such rise is accompanied by a permanent rise in the price of that bullion which is the standard of currency, then to the amount of that rise is the currency depreciated. During the American war the rise in the prices of commodities was not attended with any rise in the price of bullion and was therefore not occasioned by a depreciation of currency' (Ricardo, 1811, p. 251), quoted by Marcuzzo and Rosselli (1991), p. 41. The international movements of gold bullion also affect the domestic value of money, but in a way which is disputed. The traditional view is through the 'price-specie flow mechanism' inherited from Hume (1752), where an influx of gold increases the domestic quantity of money and thus the money price of all traded commodities, provoking a self-adjustment of the balance of trade; then the equilibrium quantity of money and the associatied value of money prevail sooner or later. Marcuzzo and Rosselli contend that no concept of equilibrium quantity of money is required in Ricardo, and that the increase in the quantity of money provoked by an influx of gold directly affects the domestic value of money, without having to rely on a balance-oftrade mechanism. Both explanations rely on the Quantity Theory of money. In fact, adherence to that theory is not logically necessary for the relevance of Ricardo's proposition concerning the stabilization of the domestic value of money and the exchange rate. See Deleplace (1994a, I 994b). See Ricardo (1816), p. 59. 'To secure the public against any other variations in the value of the currency than those to which the standard itself is subject, and, at the same time, to carry on the circulation with a medium the least expensive, is to attain the most perfect state to which a currency can be brought. .. In other words, the Bank should be obliged to purchase any quantity of gold that was offered them, not less than twenty ounces, at £3.17. per ounce and sell any quantity that might be demanded at £3.17.10.112. While they have the power of regulating the quantity of their paper, there is no possible inconvenience that could result to them from such a regulation ... That regulation is merely suggested to prevent the value of money from varying from the value of bullion more than the trifling difference between the prices at which the Bank should buy and sell, and which would be an approximation to that uniformity in its value which is acknowledged to be so desirable' (Ricardo, 1816, pp. 66-7; repro in Ricardo, 1817, p. 356-8). Robertson defines a gold standard as 'a state of affairs in which a country keeps the value of its monetary unit and the value of a defined weight of gold at an equality with one another' (1928, p. 64). This definition allows him to use the expression 'gold standard' whatever the direction of the causality in the quotation referred to in note (9) above. But the example he gives of the United States in the 1920s, who 'has been deliberately trying to treat gold as a servant and not as a master' (ibid., p. 81), creates some confusion which he acknowledges, because this country is at the same time labelled as being on a gold standard and on an arbitrary standard. 'Thus, the adjustment mechanism which restores the currency to its value acts on the quantity of money. never on the quantity of gold: the production of gold depends on other criteria' Marcuzzo and Rosselli (1991), p. 56. Even in the case of England, where since the sixteenth century the tax of seigniorage has been abolished and the cost of fabrication paid for by the state, there is a cost of minting, equal to the loss of interest during the fabrication of the coins. The title of Ricardo's pamphlet in 1810 is explicit: The High Price of Bullion, a Proof of the Depreciation of Bank Notes. On the link between the domestic value of money and foreign exchange rates, in the context of competing metallic standards in international relations, see BoyerXambeu, Deleplace and Gillard (1994b). 328 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. Does Circulation Need a Monetary Standard? See Boyer-Xambeu, Deleplace and Gillard (1994a), pp. 189-193. Marcuzzo and Rosselli (1991), pp. 49-50. See also Marcuzzo and Rosselli (1994a, I 994b). Sraffa (1960), p. 16. Jevons (1875), chapter XXV. Jevons credits Joseph Lowe in 1822 and Poulett Scrope in 1833 with the invention of this notion. But it may already be found in previous works, in Condorcet (1792) for example: 'Therefore one would be mistaken if one judged of the real loss of assignats by the ratio of their value to that of coined silver. And it is only by the price of particular goods that, thanks to a rather complicated calculation to which it would even be difficult to provide sure foundations, one could determine that depreciation with some precision' (Condorcet (1790-92), p. 122) (author's translation). I thank Lucien Gillard for that reference. Hawtrey (1927), p. 30. Ibid., p. 9. Recent literature on monetary laissez faire has resurrected the question of the standard in designing a regime which would allow to achieve greater monetary stability than in the existing ones. However, this 'BFH system' (as coined by Greenfield and Yeager, 1983) rests on a complete separation between the unit of account and any medium of exchange, redemption or settlement used by private agents (including financial intermediaries). As a consequence, there is neither standard money (base money) nor monetary standard; the unit of account is just a commodity num6raire. This literature is then outside the scope of the present paper (see the Afterword by Deleplace and Nell to this volume). For a critique of the general equilibrium approach to money, see the contribution of Benetti and the Afterword by Deleplace and Nell to the present volume. See the contribution of Moore to the present volume. See the Introduction to the present volume by Deleplace and Nell. Including Marx (1867), under the name of 'general equivalent', Keynes (1930) under the name of 'money of account' and Sraffa (1932), under the name of 'standard'. One may suggest that the rules designed by Ricardo to make the legal and the commercial values of the pound in gold coincide aim at ensuring the unicity of the unit of account. See Deleplace (1994a). In contrast with the coinage of gold in England, the coinage of gold and silver in France bears a fixed (different) seigniorage, so that the official price at which bullion is purchased by the mint is in each case below the official value of the metal coined. The role kept by gold in the Bretton Woods system does not concern the monetary regime of those countries, but the US regime and the international relations between those countries and the United States. On the distinction between money as purchasing power and money as a stock of wealth, see Graziani's contribution to the present volume. See for example Cartelier's and Schmitt's contributions to the present volume. References Benetti, C. (1992), 'The Ambiguity of the Notion of General Equilibrium with a ZeroPrice for Money', this volume. Boyer-Xambeu, M.-T .• G. Deleplace and L. Gillard (1994a), Private Money and Public Currencies. The 16th Century Challenge (Armonk, NY: M.E. Sharpe). Boyer-Xambeu. M.-T., G. Deleplace and L. Gillard (1994b), 'R6gimes mon6taires, points d'or et "serpent bim6tallique" de 1770 h 1870'. Revue economique. 45 (5). septembre. Ghislain Deleplace 329 Cartelier, 1. (1990), 'Payment Systems and Dynamics in a Monetary Economy', this volume. Condorcet, M. de (1792), 'Discouzs sur les finances', in B. Courbis and L. Gillard (eds), M6moires el discouzs sur les monnaies et les finances (1790-1792) (Paris: L'Harmattan, 1994). Deleplace, G. (1994a), 'Les diff6rents usages de 1'6talon mon6taire', Cahiers d'economie politique, 23, avril. Deleplace, G. (l994b), 'Au x origines de la pens6e mon6taire moderne', Revue economique, 45 (5), septembre. Fisher, I. (1911), The Purchasing Power of Money (New York: Macmillan). Friedman, M. (1990), 'Bimetallism Revisited', Journal of Economic Perspectives, 4 (4). Graziani, A. (1990), 'Money as Purchasing Power and Money as a Stock of Wealth in Keynesian Economic Thought', this volume. Greenfield, R.L. and L.B. Yeager (1983), 'A Laissez Faire Approach to Monetary Stability', Journal of Money, Credit, and Banking, 15 (3), August. Hawtrey, R.G. (1927), The Gold Standard in Theory and in Practice (London: Longmans, Green). Hume, D. (1752), 'On Money', in E. Rotwein (ed.), Writings on Economics (Edinburgh: Nelson, 1955). levons, W.S. (1875), Money and the Mechanism of Exchange (London: Kegan Paul). Keynes, 1.M. (1930), A Treatise on Money: The Pure Theory of Money, in D. Moggridge (ed.), Collected Writings (London: Macmillan, vol. V, 1973). Marcuzzo, M.C. and A. Rosselli (1991), Ricardo and the Gold Standard (London: Macmillan). Marcuzzo, M.C. and A. Rosselli (1994a), 'The Standard Commodity and the Standard of Money', Cahiers d'economie politique, 23, avril. Marcuzzo, M.C. and A. Rosselli (1994b), 'Ricardo's Theory of Money Matters', Revue economique, 45 (5), septembre. Marx, K. (1867), Capital (New York: International Publishers, Vol. I. 1967). Mason, W.E. (1963), Clarification of the Monetary Standard (University Park: Pennsylvania State University Press). Moore, B.l. (1992), 'The Money Supply Process: A Historical Reinterpretation', this volume. Ricardo, D. (1810), The High Price of Bullion, a Proof of the Depreciation of Banknotes, in P. Sraffa (ed.) with M. Dobb, Works and Correspondence (Cambridge: Cambridge University Press, 1951-73, vol. III). Ricardo, D. (1811), Reply to Mr Bosanquet's Practical Observations on the Report of the Bullion Committee, ibid, vol. III. Ricardo, D. (1816), Proposals for an Economical and Secure Currency, ibid., vol. IV. Ricardo, D. (1817-21), Principles of Political Economy and Taxation,ibid., vol. I. Robertson, D. (1928), Money, 3rd edn (Cambridge: Cambridge University Press). Schmitt, B. (1992), 'A New Paradigm for the Determination of Money Prices', this volume. Sraffa, P. (1932), 'Dr Hayek on Money and Capital', Economic Journal, March. Sraffa, P. (1960), Production of Commodities by Means of Commodities (Cambridge: Cambridge University Press). 10 The Relevance to Modem Economics of the Banking School View Perry Mehrling This chapter attempts a rational reconstruction of the banking school view from a modern perspective.' The greatest difficulty in such a project is encountered immediately: there exists no systematic account of the monetary theory underlying the specific policy arguments of the banking school. As one historian of monetary thought summed up regarding the seminal texts of Fullarton (1845) and Tooke (1844), 'the banking school was not trying to develop monetary theory for later generations, but was attempting to make effective arguments against specific proposals for regulating notes' (Fetter 1965: 191). Much the same could be said of those authors who have maintained and developed the banking school tradition through the years. Monetary theory, unlike value theory, has been the province of practical men with concrete proposals for solving the current crisis and staving off the future one, men less interested in consistency and logical coherence than in persuasion and political effectiveness. Nevertheless, the banking school was more than a convenient and' contingent political alliance of all those who supported a particular proposal for reform. The characteristic banking school view derives from a set of conceptual distinctions and a style of argument that is shared by all prominent writers. It is this shared foundation that the following seeks to exhume and reconstruct, not the specific proposals that are anyway more suited to the specific monetary problems and institutions of the time. The legacy of the banking school for modern economics lies in the application of that analysis to modern problems and institutions. By bringing to light the essential features of that analysis, this chapter seeks to contribute to a greater recognition of common features in a number of modern schools of thought which carry on the banking school tradition. 2 1 TWO PARABLES OF PRIMITIVE EXCHANGE In an attempt to identify the most basic and vital functions of the monetary system, all monetary theorists at some stage ask themselves the question: how would exchanges be made in a market economy which lacked the elaborate mon- 330 Perry Mehrling 331 etary institutions we take for granted? Impressed by the inefficiency of barter, economists have answered the question by positing one of two possible primitive systems of monetary exchange. In one tradition, associated with the currency school, the most primitive system imaginable would use a particular commodity such as gold as money. From this perspective modern monetary institutions improve on the primitive system by substituting paper tokens and ledger entries for the actual physical commodity, thereby freeing up real resources for more productive uses. This currency school perspective is perhaps the dominant one today, transmitted to new generations by means of the parable telling how cigarettes were used as a medium of exchange in a World War II prison camp. In the second tradition, associated with the banking school, the most primitive system imaginable would use bills of exchange as a circulating medium. The founding parable of the banking school is not the prison camp but the Irish economy of 1800 where exchanges were made using a system of circulating shopkeeper IOUs (Fullarton, 1845: 53-5; Tooke, 1844: 21-2). From this perspective, modern monetary institutions improve on the primitive system by substituting higher quality government and bank credit for the credit of small shopkeepers, thereby enhancing the acceptability of the currency and so lubricating the gears of trade. This second perspective is attracting more attention today, as an intellectual consequence of financial innovation and deregulation, but it is stiII a minority view. This different foundational parable is responsible for the proposition of the banking school that most sets it apart from orthodoxy, that the central bank cannot, does not, and should not control the quantity of money. Even in primitive economies, the argument goes, a pure commodity money system is institutionally unstable on account of the inflexibility of the money supply. The primitive system of credit money always lies dormant just beneath the surface, ready to re-emerge whenever available currency fails to meet the needs of trade. In a modern financially developed economy this supply of potential credit money is especially accessible. As a consequence, the central bank cannot ordinarily control the quantity of money. Any attempt to do so succeeds only in temporarily disrupting the smooth workings of the system until the dormant credit money system can be activated. This proposition of course brings the banking school view into fundamental conflict with the Quantity Theory of money, which derives from the currency school tradition. Recognizing the difficulties caused by inelastic supply, the currency school tradition emphasises changes in the price level as a mechanism for adjusting the real supply of currency. Spe~ifically, under a gold standard system, gold flows to the country with the lower commodity prices, and this flow increases the supply of currency in that country. Recognizing that price adjustment is slow and disruptive in a world of nominal contracts, the currency school proposes instead active manipulation of the money supply to mimic the changes that would have been brought about by the slower price changes. From a banking school view such a monetary policy is at best impotent and at worst actively 332 The Banking School View harmful. To the extent that the central bank expands and contracts the money supply in response to the fluctuating needs of trade, it is only doing actively what the banking system would tend to do on its own anyway. To the extent that the central bank tries to do anything else, it will either fail or it will disrupt the financial system for as long as it takes to establish alternative mechanisms. Rather than focus on controlling the quantity of money, the banking school tradition focuses on controlling the quality of money by making sure that the credit-backed currency is backed by credits of the highest possible quality. The idea behind the Real Bills Doctrine was to back the currency with high-quality self-liquidating bills of exchange in order to ensure an elastic supply of high quality currency. According to proponents of the doctrine, an increased demand for currency would be met by discounting a larger proportion of the pool of available bills, while a decreased demand would be met by discounting a smaller proportion. By confining the discount to real bills, the banking school hoped to make the currency fluctuate more or less automatically with the 'needs of trade', with expansion driven by the profit motive of discount banks and contraction driven by the repayment of maturing bills. By controlling the quality of money, the banking school hoped that the quantity would take care of itself. The banking school developed their ideas in the context of the economy of their day, a gold standard system with fixed exchange rates internationally and a domestic banking system with a highly developed market for short-term commercial paper. In order to apply their analysis to a modem context it is necessary to begin by reconstructing the conceptual basis for that analysis. 2 CREDIT VS. MONEY Central to the banking school approach is the distinction between credit in its many forms and money. By 'money' is meant a circulating medium of exchange whose exchange value is determined by its value as a commodity, e.g. gold coin. By 'credit' is meant an asset whose exchange value is determined by the current value of the future money payments that it promises to make. Because credit is payable in money, its value is more or less closely tied to the value of money. And in a gold standard system the value of money is determined by the technol' ogy of gold production. Among credits used to make exchanges, two broad categories can be distinguished. 'Simple credit' is used to make a single transaction, and then held until it is canceled by an offsetting credit or until maturity, whichever comes first. 'Currency of credit' circulates to make several transactions during its lifetime. The distinction between simple credit and currency of credit has to do with the use made of the credit between time of issue and time of cancellation. In principle, a simple credit is transformed into a currency of credit simply by circulation, and a currency of credit is transformed into simple credit simply by hoarding. In Perry Mehrling 333 practice, because simple credits vary widely in quality, it is only the better quality credits that circulate. These two categories of credit along with money proper exhaust the possible ways in which transactions can be made. Thus we have the identity SC + CC . Vee + M . VM == PT, where SC is the flow of simple credit, CC and M are the average quantities of circulating credit and money respectively, Vee and VM are the average number of transactions effected by a unit of circulating credit and a unit of money respectively, and PT is the flow of nominal transactions. This identity says simply that all transactions are made using one of the three possible media of exchange. The purpose of the distinction between credit and money can be made clear using this identity. Whereas the quantity of money (and perhaps also its velocity) is determined independently of the other variables in the equation, the flow of simple credit and more important the quantity and velocity of circulating credit arise along with the flow of nominal transactions which they help to effect. If sellers sell goods to buyers by extending simple credit then SC and PT are both augmented by the same amount. If sellers accept a pre-existing credit as payment then CC . Vee and PT are both augmented by the same amount. The importance of the distinction between credit and money lies in the consequence that, because simple and circulating credit can be used to make exchanges, the medium of exchange is in principle never constrained by the quantity of money proper. 3 LENDING CAPITAL VS. LENDING CREDIT The simplest form of credit-financed exchange involves the seller of goods 'lending money' to the buyer. Lending money comprises two analytically distinct operations, 'lending capital' and 'lending credit' (neither of which necessarily involves money proper at all). Lending capital involves parting with current exchange value in return for a promise of future money payment. Lending credit involves guaranteeing a promised future money payment. At the most primitive level, in simple trade credit arrangements, the seller of commodities lends capital while the buyer lends credit. Lending capital, the seller alters the composition of his working capital, replacing saleable inventories with a trade receivable. Lending credit, the buyer guarantees future payment, promising to make good out of his own capital if need be. In higher forms of credit, the original lender and borrower fade into the background and the functions of lending capital and lending credit are taken over by other agents. For example, consider a particular currency of credit, a ninety-day bill of exchange which circulates among businesses collecting endorsements as it functions as a medium of exchange. Here successive recipients of the bill lend their capital to the initial issuer for as long as they hold the bill. And they lend 334 The Banking School View their credit for the rema~ning lifetime of the bill by endorsing it when they pass it along. Thus the initial lender lends capital only for the time he holds the bill, and the initial borrower lends only a small part of the credit the bill eventually carries. The functional disappearance of the original borrower and lender proceeds even further with the mechanism of bill-broking (e.g. Bagehot, 1906, ch. 11). Here the single endorsement of the broker guarantees payment. Bill-holders lend their capital to some ultimate borrower but the identity of that borrower is irrelevant since the credit of the broker guarantees repayment of the capital loan. From the point of view of the bill-holder, he lends capital to the broker not the ultimate borrower. Furthermore, because the credit of the broker is widely recognized, the bill-holder is free to pass along the bill without lending his own credit to it. The final stage in this development is the banknote, a 'convertible currency of credit'. Here the illusion that the note-holder lends capital to the bank rather than the ultimate borrower is further reinforced by the bank's promise to repay the principal on demand in money proper. Even more, because of convertibility, note-holders are free to imagine that they have not lent capital at all, that the note they hold is itself capital, that the banknote is in fact money not credit. But all this imagining does not change the fact that the banknote is not money but only a token of the bank's credit. Although the banknote may in all practical respects function exactly as money, it remains distinct from money on account of the mechanism of its supply; the banknote enters circulation as the bank of issue discounts some capital loan. 4 CONVERTIBILITY AND THE LAW OF REFLUX The Law of Reflux states that an overissue of a convertible currency of credit is impossible, in the sense that the quantity of a convertible currency of credit adjusts to meet the needs of trade. Applied to circulating bills of exchange the law is almost a truism. Such forms of credit clearly circulate only so far as they are needed, and any excess supply simply reverts to simple credit. The endorsements on these credits may enable them to circulate, but cannot force them to do so. Applied to convertible currencies of credit, the Law of Reflux says that banknotes mimic their more primitive cousins in reverting to simple credit when they are no longer needed. But this proposition is by no means obviously true, since the simple credit which corresponds to the banknote is physically disjoint from it; thebanknote circulates in the world while the simple credit underlying the note remains locked in the bank vault. The question of whether the Law of Reflux holds thus comes down to a matter of whether market forces are adequate to impel banks to retire excess banknotes and yield up the simple credits in their vaults. The crux of the banking school argument is that so long as banknotes are convertible, the market forces banks to do precisely this. The argument is straightforward. Suppose that, at given rates of interest, asset-holders find themselves with Perry Mehrling 335 more notes than they would like and fewer simpler credits. In the banking school view, this portfolio disequilibrium is resolved as agents accepting new bills decide to hold them rather than present them for discount, while other agents pay off their maturing bills by sending excess notes back to the bank which had discounted them. In this way, the quantity of circulating bank notes fluctuates in response to the desires of asset-holders just as the quantity of a more primitive circulating credit would. The importance of convertibility in this process becomes clear when it is realized that the Law of Reflux is an early application of the modem Principle of No Arbitrage. In the portfolio disequilibrium described above, banknotes tend to depreciate against money and the note price of securities tends to rise above their money price. To the extent that these tendencies are realized, there is an opportunity for arbitrage profit at the expense of the issuing bank. The arbitrageur need only sell a security for the note price, present the notes to the issuing bank for redemption in money, and buy the security back for the lower money price, keeping the difference as pure profit. To meet the demand for note redemption the bank must either draw down its reserve holdings of money or sell a security for the money price. Drawing down money reserves merely replaces excess notes with money, leaving asset-holders still holding a greater quantity of the circulating medium than they would like, and leaving open the opportunity for additional arbitrage profit at the bank's expense. To protect its limited reserves, not to mention safeguarding its capital, the bank must soon abandon this course, sell securities from its portfolio at the money price and redeem its notes. This same arbitrage opportunity underlies the Law of Reflux since, to the extent that an excess note issue causes a depreciation of notes against money, debtors will prefer to repay their maturing bills with banknotes rather than money. In the classic channel for reflux it is the bank's debtors who exploit the arbitrage opportunity by paying depreciated notes to the only economic agent willing to accept them at par, the bank of issue. Emphasizing this channel for extinguishing excess notes, the banking school naturally focused attention on the way that reflux is affected by the maturity of bank assets. 5 THE REAL BILLS DOCTRINE AND THE ART OF CENTRAL BANKING To ensure that the channel for extinguishing excess currency was kept open, the banking school advocated confining discount to short-term bills so that banks would have a constant inflow of money which they could use to payoff depositors and defend par. This is the so-called Real Bills Doctrine. Although this doctrine has been widely criticized, some of the most common criticisms are based on misunderstanding. The supposed inflationary consequences of the doctrine follow from the mistaken idea that the doctrine would expand the currency by discounting all 336 The Banking School View real bills, whereas in fact it was intended to ensure the easy and automatic contraction of any excess currency by allowing rapid change in the proponion of all eligible bills being used to back the currency. The supposed destabilizing consequences of the doctrine follow from the same misunderstanding combined with the fact that the pool of eligible bills fluctuates widely with the business cycle. More compelling is the criticism that, even supposing that the Law of Reflux operates to regulate the quantity of currency for given interest rates, banks have the ability to influence interest rates, and hence the quantity of currency, through their ability to lend credit to capital loans (Viner, 1937). The idea that this control over lending rates can cause overissue of currency rests on the presumption that there is some natural or neutral rate of interest associated with the correct quantity of money. If banks offer to discount bills at a lower rate, then overissue is possible because merchants use the discount window to raise additional funds, and when they spend these funds they tend to raise prices. Although this criticism does not vitiate the banking school analysis, it does imply that, as a policy for regulating the monetary system as a whole, the Real Bills Doctrine is incomplete and needs to be augmented by some policy for controlling the money rate of interest. Given their commitment to laissez{aire in banking, the original banking school were reluctant to accept the need for active control of credit as a logical implication of their views on the impossibility of controlling the quantity of currency. It was left to later authors (e.g. Bagehot, 1906; Hawtrey, 1932; Wicksell, 1898) to develop the theory of credit control through manipulation of the central bank lending rate. Thus it came to be that modern students of the banking school focus on interest rate policy while modern students of the currency school focus on controlling the quantity of money. In the modern technical debate about appropriate targets and instruments for monetary control one can see the shadow of the older more fundamental debate. To repeat, from a banking school view it is impossible to regulate the quantity of a circulating currency of credit. Attempts to do so by imposing reserve requirements on banks either do not bind or they disrupt the monetary system and stimulate financial innovation. Even worse, the misguided focus on the quantity of money turns attention away from the regulation of credit through control of the interest rate, and so opens the door to uncontrolled speculative booms and their inevitable aftermath. 6 MODERN BANKING INSTITUTIONS Application of the banking school view to modem banking institutions requires attention to several important differences from the world of nineteenth-century British banking. First, depository institutions hold a much wider array of assets, including long-term and highly illiquid assets. Second, bank money is convertible into fiat money not gold coin. And third, deposit insurance shifts onto the central Perry Mehrling 337 authorities some part of the commitment to maintain banknotes at par. In the space remaining, it is possible only to suggest the lines along which the classic banking school view would need to be modified to take account of these institutional changes. In the modern US economy, banks of issue lend long term to a wide array of businesses and households. They are also important holders of government debt of various maturities. Because of the longer maturity of bank assets the flow of loan repayments is small relative to the size of the bank's deposit liabilities. As a consequence, the classic channel for reflux through debt repayment is less important and the regulation of the currency by arbitrage takes other forms. On the asset side, since some portion of bank assets are readily saleable in highly organized secondary markets, an excess issue can be extinguished simply by selling a security for the unwanted deposit. The expansion of the pool of marketable securities in recent years (through securitization of mortgages, auto loans and credit card receivables) has expanded this asset-side channel for reflux. On the liability side, since some portion of bank liabilities do not circulate, an excess issue can be absorbed simply by replacing deposit and checking liabilities with longer term time deposits, negotiable certificates of deposit and commercial paper issued by bank holding companies. From a banking school view, sophisticated forms of asset and liability management have replaced the relatively narrow classic channel of reflux with a number of wider channels more suited to meet the fluctuations of modern trade. The Federal Reserve initially resisted all these financial innovations, noting correctly that they would allow banks to expand lending more rapidly than their deposit base was growing. Thus, the mechanisms which today regulate the supply of currency also have the effect of reducing central bank control over credit. Increasingly, the central bank is forced to fall back on the classic mechanism of manipulating its lending rate, but it remains to be seen whether the art of central banking through interest rate policy will be adequate to stem speculative expansions and curb deflationary contractions under modem conditions. Second, in the modern US economy, bank money is maintained at par with an inconvertible government-issued fiat currency whose value is not fixed by the technology of its production. This means that the Law of Reflux can only be said to prevent an overissue of the circulating medium for a given price level (as well as given interest rates). If prices rise, so that agents find themselves holding smaller real balances than they desire, the profit motive will induce the banking system to provide additional balances, and if prices fall the reflux of excess balances will reduce the quantity of currency. The question, of course, is whether, in the case of an inconvertible currency, changes in the price level are in any sense caused by changes in the money supply. In the banking school view, the answer to this question seems to depend on whether there is a channel for reflux of an excess issue of money (Le. fiat currency) in addition to the channel for reflux of excess deposits. Because private asset-holders can always rid themselves of excess money balances by making a The Banking School View 338 bank deposit, the question can be narrowed to one of whether there exist channels through which banks can rid themselves of excess reserves. In the US banking system there are two such channels of importance, the discount window and the foreign exchange window. Any bank with borrowed reserves can payoff its debt to the central bank with money and so rid the banking system of excess supply. And any foreign bank with excess dollars can demand its own currency in exchange from the Federal Reserve. Both these mechanisms are variations of the channels which have already been discussed for the reflux of bank money. Obviously, the central bank can block these channels if it wishes. Once all borrowed reserves have been repaid, reflux through the discount window comes to an end and excess reserves begin to slosh around inside the banking system. Similarly, the reflux through the foreign exchange window depends on the continued ability and willingness of the central bank to defend the exchange rate. If the central bank fails to absorb excess dollars and instead allows the dollar to depreciate, then reflux through the foreign exchange window comes to an end as well. From a banking school perspective, it is of vital importance for the smooth working of the financial system to keep these channels open. One would expect to find the modem day banking school proponent arguing in favor of stabilizing borrowed reserves domestically and the exchange rate internationally. Finally, in the US economy, measures that have been put in place to support bank credit in times of crisis have tempered to some extent the ordinary discipline of reflux. Most important, the widespread provision of deposit insurance in effect allows insured banks to lend not their own but the government's credit. The responsibility to defend the par value of bank liabilities is thus shifted in part from the bank to the government, and the bank feels correspondingly less pressure to absorb any excess issue. Instead, the tendency of excess banknotes to depreciate against money puts pressure on the central bank to increase the supply of its own money in order to restore par with banknotes. 3 Only the tenuous discipline of central bank reflux through the discount window and the foreign exchange window stands in the way. The point is not that the central bank regularly yields to this pressure - as I read the record it does not - but rather that the weakened pressure on the banks tends to weaken all channels for reflux of an excess bank issue. From a banking school perspective, deposit insurance is not a good way to safeguard the quality of the currency. In the spirit of the Real Bills Doctrine, it would be better to regulate the quality of bank assets eligible to back deposits. 7 CONCLUSION Application of the conceptual framework of the banking school to modem banking institutions suggests a modification of the classic policy views of the banking school. When bank deposits are backed by longer term assets, the classic channel for reflux is largely supplanted by the banks' ability to manage the structure of their Perry Mehrling 339 balance sheets in active secondary markets. When money is a fiat currency, it is not enough to ensure the easy reflux of excess.bank deposits and it becomes important to maintain channels through which excess fiat currency can be extinguished when it is not needed, specifically through the discount window and the foreign exchange window of the central bank. Finally, when government credit is used to support the entire monetary system, it is important for reflux that the obligation to maintain bank liabilities at par with money continue to rest with the individual bank. But just as the Real Bills Doctrine proved incomplete as a policy for regulating the system as a whole, so too these modem doctrines fail to address the question of regulating credit. The classic banking school view teaches that money takes care of itself provided credit is maintained on a sound basis, but it goes on to focus on policies to help money take care of itself rather than policies to ensure the stability of credit markets. The theory of central banking was developed to fill the gap, but today the bank lending rate seems to many an awfully thin reed to depend upon. The main challenge then for the modem researcher working in the tradition of the banking school is to reconstruct the theory of central banking. It may be that the central bank cannot, does not and should not control the supply of money. But if so, then it becomes all the more important to understand what it can and should do to control credit conditions more broadly. Notes I. 2. 3. Sources which focus on a more historical understanding of the banking school include Doherty (1942,1943), Mints (1945), Viner (1937), and Wood (1939). I have in mind here particularly the Post Keynesians and circuitistes whose views may be found elsewhere in this volume. Most influential for the perspective outlined in this paper is the work of Aglietta (1976), Earley (1976), Foley (1988), Hicks (1982), Kaldor (1982), Minsky (1986), and Parguez (1984). The importance of the Principle of No Arbitrage is particularly evident in the work of financial economists such as Black (1970; 1987) and Wojnilower (1980; 1985). The lender of last resort function of the central bank has a similar but weaker effect. Because the central bank lends at its own discretion, and always demands good security, no bank can be sure that the central bank will endorse its credits. References Aglietla, M. (1976), A Theory of Capitalist Regulation (London: New Left Books). Bagehot, W. (1906), Lombard Street: A Description of the Money Market (New York: Scribner's). Black, F. (1970), 'Banking and Interest Rates in a World without Money', Journal of Bank Research (Autumn), 8-20. Black, F. (1987), Business Cycles and Equilibrium (New York: Basil Blackwell). Doherty, M.R. (1942), 'The Currency-Banking Controversy', Southern Economic Journal. 9 (October), 140-55. 340 The Banking School View Doherty, M.R. (1943), 'The Currency-Banking Controversy: II', Southern Economic Journal,9 (January), 241-51. Earley, J.S., RJ. Parsons and F.A. Thompson (1976), Money, Credit and Expenditure: A Sources and Uses of Funds Approach (New York: Center for the Study of Financial Institutions, New York University). Fetter, F.W. (1965), The Development of British Monetary Orthodoxy (Cambridge, MA: Harvard University Press) Foley, D.K. (1988), 'A Microeconomic Model of Banking and Credit', in B. Jossa and C. Panico (eds), Teorie Monetarie e Banche Centrali (Naples: Liguori Editore), 11-29. Fullarton, J. (1845), On the Regulation of Currencies, 2nd edn (London: John Murray). Hawtrey, R. (1932), The Art of Central Banking (London: Longmans, Green). Hicks, J. (1982), Money, Interest and Wages (Oxford: Basil Blackwell). Kaldor, N. (1982), The Scourge of Monetarism (London: Oxford University Press). Minsky, H. (1986), Stabilizing an Unstable Economy (New Haven: Yale University Press). Mints, L.W. (1945), A History of Banking Theory (Chicago: University of Chicago Press). Parguez, A. (1984), 'La dynamique de la monnaie', Economie et Sociere, l (4). Tooke, T. (1844), An Inquiry into the Currency Principle (London: Longman, Brown, Green and Longmans). Viner, J. (1937), Studies in the Theory of International Trade (New York: Harper). Wicksell, K. (1965), Interest and Prices (New York: Kelley, 1965; originally published by Gustav Fischer, 1898). Wojnilower, A.M. (1985), 'Private Credit Demand, Supply, and Crunches - How Different are the 1980s?', American Economic Review, 75 (2), 351-6. Wojnilower, A.M. (1980), 'The Central Role of Credit Crunches in Recent Financial History', Brookings Papers on Economic Activity, 2, 277-339. Wood, E. (1939), English Theories of Central Banking Control, 1819-1858 (Cambridge: Harvard University Press). 11 The National Economy Studied as a Whole: Aspects of Circular Flow Analysis in the German Language Bernard Schmitt and Spartaco Greppi In 1967 Helmut Reichardt published Kreislaufaspekte in der konomik, a book in which he developed the theory propounded by one of his teachers at Ttlbingen, Hans Peter. A small group of young economists led by Peter pursued the task of arriving at a clear and definitive definition of the 'true' nature of the circular flow which rules each national economy. Short of this ambitious goal, the group purported to establish a formally correct concept of the working of the national economy in a circular flow. To an Anglo-Saxon economist, at least one who is not conversant with the works of Morris Albert Copeland (1952), the very idea that national economies work in 'circles' seems strange, not to say odd. Yet, as Reichardt reminds us, in 1928 Wasily Leontief had adhered to the representation of national economies as circular flows in 'Die Wirtschaft als Kreislauf (1928). There is an easy way from the start to alleviate the impression of oddity which the English-speaking economist may have when first confronted by circular flow analysis; to this end it may suffice to point out that what we are really talking about here is simply macroeconomics in contradistinction to microeconomics. An important contemporary school of German-speaking economists holds that macroeconomic phenomena can only be understood in terms of circular flow analysis, a method whose 'discovery', in our field of research, goes back to Fran~ois Quesnay, Adam Smith and Karl Marx. But did the German school succeed in its endeavour? We should certainly answer in the affirmative with regard to the formal definition of circular flows afforded by a number of German-speaking economists. Their theory is no less convincing when they demonstrate that circular flows actually exist in the real world of economic facts. But further research is still necessary if macroeconomic theory is eventually to be based on the firm foundation which only a perfect 'conflation' offormal and real economic circular flows will provide. 341 342 The National Economy Studied as a Whole FORMAL CIRCULAR FLOWS Definition of Circular Flows Circular Flows as Simple Exchanges between Pre-existing Commodities Any exchange comprises two opposite flows: 'X gives Y commodities to the amount of U and Y some other commodities to the same amount or else pays out the sum U in a generally accepted or in a contractually agreed upon means of payment' (Peter, 1954: 3). But if circular flows were thus merely alternative expressions of Walrasian exchanges, then surely it could hardly be claimed that their study provides an original theoretical approach to economic events. A careful interpretation of the definition given by Peter, however straightforward and unequivocal it may seem, shows that some important point is at stake which takes us much deeper than would be necessary if we pursued the conservative aim of merely providing a parallel method to Walrasian analysis. In fact, Peter explicitly introduces a sum of money, equal to U units, which plays a crucial part in his reasoning. There is no doubt 'that U is a sum of an actually existing means of payment, that is a sum of money - and not just an amount of a given abstract numeraire. As soon as a sum of money is allowed to enter into and interfere with its very definition, an exchange is correctly construed as a circular flow. Circular Flows Strictly Defined as 'Monetized' Exchanges between Commodities In the theories of Walrasian and non-W alrasian equilibria, prices are expressed in terms of an arbitrarily chosen commodity, whose price is accordingly supposed to be given, only n-l independent prices remaining to be determined. The logic of circular flow analysis is basically different: in an exchange economy endowed with n goods, n independent prices - and not merely n-l - are unknown. Suppose 2 pears are exchanged for 3 apples: - In any theory accepting the paradigm of relative prices the following question is nonsensical: do the pears and the apples have the same value? The 'value' of the 2 pears is defined by the 3 apples and the 'value' of the 3 apples by the 2 pears: how could pears be the same as 3 apples? According to the paradigm of relative prices, an exchange between pears and apples defines only one price, namely the ratio of the physical quantities exchanged, 2/3 or 312 in our example. As soon as the Kreislauftheorie supersedes the paradigm of relative prices, the aforementioned exchange of 2 pears against 3 apples is seen to involve two independent prices, i.e. the price of each commodity in the chosen unit of payment. Under the new paradigm it would serve no useful Bernard Schmitt and Spartaco Greppi· 343 purpose to assume a direct exchange between the apples and the pears; on the contrary, 3 apples are sold for U units of money while, quite independently, 2 pears are sold for a number, U ., of monetary or 'circulating' units. A circular flow encompasses and 'formalizes' the given exchange if and only if the equality or 'equivalence' of U and U /• is verified. Even when Peter (1954, chapter 3) deals with material circular flows his arguments apply to monetary flows. Peter holds that only statistical magnitudes provide measurements and that such magnitudes 'always represent sums of money' (ibid.: 94). Circular flows apply to 'absolute' prices. Peter points out that from the very beginning of our science the concept of circular flows formed the core of analysis. 'In every society where the economy functions according to the division of labour, valued goods flow from one person to another' (ibid., 1953: 14). Although, as Peter was well aware, the new paradigm (new but well embedded in the highest historical tradition) finds difficulty in being embraced by the general profession of economists, the last word on the matter has yet to be said. After all the paradigm of relative prices breaks down at the moment when theorists decide at last to take money seriously. In pure logic, prices are numbers and not ratios of physical quantities. Asa unit of account (number) and a unit of payment (flow), money is present in all exchanges which are of any interest, in theory and in practice. Therefore economic science of the (not so distant) future may welI be founded on circular flow analysis, just as Peter had predicted. On the topic of circular flows in economics many books and articles have appeared in German since the beginning of this century; in most of them analysis is explicitly developed in relation to national accounting and to money. It is precisely because money is at the center of circular flow analysis that relations between X and X, namely between one and the same person, are of no interest in the eyes of Hans Peter. 'Much like in the theory of games it is expedient to proceed step by step in order to cover circular flows upwards starting from the smalIest n. When n = 1 the circular flow remains actualIy undefined ... n = 2 seems to be the next simple circular flow devoid of interest ... All the same, already the picture becomes more colourful: Robinson and Friday live in a society governed by the division of their labours. With n = 3 appears the first non-trivial structure of the circular flow' (ibid.: 17). No monetary relation can exist between X and X since money flows are expenditures; n = 1 is thus ruled out. Now, in an economy constituted by only two distinct persons, X and Y, no unit of payment has currency and 'exchanges' are reduced to swaps; for n = 2 the equivalence between the terms of trade can therefore be assumed without proof. Says J>eter: 'I take this equality for an axiom.' When n ~ 3, there can be no a priori certitude that each person involved in exchanges receives and spends in any given period exactly the same amount of 344 The National Economy Studied as a Whole money. It is therefore interesting to probe further into circular flow analysis once it is granted that n is at least equal to 3. On the other hand, n must be kept as near as possible to 3 if analysis is to remain on a manageable level of complexity. We may now underline the fundamental criterion which separates the circular flow paradigm from the paradigm of relative prices. (a) Economists who think that relative prices can logically be derived from a series of 'exchange equations' have no basic need to take money into consideration. Indeed, the Walrasian school of thought is quite consistent when it doubts the very existence, or at least the usefulness, of money. Accordingly, if the consideration of money should arise nevertheless, by way of an afterthought, the only problem to be taken at all seriously is how to add a monetary equation to the range of already determined real equilibria, which are quite immune, by their very nature, from any effect that could be exerted by monetary factors, taken to be purely nominal magnitudes, belonging in the undignified domain of applied economics. (b) Circular flow analysts are prone to think not only that money matters but that money is all that matters. Michael Krtm, for example, writes on the Zirkulartausch, every exchange transaction defining a perfect circle. Th. Piitz (1948) and W.A. Johr (1952) concur with Kroll in considering production (+) and consumption (-) as the two opposite 'forces' which must replace supply and demand analysis if prices are effectively to be determined. This end can only be achieved if production and consumption are first expressed in terms of monetary flows. Many KreislauJtheoretiker, particularly Neisser (1931) and Fohl (1955) focused their studies directly on the 'circulation' of money. And money does not seem to be a formal phenomenon. There is no doubt, however, that the construction of purely formal flows was, at first, the main task which the authors of the new paradigm attempted to accomplish. Circular Flows as Mathematical Objects In the works of Leon Walras mathematics had become the queen of economic science. The theory of circular flows is no less abstract. The very title of Hans Peter's main book, Mathematical Structures of the Circular Flow in Economics, is quite revealing in this respect; published in 1954, it extensively - to a fault? applies the modem mathematical skills and techniques of graph theory, matrix algebra and topology. The reader is left to wonder whether he did not stray from economics into pure mathematics. What is more interesting than the method employed is the purpose followed by the author. In the second (and last) part of his book, it becomes clear that what is really at stake is nothing less than the explication of the intricate events taking place in a modem market economy. Circular flows are defined between 'poles' of which, as we have already said, there must be at least three if analysis is to be at all fruitful. A pole can be any person Bernard Schmitt and Spartaco Greppi 345 whatever, who disposes of some real goods, commercial (produced or producing) or financial (assets). In Figure 11.1 letA, B and C be the poles of a circular flow. ;A\ B x ~C Figure 11.1 It is not necessary to construe x as a sum of money units; more abstractly, x is some objective 'value' (see Peter, 1954: 78-9) whose definition remains in the dark. The circular flow depicted in Figure 11.1 implies that in some single and undefined event and, furthermore, in a non-specified time interval, each 'poleperson' receives and 'spends' an equal 'value', x. Peter adorns this 'truth' with a continuous series of mathematical niceties of increasing difficulty; the first 100 pages of his book could be read as a chapter taken out of a mathematical treatise. When Peter then proceeds to apply all these formal intricacies to the pure theories of Fran~ois Quesnay and Karl Marx, one cannot help but wonder whether the theory of the Kreislauf really lends economics a new impetus towards an original departure or whether it merely provides an nth interpretation, equally as arbitrary as the rest of them, of some of the great forgotten texts of our unequalled forefathers. The theory of circular flows ceases to be purely formal and acquires a truthcontent from the moment when it is aimed at establishing the economics of a production economy versus the Walrasian economics of an exchange economy. In this respect it may be worth noting that the distinction between a production economy and an exchange economy was first drawn by the German 'historical school'. It may be deemed somewhat unfortunate, therefore, that in the state in which it was left by Peter the Kreislauftheorie neither presupposes nor precludes the existence of a production economy. Peter's theory is formal in yet another sense: the equality of inputs and outputs at each of the poles is not a condition of eqUilibrium but an axiomatic given. The Kreislaufaxiom is omnipresent in the relevant German literature. How can a theory be factually meaningful if it rests on an axiom which covers (or covers up) the point which is the very gist of the matter? Here we feel just as deprived as by Debreu's axiomatic treatment of value: what a great theorem which established the fact that goods belong in the set of real numbers but when told that no proof is needed on this all important point, provided it be given the status of an axiom, we cannot help feeling frustrated. Prima facie, goods are not numbers; and if in the final analysis goods should tum out to be numbers after all, this unlikely result can only be meaningful if it comes at the conclusion of a train of thorough thoughts and not 'cheaply' at the very beginning of research, in the shape of an artificial axiom, whose hidden purpose is to mislead economic investigation into 346 The National Economy Studied as a Whole the intellectually more comfortable domain of pure mathematics. The whole body of theoretical constructs within the Kreislauftheorie is likewise disappointing to the extent that it rests on the Kreislaufaxiom. How can there be such a thing as an 'open' circle? If it is to make any sense, the Kreislaufmay be a condition of equilibrium or else a demonstrable identity; but it is and will remain of no consequence at all, either to the theorist or to the practitioner, if it is a mere axiom. True, general equilibrium analysis also rests on an axiomatic basis. But relative prices, defined as ratios of physical quantities, can still be determined (or so it seems). But how can the Kreislauftheorie retain any determining power (or heuristic value) if it deals with circles that may be 'open', that is with circles that mayor be circles? No relief is gained in this respect when a 'non-circle' is artificially constrained to be a circle by virtue of an ad hoc axiom. There is nothing to be discovered or gained when the petitum is neatly conjured up as afait accompli. The theory of circular flows comes into its own when Peter's axiom dissolves to be replaced by hard - that is, in our field, nonmathematical - thought. 2 REAL CIRCULAR FLOWS General Considerations At the beginning of his book Helmut Reichardt (1967: 1) writes that no circular flows exist in actual economies; like Peter, Waffenschmidt and Henn, Reichardt develops, among other abstract and purely formal methods, a topological analysis to the Kreislauftheorie. It may seem surprising, therefore, to find that this author directly relates circular flow analysis to concrete methods of accounting, which are in general use everywhere in the world. 'Each elementary economic action can be described by entries in accounting books' (ibid.: 3-4). A closed system of accounting - and in the real world accounting systems are always closed, by 'construction' - deals mechanically with circular flows: the underlying 'force' which, in any period of time, equates the receipts and the expenditures of each person, stems from an 'arithmetic trick' inherent in the principle of 'double accounting' (ibid.: 101). Erich Schneider (1969), like several other authors (StOtzel, Flihl, Neisser ... ) also points out this fact, which Reichardt (1967: 101) essentially links to the definition of capital. Circular Flows Explained by Two Mutually Inconsistent Factors, Namely by an Axiom or, on the Contrary, by the Balancing Effect of Variations in the Capital Holdings of the Contracting Parties We have already mentioned the fact that the Kreislaufaxiom appears recurrently in a whole range of books and articles. But we have just come across an argument Bernard Schmitt and Spartaco Greppi 347 of quite another kind, the formation of circular flows now being ascribed to a definition - clearly, definitions are not axioms. Let A be a purchaser, B the corresponding seller, while C is a financial institution where the capital liabilities and assets of A and B are held. A payment to the value x occurs at instant t. How does this payment relate to the Kreislauftheorie? Conventional theory teaches that at instant t A is a net purchaser and B a net seller; accordingly, at t a straight flow 'leaves' A and 'reaches' B. Circular flow analysis seems to yield the same result. The novelty can only be grasped - to repeat, so it would seem - if a second payment is considered, occurring at instant t', when A is now a seller and B a purchaser. Does this second transaction necessarily take place and, furthermore, if it actually occurs, is it of necessity arithmetically equal to the value of the first transaction? The new theory answers both questions in the affirmative, in opposition to conventional wisdom. But where does the required proof lie, justifying this new line of thought? We have just mentioned that two quite separate arguments are offered. The Kreislaufaxiom If an axiomatic law forces A to sell at instant t goods equal to x, the Kreislaufth'eorem holds for the period t to t' but not for any smaller time interval starting from t. Therefore, the new theory is only intellectually binding if period t to t' is reduced to zero, that is when t' coincides with t. This conundrum of thetime-dimension of circular flows has not yet been solved within they German literature. The Accounting Constraint At instant t A pays out a sum of money, equal to x; financial intermediary C forwards this payment to seller B. If every such intermediation elicits the dual formation of a liability of the purchaser and of an equivalent asset of the seller vis-a-vis the financial institution C, then no axiom is needed to ensure that every outgoing monetary flow defines, at the very same moment of time, an equal incoming monetary flow accruing to the purchaser. At the moment when he pays for his purchases - x units of money flowing out of his holdings - A recoups this same sum of money in payment of an equivalent sum of financial obligations which he is 'mechanically' obligated to sell to financial intermediary C. We have just described a circular flow which, grounded in the 'mechanics' of double-sided accounting, has nothing whatever to do with any axiom. Reichardt (1967) comes very close to formulating the Kreislauf as the technical result of a generally applied method of accounting. In the real world evolving before our eyes, all payments effected in any bank money are circular: at instant t purchaser A is not only debited but also credited to the amount of x 348 The National Economy Studied as a Whole monetary units. Here the deep opposition which runs between the new paradigm and conventional theory is highlighted; it is a generally accepted view that two distinct transactions are required if one and the same person is to be defined as a purchaser (in one transaction) and a seller (in another transaction) - each payment therefore appears to relate to one purchase and, correspondingly (or tautologically), to one sale; circular flow analysis takes each single transaction as a 'purchase-sale' of the relevant purchaser and a 'sale-purchase' of the corresponding seller. In other words, each single payment relates to two purchases and to two sales. Suppose that at instant t A pays for a car which he has bought; since the sum paid out by A is conveyed to the seller (B) by a financial intennediary (C), A is simultaneously - namely at instant t - the purchaser of a car and the seller of a financial obligation (bought by C); it follows that at the same instant B is the sellet of a car and the purchaser of a financial obligation (ceded by C). As soon as it is realized that circular flow analysis depicts the daily working of a monetary economy, no misgivings need linger on concerning the status, formal or otherwise, of the new paradigm. Nothing is more concrete or real than the actual functioning of banks. Reichardt (1967: 101) states the accounts relating to the fonnation and to the expenditure of national income; when exports and imports are left out, they simply read as follows: Income Fonnation Banking system assets liabilities y C I Income Expenditure Economy at large assets liabilities y c s Real income is a quantity which, from its origin, is both positive and negative. Positive real income is a liability of the banking system, i.e. of the banks which are in charge of all the 'figures' constituting national accounting. Negative real income is a liability of the economy at large relative to the banking system. In the aggregate or, more accurately, for the economy as a whole, the fonnation of positive real income - C + I - and the destruction or expenditure of positive real income - C + S - each define a 'semi-circle', while the conjunction of these two opposite flows yields a circular flow. Outlays on consumption (C) and on investment (I) are sums of money flowing out of banks; the corresponding real output (y) accrues to the same banks. (b) Expenditures financed out of income, on consumption goods (C) or on 'savings' (S), are flows of money accruing to banks; the corresponding real goods enter households and finns. Conventional theory grants a degree of freedom to one set of expenditures relative to the other; aggregate supply and aggregate demand (expres(a) Bernard Schmitt and Spartaco Greppi 349 sions which are unexceptionable in this context) need not be equal to each other. Circular flow analysts, on the other hand, point out the simple fact that, for each person or group of persons, the equality of liabilities and assets must hold at all times. It is therefore logically impossible, given the rules which apply indifferently to private and to national accounting, to find, at any moment of time, a positive difference between global supply and global demand, or conversely. The fundamental question is still with us: if circular flows stem simply from accounting rules, why is it that they should still be based on an axiom? Yet another, equally important point must be raised: if circular flows coincide with accounting identities, then surely they can hardly be reconciled with equilibrium analysis. Along theoretical lines also developed by Eberhard Fels (1956), Reichardt (1967) provides the following attempt at reconciling the existence of Kreislauf identities with macroeconomic equilibrium analysis. At the same time, Reichardt endeavours to solve the question of the dual foundation of circular flow analysis, axiomatic (that is logically formal) on the one hand and/actual on the other (in so far as circular flows are simply brought about by the practical formal rules of accounting). 'In the final analysis the axiom establishing circular flows plays a role in the sphere of macroeconomic equilibrium to the extent that this axiom expresses eqUilibrium conditions as relations pertaining to the mechanics of circular flows, (ibid.: toO). Reichardt then proceeds to explain that values which finally appear in accounting books are the outcome of a process, which has been allowed to work out its full effect, by which they have all been brought into conformity with equilibrium analysis. It is quite obvious in this respect that circular flow analysis is closely akin to conventional ex ante, ex post reasoning. Most authors have followed suit, like Wilhelm Krelle (1967), Erich Schneider (1969), Franz Haslinger (1984), Alexander Blimovich (1944), Gerhard Mangold (1953) and Michael Kr611 (1954, 1956). The theory held by Kr611 is none the less somewhat original in so far as he stresses the predominant importance of identities over eqUilibria. Indeed, according to Kr611 circular flow analysis is inherent in the identity of Production and Income (1954: 212). And how could one question the validity of this identity even in conditions which did not satisfy the eqUilibrium criterion of supply and demand? The identity of Production and Income holds irrespective of the state the national economy may be in, not only in full employment but also in conditions of underemployment (deflation) or orevemplogment (inflation). KrlSll sees the validity of the Kreislauf even in an economy functioning without a circulating medium. Social product and social income can never be balanced against each other; they are not the terms of a mere equilibrium; on the contrary, the identity which holds between them 'is inherent in any economy which works according to Production and Consumption' (ibid.: 215-16). When money is brought into the 350 The National Economy Studied as a Whole picture nothing is fundamentally changed. 'The circular flows of money and of capital are mere complications of the simple circular flow' (ibid., p. 219). If Krlill had strictly adhered to these views, his scientific standing would be secure. When he argues that accounting methods are derived from underlying real phenomena his merit is all the greater; the standardized system of national accounting, which only improves previously accepted techniques - without changing their fundamental logic - rests entirely on Production and Consumption. Therefore, no 'trick' is involved. Reichardt clearly does not realize that the identity of Production and Consumption defines the economic circular flow quite irrespective of accounting methods. The truth is that neutral accounting reveals the prior real existence of circular flows. Again, if Krlill had been successful in working out a complete theory of the real circular flow encompassing the whole range of events which make up a real national economy, his theory would have turned the page of Walrasian equilibrium analysis, leading economics on to the really novel paradigm of genuine identities. In fact, Krlill cannot help reverting to the consideration of equilibrium conditions. Money is only ideally neutral; and in all circumstances which, in the actual world, do not verify the neutrality of money, the identity between S (savings) and I (investment) does not hold. 'In the standardised system it is assumed that all savings are invested ... But we have to take into account the fact that some savings can lie idle in households or in banks' (Krlill, 1956: 90); when this is the case there is no other way but to deny the very existence of circular flows. It is surely a great pity that such a considerable intellectual construction as circular flow analysis should thus end up in loose, inconclusive thought. It is warranted, all the same, to conclude on a more optimistic note. In 1989, Adolf Wagner still writes that 'a circular flow system is at equilibrium when each of the contracting parties receives and surrenders the same value' (ibid., p. 514). Taken literally this passage again points to an obvious illogic for an unbalanced circular flow is not even circular. To some extent, however, the writers did succeed in sorting out and neatly combining the three foundations of circular flow analysis, namely the Kreislaujaxiom, equilibrium, and induced variations of the capital holdings of contracting parties. The underlying logic of all this runs as follows: first, equilibrium is reached axiomatically (excess demand functions thus being rendered obsolete); secondly, equilibrium between two successive payments - negative (purchases) and positive (sales) - of one and the same person is superseded by instantaneous equilibria which are observable in both phases of income formation (sales) and of income expenditure (purchases). Thus each individual payment defines a full circular flow. When a sum of money is spent by A in favour of B, financial intermediary C purchases obligations from A (who automatically or mechanically gains the corresponding advanced income) and sells them to B (who loses for the time being the income which he advances to A via financial intermediary C; when such sales and purchases of financial obligations inherent in each and every payment are taken into account, it appears that two Bernard Schmitt and Spartaeo Greppi 351 successive payments involving the same person define two successive circular flows, each of which occurring 'within' one single instant. An axiom positing instantaneous circular flows would hardly be theoretically acceptable; but what is at stake is quite different: circular flows are instantaneous by virtue of the concrete working of financial intermediations which are implicit in each single payment effected through the means of a 'circulating medium'. Kreislauftheorie conforms to facts in yet another important respect. Real Production and Real Consumption Are the Two 'Legs' of the Circular Flow 'Ordinary' theory considers production as a flow which can be studied quite separately from consumption flows; in accordance with this assumption of 'separability' a national economy could still witness positive production flows even if its outputs were never absorbed by any consumption, either domestic or foreign. Non-invested savings, in Krlill's sense (that is non-expenditures), could replace consumption; stocks of produced, positively valued goods, would simply pile up in households and firms. Circular flow analysis links production and consumption flows so strongly to one another that no production could possibly take place if it were not simultaneously defined as an equal consumption. To produce means to 'produce-consume'. Production and consumption are the two sides of the same 'coin'; if one side is nil so is the other. Kreislauftheoretiker emphasize the distinction between production as it is conceived by engineers and production which is the core of macroeconomic studies. Only technical production - say the output of so many cars per unit time - is a process which is measured irrespective of subsequent consumption. The trivial fact remains, of course, that it would not be worth the while of entrepreneurs to produce goods which- no one wanted to buy. But any output which remains unconsumed is still positive and points to an equally positive production process; the corresponding loss which burdens the producer is positive forthe very reason, precisely, that production stands unchanged whatever the level of subsequent consumption. All circular flow theorists - let us cite a few more among them, Jlihr, Blihler, Fohl, GrUnig, Hasselblatt, Koopmans - teach that macroeconomic production and consumption flows are given uno actu since each act of production taking place in the real world is in fact a 'production-consumption'. At first (but possibly for prolonged periods of time) new theories are always dismissed as being trivial or false. Now we are at least aware of the exact criterion on which the circular flow paradigm is to be judged. Economists - and there are still a good many of them - who think that production and consumption are separate flows which may be equal or unequal to each other, implicitly disregard the strenuous efforts which a whole school of thought has exerted over a period 352 The National Economy Studied as a Whole of nearly a century. But even then the crucial point remains open: do actual facts confirm (or otherwise) that each act of production is a circular flow? And, surely, this fundamental question is not addressed by mainstream economists who simply go on building deft but imaginary production functions. 'The whole macroeconomic process is set between production and income expenditures' (Gerhard Mangold, 1953/1: 38). The same author argues that the theories developed by Walras, Pareto and Cassel can never be verified by statistical data since their foundation lies in microeconomics. To the contrary, the circular flow encompassing production and consumption is a macroeconomic phenomenon whose existence and measure can be statistically ascertained. In any economy which is still in such a 'rude state' as to forgo the circulation of a medium of exchange, the identity of production and consumption is timeless. The main task of Kreislauftheorie is formidable indeed, for how is one to explain the existence of an identity whose terms belong in two different moments of time? To this day no writer in the German language has been successful in solving this dilemma. Money is the 'culprit', for Say's law does not seem to work in a monetary economy. But as soon as Say's law is negated, circular flow analysis loses all its substance. It still remains true to say that circular flow analysist have made significant contributions to monetary theory. The Circular Flow Approach in Monetary Theory In this field of research one author stands out, namely Carl F1>hl, whose great work, GeldschiJpfung und Wirtschaftskreislauf (1955), was highly praised by Hans Peter who thought it was the best German contribution to the theory of money since Knapp. To this day F1>hl is widely hailed as the German Keynes. Like most, if not all, Kreislauf theoreticians, F1>hl translates the circular flow defined by production and consumption into a circular flow whose 'poles' are firms on the one hand and factors of production on the other. In a given period of time entreprises (E) disburse x units of money in order to 'remunerate' factors of production (F) who earn wages, rents and interests (Figure 11.2). E x -------II~~ F Figure 11.2 No circular flow is yet apparent. But FBhl pushes analysis one step further by taking into account expenditures flowing from the given incomes. Factors of production now disburse x' units of money in purchases of current output (Figure 11.3). Bernard Schmitt and Spartaco Greppi F 353 x' - - - - - -....~ E Figure 11.3 In his book F6hl pursues but one aim: he purports to prove that in the actual world we live in x' is identically equal to x. Although he was a mathematician by training, F6hl uses a purely 'conceptual' method of investigation. In mathematics it may suffice to derive identity x = from an axiom. In the eyes of an economist, however, replacing the exercise of 'conceptual thought' by the wanton formulation of an ad hoc axiom amounts to having recourse to a petitio principii. Proving identity x = x' may be the work of a lifetime - it was for Fohl. And the question whether he fully succeeded in reaching his goal loses much of its relevance if it should turn out in the end that Fohl's insight was correct. Clearly, if x> x' or if x < x', circular flow analysis has no meaning. But suppose we find that equality x = x' holds under certain conditions; even then circular flow analysis remains totally unproven. In strict logic we should go much further: the new paradigm is worthless if equality x = does not hold under all conceivable circumstances. Only one door remains open; if analysis establishes that no income can be created or formed except by its own expenditure or destruction, then - and only then - is x indeed identical to x. Even back in 1937 F6hl was much less interested in equilibrium analysis than are many circular flow analysts today. The Kreislau/schema is not a Soll-Begriff but an 1st-Beg riff; the set of all firms simply cannot create even one unit of income in excess of total expenditures resulting from the sum of created incomes; conversely, the set of all factors of production cannot spe'nd even one unit of income in excess of incomes created or 'injected' by firms. Once it is thus granted that x = x' is an identity, analysis cannot help being confronted with one of the most frequently discussed macroeconomic problems in our century, namely the identity between saving (S) and investment (I). On this very important point Ft>hl has much to offer. If consumption (x) measures production (x) it follows that savings measure investments. 'The person who saves ... provides the funds to be invested ... Entrepreneurs decide which part of the savings will actually be invested' (Ft>hl, 1955: 66). But how are we to interpret that part of savings which is not invested? We are strongly tempted to call such savings hoardings. But whenever hoarding is positive or negative x is greater or smaller than x' and, as a consequence, the Kreislauftheorie falls apart its centre does not hold. Erich Preiser, a member of the 'school', clearly points out the illogicality we are facing - by falling prey to it! In his article 'Sparen und Investieren' x x 354 The National Economy Studied as a Whole (1944: 300) he writes: 'When savings exceed investments, nominal capital is increased but real capital is not.' Gone is the Kreislauf1 Albrecht Forstmann (1955: 443) offers a more profound analysis: the investment of savings yields an equivalent amount of fixed capital; and all monetary savings which stand over and above investment proper finance an equal increment in working capital. Even consumption goods which are not yet sold belong in the category of capital goods (see F5hl, 1955: 57). There is only exception to the identity of x and x' purely apparent at that - that is the creation of an additional sum of money: creations of money reduce to creations of units of capital which are purely nominal (ibid.: 137). What F5hl calls money creation, GeldschOpfung, is an event which is quite unrelated to production and consumption. If dx is an addition to the available quantity of money, derived from money creation, then the Kreislauf, unchanged in principle, is simply quantitatively redefined, the given output now being measured by x + dx units of money, under the rule of identity x + dx = x I + dx. Was F5hl entirely successful in solving this most difficult question of identity S = I? In all fairness we must answer no. When I exceeds S, so argues F5hl, production tends to increase and S with it. Here F5hl comes very near to Keynesian multiplier analysis. But identity S = I categorically precludes the existence of any time interval, however short, during which I and S could be made equal to each other. By deriving an identity from an 'equating process', F5hl lends wider credence to a blunder which severely tarnishes Keynes's General Theory. But isn't F5hl's error even greater? If I and S can be unequal in real time, economic flows can never be circular. Denying identity S = I amounts to an outright rejection of circular flow analysis. Peter Bernholz (1960), in a comment he offers on Wolfgang Stutzel's theory of money, is more logical when he links identity S = I to national accounting. No casual analysis can be associated with accounting identities - it is therefore obvious (selbstversttindlich) that the two following categories of dynamic theory must be false: savings call forth investments; investments call forth savings. But is economic reality constrained by mere accounting identities? It is well known that most economists, even those who profess to study macroeconomic phenomena, feel entitled to neglect accounting identities and even to infringe upon them. Identity S = I is only a fact; why should high theory be encumbered with facts? F5hl acts like a great scientist when (and only when) he endeavours to make his thought conform to reality - instead of the other way around. Identity S = I must be explained, it is not good enough just to explain it away. At his best Fl)hl tackles the real problem; but does he solve it? Hardly. The argument which he advances, namely that working capital should be added to fixed capital if we are to obtain a complete measure of investment, contradicts his own theory of circular flows. When households refrain from consuming part of their incomes, consumer goods remain piled up, still available to be sold by firms; in this fashion I Bernard Schmitt and Spartaco Greppi 355 is brought to equality with S. But circular flow analysis is all about flows; flows of investment equal flows of savings. The accumulation of working capital can therefore offer no clue; if x is the amount of earnings paid out by firms to their factors of production and if x' is the amount of x spent by households, then circular flow theory, far from being corroborated, finds its demise in any inequality which may appear between x and x'. In particular, when x' < x, it serves no useful purpose to argue that a corresponding increase in working capital must take place somewhere; even if it does, the equality of flows x and x' remains broken beyond repair and the whole body of circular flow analysis with it. Fohl's monetary theory of profits merits the same praise and suffers from the same shortcomings as his theory of saving and investment. The Existence of Macroeconomic Profits In Peter's theory profits flow between 'poles' just like wages, rents and interests. Formal circular flows only consider abstract expenditures; and profits are spent like any other income. Theorists who were more inclined to relate circular flow analysis to the observation of actual national economies were more attentive to two particular kinds of expenditure, income/orming and income using. Foremost among them is Foh!. If x units of money income are produced in a given period of time, only x units of money can be spent on the equivalent output. Wages, rents and interests are paid out by firms. Within each national economy the set of all firms thus 'creates' the monetary incomes accruing to households. But can the set of all firms 'create' its own monetary income, namely profits proper? In answering this question, Fohl again equals Keynes. More 'ordinary' minds are apt to satisfy themselves with mere appearances. How could profits re~lized by the set of all firms be positive? Wolfgang StUtzel (1979: 50), a banker and a professor, certainly knows (or does he?) what he is talking about: profits can be positive for some firms (Partialsatz) but not for all firms (Globalsatz). In an economy which comprises only firms EI and E 2, profits made by EI or E2 may be positive; but how could the economy as a whole procure a positive profit to the set of firms EI and E2? Nevertheless, Fohl's theory remained firmly committed to I.he existence of a positive macroeconomic profit accruing to the set of all firms of a national economy. 'When dealing with the profits of the set of firms we were referring to profits which are not compensated by any corresponding losses' (Fohl, 1955: 155). It is important to note that the profits Fohl had in mind were monetary magnitudes. A global positive profit could arise in spite of identity x = x' if firms were able to employ some factors of production 'for nothing'. That is a 'cheap' assumption and circular flow analysis would have nothing to do with it. Like any other category of income, profits are formed in money units. Suppose firms payout 10 thousand million dollars in 'factor costs' (x = 10); households correspondingly spend 356 The National Economy Studied as a Whole 10 thousand million dollars (x' = x); if global profits are equal to 2 thousand million dollars, this sum is neither a part of x (wages, rents and interests) nor a part of x' (expenditures which just cover 'factor costs'). Now, circular flow x = x' comprises all incomes, without any exception, that is wages, interests, rents and profits. When facing this dilemma Fohl is able to resist jumping at the obvious and utterly flawed - conclusion offered by StOtze\. Fohl holds on firmly to positive global profits, which is no small accomplishment. If Fohl finally failed to make his most remarkable point on the existence of macroeconomic profits in an entirely convincing manner, his method was not at fault; it only stood in need of being brought to a greater degree of maturity. When formal and real circular flows are eventually 'merged' together by the means of a renewed theoretical effort, into one and the same concept, relating to one and the same concrete fact, fundamental problems which the Kreislauftheoretiker had clearly perceived and tackled will be solved - and macroeconomics will come of age. 3 CIRCULAR FLOWS: A FORMALLY RIGOROUS DESCRIJYfION OF REAL PHENOMENA Leon Walras and Hans Peter StUtzel (1979: 127) is right when he emphasizes the Walrasian tradition in Peter's writings. General equilibrium and circular flow analyses both have very abstract foundations. But equally abstract theories may differ in an important respect; some relate to facts while others depict a purely imaginery world. General equilibrium analysis is purely abstract in a derogatory sense. Pure competition simply cannot exist in the real world. The existence and stability of equilibria are mathematical problems, unrelated to any real economic needs of people or society. But 'general competitive analysis' is open to a much more serious reproach for it is fundamentally illogical to express prices as ratios between physical quantities. In other words, even in the realm of pure logic relative prices cannot be determined. Consider an economy where n = 2; A offers a quantity of commodity 1 and B offers a quantity of commodity 2, each person demanding a quantity of the commodity held by the other person. If a price is a ratio between physical quantities, then only one price is to be determind. The price of one commodity relative to the other could not be computed if it were not for the existence of a deus ex machina known as Walras's law which states that supply and demand of commodity I are identically equal to demand and supply of commodity 2. But logically this 'law' is only correct if it expresses the separate - and even diverging - viewpoints of the parties involved, A and B. During the tatonnement, two independent supply and demand equations coexist; as a result the price determin- Bernard Schmitt and Spartaco Greppi 357 ing process cannot converge towards the expected solution. And theory can derive no help from the numeraire. Say the price of the 'numeraire-commodity' is, by convention, made equal to number I; the precise physical quantity of the 'numeraire-commodity' which will actually enter into exchange at equilibrium still remains to be determined: the equilibrium price of the numeraire is therefore unknown. General equilibrium analysis is thus a system of price calculation which is either overdetermined or underdetermined. Circular flow analysis reaches much beyond Peter's modest ambition for it replaces relative prices by absolute prices. When n = 3 (including a bank) two distinct prices are determined by two independent equations; the price of commodity I is determined by equation xl::: x'I while the price of commodity 2 is determined by equation x2 = x'2 - where x I, x2 are the 'factor costs' of the production of commodities I and 2 and x'I, x'2 the corresponding values of expenditures effected by households in purchasing commodities by I and 2. Circular Flows: Identities or Equilibria? We have already mentioned a curious contradiction into which several Kreuslauftheoretiker (like Reichardt, Mangold, Kroll and Krelle) have fallen by stating that circular flows are defined by the conjunction of two identical (but opposite) flows which have previously been brought into mutual balance or equality. If x and x' are the terms of an identity, by no stretch of imagination can these two magnitudes be made independent of one another. The fact that the two sides of the same coin are identically equal precludes the existence of any t{Jtonnement or 'equalizing process' by which they were eventually made equal. In other words, identities are timeless or permanent phenomena. If, at any moment of time whatever, x should be unequal to x', then at no moment of time at all would it be logically correct to construe equality x = x' as defining a circular flow. Conversely, if equality x::: x' positively depicts a circular flow, then neither x nor x' can be found at ANYTIME to differ from its corresponding term, x' or x. But if that much is granted we are landed with the following, most awkward problem: if x and x' are no longer to be measured against each other (as they were supposed to be in Walrasian analysis), by what method is their measure nevertheless ascertainable? Circular Flows and the Validity of Say's Law In an economy where consumption is directly linked to production - that is, in a Naturwirtschaft - Say's law is valid. In a monetary economy - Geldwirtschaftthe free behavior of income-holders takes substance out of Say's law which, as a consequence, is degraded from the status of an identity to become a mere condition of equilibrium. We have not found a single instance in the German language in which circular flow analysis was said to hold, without a fault, in strict conformity to Say's law, with respect to monetary economies. The National Economy Studied as a Whole 358 The reason why Reichardt (1967), in particular, considered circular flow analysis to be a purely formal construct lies in the fact (in his mind indubitable) that Say's law strictly applies only to a non-monetary economy. If it were true that Say's Law is irreconcilable with the working of a monetary economy, circular flow analysis would belong in the same category as the theory of general equilibrium, both paradigms being abstract in the pejorative sense of the word. In reality, however, the more abstract it gets the nearer circular flow analysis comes to actual facts and events. Finally circular flow analysis is nothing but the formulation of Say's law, undiminished in its strict validity, in the context of actual monetary economies. Five Problems Are Simultaneously Solved: the Determination of Prices, the General Validity of Say's Law, the Time-dimension of Circular Flows, the Definition of Macroeconomic Magnitudes and, Finally, the Existence of Macroeconomic Profits What Is a Payment? 1. The NOMINAL aspect of each payment: Following and improving upon what we call the 'collective works' of the Kreislauftheoretiker - as published especially by Peter, Reichardt, Mangold, Fohl and SUitzel - we may represent as follows a single payment of x units of bank money made out by firm E in favour of 'factor of production' F (Figure 11.4). Here we have three 'poles', a bank, a firm (F) and a factor of production (F). (a) (b) On E's request a bank lets x units of money 'flow out of its accounting books', thus eliciting flow (1). Flow (I) is null and void unless it is 'continued' beyond E. x____ E a bank ~I) ~x' (4)~ ~ x the same bank [A] Figure 11.4 Bernard Schmitt and Spartaco Greppi 359 No bank can issue a sum of money without instantaneously taking it back; the necessary equality, or equivalence, in banks' accounting books, of money 'outputs' and money 'inputs' makes it logically impossible for a bank to issue a positive sum of money (+x) when n = 2; we already know that Peter has not failed to highlight this crucial fact (another member of the 'school' even argues that bank money can never leave the banking system (Binder, 1975: 28). If E was tempted to keep, instead of handing on, the 'current' (or Howing) sum of x units of money, he would land up with a zero sum of money for he would posses (+x) the very object he owes to the bank that issues x units of money out of E's deposits, which, in consequence, are drawn down by that amount (-x). When factors of production F concur, the desired result is achieved; n now being equal to 3, a positive payment of x units of money actually emerges in the depicted process. Nevertheless, the accounting mechanism is in force for any value of nand bankers can only act in accordance with it. In other words, flow (4) instantly counteracts flow ( 1). We have just 'discovered' - and remember, Reichardt wondered whether any discovery could be made in this field of research - the existence of a perfect circular flow which is pure both in theory and in fact. The REAL aspect of each payment: Nominal Hows in [A] correspond to the following real 'motions' (Figure 11.5). In a NaturwirtschaJt, factors of production take instant possession of their current output, at the moment when it is produced. Even then the process is aptly construed as a circular How. In economics (as opposed to biology) a commodity is consumed precisely at the moment it is appropriated by households. In the absence of a paying medium each act of production is therefore, without any delay, also an act of consumption. And each 'production-consumption' is a circular How. In a Geldwirtschaft factors of production lend their current output to banks, which lend them on to the relevant firms. The newly produced commercial goods are thus first replaced by financial obligations: via the banking system firms become financially indebted towards their factors of production. 2. current output ~ lent ~~~~~~~~ .A~~~~ E ~~~ """," F obligations ""..... ~/~/~biigations ceded the same bank ceded [B] Figure 11.5 360 The National Economy Studied as a Whole In a monetary economy factors of production instantly purchase their own output - not in kind but in money - at the very moment when they perceive their money wages. In Figure 11.4, Flow 4 defines a purchase of output which logically and instantaneously prolongs flow 3: money wages are therefore spent - in money, not in kind - as soon as they are received. This necessary· equality of incomes received and incomes spent is the very essence of the Kreislauftheorie. We may now briefly tum our attention to the five aforementioned problems. The Determination of Prices Suppose that for a given output equality x =x' is a condition of equilibrium; x' is then the eqUilibrium price of that output. Now, circular flow analysis establishes the unquestionable fact that whatever the value of x, x is always equal to x' or, more exactly, x' always measures up - neither more nor less - to the value of x. As a result, money prices seem to be indeterminate. Further reasoning leads to the determination of money prices, or absolute prices - while, in the Walrasian system relative prices must logically remain undetermined. Circular flow analysis is the only available method capable of yielding the knowledge of relative prices, which, in the real world and in pure theory, are nothing but ratios of absolute prices. True, x' is a mere consequences of x, or a 'mechanical' prolongation of x (as the Kreislauftheoretiker would say). But it is important to keep in mind the fact that x' does not define the final purchase of the given output; flow x' defines the purchase of current output at one remove, that is, financially and not (yet) commercially. Firm E will incur a loss if a positive fraction of x should finally be found to be lacking in the sum of commercial expenditures decided upon by households (I'). If, given x, x" is the value of expected final commercial purchases, x =x" defines the condition of eqUilibrium by which the price of current output is determined: x is the 'equilibrium price' if and only if households decide to spend x" = x units of money in purchasing the physical or commercial output whose 'factor cost' is x. Here the reader may have the uncanny feeling of being on familiar ground; indeed, if we construe effective demand as a concept including supply, Keynes emerges as a central figure among flow theorists. The General Validity ofSay's Law Equality x =x' is not a condition of equilibrium, x is the value of current supply; x' is the corresponding value of current demand; demand has no degree of freedom relative to supply; any output whose 'factor cost' is actually met by firms is the object of an equal demand exerted by households; any hoarding, whether positive or negative, is the fruit of a misconception; in reality no fraction of incomes can ever be hoarded since all incomes formed in bank money are, from the moment they spring into existence, entrusted to the banking system where they remain deposited. Flow x' is the expenditure of 'factor earnings' which automatically (that is whether jJernard Schmitt and Spartaco Greppi 361 households like it or not) transmogrified into bank deposits. Furthermore, the purchase (by households) of bank deposits to the amount of X'= x units of money denotes the instantaneous purchase of the corresponding physical output which assumes the temporary form of financial obligations (owed deposits). Say's law is valid under all conceivable circumstances, irrespective of the state the national economy may be in (fully employed or otherwise; stationary or growing). But does Say's law hold when x' x"? It certainly does; when x" < x firms E incur a loss (equal to x - x"); this loss is a direct result of Say's law - which suffers no possible exception. *' The Time-dimension of Circular Flows Suppose that 'at first' and for an unspecified duration, available incomes (x) are only partly spent (x'). Then the existence of circular flow x = x' may still be confirmed some day; facetiously, it could even be argued that at the end of times circular flows will at last prove their truth-value. More seriously, circular flows are timeless identities. If x' were not equal to x from the beginning, that is from the moment flow x creates current income, circular flow analysis would have no object. When households finally convert their money deposits into physical consumer goods, by retrieving current output kept 'waiting' in the holdings of firms, additional circular flows appear with the expenditure of each unit of income; all these flows, like all circular flows, are instantaneous identities. Consider now the conjunction between one unit of an 'income-forming' flow and one unit of the corresponding 'income-destroying' flow, both flows relating to physical output. Each of these two opposite flows is a payment and, as we know by now, each payment is in itself a perfect circular flow. But, relative to a given physical output, 'income-forming' (IF) and 'income-destroying' (ID) payments are successive events in chronological time. At this point it is expedient to consider a 'stroke of genius' to be found in Peter's work: circular flows are not defined in Euclidian space (1954: 7). Since payments IF and ID relate to the same physical output, IF and ID constitute a timeless identity, that is a circular flow, whose terms are two 'circular payments'. The Definition of Macroeconomic Magnitudes Let us reconsider the paradox posited by Stiltzel (1979). Relative to the banking system, factors F become creditors (x'); at the same time, though, firm E assumes an equal debt (x =x'); it can readily be seen, therefore, that no positive macroeconomic value emerges from the production process. But StUzel's argument is flawed. Flow x' defines the financial purchase - not the commercial purchase - of current output. It follows that physical output is still there, waiting to be pur- 362 The National Economy Studied as a Whole chased. The value of these newly produced saleable goods is x. We are thus confronted with: -- A monetary + x (F's) credit towards the banking system; - A physical + x (saleable current output); - and a single - x (E's debt towards the banking system). The + x which stands out defines a macroeconomic value. Circular flow analysis deals with macroeconomic values, that is, with newly produced goods, defining real income and real capital. In contradistinction to circular flow analysis, the theory of relative prices formalizes exchanges between already existing goods; even exchanges involving productive services follow the same - 'conservationist' versus a 'creationist' pattern of thought. Circular flow analysis sheds light on the distinction between microeconomics and macroeconomics. Microeconomics is the science of the conflicting and reconcilable behaviours of economic agents. Macroeconomics is the mechanics of value determination via the production process: x =x' is an identity irrespective of the behaviours of producers and consumers. Even a single unitary circular flow results in the definition of a macroeconomic value. Conversely, not even the sum of all relative prices (supposedly) determined in a national economy constitutes a macroeconomic value. In the textbook published by Wilheim Lexis (1910), Volkswirthschaft and Kreislaufare therefore rightly considered to be synonymous expressions. Macroeconomic Profits Fiihl was fundamentally right; macroeconomic profits do exist. In contrast to Walrasian theory, circular flow analysis, if pushed to its logical conclusion, clearly establishes the existence of positive profits to which no negative profits correspond anywhere. When a circular flow consumes none of the incomes which it cr~ates, all resulting earnings are wages. Macroeconomic profits arise whenever circular flows reproduce wages which had previously resulted from other circular flows. Macroeconomic profits imply a concatenation of circular flows in chronological time. Ftihl stopped short of recognizing this fact - so here ends our brief survey of circular flow analysis in the German language. References Bernholz, P. (1960), 'Volkswirtschaftliche Saldenmechanik' , Schweizerische Zeitschrift fUr Volkswirtschaft und Statistik, 96, pp. 228-34. Binder, P. (1975), Kritik der tradition ellen Wirtschaftstheorie und der herkommlichen Wirtschaftspolitik (Berlin: Duncker & Humblot). Bernard Schmitt and Spartaco Greppi 363 Blimovich, A. (l944), 'Das allgemeine Schema des wirtschaftlichen Kreislaufs', Zeitschrift fiir Nationalokonomie, 10, pp. 199-241. B6hler, E. (l957), NationalOkonomie: Grundlagen und Grundlehren (ZUrich: Polygraphischer Verlag). Burchardt, F. (l931), 'Die Schemata des stationliren Kreislaufs bei B6hm-Bawerk und Marx', Weltwirtschaftliches Archiv, 34, pp. 525-64. Burchardt, F. (1932), 'Die Schemata des stationliren Kreislaufs bei B6hm-Bawerk und Marx', Weltwirtschaftliches Archiv, 35, pp. 116-76. Copeland, M.A. (1952), A Study of Money Flows in the United States (New York: National Bureau of Economic Research). Fels, E. (1956), 'Katallatik', in Handworlerbuch der Sozialwissenschaflen, vol. 5 (Stuttgart: Gustav Fischer; TUbingen: J.C.B. Mohr (Paul Siebeck); and Gtlttingen: Vandenhoedk & Ruprecht), pp. 569....75. Ftlhl, C. (1955), GeldschOpfung und Wirtschaftskreislauf(Munich and Berlin: Duncker & Humblot, Istedn, 1937}. Forstmann, A. (1955), 'Vom Paradoxon der Ausschliesslichkeit der Kreditsch6pfung', Zeitschrift/iir die gesamte Staatswissenschaft, III, pp. 438-72. GrUnig, F. (l933), Der Wirtschaftskreislauf(Munich: Beck}. Haslinger, F. (1984), Volkswirlschaftliche Gesamtrechnung (Munich and Vienna: Oldenbourg). Hasselblatt, W.B. (l952), 'Einkommen und Prodution', Zeitschrift fiir die gesamte Staatswissenschafl, 108, pp. 101-25. Henn, R. (l957), 'Modellbetrachtungen in der Wirstchaft', Zeitschrift fiir die gesamte Slaatswissenschaft, 113, pp. 193-204. J6hr, W.A (l952), Theoretische Grundlagen der Volkswirtschaflspolitik, vol. II: Die Konjunkturschwankungen (TUbingen: J.C.B. Mohr (Paul Siebeck), ZUrich: Polygraphischer Verlag). Koopmans, J.G. (1933), 'Zum Problem des "neutralen" Geldes', in F.A. von Hayek (ed.), Beitriige zur Geldtheorie (Vienna: Verlag von Julius Springer), pp. 211-359. Krelle, W. (l967), Volkswirschaftliche Gesamtrechnung einschliesslich input-outputAnalyse mit Zahlen flir die Bundesrepublik Deustchlalld (Berlin: Duncker & Humblot; 1st edn, 1959). Krtlll, M. (l954), 'Der Zirkulartausch. Eine Kreislaufanalyse', Zeitschrift fiir die gesamte Staatswissenschaft, 110, pp. 212-38. Kr611, M. (1956), 'Sinn und Wert der volkswirtschaftlichen Gesamtrechnung', lahrbiicher fiir NationalOkonomie und Stalistik, 168, pp. 81-105. Kroll, M. (1956), 'Sinn und Wert der volkswirtschaftlichen Oesamtrechnung', lahrbiicher fiir NationalOkonomie und Statislik, 60, pp. 577-623. Leontieff, W.W (1928), 'Die Wirstchaft als Kreislauf, Archiv fUr Sozialwissenschaften und Sozialpolitik. 60, pp. 577-623. Lexis, W. (1910), Allgemeine Volkswirtschaftslehre (Berlin and Leipzig: B.O. Teubner). Mangold, O. (1953a), 'Die Strukturanalyse des wirtschaftlichen Kreislaufs', Schmollers lahrbuch, 73/1, pp. 27-72. Mangold, O. (1953b), 'Die Strukturanalyse des wirtschaftlichen Kreislauf', Schmollers lahrbuch, 73111, pp. 161-94. Neisser, H. (1931), 'Der Kreislauf des Oeldes', Weltwirtschaftliches Archiv, 33, pp. 365-408. Peter, H. (1938), 'Kreislauftheorie. Bemerkungen zu Carl Fijhls "Oeldsch6pfung und Wirtschaftskreislaur", 'Archiv /iir mothematische Wirtschafts-und Sozialforschung, 4, pp.251-59. Peter, H. (1953a), 'Zur Oeschichte, Theorie und Anwendung der Kreislaufbetrachtung', Schweizerische Zeitschriftfiir Volkswirtschafl und Slalistik, 89, pp. 1-24. 364 The National Economy Studied as a Whole Peter, H. (1953b), 'Zur Geschichte, Theorie und Anwendung der Kreislaufbetrachtung', Schweizerische Zeitschriftfiir VolkswirtschaJt und Statistik, 89, pp. 160-70. Peter, H. (1954), Mathematische Strukturlehre des Wirtschaftskreislaufs (Gottingen: Verlag Otto Schwarz). Preiser, E. (1944), 'Sparen und Investieren', Jahrbiicher fiir NationalOkonomie und Statistik, 159, pp. 256-309. PUtz, T. (1948), Theorie der allgemeinen Wirtschaftspolitik und Wirtschaftslenkung (Veri, fUr Geschichte und PoUtik). Reichardt, H. (1967), Kreislaufaspekte in der ()konomik (TUbingen: J.C.B. Mohr (Paul Siebeck». Schneider, E. (1969), Einfiihrung in die Wirtschaftstheorie. I: Theorie des Wirtschaftskreislaufs (TUbingen: J.C.B. Mohr (Paul Siebeck». StUtzel, W. (1958), VolkswirtschaJtliche Saldenmechanik (TUbingen: J.C.B. Mohr (Paul Siebeck». StUtzel, W. (1979), Paradoxa der Geld-und KonkurrenzwirtschaJt (Aalen: Scientia Verlag). Waffenschmidt, W.O. (1956), 'Mathematische Strukturlehre des Wirtschaftskreislaufs: Bemerkungen zu dem Buch von Hans Peter', Jahrbiicher fiir Nationaliikonomie und Statistik, 168, pp. 116-25. Waffenschmidt, W.O. (1957), Wirtschaftsmechanik (Stuttgart: W. Kohlhammer Verlag). Wagner, A. (1989), 'Makrookonomisches Kreislauf-undloder Marktgleichgewicht', Jahrbiicher fiir Nationalokonomie und Statistik, 206, pp. 510-16. B. Confronting the Mainstream 12 The Ambiguity of the Notion of General Equilibrium with a Zero-Price for Money Carlo Benetti INTRODUCTION The first problem of monetary theory is the relationship between money and the theory of value. Until now, two approaches have been proposed to deal with this problem. The first approach characterizes the dominant orthodox theory. Its foundation is the theory of value, that is to say, a theory of an economy composed of individuals and an a priori given list of physical goods known by each individual. In this framework, the monetary question necessarily deals with the integration of money into the theory of value. After the essential contributions of Walras, Hicks and Patinkin, Hahn deals with this problem in a more rigorous way. If a few opinions which were given in a moment of enthusiasm (although ephemeral)· are put aside, it is generally recognized that the vast research on this question has not led to satisfactory results. Within the orthodox approach there still does not exist a monetary theory which receives, by far, as much approval as that given to the theory of value developed by Arrow and Debreu. The second approach - often called heterodox - rather than attempting to overcome the difficulties of its rival, considers it is possible to avoid them. It has then tried to develop a theory of a market economy in purely monetary terms, without explicit reference to a theory of value. A decisive impulse in this direction has been given by Keynes in some parts of his works. But the problem raised by the theory of value (which should not be confused with the orthodox method) is fundamental to the understanding of the market society, as we will see later. As a result, the heterodox approach has failed to develop a general theory and concentrates on particular themes (the most important is the monetary and financial circulation). Whichever opinion is adopted, in accordance with a well-established tradition in this subject, we must admit that the present state of monetary theory is highly unsatisfactory. It seems to be due to the union of two negative tendencies embodied within these approaches: the second one tries to ignore an un 366 Carlo Benetti 367 avoidable problem, the first one confronts the problem, although incorrectly. In this chapter, one will find arguments in favor of the second conjecture. The orthodox theory conceives the monetary economy as an extension of the non-monetary economy which is obtained by adding a particular commodity fiat money - to the list of commodities in the initial endowment. The procedure leads to two very distinct problems, both clearly identified in Hahn's pioneering contribution. I. The first problem deals with the determination of monetary equilibrium. Since money is a commodity, the essential property of monetary equilibrium is the existence of a positive quantity of this commodity in the eqUilibrium allocation. In this respect, the monetary theory has become identified almost completely with the theory of the demand for money. This tendency goes well beyond the frontiers of the neo-Classical school. We also find this view in a considerable part of the Keynesian literature, in so far as it rests on suggestions made by Keynes with respect to money as a particular asset or wealth. The difficulty of this approach, although it remains unsolved, is clearly identified, primarily by Hahn. No matter which kind of eqUilibrium is being considered (temporary or stationary), no matter what the demand for money is (transaction, precautionary or speculative), and no matter what the underlying hypothesis is (uncertainty, risk or certainty), the demand for fiat money - contrary to the demand for any other commodity - is only defined if its price, anticipated and present, is positive. In other words, the theory of the demand for money presupposes the positive price of money and cannot be used in order to explain it (and therefore explain the eqUilibrium with monetary prices). As yet, none of the numerous attempts has succeeded in solving the problem in a satisfactory way.2 2. No matter what the answer to the first question turns out to be, we are faced with another problem which consists of Hahn's famous proposition put forward in his critique of Patinkin's theory. Hahn argues that the set of equilibria of an economy with fiat money contains a non-monetary equilibrium, that is an equilibrium where the price of money is nil. It is known that if the horizon of the economy is finite, such equilibrium is the only possible one. Theoretical research has therefore been oriented in either building models where, in addition to non-monetary eqUilibria, there also exist monetary eqUilibria, or in introducing hypotheses, generally exorbitant, such that non-monetary equilibria are eliminated. To our knowledge, Hahn's proposition has been unanimously accepted. The only differences which we have observed are in its interpretation. This proposition has been interpreted, on the one hand, as the natural property of all monetary economies (this seems to be the most widely accepted), on the other, as the consequence of the weakness in monetary theory or of an insufficient definition of a monetary economy. The most eminent representative of this second tendency is Hahn. 368 General Equilibrium with a Zero-Price/or Money Independently of the interpretation which can be given to Hahn's proposition, it has deep implications with respect to the relationship between the monetary and the value theory. Let us recall the central question which the theory of value tries to answer: is there a state of reciprocal compatibility between the individual activities which are based on the price information given by the markets, and if so, how is this state of compatibility achieved? The theory of value has succeeded in answering positively to the first part of this question, which is, in fact, that of the coherence of formal systems used in representing general equilibrium. It is, however, well known that it has failed in its attempts to answer the second part. Let us now consider an economy with money. The orthodox approach considers that, the problem of the existence of equilibrium prices being assumed as solved once and for all by the theory of value, the point of interest then is whether these prices can be expressed in terms of money (or, similarly, whether the equilibrium price of money is positive). In the worst case, it will be concluded that the monetary theory is bad in this sense: if the economy with money is not well specified, then it will behave like an economy without money. Being the result of a demonstration, Hahn's proposition reinforces the traditional idea that the theory of value is the foundation of economic theory, or that the 'real economy' is the core of the entire structure of the economic theory. Any other approach is hence discredited. This chapter proposes a critique of this proposition. 3 No attempts will be made to examine either disequilibrium positions and the adjustment of prices (although these constitute the best setting for studying money), or the orthodox conception of money. The purpose at hand is more limited. We will consider an exchange economy with fiat money where the 'inside money' is excluded.4 We will try to show that, in general, there does not exist in this kind of economy - well or not so well specified - a non-monetary equilibrium. This result is explained by the fact that a monetary economy is an object structurally heterogeneous to that of the theory of value. The latter is therefore not a subset always present in the former. As a consequence, if the object of economic theory is a monetary economy, the theory of value is not the most appropriate foundation. In other words, money must be an integral part of the initial description of an economy. The idea that, in an economy with money, private activities can be compatible in the absence of money is very far from common sense. Such are the kind of surprises we have come to expect from economic theory. Fortunately, on the point examined in this chapter, theory and common sense meet. The critique of Hahn's proposition is quite simple. First, a few notions will be introduced in section 2 to facilitate the presentation of the argument in section 3. The most delicate question is to find out how such a proposition has been demonstrated. In section 4 we will show how Hahn's reasoning, while formally correct, does not prove the existence of an eqllilibrium with a zero-price for money in a monetary economy. He demonstrates something very different: in an economy without money, which Carlo Benetti 369 allows for an equilibrium of prices. there does not exist an equilibrium where the price of money is positive. If our conclusions are correct. they will shed light on the relationship between money and the theory of value by specifying the limits of the latter which are more stringent than is generally admitted. The interest of alternative approaches to the dominant theory is then confirmed. 2 THE GENERAL EQUILIBRIUM OF THREE TYPES OF ECONOMIES Let us consider the simplest case of an exchange economy. Commodities are noted as i = 1•...• n. and individuals as h = 1•...• H. A general eqUilibrium is a vector of non-negative prices p*. measured in an abstract unit of account. such as. all individuals reach their target: they obtain their desired allocation subject to the budget constraint at p*. Except for free goods. all individual excess demands are zero. that is are fulfilIed. s In order to simplify the discussion we will accept the bad. although the current. method of treating separately the determination of eqUilibrium prices and exchanges. Although a bit tedious. it is essential here to distinguish three types of economies. 1. Let us call Wan economy composed of individuals and of commodities which verify all the conditions of existence of a general equilibrium as defined above. These conditions are twofold. HI: Conditions related to the determination of non-negative prices. such as. the aggregate excess demands are equal to zero. except for free goods. H2: Conditions related to the trading procedure by which all individual excess demands are nil. The conditions HI are well known. 6 Let us. however. clarify the conditions H2 which are central to monetary theory as well as to this discussion. In an economy such as W. the individual eqUilibrium allocation at prices p* is achieved without cost by means of a central clearing house. Each agent commits himself to delivering commodities which he supplies and to accepting the commodities which he demands. The value of each transaction is then credited and debited respectively to his account. This is the hypothesis which Debreu introduces at the very beginning of his book. Exchange is excluded from the individual activities considered by the theory. by substituting to it a procedure through which the economy can work 'without the help of a commodity acting as a medium of exchange·. 7 An important consequence is that H2 allows for the usual definition of general eqUilibrium in aggregate terms. that is: the vector of prices p* ~ 0 such that Zi(P*) = LhZhi(P*) ~ O. where Zi(P*) < 0 only if Pi = O. Given H2. this definition is identical to the general definition above. This feature is the source of a widespread ambiguity which considers that conditions HI are necessary and sufficient for the existence of 370 General Equilibrium with a Zero-Price for Money a general equilibrium of the W-type economy. As we have just seen, HI only allows for the determination of prices which equalize the aggregate supplies and the aggregate demands. These prices are general equilibrium prices only once H2 is added. 2. Let us call M the previous economy where H2 is suppressed and replaced by H'2: there exists a good (n + I), denoted 0, with no intrinsic or direct utility. If it is adopted as a medium of exchange, then the individual excess demands at prices p* will be fulfilled through bilateral trades. Contrary to the previous case, in an M-type economy, exchange is an integral part of individual economic activities; exchange is decentralized. The immediate consequence is that general equilibrium can no longer be defined in aggregate terms. We must still make sure, however, that the equilibrium price of money is positive. As we have already mentioned it, there is no satisfactory answer to this question as yet. 8 But it is not an essential point here. . 3. Let us call B the previous economy where H'2 is eliminated. There does not exist any commodity universally agreed upon which can be used as a medium of exchange. This implies that individual equilibrium cannot be reached by the use of a clearinghouse or of money. A B-type economy is clearly identified as a barter economy, namely the fiction, inspired by the 'state of nature' of the seventeenthand eighteenth-century political philosophy, which economists use to justify the existence of money. The strategy of demonstration is simple: start from an Mtype economy, suppress money and obtain a B-type economy where individual excess demands cannot generally be fulfilled, even at prices such that aggregate excess demands are nil. As a consequence, a general equilibrium does not exist in an economy only defined with respect to HI. Since Smith, such 'inconveniences of barter' are taken as the starting point of monetary theory. 3 THE NON-EXISTENCE OF GENERAL EQUILIBRIUM WITH A ZERO-PRICE FOR MONEY Let us recall what Wicksell considered as an obvious fact: 'money, evidently, only performs its function [as a medium of exchange and store of value] to the extent that it possesses exchange value'.9 In the modem theory of general equilibrium this means that a zero-price for money is always an equilibrium price. Indeed, no matter what the preferences and technology are, when the price of the fiat money is zero, its demand is zero. As a result, the excess demand for money is negative at Po = 0.10 Consider now an M-type economy. If Po = 0, it follows from Wicksell's 'obvious fact' that this economy then demonetizes itself, and behaves like a Btype economy where, in general, fulfilment of individual excess demands is precluded. It follows that equilibrium does not exist. More precisely, equilibrium can only exist if we assume the double coincidence of wants in all markets; or the Carlo Benetti 371 hypotheses underlying the overlapping generations models. These assumptions are very restrictive and are evidently absent from Patinkin's model. According to Hahn, the existence of an equilibrium with zero-price for money is the consequence of an 'ill-formulated monetary theory' which means a theory of a "monetary economy" loosely specified' .11 This explanation is, however, not satisfactory as we will show by examining its various aspects. Hahn is unsatisfied, and rightly so, with the integration of money in the theory of value through a 'theory in which the demand for money is simply derived by putting "real cash balances" into the utility function, not as a proxy for something else, but to account for conveniences and the like'. We recognize Patinkin's model. Relying on Wicksell's 'obvious fact', Hahn concludes that this model 'always has an equilibrium in which money plays no part' .12 In our opinion, Hahn's statement is fully justified in a very different context. In overlapping generations models the economy can settle in a non-monetary equilibrium. As Hahn points out, 'this is a highly indesirable result. When I first noticed this phenomenon some 15 years ago (Hahn, 1965), I argued that it arose from the fact that we gave money no work to do that could not equally well have been performed by some other asset. This view I still hold. '13 As we have already mentioned, in an M-type economy with one commodity, money and individuals who can only be distinguished by their date of birth, there exists an eqUilibrium with a zero-price for money. According to Hahn, if the functions of money are to be correctly stated, this would need, at least, a well-specified 'technology of exchange'. It is not necessary to develop this point; we can limit ourselves to the simplest case. Clower's constraint, while inadequate to explain the positive price of money, 'has the virtue that it gives money something to do, and thereby says something about the "technology" of exchange' .14 Patinkin's theory is, however, silent on this point. The question is the following: what conclusions can be drawn from these correct premises? To answer this question, let us compare the effect of a zero-price for money with and without Clower's constraint. In the first case, individual demands can no longer be financed. As a result they will fall to zero, and so will individual supplies. Equilibrium hence does not exist or, more precisely, for each commodity supply and demand are equal at the zero level, no matter what the prices of commodities are. IS In the second case, Patinkin's economy is reduced to a B-type economy. While it is possible that some bartering may occur, which justifies Clower's criticism of Patinkin, we have seen that, in general, individual excess demands cannot be fulfilled. In both cases, therefore, at a zero-price for money there cannot be any eqUilibrium. Let us proceed with our analysis by asking which conditions must be fulfilled for Hahn's proposition to hold. In Patinkin's model, if the price of money is positive, then, in certain conditions (shown by Hahn) a monetary eqUilibrium does exist: through monetary trade individual excess demands can be fulfilled. Assume now that on a given 'monday' the auctioneer calls a zero-price for money. Hahn's 372 General Equilibrium with a Zero-Price/or Money result holds if: first, the conditions for existence of prices that equilibrate aggregate demand and aggregate supply still hold,I6 second (which is the most important with respect to monetary theory), the clearing house is miraculously put in place on 'tuesday', allowing fulfilment of individual excess demands. 17 Hahn states that, in Patinkin's model, the equilibrium with a zero-price for money 'of course, is due entirely to the fact that a proper function for money has not been formulated, so that when money is worthless, transactions can be carried out as before' .18 The main condition for Hahn's result is then the reutilization of hypotheses H2 of the W-type economy that had to be eliminated in order to introduce money and study the consequence of its zero-price. This leads to an incoherence. The foregoing discussion leads one to believe that the problem of eqUilibrium with a zero-price for money has nothing to do with the quality of the monetary theory under discussion. Patinkin's 'ill-formulated monetary-economy' is none the less an M-type economy. Whether this economy is or not 'loosely specified', the clearing house is necessarily absent. In the opposite case the question of money cannot even be asked (see section 4). This statement is implicitely accepted by Hahn when he proves the existence (under very strong restrictions on Engel's curves) of an equilibrium in Patinkin's model, if it is assumed that the price of money is positive. If it is true (as in the orthodox approach) that a 'good' monetary theory must prove the existence of a monetary equilibrium, in no way does this imply that a 'bad' monetary theory gives the same result as a good theory of a W-type economy. We do not see why 'a paradox ... may arise from an ill-formulated monetary theory' (that is the presence of a non-monetary equilibrium).19 With respect to the equilibrium concept, monetary theory is not different from value theory. In both cases, the proof of the existence of an eqUilibrium (that is of an economically significant solution) is the test of the model's coherence. If the theory is ill-formulated, the test is negative. The first part of Hahn's conclusion with respect to his criticism of Patinkin's model should therefore be modified in the following way: in this model, there does not exist an equilibrium if the price of money is nil. This will be confirmed by studying Hahn's demonstration in the next section. The second part of Hahn's conclusion seems to us entirely correct: 'it is not at once clear how we could establish [that Patinkin's model] contains a solution with [a positive price of money)' .20 4 A CRITIQUE OF THE DEMONSTRATION OF THE EXISTENCE OF AN EQUILIBRIUM WITH A ZERO-PRICE OF MONEY Hahn's demonstration 21 can be divided into four steps. 1. First, he considers an economy with (n+ 1) commodities where there exists an eqUilibrium price vector p*. This is a W-type economy as defined above. Suppose now that one of the components of the aggregate excess demand Carlo Benetti 2. 3. 4. 373 vector, for instance the first one, has the following property, which will be called (m): Zo (0, PI, ... , Pn) ~ 0, for all price vectors (0, PI' ... , Po) ~ 0. Then there exists an equilibrium price vector (0, pi, ... , p~) ~ 0, such that its first component is nil. Hahn then describes a model which is 'in all respects similar to Patinkin's except that it as yet contains no money', where 'the solution p* gives the eqUilibrium prices for the current period'. 'Let us designate the good with the label as fiat money.' We have already seen that at Po = the demand for money is nil. Hence the function Zo (0, PI' ... , Pn) has the property (m). From this, Hahn concludes that Patinkin's model 'always contains a "nonmonetary" solution'. ° ° The ambiguity starts with the second step. What is a Patinkin's economy without money but which still allows for an equilibrium p*? We can see only one possible answer: a W-type economy where, owing to HI and H2, individual excess demands are satisfied at prices p*. In the third step, Hahn specifies that the commodity has the property (m). From this he concludes that, in this economy, there exists an equilibrium with a zero-price for this commodity. This is absolutely right, but it has nothing to do with the non-monetary eqUilibrium in Patinkin's model. Hahn only proves that if, in a W-type economy, we introduce a commodity whose utility depends on its price and if we set this price equal to zero, then the component of the eqUilibrium price vector which corresponds to this commodity will be zero. We see no link between this result, which is right, and the quite different statement by which an M-type economy always has an equilibrium with a zero-price for money. Proposition (4) does not follow the preceding ones and consequently is not proved. If Hahn does not succeed in proving that Patinkin's model has a non-monetary solution, what does he prove? To answer this question let us go back to (3) and ask what would happen if Po > 0. It cannot be an eqUilibrium price. Indeed, we know that in a W-type economy, transactions are executed without cost through a clearing house. The demand for a commodity which has no intrinsic utility and whose only use is exchange, is necessarily zero and Po = 0 is then the only equilibrium price. In this kind of economy, there does not exist an equilibrium with a positive price for money. At no step of his demonstration does Hahn consider an economy distinct from W. The result which he obtains is not the existence of a non-monetary eqUilibrium in a M-type economy, but the impossibility of a monetary eqUilibrium in a W-type economy. This is surely not surprising since an equilibrium with a zero-price for money is the only one permitted in a W-type economy where fiat money is absent and cannot be added to the initial list of commodities. Gale accepts Hahn's proposition which he interprets as a general property of models with fiat money, independently of the quality of the monetary theory which is used. 22 The critique of his presentation, which is particularly clear, ° 374 General Equilibrium with a Zero-Pricejor Money allows us to confirm the ambiguity of the notion of equilibrium with a zero-price for money. Let us consider the problem 'from the point of view of the equilibrium concept'. Gale's reasoning is as follows: (a) (b) 'Any monetary economy contains within it a non-monetary economy, namely, the one obtained by putting the price of money identically equal to zero'. 'Suppose the non-monetary economy possesses an equilibrium. Then this is also an equilibrium for the monetary economy since with the price of money equal to zero there will be no demand for it (more precisely, agents are indifferent to the quantity they hold), . In our opinion, these two statements are not compatible. According to the first statement, the 'non-monetary' economy is of the B-type which, in general, does not allow for an equilibrium. Statement (b) is therefore impossible: an equilibrium with a zero-price for money does not exist. According to the second statement, the 'non-monetary' economy is of the W-type where fiat money cannot be introduced. It thus follows that (a) is impossible: there cannot exist a 'monetary' economy which contains a W-type economy. The difficulty of Patinkin's model, and more generally, of general eqUilibrium models with fiat money, is not the presence of an eqUilibrium with a zero-price for money, that is to say a state of coherence which is obtained outside money. Rather, it deals with the absence of equilibria and therefore of coherence. The reason for this result is that, even if we accept the notion of money as a particular commodity, an economy with money cannot be conceived as an extension of an economy without money. These two economies are heterogeneous and therefore no equilibrium in the second economy can be contained in the first one. Notes This chapter is a development of a section of my book Moneda y leorfa del valor, U.A.M. and F.C.E., Mexico, 1990, and has benefited from discussions with J. Cartelier, G. Deleplace and A. Rebeyrol. 1. 2. 3. 4. For instance, see Wallace (1980: 77) who affirms that overlapping generations models 'do successfully integrate value and monetary theory'. A non-exhaustive list will be given later (see note 8). And not of Hahn's works in monetary theory. We fully agree with some of his ideas, for instance, when he states that 'the challenge of monetary theory is not the neutrality theorem or related results; it is the required reconstruction of our paradigm if we are to make sense of money' (1973: 172). In the first chapters of his book, Patinkin's model, which is criticized by Hahn, deals with this type of economy: 'there is a fiat money, which serves as the actual, physical medium of exchange and store of value. It is assumed to be an "outside money" ... It is only of this money that balances can be held'. 'Inside money' is Carlo Benetti 5. 6. 7. 8. 9. 10. II. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 375 excluded and only introduced later in the second part of the book devoted to macroeconomics. See Patinkin, 1965: 15. Although it is not standard, this definition comesponds to the one which is generally accepted. It has the advantage of explicitly eliminating the nonsense of a definition of equilibrium which would not allow the possibility for individuals to obtain the equilibrium allocation. These hypotheses are recalled in Hahn, 1965: 148. See Debreu (1959). Hypothesis H2 may also be interpreted as a centralized system of trading 'chains' by which barter exchanges can be completed (see Ostroy and Starr, 1974). This procedure is assumed to be as costless as the central clearinghouse. A negative result is obtained: no general equilibrium can exist in a finite-horizon M-type economy, except if we allow the ad hoc hypothesis of tax payments in money. In a finite-horizon M-type economy, the existence of equilibrium depends either on conditions relative to the elasticity of the expectations on the future price of money (monetary temporary equilibrium), or on the conditions on initial endowments and preferences (monetary stationary equilibrium in the basic overlapping generations model). Other possible stratagems in order to obtain a short period equilibrium in an M-type economy can be found in Hahn, 1975: 35-6. Wicksell, 1926: 142. This property plays an important role in the demonstration of Hahn's proposition. See below, section 4 and Hahn, 1965: 150. See Arrow and Hahn, 1971: 361-2. Ibid., p. 362. Hahn, 1982: 11. Ibid., p. 21. This 'equilibrium' has therefore no significance other than a formal one. Which is not guaranteed. In particular, the condition for continuity is not satisfied when the price of money falls to zero. See Hahn, 1971: 100 and Starr, 1974: 48-9. Except if another commodity replaces money as a medium of exchange. In this case, equilibrium, if it exists, is monetary. F. Hahn admits it: 'the zero exchange value of money equilibria are pseudo-constructs: they are equilibria of a system based on the implicit assumption that the process of exchange proceeds as it does when aided by the device of money' (see F. Hahn, 1982: 20). But he does not seem to see that this 'implicit hypothesis' (in fact H2) is incompatible with the result he obtains in his 1965 paper. Arrow and Hahn, 1971: 362. Ibid., p. 361. Hahn, 1965: 148. Ibid., pp. 149-50. Gale, 1982: 290-1. See also Starr's recent observation: 'A model of fiat money may be expected ... to include an equilibrium where price of money is nil' (R. Starr, 1989: 293. References Arrow, K. and F. Hahn (1971), General Competitive Analysis (San Francisco: HoldenDay). Debreu, G. (1959), Theory o/Value (New York: Wiley). Gale, D. (1982), Money in Equilibrium (Cambridge: Cambridge University Press). 376 General Equilibrium with a Zero-Price for Money Hahn, F. (1965), 'On Some Problems of Proving the Existence of Equilibrium in a Monetary Economy', in F. Hahn, Equilibrium and Macroeconomics (Oxford: Basil Blackwell,1984). Hahn, F. (1971), 'Equilibrium with Transaction Costs', in F. Hahn, Money, Growth and Stability (Cambridge MA: MIT Press, 1985). Hahn, F. (1973), 'On the Foundations of Monetary Theory', in F. Hahn, Equilibrium and Macroeconomics (Oxford: Basil Blackwell, 1984). Hahn, F. (1975), 'Money and General Equilibrium', in F. Hahn, Money, Growth and Stability (Cambridge MA: MIT Press, 1985). Hahn, F. (1982), Money and Inflation (Oxford: Basil Blackwell). Ostroy, J. M. and R. M. Starr (1974). 'Money and the Decentralization of Exchange', Econometrica, 42(6). Patinkin, D. (1965), Money, Interest and Prices, 2nd edn (New York: Harper & Row). Starr, R. (1974), 'The Price of Money in a Pure Exchange Monetary Economy with Taxation', Econometrica, 42. Starr, R. (ed.) (1989), General Equilibrium Models of Monetary Economies (Boston: Academic Press). Wallace, N. (1980), 'The Overlapping Generations Models of Fiat Money', in J. Kareken and N. Wallace (eds.), Models of Monetary Economies (Minneapolis: Federal Reserve Bank of Minneapolis). Wicksell, K. (1926), Lectures on Political Economy, vol. II, Money (New York: Augustus M. Kelley, repro 1978) 13 Basic Choices in Keynesian Models of Credit Gary A. Dymski 1 INTRODUCTION Borges once described a Library of interconnected rooms, each containing a fixed number of books on whose pages appeared randomly placed spaces, punctuation marks, numbers and letters. There were as many rooms in the Library as there were combinations of these elements - a finite number, but a maddeningly large one. The lonely denizens of this place wandered from room to room, seeking books whose pages conveyed meaning. The decade of the 1980s offered monetary economists an experience in disorientation akin to being lost in Borges' Library. In the United States and the United Kingdom, the 1980s opened in the midst of sustained contractions in monetary policy, which threw these nations' banking systems into disarray. The historically unusual combination of sustained high real rates of interest and economic expansion had diverse effects in the center and the periphery. In the integrated financial markets of the advanced economies, new instruments were created, sold, and traded by exploiting the market potential of previously unmarketed assets. In the isolated economies of the periphery, a debt crisis broke out and lingered despite the medicines of free-market doctors. Two stock market crashes dampened Wall Street's hothouse atmosphere of speculation and fee-taking, but caused no recession. Entire industries - both the new (junk bonds) and the old (thrifts) - were crippled by the pace of change, which made the magnitude of risks impossible to judge for both the taker and the taken. The US recession, when it came, was not preceded by monetary tightening in the face of inflation, as had all previous postwar peaks. Instead, the Federal Reserve's efforts to stimulate bank credit supply failed, as institutions with heavy loan-loss reserves and damaged capitaVasset ratios pulled back. This rich menu of financial events renewed the interest of mainstream macroeconomists in financial questions. What began as a search for missing money in Goldfeld money-demand equations led on to models showing how market failures between lenders and borrowers could affect the aggregate level of economic activity. The 'New Keynesians' have drafted these credit-market theories for their war with the new Classical paradigm of Lucas et al. But the new Keynesians 377 378 Basic Choices in Keynesian Models of Credit were not the first Keynesians to imagine that there was more to monetary economics than the quantity equation. At least two decades earlier, Davidson, Joan Robinson and Shackle, among others, had begun to explore the profound implications of fundamental or Keynesian uncertainty for monetary analyses. The term 'Keynesian' here recognizes Keynes's analysis of fundamental uncertainty in chapter 12 of his General Theory. Building on this foundation, Hyman Minsky developed his financial instability hypothesis. According to this hypothesis, increasing financial volatility and debt overaccumulation will occur endogenously as an expansion proceeds; at some limit point, these trends will result in an outbreak of welfare-reducing financial instability, even given the presence of a lender of last resort. These Post Keynesian views were first framed as a critique of the financial conception embodied in the neo-Classical synthesis model. It ignored credit in favor of money, but its analysis of money was too simplistic, because it ignored the impact of Keynesian uncertainty and assumed the money supply could be exogenously controlled by the monetary authority. The rise of New Keynesian analysis raises anew the question of whether this new set of models is compatible with the Post Keynesian framework. Both New Keynesian and Post Keynesian models proceed by postulating some important deviation from the complete-information assumption of Walrasian theory. New Keynesians have explored how informational asymmetries may induce credit rationing, price stickiness and unemployment; Post Keynesians have explored how real time and 'Keynesian uncertainty' undermine agents' ability to make longer term plans with confidence. Asymmetric information and Keynesian uncertainty represent conceptually distinct deviations from Walrasian general equilibrium: one does not imply the other (Dymski, 1994). The intersection of these two perspectives has been largely unexplored. The absence of a synthetic New/Post Keynesian model is attributable to two factors: (i) the insistence of New Keynesians on microfounded explanations; (ii) the analytical complexity inherent in relaxing Walrasian general equilibrium in two ways simultaneously. New Keynesian microfoundational models are equilibrium based, and center on the analysis of either stable equilibria (for example, adverse selection models of the bank credit market) or adjustments to adverse shocks (for example, Bemanke's (1983) article about how price deflation raised the cost of intermediation). Conversely, Post Keynesian macro models have rejected equilibrium in favor of the idea that financial instability builds up endogenously past the point of crisis. Can New and Post Keynesian insights be combined in synthetic economic models?1 This chapter argues that they can and lays some groundwork for such models. The pages that follow examine the implications of some crucial building blocks in macroeconomic theories encompassing money and credit: first, the assumptions of whether the outcomes of time-using processes are predictable in a probabilistic sense; second, whether an important fraction of real activity is finance constrained; third, whether knowledge about agents' capacities and intentions is asymmetrically distributed. The assumptions made in these three areas are Gary A. Dymski 379 termed 'macroeconomic' because they directly affect the behavior of every agent in a model, bank or non-bank. We also investigate several 'microeconomic' aspects of financial behaviour - attitudes toward risk, the degree of competition in intermediary markets, and the bearing of risks. So instead of building a model per se, this chapter separates the elements of macroeconomic banking theories into their atoms for closer examination. The purpose of the exercise is to understand the analytical importance of each of these elements individually. It is everywhere true in social science that conclusions are driven by assumptions. The area of money and credit is a particularly interesting case. It turns out that the link between money, credit, financial structure and ('real') economic outcomes depends on the theoretical treatment of finance constraints, uncertainty, and asymmetric information. It is shown below that whether money and credit 'matter' depends on whether they perform essential economic functions; and this in tum depends on whether finance constraints or uncertainty (especially of the Keynesian type) are present. 2 The presence of uncertainty and finance constraints, in tum, depend on the primitive assumptions made about information, time, and market coordination. If information is complete and perfect, and if economic processes are timeless and market transactions precoordinated, then money and credit can only be introduced artificially; it is then only reasonable to conclude they are redundant in this macro setting. It is also shown that whether intermediary structure matters depends first on whether finance constraints exist, and then on whether financial intermediaries have any special ability to counteract the effects of asymmetric information with information gathering or monitoring. Turning to the micro analytics that underly macro assertions, it is shown first that different macro treatments of money and credit have profound implications for the nature of interactions at the micro level. Shifts in macro environments can reduce the scope for certain types of micro behaviors, and can lead to micro/macro disequilibria wherein micro agents are maladapted to their macro environments. 2 MACRO ASSUMPTIONS; FINANCE CONSTRAINTS, INFORMATION, UNCERTAINTY This section considers the implications for conceptualizing money, credit and intermediation of three root characteristics of economic models: (i) the predictabilityof time-using processes; (ii) the need to procure means of exchange in advance of making expenditures; and (iii) the dispersion of information about agent intentions and capacities. Section 2.1 first reviews the benchmark case of Walrasian general eqUilibrium, in which money, credit and financial intermediation are all redundant. Section 2.2 discusses the implications of introducing uncertainty or finance constraints and uncertainty (items (i) and (ii) above), either singly or in tandem. Section 2.3 considers two combinations of these elements, 380 Basic Choices in Keynesian Models of Credit termed Schumpeterian and financially fragile macro environments. Section 2.4 then explores asymmetric information (item (iii) above), and especially its complementarity with Keynesian uncertainty. 2.1 Walrasian General Equilibrium: a benchmark Case As a reference point for discussion, consider the Walrasian general equilibrium (WOE) with complete markets. Agents in this fictitious world maximize utility by using their resources to engage in consumption (the pure exchange case), or production and then consumption. All actions are pre-coordinated through the price mechanism: an equilibrium price vector insures that demand equals supply for each good. Money plays no important role here. All resources and goods can be shifted as required among agents, at a zero discount to equilibrium prices, until an optimum is achieved. So all goods are equally useful as stores of value. Further, since goods exchange against goods, money is not required as a specialized means of exchange. A nominal quantity of no inherent worth, it plays only the trivial role of numeraire. Foley writes of WOE models. that: 'We can see why it becomes very awkward to reintroduce money into these models when it is, in a sense, already there but has been removed by abstraction from the start' (1986; p. 20). 3 Credit may be present in WOE if it is optimal for agents to buy and sell endowments in advance of production. Since the output volume of production processes is certain, and the prices of that output known ex ante, no risk of default is entailed. Credit of this kind is self-liquidating; its nearest real-world analogue is trade credit that is fully collateralized with safe assets. Further, if credit does exist, it is simply part of the precoordination mechanism - the exchange of present services for future goods, at zero risk. Another social device for transferring goods across time could be imagined - so while a 'credit market' exists, it is inessential. Money and credit are inessential in WOE because of this framework's strong assumptions about market precoordination, time and information. Agents' plans are fully coordinated before market transactions ever occur, so actual economic behaviors simply fulfill ex ante plans. Time is unimportant because anything with economic value is assigned a market; in particular, there is a complete set of ex ante futures markets, so all events are effectively simultaneous. The model is timeless. Information is perfect: agents have complete knowledge of their own preferences and resources and, through the price vector, perfect knowledge of all other agents' aggregated preferences and resources.4 Nature - that is, all exogenous natural or social behavioral forces - is benign, and undoes no agent's plans. There is no role for financial intermediaries to play, for agents need no money and no credit - they can already buy and sell whatever they want (consistent with the relevant solvency constraints governing their outlays), at each point in time and across time. Gary A. Dymski 2.2 381 Uncertainty, Finance Constraints and Macro Outcomes If the Walrasian assumptions of timelessness, perfect information and precoordination are relaxed, money and/or credit take on essential economic functions. Suppose first that agents' information set is imperfect because some outcomes are subject to what Pesaran (1987) terms exogenous uncertainty. That is, timeusing activities exist, wherein ex ante commitments and expenditures at some early stage necessarily precede ex post output at a later stage in time. Anticipated and realized outcomes can differ due to shocks whose source is outside the realm of explanation. The existence of slippage between anticipation and realization implies a failure in market precoordination: ex ante markets cannot take any given agent at any point x to any later point y. This market failure would have no impact if there existed, at the same time, a complete set of futures markets for each future outcome of every stochastic process. But suppose such an extensive set of futures markets does not exist. In this situation, market precoordination does not suffice to convert time-using processes back into timeless marketcoordination problems, and analytical roles for both money and credit may arise. Exogenous uncertainty creates a need for a store of value: agents' willingness to undertake time-using activities may entail a desire simply to hold resources in reserve. This desire was termed liquidity preference by Keynes. A variety of assets might serve this role of storing value. Money is an immediate candidate for this role, since it is purchasing power. Its use in this role, however, depends in part on whether more attractive stores of value are available. For example, in periods of high price inflation, non-monies may be better stores of value. In any event, the rate of price inflation, money and asset-holding are linked together when outcomes are stochastic. How important this role for money is depends on how likely agents are to shift between time-using activities and stores of value. This, in turn, turns on how much is known about the properties of stochastic outcomes. At one extreme, the range and relative likelihood of each possible realization may be well understood. In this situation of risk, the probability calculus is readily used to discount the future, and agents are unlikely to make large or frequent shifts in asset holdings. But if instead realizations cannot be bounded meaningfully through analyzing prior experience, and are instead broadly indeterminate, as described by Keynes in chapter 12 of his General Theory, then this uncertainty is not simply exo genous but Keynesian - that is, fundamental. In this event, agents must rely on estimates of future outcomes which may shift rapidly, and in which they may have varying degrees of confidence. In a Keynesian uncertain environment, large gaps between ex ante anticipations and ex post realizations are much more likely. This implies an analytical landscape in which financial stock/flow imbalances will occur. The reasoning behind this conclusion is straightforward. In an economy with external financing, 382 Basic Choices in Keynesian Models of Credit many time-using real processes leave a financial 'trail' - a stream of time-dated or contingent repayment commitments that must be met regardless of what happens to the assets this financing supports. If a bad realization is had on these assets, then the unit that has emitted these liabilities will have a larger stock of debt than is appropriate given the (new ex post) value of her income stream. In this context, agent behaviour will necessarily be more sensitive to the possibility of stock/flow imbalances, and stores of value - reserves of value held in the event of such imbalances - will be of crucial importance. This includes, especially, the store of value aspect of money. In addition to this non-redundant role for money as store of value, a further failure in the coordination of markets over time may give rise to a non-redundant role for credit. Suppose agents can exchange goods only for money: the promise of receipt of money later in times does not substitute for the ex ante possession of money. Suppose further that agents' ideal expenditure volume ex ante does not necessarily equal their ex ante money holdings. Given any set of ex ante and ex post prices, some units would prefer to spend less money ex ante than they initially hold, so as to spend more ex post. These are termed, following Gurley and Shaw (1960), deficit units; units with the opposite characteristics are termed surplus units. Expenditures by deficit units - and hence the level of ex ante and ex post market exchanges - depends on whether they can secure credit or external financing from surplus units. s So credit precedes production (and exchange): it is the social mechanism whereby an economy overcomes finance constraints on its aggregate activity level. The question is whether all units obtain all the (warranted) ex ante credit supply they need. If so, then credit is a completely passive aspect of the economy: inadequate purchasing power never prevents units from undertaking activities. But if not, then some units seeking to undertake time-phased activities will be finance constrained - the volume of their planned expenditures will exceed their available financing. Binding finance constraints reflect a coordination failure because ex ante deficit units' demand for purchasing power exceeds the supply which ex ante surplus units possess. 6 2.3 Schumpeterian and Financially Fragile Macro Environments Every theorist designing a macro environment must choose some combination of the elements discussed above: uncertainty or certainty; finance constraint or precoordinated financing. As discussed in section 2.1, the combination (certaintyI precoordinated financing) appears in Walrasian general eqUilibrium. We might note in passing that the combination (exogenous uncertainty/precoordinated financing) underlies the capital asset pricing model of finance theory (see Fama, 1976). The remaining two possibilities are of more interest. Gary A. Dymski 383 A Schumpeterian Macro Environment The combination (certainty/financing constraint) might be termed a 'Schumpeterian' macro environment after Schum peter (1910). In this environment, the importance of liquidity preference is de-emphasized in favor of careful attention to credit constraints on economic development. Schumpeter asserts that economic equilibrium is a degenerate state, from which an economy could escape only if it transferred capital to those capable of achieving technical advances and hence of sustaining growth - that is, to 'entrepreneurs'. This transfer of capital is accomplished through the credit market, which puts purchasing power into the hands of (deficit-unit) entrepreneurs. In the absence of such credit flows, technical innovations are blocked. The focus then is on how credit flows augment the pace of growth; money is merely the purchasing-power face of credit. There is little attention to the store-of-value aspect of money that is associated with exogenous uncertainty. Money is simply a means of exchange (MME) - units need money to make expenditures, and ex ante deficit units need credit to obtain money. Credit alone is essential; the role of money is derivative. This type of macro environment is of special interest because it has been emphasized in recent on theories of 'money in motion'. In particular, several French economists have been associated with the 'Regulationist' approach, which combines a 'circuitist' conception of financial relations with a conflict theory of class relations. The appendix to this chapter argues that 'Regulationist' writings, including some important papers penned by Duncan Foley, assume the existence of what is termed here a Schumpeterian macro environment. A Financially Fragile Macro Environment The fourth and final combination is (exogenous uncertainty/financing constraints). This combination forms the basis for rich analyses of money and credit - for example, it underlies both Post Keynesian models of financial fragility and New Keynesian models of credit rationing. Note that if both Keynesian uncertainty and finance constraints are present, both credit and money perform essential roles - the former as a means of making expenditures, the latter as a means of storing value in the face of exogenous uncertainty. Agents in this environment will feel tension between its money and credit dimensions: the volume of liquidity available to finance the activities of deficit units varies inversely with surplus units' demand for money as a store of value. Agents operating in this environment need 'cash in advance' to undertake investment or other time-using projects, without being certain that those projects' realized outcomes will accord with those anticipated. For agents who finance activities from their own reserves, this environment is no different from that with no financial constraints. But for agents who finance their activities externally the situation is altered, because any debt they take on is not self-liquidating. A bad 384 Basic Choices in Keynesian Models of Credit realization on an externally financed project may result in losses for both borrower and creditor. Creditors have ways to mitigate their exposure in particular, by having borrowers pledge collateral. The presence or absence of collateral then becomes a new barrier which may cause some agents to be more tightly financeconstrained than others.7 The value of collateral may itself, of course, vary with market conditions and other factors. This scenario of external financing without pregiven returns sets up an ongoing financing dimension for deficit units: these units must meet their financial obligations as a precondition of achieving their other goals. Balance sheet relations become as important as real activities in the analysis of economic agents. This is the context within which Minsky has developed his 'Wall Street paradigm'. Minsky has, of course, proposed in his 'financial instability' hypothesis that the conditions required for economic units to make and sell real-asset and financial positions smoothly will be eroded endogenously over time. In proposing this hypothesis, it should be noted that Minsky assumes that exogenous uncertainty is Keynesian, not risk-like: agents' financial and real positions can become disorganized because they are operating in stochastic worlds without parameters, in which the gaps between anticipations and realizations can be large and persistent. 2.4 Asymmetric Information in the Macro Environment Until now, we have deviated from the informational assumption of Walrasian general equilibrium by assuming the existence of exogenous uncertainty. But information can be non-Walrasian in a second sense. If each agent in the model does not costlessly know everything that might affect every other agent's economic behavior, directly or indirectly, then the information set is incomplete. Incomplete information creates (at least) two classes. of agents, the informed and uninformed; the latter are unaware of either the capacities or intentions of the former. This is the case of asymmetric information. Pesaran (1987) observes that asymmetric information implies endogenous uncertainty. That is, in the presence of asymmetric information about some agents' capacities or intentions to perform under contract, uninformed agents cannot rely on the price vector (alone) to differentiate goods by quality (Stiglitz, 1987). With endogenous uncertainty, the price vector will no longer summarize all economically relevant information; such as economy has multiple equilibria, because there are fewer markets than there are economically important behaviors.8 This problem of asymmetrically informed agents only matters in a macro environment wherein at least some agents are finance-constrained. Indeed, New Keynesian analysis is based on interpreting asymmetric information models as microfoundations for macro settings wherein finance constraints bind on some units. When all pertinent information concerning the creditworthiness of financeconstrained units is not costlessly observable, surplus units' decision to provide external financing takes on an allocational dimension. The possession of 'private Gary A. Dymski 385 information' about their creditworthiness by deficit units (potential borrowers) gives these units an incentive to cheat; in this context, market signals may function perversely. In such an economy, the flow of credit - and consequently, the level of economic activity - depends on whether social devices are available for restricting the amount of cheating by borrowers. Then surplus agents will extend credit to the extent that they can reduce their endogenous uncertainty about borrowers through monitoring, extensive use of short-term renewable contracts, credit rationing, and so on. Usually, the 'social devices' in question are assumed to be financial intermediaries functioning as information specialists or costly monitors: that is, asymmetric information among borrowers and lenders creates an analytical role for financial intermediation. The equilibria that financial intermediaries achieve under these circumstances often entail credit rationing and 'sticky' loan rates; these phenomena exemplify, respectively, the quantity constraints and price stickiness associated with Keynesian macro models. The Intersection of Exogenous and Endogenous Uncertainty The existence of two types of uncertainty in a model adds to its realism. At the same time, some caution in combining exogenous and endogenous uncertainty is warranted. Asymmetric information models often postulate the existence of risk that is, of one or more stochastic variables whose behavior is well described by stable probabilities over well-defined ranges. In fact, the presence of risk often explains why asymmetric information arises. For example, a lender will have difficulty sorting good from bad borrowers when their default can be due to their own inadequate effort or to a bad draw from nature. It is often postulated that lenders know the probabilistic characteristics of an exogenously uncertain event, and use this knowledge to find an equilibrium point. But agents can only make effective use of probabilistic information if that information is based in stable parameters that underlie stochastic fluctuation. If this is not the case - if Keynesian uncertainty arises, instead of risk, in a non-parametric stochastic world - then assertions that parameters can be known are empty. The presence of Keynesian uncertainty would require at least a reinterpretation of the status of any agent's (or principal's) 'knowledge' about stochastic events, and might also require some specification of how estimates of future variables are constructed.9 3 MICRO ASSUMPTIONS: MARKET STRUCTURE, RISK-TAKING, AND RISK-BEARING Thus far, we have focused on those common characteristics of the economic environment which all agents must confront. But in modelling money, credit and financial intermediation, microeconomic aspects of financial behavior may be important because they interact with macro conditions. This discussion singles 386 Basic Choices in Keynesian Models of Credit out three micro aspects: the degree of competition in money and credit markets; the attitudes toward risk of financial intermediaries and of other agents; and the closure rule under which financial intermediaries operate. 10 11te two overarching questions vis-a-vis each micro aspect are these: how is this micro aspect affected by the macro assumptions that are made? and how are macro analyses affected by micro-level dynamics in the markets for money and credit? This discussion cannot hope to be complete and strives only to be suggestive. Section 3.1 introduces intermediaries' attitude(s) toward risk, with an emphasis on whether intermediaries are operating under risk or under Keynesian uncertainty. Section 3.2 discusses how the degree of competition in intermediary markets affects intermediaries' credit-supply behavior; section 3.3 then investigates the relationship between competitive structure and risk bearing. Section 3.4 sums up this discussion by exploring the possibility of micro/macro disequilibrium. 3.1 Intermediaries' Attitude Toward Risk and Exogenous Uncertainty This section investigates how the nature of exogenous uncertainty affects intermediaries' expectation formation process and their attitudes toward risk. Attention centers on the difference between a merely risky environment and a Keynesian uncertain one. The key difference is that in a Keynesian uncertain environment, micro behaviors become much less determinate ex ante than in a risky environment - the possibility of feedbacks from unstable micro behaviors to macro outcomes and conditions arises. Expectations and Risk Aversion Amidst Risk Agents' expectations are a bridge between the outcomes that may happen and the outcomes that agents can imagine happening. Agents placed in a risky environment know the set of possible outcomes for stochastic processes, and they know or can learn the probabilities associated with these outcomes. Enough is known about the stochastic environment that it can be discounted - that is, treated by decision-makers therein as akin to a certain environment. Gaps between ex ante anticipations and ex post realizations are usually small; but in any event, these realizations do not shake the intermediary's faith in its ability to comprehend the nature of its stochastic environment. Given some well-defined sets of parameters for stochastic outcomes, the intermediary's decision problem is simple: it must choose some combination of asset instruments (and possibly of liability instruments) which will maximize its expected utility. If the intermediary is risk-neutral, it completely ignores the variance associated with stochastic returns, and makes its portfolio decision based exclusively on the basis of expected values per se. In this event, the entire effect of exogenous uncertainty on this ex ante decision problem is washed away. All possible ex post outcomes are accorded equal status: each appears as a certain Gary A. Dymski 387 outcome, with some appropriate weight (the cumulative probability of that outcome) attached. If the intermediary is risk-averse, exogenous uncertainty has some effect - the intermediary will hold proportionately less of any given instrument, ceteris paribus, as the variance of its return increases. In sum, the behavior of intermediaries (or, indeed, of other agents and firms) in a merely risky environment is continuous and driven entirely by parameters that obtain for money and credit markets as a whole. Expectations and Risk Aversion Amidst Keynesian Uncertainty Replacing risk with Keynesian uncertainty, of course means shifting into a world in which parameters do not exist for stochastic outcomes. This raises two questions: first, do agents view themselves as operating in a world without parameters? and second, if so, how does this affect their behavior? We consider first the general case of any agent, and then the specific case of financial intermediaries perse. A variety of responses has been suggested for the first question; see Vickers (1987) and Davidson (1991). My own response is that agents may not understand the world in this way. At one extreme, agents may construct subjective probabilities over a finite range of anticipated outcomes (which might even include a disastrous outcome, like a bank run). In a more turbulent setting, the willingness to believe that future outcomes can be so readily weighed and discounted may be shaken; so a method like 'potential surprise' may be substituted, with attention centered on a smaller set of salient outcomes (Vickers 1987, ch. 12). In a yet more turbulent world, expectations about future outcomes may collapse onto points, instead of ranges, and become more wish-like or wilful than scientific. In this case, the stop-go decision may rest almost purely on unconstrained animal spirits (Davidson, 1991). Economic agents may shift among these methods of dealing with uncertainty. During periods in which realizations closely track anticipations, and fall into a relatively narrow range, decisions will more likely involve subjective-probability calculations; as realizations and anticipations diverge, and vary more widely, such calculations will lose credibility and be replaced by animal spirits. In effect, as outcome turbulence increases, entities will place less trust in the past as a guide to the future. This might also cause them to change their degree of 'risk aversion', and to become more wary of irreversible commitments that increase the threat of illiquidity. Now, what are the implications of a Keynesian uncertain environment for a conception of financial intermediary behavior? Intermediaries face stochastic outcomes on both sides of their balance sheets. In the context of Keynesian uncertainty, intermediaries function is to assess and absorb default and liquidity risks in financial markets. How much asset- or liability-side risk they are willing to absorb depends on how they form expectations about stochastic outcomes, on 388 Basic Choices in Keynesian Models of Credit how much they trust those expectations, and how willing they are to bet on less likely events occurring. Intermediary behavior can be viewed as predictable only if outcomes and anticipations closely match one another for some period of time. But if instead these firms operate in a turbulent period, their behavior may be less predictable as they alter their methods of assessing risks and alter the composition of their portfolios. Clearly, in this environment, micro intermediary behaviors per se may have important effects on aggregate economic activity. Specifically, the overall level of credit flows (and hence of economic activity) will increase in so far as banks are willing to create credit by increasing their own exposure to default and liquidity risks. If banks shed these risks, then the volume (and types) of credit will resemble more closely that in a world without financial intermediaries. There may be important feedback effects from micro behaviors to macro conditions. In sum, the presence of Keynesian uncertainty makes it more crucial to examine micro behaviors, even as it makes micro analysis more difficult by removing analytical props that allow the discounting of exogenous uncertainty. 3.2 Competitive Pressure and Intermediaries' External Financing Roles We now tum to the question of how the degree of competitive pressure affects macro analyses of money, credit and intermediation. At one level, the impact of the degree of competition on outcomes in money and credit markets is straightforward: the degree of monopoly in these markets will affect the quantity and prices of monetary and credit instruments. But this subsection and the next take up two more subtle implications of the degree of competitive pressure in money and credit markets. The focus of our attention in this subsection is on how the degree of competitive pressure in asset and liability markets affects the creditsupply behavior of financial intermediaries. It is convenient to proceed by constructing two polar examples of perfectly competitive and monopolistic banking markets. Consider first a monopoly bank in a market area composed of both large companies with publicly traded equities and small businesses that must borrow from banks to obtain external financing. The monopoly status of this bank means it can set prices in both its loan and its deposit markets. It thus prices its deposit instruments so as to extract maximum net cash flow, buying the remainder of its funds in an interbank market. On the asset side, this intermediary earns money in two ways: (i) it sells loans in a monopolistic market, setting its own loan rate and rationing credit to borrowers; (ii) it packages and holds securities issued by large companies. This bank's lending to small firms is well described by the scenario of asymmetric information: these firms face extremely bad survival odds; they are so small that they cannot sell paper, and thus must borrow; and the bank must retain any loans made to them, since these are idiosyncratic and cannot be standardized into securities. Any loans this bank makes to small firms must remain on its Gary A. Dymski 389 balance sheet; there is no secondary market for them. By contrast, asymmetric information is irrelevant to its relationship with large companies. Indeed, the bank's status as an information specialist is itself largely irrelevant in this arena: other firms will step in to provide security-related services for large companies, if it does not. Competition keeps fees low, so that large volumes are required to generate the same net cash flow that would have followed from far fewer small business loans. This contrast is of interest because it suggests that the relevance of asymmetric information as a macroeconomic conception is proprotional to the amount of lending done to small private firms, compared to the amount of securities and loans floated by large, publicly traded companies. Increased competitive pressure on either side of the balance sheet reduces the volume of loans to small businesses. A shift of depository funds out of banks and into securities firms will have the same effect. Indeed, banks losing their deposit bases to securities firms might attempt to regain profits by dramatically increasing their fee-based, offbalance-sheet commitments. Changing Degree of Competition and Intermediaries' Franchise This analysis of competitive structure has been presented in static terms; but its dynamic aspects may also be of interest. In particular, note that one impact of deregulating an intermediary sector - as has been occurring in the US over the past decade - is, arguably, that intermediary markets shift from being monopolistic (or at least imperfectly competitive) toward being perfectly competitive. When exposed to outside competition, intermediaries that want to buy short-term liabilities can do so only if they offer a small premium over the 'going' rate. One key question, of course, is whether this shift toward a greater degree of price competition alters two of the key macro elements discussed in this chapter the importance of finance constraints and asymmetric information. If these elements (and especially asymmetric information) are unimportant, then reduced monopoly power in intermediary markets should simply means that more credit is extended at more advantages loan rates (for the borrower). Recall that in the absence of asymmetric information, there is no clearly defined or unique role for financial intermediaries to play in the macro environment. So intermediaries' loss of monopoly power causes no problems - indeed, new intermediaries may take business from old ones. The only 'franchise' of the old intermediaries was, indeed, their monopoly power over local banking markets; and this franchise was only a means of earning economic rents. A different conclusion might be reached if asymmetric information and finance constraints are adjudged to be quantitatively important in credit relations. For then, the loss of pricing power - particularly in the credit market - implies that intermediaries can no longer respond to asymmetric information by setting some appropriate loan rate and rationing credit. It might be more sensible for interme- 390 Basic Choices in Keynesian Models of Credit diaries to withdraw entirely from credit relations with borrowers whose creditworthiness is suspect. In this event, a transition from imperfect to perfect competition may well imply that financial intermediaries will lose their 'franchise' their economic role as information specialists. Credit flows to borrowers who lack direct access to credit markets may be eroded. Deregulation may well then create a larger pool of credit-rationed units, unable to access financial markets directly and unable to find intermediaries willing to lend to them. 3.3 Competitive Structure and Risk Bearing Exogenous uncertainty, of either the risk or Keynesian uncertain variety, drives a wedge between ex ante anticipations and ex post realizations. Ex post outcomes may yield either windfalls or shortfalls. Even wher. ex ante and ex post outcomes can be reliably equated via statistical approximations (the case of risk), ex ante/ex post wedges can occur for every individual draw. This wedge of uncertain size constitutes a risk inherent in the initiation of this process, a risk that some agent must bear. The possibility of ex ante/ex post deviations requires that the risks of such deviations be taken and, subsequently, borne. In general, the owner of an economic enterprise 'takes' this risk - a bad draw will reduce the net worth of that enterprise, while a good one will enhance it. The rationale for owners to bear this risk to straightforward: this risk is attributable to the firm's decisions (above, about whether to sell abroad), which is at the same time independent of social relations within the firm. But risk bearing is another matter. In general, when any firm's input and output markets are imperfectly competitive, that firm's agents may bear a portion of the risks taken by its principals. For example, suppose a firm faces an exogenously uncertain output price, and hence exogenously uncertain net profits. If that firm is a price-maker in its labor market, then a portion of the firm's risk of reduced income flow is passed along to its agents (workers) in the form of lower wages and fewer hirings than if the firm's output price were determinate ex ante (beyond the effects associated with imperfect competition per se). In general, the risk of loss any principal bears varies inversely with firm market power. There is one important distinction between risk bearing by principals and by agents. When borne by agents, these costs of fundamental uncertainty are absorbed ex ante; when borne by principals, these costs are absorbed ex post. In the firm-hiring example, workers' portion of the costs associated with exogenous uncertainty are manifest as less favorable contractual terms and fewer hours hired; but owners bear costs only when there is a reduced ex post residual income flow (which may also reduce the value of the firm's equity shares). This construct is readily mapped onto the case of a bank credit market. A financial intermediary that can set prices in its loan market will set a higher rate and supply less credit, ceteris paribus, as it faces more exogenous uncertainty about default risk (and a higher expected rate of default risk). Similarly, deposit Gary A. Dymski 391 rates will be lower than otherwise as intermediaries face more exogenous uncertainty about their earnings prospects on assets. And as an intermediary's markets become more perfectly competitive, risk is shifted onto intermediary owners, ex post, and off of its deposit and loan customers, ex ante. 3.4 MicrolMacro Disequilibrium This section has discussed how particular 'macro' elements have implications for 'micro' behaviors. Asymmetric information at the macro level requires imperfect competition at the micro level. Keynesian uncertainty in a macro environment is inconsistent with firms that unfailingly evaluate that environment as being merely risky. This approach to the interrelation of micro and macro analytic elements may enrich the notion of 'disequilibrium'. In the usual economic context, a disequilibrium is a condition in which one or more markets do not clear. Another type of disequilibrium may exist as well, however: a situation in which the behavior of micro agents is misaligned with their macro environment. Two examples from the above discussion will illustrate. First, consider a set of firms and wealth-owners in a particular market, all of them believing that they are in a parametrically stable, risky environment. If the environment is Keynesian uncertain instead, there could be substantial carnage among these market participants. The Keynesian uncertain aspect of this market would result in one or more large shifts in ex post outcomes for any given set of ex ante behaviors. The agents in this setting would, of course, use these new ex post results to update their understanding of their stochastic environment. But if they continued to believe that the outcomes they observed were rooted in parameters. they might fail to make adjustments as rapidly as needed to survive shifts in the circumstances of this market. An unusually high rate of failure might result. The idea of a mismatch between agents counting on 'normal times' and an environment that has been fundamentally transformed might have some applicability to the recent cases of the junk-bond market in the US, and to the plight of the US savings and loan industry, among others. Second, consider a financial intermediary of a new Keynesian type, which has been operating as an information specialist in an imperfectly competitive banking market characterized by asymmetric information. This bank will have developed an apparatus for evaluating the creditworthiness of borrowers about whom they have incomplete information, and it will characteristically have set loan and deposit rates so as to maximize its profits. The survival of this intermediary is then challenged if shifts of two kinds occur. First, an increasing degree of competition in its asset and liability markets will reduce and possibly even eliminate its ability to set loan and deposit rates according to some optimum. In the extreme, it will no longer be able to make loans with private-information characteristics with the expectation of earning positive net profits. Second, this firm's viability will also be challenged if its macro environment is Keynesian uncertain and becomes 392 Basic Choices in Keynesian Models of Credit more volatile. The share of its net profits that are attributable to overcoming asymmetric information may be overwhelmed by the share that are attributable to stochastic effects beyond the firm's control. There is no reason to assume that micro conditions will speedily adjust to changed macro environments. Micro constructs are embedded in institutional structures - in the case of financial intermediation, including departments of risk analysis, methods of assessing creditworthiness, and so on - which cannot usually adapt quickly to changes. So mismatches between macro environments and micro constructs might arise and persist. One result of such a mismatch would be high rates of insolvency and, in turn, of failure by the affected firms. Indeed, this is precisely what has occurred in the 1980s and 1990s in the US commercial banking and thrift industries. Both industries are experiencing rates of failure whose only historical equivalent is the Great Depression. As a macro environment changed, then, one would expect that micro agents that failed to adapt would be eliminated. This expectation is based on the assumption that firms with negative net worth will be closed; and this assumption is warranted only if a closure rule is operative. This need not be the case. If it is not, micro constructs may persist well after the macro environment required for their viability has disappeared. 4 CONCLUSION Differing conclusions about whether money, credit and intermediation matter flow from prior analytical choices in every economic analysis about how time, uncertainty and information should be treated. Deviations from WGE along these axes gives rise to roles for money, credit, and processes of credit allocation or intermediation. Money will almost certainly be essential if exogenous uncertainty is present; credit, if finance constraints exist, and intermediary structure, if endogenous uncertainty permeates the credit market. It is pointless to consider the importance of financial fragility in the context of a model premised on WGE. The key step in determining whether money and credit matter is, in effect, preanalytical. If we posit that Keynesian uncertainty, finance constraints and private information exist in the 'real' world, then only analyses encompassing all these features are 'realistic'. But such realism is not bought cheaply. Each shift away from the WGE origin point increases analytical complexity: financial realism comes at a high price. The above discussion has shown, in particular, that different macro assumptions vis-a-vis money, credit and intermediation have clear implications for micro behaviors. The richer the macro environment, the harder it will be to avoid micro/macro disequilibrium. Of course, inconsistency between macro environments and micro behaviors occurs in the 'real world'; the massive failures Gary A. Dymski 393 of financial institutions in the US over the past decade indicates that disequilibria of this kind can arise and persist. Usually, however, the presence of micro/macro inconsistency within a theoretical model is interpreted as implying that the model is flawed in some fundamental way. So when should non-redundant roles for money, credit and financial intermediation be explicitly incorporated into economic models, through appropriate assumptions about the nature of uncertainty and of information? When the scope and conclusions of analysis are partial, a money, credit or financial intermediation dimension mayor may not be important; their inclusion or exclusion is a matter of analytical convenience. But as theoretical intention becomes more general, inclusive treatment of all of these dimensions becomes more imperative, even as it increases the difficulty in reaching analytical conclusions. It is important to keep in mind that restricting the range of outcomes and relations in the model does not restrict those in the world; and policies are promulgated in the real, not the theoretical world. Appendix: Money, Credit and Economic Crisis in the Writings of the Regulation School The term 'Regulation school' or 'Regu\ationism' originated with works written by a number of French authors in the 1970s and 1980s (see, especially, Aglietta, 1979, and Lipietz, 1986). The basic premise of the Regulation school is that the state, as an embodiment of a collective capitalist interest, broadly 'regulates' social transactions and the societal institutional structure so as to maximize the sum of societal surplus value. A regime is defined as a stable institutional framework, orchestrated and regulated by state oversight, which allows successful capitalist extraction of surplus labor. A regime continues until it experiences insurmountable crises in accumulating capital, leading to its transformation and reconstitution. II These crises not only take a financial form, they are financial in their root. Indeed, what makes Regulationism of particular interest in this context is that credit is seen as a key locus of control by the state and capitalist interests over the economy as a whole. This emphasis is distinctive among radical political economic approaches to capitalism, for other approaches regard financial crises as merely the shells of more fundamental imbalances. Instead, financial crises are seen in this perspective as emanating from and embodying the value relationships of the system; as such, the manner of their resolution is critical for the system's future trajectory. Value magnitudes are linked with credit and money because the latter are central in accumulation and regulation, respectively. In the terminology of this chapter, Regulation school analyses assume the existence of a 'Schumpeterian' macro environment (as per section 2.3), wherein credit has an essential role due to finance constraints, but money has no important role because the importance of exogenous uncertainty is discounted. Credit originates in the circuit of capital and facilitates accumulation; credit relations arise principally among capitalists for productive purposes. Only the means of exchange role of money is acknowledged and this function is performed by the credit system. So money is always and only in motion as a response to needs for credit. What happens when money or other assets fall out of motion due to ex ante/ex post deviations is left out of the picture. 394 Basic Choices in Keynesian Models of Credit This distinctive Regulationist analysis of capitalist crisis has two central strands. Aglietta and Lipietz, in particular, have formulated a model of secular capitalist momentum; Foley, on the other hand, has developed a version of this model that emphasizes cyclical variation. Secular Tendencies: the Regulationist Analysis Proper Credit is essential in accumulation because both reversible time and discounting, which are at the heart of a dynamic version of passive finance, are rejected (Aglietta, 1979: 44-5).12 But credit originates outside the production sector: it is created by banks. As such, '[t]he credit system is an immense source of instability for the realization of exchange-value, because bank money makes credits liquid' (ibid.: 337). None the less, forces regulating credit creation do exist and insure that credit is functional for accumulation: specifically, banks' ability to create socially validated credit is strictly limited. This result follows from a restrictive definition of money. Money is 'one single commodity that becomes the general equivalent'; the monetary system as 'the sub-set of those conditions affecting the formation of the general equivalent and the choice of the measuring unit' (ibid.: 41). Money's role as general equivalent differs from its role as numeraire in WOE (ibid.: 329-31), for 'to enter into a relation of equivalence, money must possess the fundamental character of a commodity, that is, of being a sum of abstract labor' (ibid.: 328). The character of money, then, derives from accumulation. The notion of money as a store of value is unnecessary (ibid.: 330). The aggregate value of commodities itself, then, in its money equivalent, constitutes a monetary constraint for the economy. Thus, while money is physically created by banks, it 'acquires its attributes of general equivalent outside th[e banking] system, when these tokens of credit are utilized within commodity production' (ibid.: 331). So any gap between the volume of credit and the monetary constraint constitutes a source of crisis, for '[a]l\ private bank money Is subject to a sanction of gold equivalence, directly and without limit' (ibid.: 340). So while credit compensates for the lack of systemic coordination, '[T]he monetary constraint ... proves to be the unifying principle of the credit system. Financial crises are a reminder that the tendency of the credit system to become self contained is necessarily checked in a commodity economy' (ibid.: 341-2). Thus, the economy's monetary constraint restricts the real value of money, no matter what nominal amount of money is created by banks or destroyed in financial crises; and because this monetary constraint is always operative, financial relations are best understood as subordinate to accumulation. The consequences of a binding monetary constraint - financial crisis - differs according to the regime of accumulation in place. Within the (pre-war) 'extensive' regime based on low wages, financial crises brutally rein in credit and devalue capital. When the extensive regime exhausts itself due to overproduction, an 'intensive' (postwar) regime emerges in which workers not only generate, but consume value. In the new regime, money is prevented from autonomously regulating credit, because devaluation would destroy the relations of circulation on which accumulation depends; instead, the state pumps up liquidity during crises. This prevents any disastrous fall in the nominal profit rate. Thus the monetary constraint ceases to be operative and credit creation is not checked (Lipietz, 1983: 98). To the contrary, 'the ability to create credit at the national level was an important facilitating condition for sustained growth of real incomes' (Olyn et al., 1988: 26). However, 'credit can defer the social validation of private labors, but not suppress it' (Aglietta, 1979: 337). The recurrent gap between credit volume and monetary value is instead closed through inflation - that is, devaluing the unit of value. Using credit and price inflation to prevalidate income streams has two negative consequences, however: it shifts the income distribution in favor of workers, and causes capitalists to invest more heavily than underlying conditions warrant. Ever more credit is required relative to income to prolong accumulation; so a debt build-up accompanies inflation. But while manipulation of the divergence Gary A. Dymski 395 between credit, money and the value of money artificially restores nominal profit flows, it gradually destroys corporate balance sheets. Aglietta acknowledges the possibility of speculative finance. He writes that 'financial crisis begins with business euphoria' (1979: 358). This euphoria consists of commodity speculation within the department of circulating capital, which is spurred by the artificial rise in money prices fueled by credit expansion. This is a restricted conception of speculation, which is here attributable entirely to the (productive) expansion of credit as a means of prolonging accumulation in the intensive regime of accumulation. The idea of speculation in an environment of exogenous uncertainty does not arise. Cyclical Variation: Foley Foley's approach to money is identical to that of Regulationism. He cautions against 'the illusion of an historical emergence of money separate from ... the emergence of commodities', for money's fundamental function is 'expressing labor time' (Foley, 1983b: 8; and see 1986: 20). Money then "faithfully reflect[s] the underlying social relations of production' (ibid., 1983b: 19; and see 1982a). He ignores uncertainty: his formal models of the circuit of capital take the form of deterministic dynamical systems, which cannot endogenously represent stock/flow imbalances (ibid., 1986b: 33); in his recent work, firms have perfect foresight (ibid., 1987c). Crises arise from contradictions in the mode of production (ibid., 1986a: 145). A role for credit arises because production takes time and intertemporal precoordination is denied. Foley's formal models of accumulation emphasize the role of credit as 'a critical mediating channel .between changes in underlying parameters of accumulation like the markup and the ebbing of aggregate demand associated with the realization phase of crises' (ibid., 1986b: 54). His first set of credit models (ibid., 1982b, 1983a, 1986a) shows how time delays in sales, production and investment create a need for credit to fill in the gap between the value of capital supplied and demanded at each point in time. He uses these models to make two points: (i) lengthening any of these time delays can trigger a realization crisis; (ii) lengthening time delays in circulation activities induces lengthening time delays in production, and vice versa. These models lack any motivation for why delays should lengthen, since consumers only hold money for convenience, and Marx argues that time lags should shorten, not lengthen. in production (ibid., 1983a). Foley takes up this problem in a second set of formal models (ibid., 1987a, 1987b) rooted in a Regulation perspective. He posits that money is released at a constant exogenous rate and that firms face a cash-in-advance constraint. This, combined with time-using production processes, generates endogenous cyclical behavior. This suggests that cyclical instability is due to the interaction between firms' financial and production decisions, not to external limits on capital accumulation. Similarly, Loranger (1988) uses Foley's formal apparatus to conclude that inflation is caused not only by classical cost and productivity determinants, but also by this same interaction between incompletely coordinated financial and production activities. Notes I. 2. Dymski (1992) discusses some ideas for an approach to the banking firm that synthesizes Post and New Keynesian ideas. An illustrative model of this type is presented in Dymski (1994). Money is defined here as whatever media are commonly used as means of payment. The set of assets that functions as money is contingent on agents' attitudes to near monies issued by banks and fiat monies issued by governments. Credit (external finance) exists when one economic unit borrows other's idle resources in exchange for 396 3. 4. 5. 6. 7. 8. 9. 10. II. Basic Choices in Keynesian Models of Credit a promise to repay. Credit may be extended by 'surplus' units to 'deficit' units either directly or indirectly, through financial intermediaries. A 'stochastic' process here refers to any act whose precise outcomes is not known with certainty in advance. By 'already there', Foley means that 'the money form of value is inherent in commodity relations' (1986a: 22) - that is, the money commodity functions as the numeraire. Patinkin's 1963 work provides an apt illustration of the inessentiality of money when it lacks any function. Patinkin posited that money was an argument in agents' utility functions, and hence affected equilibrium outcomes through a Pigou (wealth) effect on consumption. But since money was a nominal good and equilibrium was precoordinated, there was no reason for any agent to hold money. Patinkin's role for money was thus analytical artificial. To have perfect information, agents need not know everything; rather, all economically relevant information must either be known directly by agents or summarized in a vector of competitive prices. Rasmusen (1989) has a useful discussion of the categories of information introduced herein. If the units buying goods from these units also finance their purchases, this scenario is one of a 'credit economy'. There has always been a temptation in conceptualizing credit economies to view the social economy as an EI Dorado of potentially infinite income possibilities. The source of this temptation is the assumption that all units need credit to operate, combined with the apparently innocuous assumption that credit is self-liquidating (alternatively, that all credit is trade credit). Finance constraints, which are equivalent to 'Clower' constraints (see Clower, 1984), are discussed in Kohn and TSiang (1988). Note that while finance constraints need not bind because ex ante surplus units wish to engage in time-phased activities, financing per se is of importance in an economy only if time-phased activities exist. Collateral is also postulated as a creditor device for disciplining borrowers in the presence of asymmetric information, the idea being that those with more to lose will be less likely to take chances with creditor funds. That is a different case from this one which encompasses no asymmetric information. Collateral here is a device for protecting against exogenous uncertainty. It may be possible for 'principals' - those who must hire the agents whose capacities or intentions are in question - to find ways of reducing the indeterminacy of agent action. However, this will be costly; so even if the problem is solved, there is a deadweight loss that would be ruled out of a WGE. Dymski (1994) argues that partial-equilibrium models do not provide sufficient micro-foundations in environments of both asymmetric information and Keynesian uncertainty. This list leaves off the various institutional mechanisms by which financial intermediaries might seek to reduce their endogenous uncertainty about borrowers. Such a discussion would be premature, since the mechanisms by which borrowers are monitored and creditworthiness is assessed have yet to be carefully dissected in the asymmetric information literature. The elements that are discussed in this literature are not particular to specific approaches to the credit market. There is some ambiguity about the root causes of accumulation crises in Regulationism. Aglietta (1979) suggests that overaccumulation of capital is invariably the problem; but more recently Lipietz (1986) has suggested that the cause may be contingent - overaccumulation of capital in the extensive regime (capitalists are too strong) and underaccumulation of capital (capitalists are too weak) in the intensive regime. In the latter case, Lipietz (1983, 1986) leaves open the possibility that both rising organic composition and wage-squeeze may be behind the result of underaccumulation. Gary A. Dymski 12. 397 Citations throughout this appendix are taken from Aglietta (1979), unless otherwise indicated. References Aglietta, M. (1979), A Theory of Capitalist Regulation (London: New Left Books). Bernanke, B. (1983), 'Non-monetary Effects of the Financial Crisis in the Propagation of the Great Depression', American Economic Review, 73,257-76. Clower, R.W. (1984), Money and Markets. ed. D. A. Walker (Cambridge: Cambridge University Press). Crotty, J. (1987), 'The Role of Money and Finance in Marx's Crisis Theory', in R. Cherry (ed.), The Imperiled Economy, Book I: Macroeconomicsfrom a Left Perspective (New York: Union for Radical Political Economics), 71-81. Davidson, P. (1991), 'Is Probability Theory Relevant for Uncertainty?: A Post Keynesian Perspective', Journal of Economic Perspectives. Winter, 5(1), 129-143. Dymski, G. (1990), 'Money and Credit in Radical Political Economy: A Review of Contemporary Perspectives', Review of Radical Political Economy. 38-65. Dymski, G. (1994), 'Fundamental Uncertainty, Asymmetric Information, and Financial Structure: "Post" verses "New" Keynesian Microfoundation', in G. Dymski and R. Pollin (eds.), New Directions in Monetary Macroeconomics: Essays in the Tradition of Hymen P. Minsky (Ann Arbor: University of Michigan Press), 77-103. Dymski, G. (1992). 'A "New View" of the Role of Banking Firms in Keynesian Theory', in Journal of Post Keynesian Economics, 14(3), Spring, 311-320. Fama, E. (1976) Foundations of Finance: Portfolio Decisions and Securities Prices (New York: Basic Books). Foley, D. (l982a), 'The Value of Money, the Value of Labor Power, and the Marxian Transformation Problem', Review of Radical Political Economy, 14 (2), 37-48. Foley, D. (1982b), 'Realization and Accumulation in a Marxian Model of the Circuit of Capital', Journal of Economic Theory, 28,300-19. Foley, D. (1983a), 'Say's Law in Marx and Keynes', miml!o, Barnard College, Columbia University. Foley, D. (1983b), 'On Marx's Theory of Money', Social Concept. 1(1),5-19. Foley, D. (1986a), Understanding Capital (Cambridge MA: Harvard University Press). Foley, D. (1986b), Money, Accumulation and Crisis (London: Harwood). Foley, D. (1987a), 'Liquidity - Profit Rate Cycles in a Capitalist Economy', Journal of Economic Behavior and Organization. 8, 363-76. Foley, D. (l987b), 'Endogenous Financial-Production Cycles in a Macroeconomic Model', mimeo, Barnard College, Columbia University. Foley, D. (1987c), 'A Marxian-Keynesian Model of Accumulation with Money', Working Paper, Department of Economics, Barnard College, Columbia University, September. Glyn, A., A. Hughes, A. Lipietz, and A. Singh (1988), 'The Rise and Fall of the Golden Age', DA,E Working Paper No. 884. Department of Applied Economics, University of Cambridge. Gurley, J. and E. Shaw (1960), Money in the Theory of Finance (Washington, DC: Brookings Institute). Kohn, M. and S.-C. Tsiang (1988), Finance Constraints, Expectations, and Macroeconomics (Oxford: Oxford University Press). Lipietz, A. (1982), 'Credit Money: A Condition Permitting Inflationary Crisis', Review of Radical Political Economy, 14(2),49-57. 398 Basic Choices in Keynesian Models of Credit Lipietz, A. (1983), The Enchanted World: l'Iflation, Credit and the World Crisis (London: Verso). Lipietz, A. (1986) 'Behind the Crisis: The Bxhausation of a Regime of Accumulation. A 'Regulation School' Perspective on some French Bmpirical Works', Review of Radical Political Economy,· 18 (1/2),13-32. Loranger, O. (1988), 'A Reexamination of the Marxian Circuit of Capital: A New Look at Inflation', Cahler 8802, D6partement de Science 6conomique, Universit6 de Montreal. Forthcoming, Review of Radical Political Economy. Pesaran, M.H. (1987) The Limits to Rational Expectations (Oxford: Blackwell). Patinkin, D. (1965) Money. Interest. and Prices. 2nd edn (New York: Harper & Row). Rasmusen, B. (1989) Games and Information (Oxford: Blackwell). Roemer, J. (1988), Free to Lose (Cambridge MA: Harvard University Press). Schumpeter, J. (1910), Theory of Economic Development (New York: Oxford University Press). Stiglitz, J. (1987) 'The Causes and Consequences of the Dependence of Quality on Price', Journal of Economic Literature, 25( 1), 1-48. Vickers, D. (1987), Money Capital in the Theory of the Firm (New York: Cambridge University Press). C. Comparing Post Keynesian and Circulation Approaches 14 Post Keynesian Fundism and Monetary Circulation Mario Seccareccia* Following the seminal writings of Sidney Weintraub, Post Keynesian economists have traditionally focused on the role played by wages and, more precisely, unit labor costs in both explaining inflationary pressures and in providing insights on the credit-driven nature of the money-supply process. Holding the view that money-supply growth is inextricably linked to the movement in money wages and employment, modem Post Keynesian writers, such as Basil Moore (1988, chapter 9), have made this Weintraubian view the cornerstone of their theory of endogenous money. Indeed, Paul Davidson (1988: 166-7) has gone as far as to suggest that this Post Keynesian theory of money can be interpreted as the modem counterpart of the 'real bills' doctrine widely espoused by the nineteenthcentury endogenous-money theorists of the banking school. The purpose of this chapter is to question the logical validity of this Weintraubian view of the money-supply process in an advanced credit system which, according to modem 'circuit' writers as well as to old-line banking theorists, is governed by what was once dubbed the 'Law of Reflux'. In their haste to break away from the neo-Classical conception of banks viewed traditionally as mere financial intermediaries whose activity is to transfer funds from savers to investors in the 'money' market, some Post Keynesian writers have taken the equally untenable position that bank credit only goes towards the financing of working capital. In an economy in which bank credit is the primary mode of financing productive activity, the exigencies of the payment system require that initial bank finance must go towards the purchases of all types of production, whether it be of fixed or of circulating capital, or, as described by Graziani (1990: 16), both of consumption and of capital goods. Any initial finance that falls short of this norm would lead to problems of structural insolvency since firms would be unable to realize fully money profits and, under specific conditions, extinguish their debts vis-ii-vis the banking system. This chapter is largely a commentary on the challenge posed by the monetary circuit approach and, more specifically, on the fundamental question historically raised inter alia by Minsky (1982: 72) about how, in a normal functioning capitalist economy, credit money is continually created and destroyed. After addressing this question within the confines of what has been described as Post Keynesian 'fundism', the analysis is then briefly followed by a discussion of 400 Mario Seccareccia 401 more traditional issues as to how the existence of interest and household saving is handled within the theory of the monetary circuit. 1 THE POST KEYNESIAN THEORY OF CIRCULATION AND THE PROBLEM OF MONETARY REFLUX The basic Post Keynesian position is perhaps best represented by the popularized version of Weintraub's original theory of the inflationary process. In his theory. he achieved the conceptual integration of the simple quantity equation with his famous wage-cost-mark-up explanation of price formation. Defining. in growth rate terms. the quantity equation: (1) (where the variables are respectively the rate of inflation. the rate of growth of the money supply. the rate of growth of velocity. and the rate of growth of output). and his wage-cost-mark-up equation: PIP = kIk + wlw + NIN-QlQ (2) (where kIk is the rate of growth of the mark-up. w/w is the rate of growth of money wages. and NIN is the rate of growth of employment). then. on the assumption that VIV = kIk = O. it follows: MIM = wlw + NIN (3) which. on account of the restrictive assumption that both the mark-up and velocity are assumed constant. cannot any longer be considered a truism as are equations (1) and (2). From this. a number of Post Keynesian writers have concluded that money-supply growth depends essentially on the movement in the wage bill. As Nell (1990) points out. however. equation (3) tells us nothing about the direction of causation. Traditional monetarists have argued that M/M must be treated as the intermediate instrument to control wages. incomes and prices. Alternatively. the customary Post Keynesian approach has been to control wages through incomes policy. whose effect would be to slow the growth of both prices and the money supply. Both approaches. however. associate some form of monetary austerity as a necessary concomitant to combating inflation. For the former. this has been very explicit and it has normally been done through the control of monetary aggregates by means of high interest rate policy; while for the latter this has traditionally entailed money-wage controls. which for a given velocity and firms' mark-up are assumed to have an impact on both the rate of inflation and money-supply growth. 402 Post Keynesian Fundism and Monetary Circulation Numerous Post Keynesians, from Weintraub to Cornwall (1984), have thus proudly defended the position that an incomes policy or, more specifically, a productivity-geared wage policy would be a far more equitable and civilized approach to controlling inflation than the alternative high interest-rate policy of creating massive unemployment. While this is undeniably true, critics (especially from a trade union perspective) have long argued that the degree to which the Post Keynesian policy option can effectively deal with inflationary pressures very much depends on whether control over wages does, in fact, have the predicted effect on prices and money-supply growth (cf. Weldon, 1991, ch. 9). There exists a significant body of literature of Kaleckian pedigree which from Steindl to Eichner has questioned the constant mark-up hypothesis that underlies the Weintraubian option in favour of incomes policy. By its assumption that firms have a net preference for 'internal financing', this literature, also within the Post Keynesian camp, points to the behavior of the mark-up, and its link to the investment process, to be as important as wage cost in explaining price formation (cf. Seccareccia, 1984). While the possible correlation between firms' mark-ups and investment spending may cast some doubt on this traditional policy perspective in favor of some form of wage controls to combat inflation, a still more serious problem arises when this Post Keynesian 'real bills' view of financing production passes more rigorous scrutiny in terms of its logical coherence. Reviving the Classical 'fundist' view of production (to use Hicks' (1974) expression) in which capital is advanced during the production period in anticipation of sales receipts, Post Keynesian writers such as Moore (1985, 1988) have argued along Weintraubian lines that firms fundamentally borrow from banks a fund for the purpose of financing working capital, of which its single most important component is wage cost. As in certain Classical writings adhering to the banking principle, perhaps best expounded by Thomas Tooke (1844), the role of banks is to provide the credit necessary to constitute a wage fund which is then advanced during the production process. Money supply growth is thus endogenous to credit demand which, in tum, depends on the 'needs of business' to hire labor and carry out productive activity. Knowledge of both nineteenth-century banking history and contemporary bank behavior would suggest that such an assumption, regarding bank credit going exclusively towards the financing of working capital, is highly questionable. For instance, in Canada, while lip service was once paid to the 'real bills' doctrine, actual banking practices frequently deviated from such principles even during the nineteenth century. Historically, railways and other corporate securities of a longer term nature constituted significant shares of total securities held by banks (Neufeld, 1972: 113), and, indeed, as recognized by numerous practioners (cf. Hood, 1959, chapter 6), banks have long played an important role in the financing of investment in fixed capital. Time series evidence is also of interest. Figure 14.1 gives some empirical evidence that both supports and, at the same time, raises certain doubts about the Mario Seccareccia 403 relevance of such an assumption. Indeed, the figure shows that while various money and credit aggregates in Canada are very strongly correlated with movements in the wage bill, they also indicate a weaker association with the flow of business spending on fixed capital (I). For instance, changes in the wage bill (in first-difference form) is undeniably very strongly related to changes in the various money and credit aggregates for the period 1970-89 in Canada. However, changes in business credit (ABC) have also moved somewhat in tandem with the flow of investment spending. As it will hopefully be better understood later, the reason why there may exist a widening gap between (ABC) and I (particularly for the post-1982 period) has more to do with the fact that other sectors, such as government, have been engaged more and more in deficit spending, thus allowing firms to finance 'internally' a greater and greater share of their investment expenditures. In earlier periods, when government deficits were less important in magnitude, there appears to be a much closer link between ABC and I. Hence, if we interpret Figure 14.1 in this light, it would suggest that empirically, at least the Weintraubian assumption is only partially correct, since it does not take into account the fact that !lBC also may go towards the financing of fixed capital expenditures. On the theoretical level, however, the problem with this specific Weintraubian assumption is far more fundamental. This has to do with what both nineteenth-century adherents of the banking principle and modem circuitistes have traditionally described as the problem of monetary reflux. To understand the problem of reflux, let us develop a simple analysis within the framework of a Wicksellian 'pure credit' economy, in which all transactions are finalized through the banking system without the public's need to hold any cash balances. In this type of economy, credit money is created at the stroke of commercial banks' 'fountain pens' (as Tobin, 1963: 415) described it) as soon as firms engage in borrowing to finance production. Moreover, to deal with the issue of financing fixed versus working capital, we shall adopt the familiar two-sector model in which production may be of investment goods (fixed capital) and of consumption goods (circulating capital). Following Keynes of the General Theory, at the beginning of the monetary circuit firms in each sector take decisions regarding production based on their short-term expectations of future sales proceeds. Given these expectations, firms approach banks for the purpose of financing their planned levels of production, ZI and Zc respectively. However, according to the Weintraubian 'real bills' assumption that ab initio firms only borrow to finance the wage bill, there must emerge an 'efflux' (to use Tooke's (1848, p. 194) expression) of credit money equal to the wage bill wNI and wNc' That is to say, the flow of credit money (or initial finance), M, originating from the banking system will allow firms to hire N workers at a given money wage rate, w, such that: (4) This fund of working capital then permits firms in each sector to undertake production, ZI and Zc. from which, given the levels of effective demands, firms are 404 Post Keynesian Fundism and Monetary Circulation 60_ 78_ 76_ 74_ 72 _ 70_ 88_ 66_ I!! 64_ ~ 62_ 'D a 60_ 58_ 56_ 54_ iii 52_ 50_ 48_ 46 _ 44_ 42_ 40_ 38_ 36_ 34_ 32_ 30_ 28_ 26_ 24_ 22_ 20_ 1 18_ 16_ 14_ 12_ 106_ 6_ 4_ 2_ 0_ -2_ -4 _ 1970 71 I AwN ABC AM2 60M3 AM4 72 73 74 75 76 77 76 79 60 81 62 63 64 85 68 67 88 89 = Business fixed capital formation = Changes in wages, salaries and supplementary labour income = Changes in total business credit = Changes in currency, demand deposits, daily interest chequable deposits, notice deposits and personal term deposits = 60M2 plus changes in other non personal fixed term deposits and foreign currency deposits of residents booked in Canada. = Changes in currency plus total Canadian dollar privately held bank deposits Source: Bank of Canada, Bank of Canada Review, selected issues; Statistics Canada, National Income and Expenditure Accounts, Cat. n°13--001 , Ottawa: Supply and Services Canada, selected issues. Figure 14.1 Fluctuations in investment and in changes in the wage bill and various measures of money and credit, Canada, 1970-89 Mario Seccareccia 405 thus assumed to be able to realize profits or, namely, net revenues 1t1 and 1tc ' But can firms realize money profits in this 'real bills' world and still reimburse the banks? This is a question posed certainly as far back as Malthus (as suggested by Bleaney, 1976: 53-4) and it remains a key problem still debated by numerous modern writers in the tradition of the monetary circuit. To answer this question, we shall assume for the purpose of simplifying the analysis extreme Classical savings functions whereby the propensity to save is zero for wage earners and unity for profit earners, so that there are no leakages from the household sector. From this it follows that: (5) (5') Hence, as is well known from Kaleckian theory, money profits in the consumption goods sector turn out to be merely wN/ the proportion of total M that went towards the financing of production in the investment goods sector. However, if Zc = M, where do firms in the investment goods sector get the money to reimburse the banks, thus closing the circuit of credit? Admittedly, one can always parachute in a deus ex machina fashion a third sector into the analysis, such as a government sector or a foreign trade sector. A fortiori, governments can be assumed to run permanently budget deficits and/or follow a mercantilist policy of engendering persistent trade surpluses, thus creating the additional amount of credit money that can give rise to a flow of money revenues also in the investment goods sector. The existence of large budget deficits, in fact, was suggested as a plausible explanation for the widening gap between dBC and I in Figure 14.1. Though empirically relevant, these would be merely ad hoc solutions to a structural problem of generating a monetary flow to the investment goods sector in an endogenous credit money world. But what would be the solution in the case of a closed economy with neither a government nor a foreign trade sector? It is obvious that in a growing pure private economy where Z/ is positive, part of the production of the investment goods sector would, or course, be sold to the consumption goods sector. In the limiting case in which all money profits in the consumption goods sector go towards the purchase of fixed capital goods, firms in the investment goods sector can, at best, get back the amount previously paid out in wages and thus would be able to extinguish their debt of wN/ with the banks. Though in this limiting case, firms in both sectors obtain just enough money revenues to pay their previously contracted loans with the banking system, the problem of generating money profits in the investment goods sector still remains outstanding. Indeed, as writers of the circuit approach, as well as American institutionalist economists such as Fagg Foster (1981), have long recognized, this problem is quite general. How can firms collect more money revenues than the amount that 406 Post Keynesian Fundism and Monetary Circulation economic agents are willing to borrow collectively from the banking system which, in this analysis, has been assumed to be sole creator of credit money? The explicit double assumption found, therefore, in writers such as Moore (1988) that the money supply is fundamentally endogenous credit money created by the banking system and that firms borrow exclusively to finance their working capital are simply incompatible with one another and do not logically add up. Moreover, contrary to the suggestion in Lavoie (1987: 80) that perhaps this difficulty arises only in our simple fully integrated two-sector model and that the disaggregation of the investment goods sector into two (or more) subsectors may solve the problem, it would appear that this is not so. For instance, adopting Adolph Lowe's (1976) subdivisions, whereby Z'a is the production of capital goods used by the investment goods sector and Z'b is the production of capital goods employed in the consumption goods sector, so that we have: (6) Z'b = 7T'b + wN'b Zc = 7Tc + wNc (7) (8) then, with the usual Weintraubian assumption that M = wN'a + WN'b + wNc (4') it follows that (9) If, as before, we further suppose that these money profits in the consumption goods sector go fully towards the purchase of fixed capital goods 2'b (so that we may infer that 7Tib = wN'a) and that sector lb • in tum, purchases capital goods from sector la' firms in this latter sector can, at best, obtain sufficient money receipts merely to wipe out their debts of wN'a' Hence, as before, while in the limiting case the problem of monetary reflux can be solved on the strict proviso that everyone spends their net receipts, there seems always to emerge a group of firms in the investment goods sector who cannot fully validate their output in money terms. As can be found in the summary statements by Lavoie (1987) and Graziani (1988), this problem has been widely recognized in the literature on the monetary circuit. One way of conceptualizing profits in the investment goods sector (or subsector la) is to consider them as pure book values whose physical amount has no monetary counterpart in circulation. This, however, would be meaningful only if the investment goods sector were analytically treated as one large firm for whom the only form of money costs are the wages paid. In this case, the residual fixed capital not sold to firms in the consumption goods sector would be accumulated for internal use by this large firm in the form of unsold inventories and would be equal to its book profits, 7T,. However, as soon as one allows for decen- Mario Seccareccia 407 tralized atomistic behavior on the part of firms in the investment goods sector, such a solution is no longer viable. An alternative solution is to imagine that in this pure credit economy there develops a parallel credit circuit, existing outside of the banking system proper, for the purpose of regulating interfirm transactions within the investment goods sector. Under this arrangement defined by Nell (1986: 30) as one of private 'selfliquidating credit', firms in the investment goods sector would be extending among themselves what Wicksell once described as 'simple credit', and which would cancel itself out in the aggregate. In fact, this would be conceptually identical to a situation in which firms undertake barter within their sector since they do not rely on any bank-created circulating medium to settle their accounts. While this might be considered a viable theoretical solution, the fundamental problem with such an alternative, however, is that it postulates an asymmetry in firms' behavior that appears to defy logic. Why should firms in one sector provide private credit for purchases among themselves and, yet, require exchanges in money terms originating from bank credit with firms in the consumption goods sector? This questionable asymmetry and the fact that this is essentially a nonmonetary solution to a problem of monetary circulation should further lead us to question its theoretical and practical relevance. Other more complicated solutions have also been offered, some of which are put forth in Dupont and Reus (1989). However, these seem to require the abandonment of the assumption of a Wicksellian pure credit economy and to necessitate the introduction of outside (or commodity) money that 'circulates' from firm to firm and/or that is carried over by households from one period to the next in the form of idle balances. Such solutions to the problem of the realization of profits that appear to dispense with the fundamental Post Keynesian assumption of an endogenous flow of credit money raise more problems than they solve and thus will not be further entertained in this context. The only satisfactory solution must be one, as mentioned by Cencini (1984: 145-7; 1988: 95-6) and De Vroey (1988: 227), in which bank loans to firms are extended so as to include the money profits to be realized in both sectors. While disagreeing with, say, De Vroey's (1988) specific proposal which seems to rest on the highly untenable assumption that firms somehow borrow and spend their 'profits' before they can be realized at the end of the circuit, Parguez's (1981: 175-6) and Nell's (1986: 31) position that, in an expanding economy, banks also finance investment spending is by far the most plausible alternative that can deal effectively with this theoretical conundrum. This competing Keynesian view would suggest that initial bank credit must go towards the financing of all business expenditures regardless of whether they go toward the purchases of fixed or working capital. Hence, returning to our simple two-sector model and disaggregating equation (4), this would imply: (10) 408 Post Keynesian Fundism and Monetary Circulation M/ = wNI + (1 - a)ZI (11) where a is the proportion of total investment expenditures originating from firms in the consumption goods sector and 1 - a is the proportion originating from those in the investment goods sector. As a result, total bank credit is: M =Mc+MI = wNc + wN1 + aZ I + (I - a)ZI (12) or simply, (12') M =Zc+ZI since we have assumed that workers spend what they get on Zc' It follows, therefore, that: (13) and 7T/ = aZ/ + (1 - a)Z/ - wN 1 (14) purchases interfirm from purchases C-sector within I-sector both of which are sectoral profits realized in money terms and financed by bank credit. In the limiting case in which profits in the consumption goods sector (equal to wN/) go fully towards the purchase of capital goods aZ/, then firms in both sectors will obtain sufficient money revenues to repay fully their loans, M, previously contracted with the banking system. Indeed, any financing in this pure credit world that would fall short of what would permit firms and households to purchase total production, Z, would entail problems of structural insolvency of the type first recognized by Nell (1967) in this analysis of WickseIl's theory of monetary circulation. Schematically, our description of monetary circulation can be depicted in Figure 14.2. Following Tooke's circular conception of credit money, there emerges an 'efflux' of newly created bank money that goes towards the financing of wN + Z/. The receipts from the sale of output, Zc + Z/, then permit firms to retire their bank loans, so that at the end of the monetary circuit, a reflux of M returns to the banking system to extinguish their debts. The monetary circuit thus comes to a close in this simple scheme without the requirement of further financing because of the incapacity of firms in the aggregate to reimburse fully the principal of their loans. Not only can this assumption regarding the financing of both fixed and working capital solve the conceptual imbroglio confronting those advocating the Post Keynesian 'fundist' principle of financing the wage bill, but it is also supported Mario Seccareccia i I I M('efflux'~ J1('ref/ux') I ~ IZIJ wN Zc 409 • HH 00 UL SD E S Figure 14.2 by historical experience as well as by some of the very limited empirical evidence provided in Figure 14.1 in which 11M is not only correlated with I1wN but also, to a certain extent, with Z,. As emphasized by Parguez (1980) and, among others, Parker Foster (1987), moreover, this assumption is certainly consistent with Keynes's more traditional position in the Treatise and the General Theory where bank credit does go towards the financing of investment, Z,. 2 SOME FURTHER REMARKS ON THE PROBLEM OF MONETARY REFLUX The macroeconomic problem of the realization of money profits, as well as the issue of monetary reflux. have been discussed within a very simplified framework in which banks neither charge interest on their loans nor do households acquire liquid assets for the purpose of saving a portion of their incomes. Though we may still want to retain the more restrictive assumption of a Wicksellian 'pure credit' economy. it will now be appropriate to introduce interest earnings as a separate and distinct category of income which is received from the holdings of bank deposits and, of course. to suppose that banks charge interest on loans advanced to firms. All other simplifying assumptions, such as that there is no household saving, will be retained during this initial discussion of the problem of interest. The traditional question posed by writers, perhaps going as far back as Major Douglas, has been as follows: how can the banking system, that has only advanced M dollars in the form of loans to firms, be reimbursed this principal plus interest? As Schumpeter (1934: 189) put it, 'within the circular flow ... it is impossible with a given money sum to obtain a greater money sum'. Surely, given the difficulties encountered in reconciling the efflux-reflux in the cases previously discussed where only the principal of the loans was at stake, it would not be unreasonable to infer that, as a group, firms are condemned to be continually engaged in Ponzi finance so that over time their cumulative debt would be growing exponentially by the amount of compound interest. Such a conclusion on the impossibility of meeting their interest payments, as can be found even in modern writers as Uonard (1987), however, is based on the questionable implicit assumption that the interest on loans has no counterpart on the deposit side of banks' balance sheets. 410 Post Keynesian Fundism and Monetary Circulation As Lavoie (1985) and Moore (1988) make it abundantly clear, in the real world of endogenous credit money, loans make deposits. We shall thus define the rate banks charge on their loans, r, while the interest rate paid on deposits, i. In our previous analysis, we had concluded from equation (12) that firms need to borrow M = wN + Z, in order for the circuit to be closed. Unfortunately, to obtain such a result we had abstracted from the whole question of the time-pattern of borrowing and spending on the part of the economic units, as if everything is borrowed at the beginning of the period. In reality, of course, we know that both borrowing requirements and spending may be unevenly distributed over time. For instance, while the wage bill may be advanced on a regular basis, other expenditures on fixed capital are undoubtedly more lumpy. Following Robinson (1956: 228-9, fn. 3) and Graziani (1987: 32), let us assume for simplicity a continuous stream of borrowing and spending so that what firms owe the banks at the end of the period of the circuit is an amount equal to M(I + 1/2r). At the same time, this value would have as counterpart on the deposit side an amount equal to M(I + 1I2i). Hence, abstracting from the problem of depreciation, business profits in each sector now become: 7f, = Z,- wN,-ll2rM, (IS) 7fc = wN, + 112iM,-112 (r-i)Me (16) with this latter result being obtained on the strong assumption that all interest accruing to those holding deposits, whether it be workers or firms spend this interest income, 1I2iM, + 112iMe' on consumption goods. That is to say, there is zero propensity to save out of rentier income. We shall naturally abandon this restrictive assumption when household saving is later introduced into the analysis. In this case, the total flow of profit, 7f, is: 7f = ZrIl2(r-i)M,- 112 (r-i) Me = Zr 1I2(r-i)M (17) where (r - i) is the interest spread or what Rousseas (1986) and Moore (1988) have appropriately defined as a bank mark-up, 1I2(r - i)M being the flow of revenue to the banking sector. In the unlikely case of zero spread, so that the loan rate just equals the deposit rate, the effect of positive interest rates on the system would be neutral since the monetary reflux to the banks is analytically equivalent to the 'zero interest' case previously discussed. That is to say, the net reflux to the banks is [1 + 1I2(r - i)]M = M, and this would be so regardless of the level of the interest rates. However, in the more realistic case in which r> i, the effect would be to put a squeeze on gross business profits and increase the required net reflux to the banks. Yet, we know that banks cannot obtain more money than they have advanced to firms (1 + 1I2i)M, with the inevitable result being that certain firms would be unable to meet their financial commitments. Since, in this case[1 + 112(r - i)]M > M, firms Mario Seccareccia 411 must either sell part of their output and/or physical assets to the banks or ask for extension in the form of new loans equal to the volume of the interest spread. In this latter Ponzi scenario, firms become progressively indebted to the banking system over time, so that the circuit of credit is never closed. In the former case, instead, two possibilities exist for the closure of the circuit. This closure can be done voluntarily or involuntarily. As recognized long ago by Joan Robinson (1956: 249-50), banks, obviously, must advance wages to their own employees and regularly payout dividends to their shareholders. If this amount that is periodically advanced and spent during the period of the circuit is just equal to 112(r - i)M, then firms would be able to pay the required reflux of (1+1I2r)M. On the other hand, if less than the full amount has been advanced during the period, firms would be unable to meet their payments and would thus have to transfer involuntarily some of their illiquid assets to the banking system as certain businesses are forced to go under. All such outcomes are possible. Indeed, to the extent that the interest rate on deposits (i), wages, and dividends paid out by the banks are more sticky in the short run than the interest rate charged on business loans (r), one should expect that during periods of rising interest rates, the gap between rand i widens and, therefore, firms would be facing progressively more difficulties in meeting their financial commitments. One such indicator of this banks' mark-up for the US economy has been provided by Rousseas (1986: 54-6). It is the mark-up of the prime rate over the Federal Funds rate for the period 1955-84. While undeniably there are more appropriate indicators, it is nevertheless interesting that Rousseas's original mark-up series peaks at approximately the same time as the NBER reference cycle troughs, which is consistent with our hypothesis on the negative effect that a widening interest spread has on business profitability and on economic activity. Moreover, to the extent that higher interest spreads are indeed normally associated with rising interest rates, there exists ample empirical evidence in the literature of the negative effect of the growing interest burden on the productive system (cf. Seccareccia, 1988), that even mainstream economists would agree that such a link might be highly significant empirically. Until now, we have discussed the problem of monetary reflux in a world in which households do not save, thus obtaining the standard 'widow's cruse' Keynes-Kalecki relation that 'TT'/ + 'TT'c = It is now time to modify this assumption by introducing household saving directly into the analysis so that 'TT'/ + 'TT'c + Sh = Z/ where Sh is the flow of household saving. With the introduction of Sh' our basic scheme (depicted in Figure 14.2) representing the monetary flows requires a slight modification. We now also need to take into account the fact that there is an outflow from the household sector either to the banks in the form of liquidity or, through the financial capital market, to firms for the purpose of what Davidson (1986) and Graziani (1987) identify respectively as 'long-term funding' or 'final finance'. This is shown by an additional arrow, depicted in Figure 14.3, originating from household sector and reflecting household saving, Sh. z/. 412 Post Keynesian Fundism and Monetary Circulation w z,J rn L 112 iM I B A N K S M i ~ J+ J12 ,)M I ~ 'Liquidity preference'(M (19) for all values of r ~ i and 'P > O. The existence of idle balances, 'PM, thus forces banks into the untenable position of accumulating illiquid corporate securities and, de facto, extending business credit over time without a complete closure of the monetary circuit. This simple model can be further refined, as numerous economists associated with the circuit approach, such as Schmitt, Parguez and Graziani, have done so historically. One may widen this theory so as to deal with business pricing, consumer credit, government, and, among many other questions, foreign trade. In all cases, the basic monetary constraints of efflux = reflux would apply. For instance, let us briefly see how the existence of a foreign trade sector can affect our basic model. As with other business enterprises producing for the domestic market, firms involved in producing internationally traded goods would also need to finance their production. In much the same way, the net receipts from the sale of goods to foreigners would allow firms in the export sector to extinguish their debts which they had contracted during the production period with the domestic banking system. The degree to which the macroeconomic circuit is closed, however, would depend on the behavior of the exchange rate and on the magnitude of the leakage due to imports. Unfortunately, these concerns, albeit of immense theoretical importance, are outside the scope of the present chapter. The same would apply if a government sector were inserted into this picture. As economists, such as Minsky, have long argued, big governments with large budget deficits make this otherwise very fragile credit system less susceptible to financial breakdown and collapse because of their broad spending powers and their impact on aggregate demand. Through their links with central bank financing, governments simply do not face the same sort of constraints as, say, private firms do in dealing with the problem of monetary reflux. It is only when policy-makers seek to subject the state sector to the same fiscal constraints, which are imposed by the financial system on the private sector, that systemic 414 Post Keynesian Fundism and Monetary Circulation instabilities arise and bring the economy closer to the type of scenario described by Keynes during the 1930s. As can be seen, therefore, the upshot of all this is that the problem of monetary reflux is a very general one that, admittedly, has not been sufficiently recognized by those of us dissatisfied with what Moore (1988) has caricatured the 'verticalist' conception of money. 3 CONCLUDING REMARKS As writers such as Schumpeter had long understood, when one seeks to penetrate the veil of money one discovers the wall of credit and, in particular, the subordination of 'real' activities to exigencies of the credit system. From the early advocates of the banking principle in the nineteenth century to the modern theorists of the monetary circuit, the primary concern has been with the analysis of a dynamic capitalist economy in which credit money is perpetually created and destroyed. While Post Keynesian economists have overwhelmingly recognized the fundamental importance of credit money, they have not always worked out its full implications. This is especially problematic when one seeks to apply the Weintraubian 'fundist' assumption within the framework of an endogenous credit money world. In response, we have tried to describe the numerous new insights that can be obtained from a more complete analysis of monetary reflux - a concept that remains central to the theory of the monetary circuit. Note * The author would like to thank M. Lavoie, A. Parguez, J. M. Smithin and I.M. Spry for their very useful comments. The usual disclaimer applies. References BIeaney, M. (1976), Underconsumption Theories: A History and Critical Analysis (New York: International Publishers). Cencini, A. (1984), Time and the Macroeconomic Analysis of Income (London: Frances Pinter). Cencini, A. (1988), Money, Income and Time (London: Frances Pinter). Cornwall, J. (ed.) 1984), After Stagflation: Alternatives to Economic Decline (Armonk, NY: M.E. Sharpe) Davidson, P. (1986), 'Finance, Funding, Saving and Investment', Journal of Post Keynesian Economics, 9 (1),101-10. Davidson, P. (1988), 'Endogenous Money, the Production Process and Inflation Analysis', Economie appliquee, 41 (1),151-69. De Vroey, M. (1988), 'II circuito della moneta: due interpretazioni', in A. Graziani and M. Messori (eds), Moneta e Produzione (Turin: Giulio Einaudi), 215-45. Mario Seccareccia 415 Dupont, F. and E. Reus (1989), 'Le profit macroeconomique monetaire', Economie Appliquee, 42 (2) 87-114. Fagg Foster, 1. (1981), 'Understandings and Misunderstandings of Keynesian Economics', Journal of Economic Issues, 15 (4) 949-57. Graziani, A. (1987), 'Keynes' Finance Motive', Economies et Soc;etes, 21 (9),23-42. Graziani, A. (1988), 'II circuito menetario', in A. Graziani and M. Messori (eds.), Moneta e Produzione (Turin: Giulio Einaudi), xi-xiii. Graziani, A. (1990), 'The Theory of the Monetary Circuit', Economies et Soc;etes, 24 (6), 7-36. Hicks, 1. (1974), 'Capital Controversies: Ancient and Modern', American Economic Review, Papers and Proceedings, 64 (2), 307-16. Hood, W.C. (1959), Financing of Economic Activity in Canada (Ottawa: Queen's Printer). Lavoie, M. (1985), 'Credit and Money: The Dynamic Circuit, Overdraft Economies and Post Keynesian Economics', in M. larsulic (ed.), Money and Macro Policy (Boston: Kluwer-Nijhoff), 63-84. Lavoie, M. (1987), 'Monnaie et production: une syntMse de la theorie du circuit', Economies et Soc;etes, 21 (9) 65-101. Leonard, 1. (1987), 'Le paradoxe de I'interet et la crise de I'economie monetaire de production', Economies et Soc;etes, 21 (9), 149-68. Lowe, A. (1976), The Path of Economic Growth (Cambridge: Cambridge University Press). Messori, M. (1985), 'Le circuit de la monnaie: acquis et problemes non resolus', in R. Arena and A. Graziani (eds.) Production, circulation et monnaie (Paris: Presses Universitaries de France), 207-46. Minsky, H.P. (1982), Can 'It' Happen Again? Essays on Instability and Finance (Armonk. NY: M.E. Sharpe). Moore. B.l. (1985), 'Wages, Bank Lending and the Endogeneity of Credit Money', in M. larsulic (ed.). Money and Macro Policy (Boston: Kluwer-Nijhoff), 1-28. Moore, B.J. (1988) Horizontalists and Verticalists: The Macroeconomics of Credit Money (Cambridge: Cambridge University Press). Nell, E.J. (1967), 'Wicksell's Theory of Circulation', Journal of Political Economy, 75 (4) (Part I), 386-94. Nell, E.J. (1986), 'On Monetary Circulation and the Rate of Exploitation', Thames Papers in Political Economy (Summer), 1-36. Nell, E.J. (1990), 'The Quantity Theory and the Mark-Up Equation', Economie Appliquee, 43 (2), 33-42. Neufeld, E.P. (1972), The Financial System of Canada. Its Growth and Development. (Toronto: Macmillan of Canada). Parguez. A. (1980), 'Profit, epargne, investissement: elements pour une tMorie monetaire du profit', Economie Appliquee, 33 (2).425-55. Parguez, A. (1981), 'Keynes et la revolution', Cahier d'economie politique, (6), 171-87. Parguez, A. (1986), 'Au coeur du circuit ou quelques reponses aux enigmes du circuit', Economies et Soc;etes, 20 (8/9), 24-40. Parker Foster, G. (1987), 'Financing Investment', Journal of Economic Issues, 21 (I), 101-12. Robinson, J. (1956), The Accumulation of Capital (London: Macmillan). Rousseas, S. (1986), Post Keynesian Monetary Economics (Armonk, NY: M.E. Sharpe). Schmitt, B. (1966), Monnaie, salaires et profit (Paris: Presses Universitaires de France). Schumpeter, J.A. (1934), The Theory of Economic Development (Oxford: Oxford University Press). 416 Post Keynesian Fundism and Monetary Circulation Seccareccia, M. (1984), 'The Fundamental Macroeconomic Link between Investment Activity, the Structure of Employment and Price Changes: A Theoretical and Empirical Analysis', Economies et Societes, 18 (4),165-219. Seccareccia, M. (1988), 'Systemic Viability and Credit Crunches: An Examination of Recent Canadian Cyclical Fluctuations', Journal of Economic Issues, 22(1),49-77. Tobin, J. (1963), 'Commercial Banks as Creators of "Money' .. , in D. Carson (ed.), Banking and Monetary Studies (Homewood, III: Richard D. Irwin), 408-19. Tooke, T. (1844), An Inquiry into the Currency Principle, 2nd edn (London: Longman, Brown, Green and Longmans). Tooke, T. (1848), A History of Prices, and the State of the Circulation, vol. IV (London: Longman, Brown, Green, Longmans and Roberts). Weintraub, S. (1978), Capitalism's Inflation and Unemployment Crisis (Reading, Mass.: Addison-Wesley). Weldon, J. (1991), On the Political Economy of Social Democracy: Selected Papers of J.C. Weldon (Montreal: McGill-Queen's University Press). Wray, L. (1990), Money and Credit in Capitalist Economies: The Endogenous Money Approach (Aldershot: Edward Elgar). 15 Investment Decisions in Circuit and Post Keynesian Approaches: A Comparison Richard Arena* Those last years, Keynesian economists paid a specific attention to the theory of investment. The distinction between finance and funding (Davidson, 1986) was especially investigated. Less emphasis, however, was given to investment decisions as such. This difference of treatment between these aspects of the theory of investment can be easily understood. The determination of the factors of investment decisions is indeed a very uneasy task for both mainstream and non-orthodox economics. As E. Malinvaud pointed out: the respective part played by the various factors capable to explain the more or less important magnitude of investment is one of the themes to which econometricians paid the major attention. The results being often rather doubtful, strenuous debates happened in various occasions, concerning the conclusions they implied. This is not surprising: our theoretical investigation showed us the multiplicity of elements capable to playa significant part, and the difficulty we may experience when we try to define observable indexes which could be used to express some of them. (Malinvaud, 1981, vol I: 199) The problem of investment decisions appears, however, to be a crucial one for Keynesian economists. It also gives the opportunity to compare both Post Keynesian and Circuit traditions, focusing upon a topic which permits to exhibit important and significant theoretical differences between them. More precisely, the present chapter will be dedicated to the respective importance of industrial and financial factors in investment decisions.! Three steps will be necessary to deal with the problem. The first will be devoted to the analysis of the differences between Kalecki's and Keynes's views of investment decisions. The second one will try to exhibit two possible synthetic approaches, offered by Minsky's and Steindl's contributions. 417 418 Circuit and Post Keynesian Investment Decisions Finally, the last part of the chapter will explain why these syntheses cannot involve the circuit approach and, therefore, help us to understand more thoroughly the theoretical differences between this approach and the Post Keynesian one. KALECKI AND KEYNES We all remember the articles Tom Asimakopulos dedicated during the last years of his life to the common features of Kalecki's and Keynes's analysis of the finance and funding processes of investment (Asimakopulos, 1983, 1986, 1990). These common features do not exclude, however, the existence of noticeable differences between both authors' treatments of investment decisions. The First Difference Lies in the Social Frameworks into which Kalecki and Keynes Respectively Inserted their Theoretical Approaches Thus, as we know, Keynes drew a sharp distinction between entrepreneurs and rentiers and, therefore, ownership of means of production was never considered by him as a necessary condition of entrepreneurship. Quite the contrary, the sep aralion of ownership and management in modem corporate business enterprise provided one of the foundations of Keynes's view of investment decisions, in so far as it permitted to take into account the influence of financial markets upon the magnitude of investment. According to this view, the prevailing state of longterm investment expectations in modem capitalist economies is reflected in the activities of the Stock Exchange. Seen by the predominant weight of 'old' securities on financial markets, their prices, governed by speculation, indeed influence the prices at which 'new' securities can be floated in the new investment market. When the price of old securities is high, the price of new ones will tend to be high too. Thus, ability to float new securities at high prices will encourage investment in new projects on a scale which might seem extravagant in other circumstances. Therefore, speculative activity of financial agents will contribute to the instability of the marginal efficiency of capital, and thus, of investment. Conversely, as Asimakopulos (1990: 37) noticed, in Kalecki, 'both' entrepreneurs and rentiers were covered by the term "capitalists", while the income category "profits" included interest payments and dividends, as well as the retained earnings of firms'. This conception of agents is especially enlightened by Kalecki's famous notion of 'entrepreneurial capital': The limitation of the size of the firm by the availability of entrepreneurial capital goes to the very heart of the capitalist system. Many economists Richard Arena 419 assume, at least, in their abstract theories, a state of business democracy where anybody endowed with entrepreneurial ability can obtain capital for starting a business venture. This picture of the activities of the 'pure' entrepreneur is, to put it mildly, unrealistic. The most important prerequisite for becoming an entrepreneur is the ownership of capital. The above considerations are of great importance for the theory of determination of investment. One of the important factors of investment decisions is the accumulation of firms' capital out of current profits. (Kalecki, 1971: 109) Moreover, Kalecki did not accept the idea that financial speculation may exert noticeable effects on investment. For him, financial markets appear indeed to be imperfectly competitive. As he wrote, 'a joint stock company is not a "brotherhood of share-holders" but is managed by a controlling group of key shareholders, while the rest of the share holders do not differ from holders of bonds with a flexible rate of interest' (ibid.: 107). Now, the desire of these key shareholders to maintain their control of the company restrain their propensity to issue further shares and therefore contribute to limiting the issue of new shares. This limit is reinforced by the increasing weight of share issue costs in relation to the growing amount of capital. Obviously, the existence of key shareholders and their cautiousness in front of the resort to financial markets is already sufficient to avoid significant effects of speculation on investment. Moreover, as Steindl (1952: 139) noticed, Kalecki implicitly assimilated savings out of dividends as internal savings. More precisely, 'Mr. Kalecki... assumes that dividend receivers are, roughly speaking, potential investors in new shares, and other savers are not. Therefore, a market for new shares will be created to the extent, and only to the extent, to which saving out of dividends takes place' (ibid.: 138). In this case, obviously, the savings out of dividends would be equal to the new shares issues. This means that, if an issue of new shares is decided by managers or key shareholders, it will increase the 'entrepreneurial capital' of the company without any influence of financial speculation. Finally, Kalecki (1971: 107-8) stresses the possible means which firms may use in order to control both the ownership of capital and the conditions of gathering new saving, referring to the creation of 'portfolio companies' or to the purchasing of new shares by the key shareholders. A Second Important Difference between Kalecki's and Keynes's Approaches Lies in their Respective Characterizations of the 'Industrial' Factors of Investment Decisions The different formulations of the solution offered by Kalecki (1933, 1936, 1937, 1954) show a unity which appears clearly in the description of 'the principle of increasing risk' (ibid., 1937). According to this principle, the risk of investment is the more because the relation between planned investment and the magnitude of 420 Circuit and Post Ke.ynesian Investment Decisions entrepreneurial capital is high. This principle was, in a way, a synthesis between Kalecki 1933 and 1936. The Polish text of 1933 indeed showed that the volume of planned investment was positively related to the volume of entrepreneurial capital. A high level of entrepreneurial capital is first associated with a high ratio between this level and the level of investment, such that possible losses only represent a lower proportion of the wealth of the producer. Moreover, a high level of entrepreneurial capital provides a better guarantee for lenders. That means the risk obviously increases with the volume of investment, for a given magnitude of entrepreneurial capital. Such an analysis is a natural complement of the 1933 text, according to which investment was only related to the ratio of profit to the existing capital and therefore might be increased without limit. The comment on the General Theory published in 1936 helped Kalecki to introduce explicitly the influence of the rate of interest, which was absent in the 1933 text because it was considered as a function of the profit/capital ratio. It also lead Kalecki to take into account the expected rate of profit on investment which was assumed to be equal to the current one in 1933. As we know, Kalecki indeed rejected Keynes's view of the relation between the marginal efficiency of capital and the rate of interest. He criticized it especially because it did not provide an ex ante conception of investment decision. He interpreted the necessary difference between the expected marginal profitability and the rate of interest as a counterpart of the existence of an increasing risk implied by the liwel of the ratio between entrepreneurial capital and planned investment (Feiwell, 1989; Meacci, 1989; Patinkin, 1989). Concerning the nature of profit expectations, little can be found in Kalecki. The predominant implicit idea is, however, that views about the future are heavily conditioned by the present and recent past. This limited part given to uncertainty in the process of expectations formation may be emphasized in considering Kalecki's interpretation of the concept of marginal efficiency of capital. Kalecki indeed excludes that investors are able to expect the variation of this efficiency. He does not accept Keynes's view according to which the level of efficiency is the collective result of all firms' decisions. For Kalecki, the marginal efficiency is given for each particular entrepreneur and provides one the foundations of investment decisions. In the General Theory, Keynes did not neglect the notion of risk: Two types of risk affect the volume of investment which have not commonly been distinguish. The first is the entrepreneur's or borrower's risk and arises out of doubts in his own minds as to the probability/minds of his actually earning the prospective yield for which he hopes. If a man is venturing his own money, this is the only risk which is relevant. But where a system of borrowing and lending exists, by which I mean the granting of loans with a margin of real or personal security, a second type of risk is relevant and which we may call the lender's risk. (Keynes, 1936: 44) Richard Arena 421 However, it is clear that these notions of risk do not involve any explicit consideration of the various possible types of investment financement, as in Kalecki. We do not find in Keynes a specific analysis of the risk consequences implied by those types, the borrower's and the lender's risks describing only the uncertainty related to the outcome of investment as such. For Keynes, the risk was much more inherent in the investment decision and to uncertain long-term expectations than to the form of financial resources. The well-known chapter 12 of the General Theory and the Quarterly Journal of Economics article published in 1937 described vividly and in detail the reasons and the nature of this uncertainty. In this framework, it is interesting to compare the different ways according to which Keynes analyzes long-term and short-term expectations, the latter being considered in chapter 5 of the General Theory. In the short run, Keynes indeed accepts that expectations are prevalently determined by the immediate past; this assumption is not maintained in the long run, where extrapolative expectations become only a very particular technique in the case of strong uncertainty. A clear dissymmetry may therefore be pointed out here, between Kalecki and Keynes. Moreover, the instability of expected profitability is weaker in Kalecki than in Keynes and related to business cycles' mechanisms. Quite the reverse, in Keynes, where marginal efficiency of capital is strongly unstable because of the volability of the state of confidence. The gradual increase of financial operators who have no specific knowledge of investment decisions; the overemphasis of 'day-to-day fluctuations in the profits of existent investments, which are obviously of an ephemeral and non-significant character'; the weight of the 'mass psychology of a large number of ignorant individuals' and the behavior of experts who prefer 'to fail conventionaly than to succeed unconventionaly' contribute to reinforce variations in the level of investment. The Third Sensible Difference between Kalecki's and Keynes's Approaches Concerns the 'Financial' Factor in Investment Decisions, i.e. the Influence of the Rate of Interest The part played by the rate of interest is not considerable in Kalecki's formulation of investment decisions. Thus, in the 1933 text, Kalecki assumes that 'the' rate of interest is an increasing function of gross profitability, varies according to the general movements of other economic variables in business cycles and, therefore, can be eliminated from the equations of investment decisions. At the end of the 1930s, Kalecki's ideas on monetary theory evolved, presumably under the influence of the General Theory, and he distinguished two rates of interest, the short-term and the long-term one. In this framework, the major role was given to the short-term rate which contributed to the explanation of business cycles. On the other hand, the long-term rate was assumed to be constant according to the observed reality. This rate was therefore considered as a parameter and Kalecki never gave it the importance it had in Keynes's approach. 422 Circuit and Post Keynesian Investment Decisions This evolution is obviously related to a progressive change in the analysis of the rates of interest. In his 1933 work, Kalecki took into account what he called the 'market' rate of interest. However, little was said about it. One is invited to assume that the competitive forces on the money market determine the level of this rate, provided a given state of confidence and neglecting the possible impact of central bank policy. The appearance of two rates in Kalecki' s works at the end of the 1930s means that the author no longer reduced monetary functions to exchange but also considered money as a store of value. Three assets (money, bills and bonds) are distinguished by Kalecki. Money and bills differ, because money is perfectly liquid but does not yield any return, while bills are profitable but illiquid. Both assets, however, exhibit a common characteristic. The value of bills being fixed, both assets appear to be quite safe. Bills differ from bonds because the former offer possible capital losses, while the latter may imply income risks. Their common characteristics are illiquidity and profitability. These analogies and differences allow the possibility of two kinds of arbitrage. Holding money instead of bills is equivalent to what Kalecki calls a 'pure liquidity preference'. The price economic agents demand to renounce this liquidity is short-term interest. Purchasing bills instead of bonds reveals a risk aversion. That is the reason why bondholders receive long-term interest: this interest is the price agents demand to compensate risk aversion. These three assets being characterized, it is now possible to describe the determination of the two corresponding interest rates. For a given quantity of transactions, the short-term rate of interest is a decreasing function of money demand. It is therefore determined by the habits of economic agents, money supply and the state of confidence which governs pure liquidity preference. The level of the long-term rate of interest is also influenced by the money supply and liquidity preference but this influence is indirect, i.e. mediated through short-term financial operations. Time is necessary for both different substitutions, between short-term and long-term assets, to take place successively. The detailed explanation of these substitutions and the introduction of the principle of increasing risk into the financial area, allow Kalecki to express the long-term rate of interest as a function of the expected short-term rates of interest corresponding to the same period. This expression first reinforces the thesis of the long-run stability of the longterm rate of interest - variations of short-term rates of interest are always larger than long-term rates and have only limited effects. The relation between the longterm rate of interest and expected short-term rates also leads to the consideration of the short-term rate of interest as the basis of the general structure of the rates of interest. Therefore, the short-term rate of interest may be relevant for investment decisions. However, this is not the case. Kalecki considers that investment is selffinancing, whatever the level of the short-term rate of interest (Kalecki, 1954, ch. 6). Moreover, investment decisions are completely independent from the long-term rate of interest (Kregel, 1989, note 5: 204) which remains stable, and therefore cannot imply any effect on firms' behaviors. Richard Arena 423 Keynes's analysis of the determination of short-and long-term rates of interest, and on their effects on investment decisions, differs drastically from that of Kalecki. Before 1936, Keynes was already convinced that the short-term rate of interest did not playa significant part among the determinants of investment. Thus, in 1935, in his correspondence with Hawtrey, he emphasized the lowness of short-term interest charges in proportion to total costs and inferred from this observation the weak influence of short-term interest on business decisions. Keynes confirmed this view in the General Theory, using a different argument. He noticed that the short term rate of interest is easily controlled by the monetary authority, both because it is not difficult to produce a conviction that its policy will not greatly change in the very near future, and also because the possible loss is small compared with the running yield (unless it is approaching vanishing point). (Keynes, 1936) Therefore, in opposition to Kalecki's views, the significant rate becomes the long-term rate of interest, which monetary authorities cannot always control. The most obvious and sudden variations of the long-term rate do not come from changes in the money supply, but from changes in the liquidity preference function, as Keynes saw it. Thus, Keynes's view is clearly different from Kalecki's: according to Keynes, the long-term rate of interest is the essential one and plays a substantial role in the determination of investment decisions. The comparison between Kalecki's and Keynes's conceptions of investment decisions, respectively, is quite informative. On one side, investment decisions are planned by capitalists, taking into account the current level of productive capacity, firms' self-financing possibilities and the risk implied by the importance of borrowing in relation to the volume of 'entrepreneurial capital'; the rate of interest, either short term or long term, does not play any prevalent role and financial markets have scarcely any influence on industry. On the other side, investment decisions are taken by entrepreneurs according to the relative levels of the marginal efficiency of capital and the long-term rate of interest, considering the importance of the borrower's or entrepreneur's risk; the rate of interest thus exerts some influence on the investment decision mechanism, so that the volume of new capital depends on both 'industrial' and 'financial' factors. 2 MINSKY AND STEINDL: TWO POSSIBLE SYNTHESES The differences just exhibited between Kalecki's and Keynes's conceptions of investment decisions may be cancelled, or at least reduced, if we try to combine 'industrial' and 'financial' factors in a unique synthetic model. That is what Minsky and Steindl may be considered to have done. 424 Circuit and Post Keynesian Investment Decisions The Solution Offered by Steindl (1945, 1952) Steindl first defines what he calls 'the rate of profit earned on the entrepreneur's capital (after deduction of interest paid' (Steindl, 1990, p. 13). We may notice 7T, such that ~ (e - r) + r (1) where C is the entrepreneur's capital; I is the amount of investment of the firm; e is the rate of profit earned on this investment; r is the rate of interest. In the case of a private entrepreneur, entrepreneurial capital is defined, as in Kalecki: it corresponds to the entrepreneur's private capital. In the case of a joint stock company, Steindl's definition differs from Kalecki's: entrepreneurial capital is identified as the sum of ordinary share capital and capital reserves. This definition excludes savings out of dividends. The reason for this exclusion is the denial of Kalecki's assumption, according to which dividend receivers might be considered as potential investors in new shares. Thus, the issue of new shares does not necessarily mean a purchase of these shares by old shareholders. In other words, Steindl does not consider that the key shareholders of a given company have the same strong power of controlling new share issues as they had in Kalecki. In addition, Steindl stresses the importance and composition of 'outside savings' (Kalecki, 1954: 115-17). Therefore, he reintroduces the possible impact of financial markets on investment, through the behaviour of outside savers (ibid.: 143). From the above equation, (1), we can introduce the notion of risk premium. Indeed we know that to induce the entrepreneur to invest, the rate of profit must cover not only interest but also a magnitude which we can identify with the risk premium for a given cost of equipment. Now, let us define 80 as the standard deviation of profit expectations errors, and 8 as the standard deviation of profit 7T expectations errors. The risk premium is then equal to: (Steindl, 1990: 14-15) (2) In this framework, the size of the investment will be determined at the point of equalization between the marginal risk premium and the marginal rate of profit: W=/ 1 - j s (3) This point is the one which corresponds to the maximum difference between the rate of profit 7T and the risk premium. 2 The solution thus offered by Steindl, however simple it may appear, provides a possible synthesis between Kalecki's and Keynes's perspectives. It includes the principle of increasing risk and the effect of entrepreneurial capital, combining Richard Arena 425 them with the influence of the long-term rate of interest on investment decisions. Moreover, Steindl stresses the necessity for joint stock companies to resort to financial markets when they desire to fund investment. In this case, the 'state of the market' (Steindl, 1952: 139) must be taken into account and it clearly depends on portfolio behaviors of outside savers, i.e of financial agents. This fact obviously involves consideration of the impact of speculation on firms' profitability expectations. However, this impact is not emphasized by Steindl, who remains more Kaleckian than Keynesian from this point of view. Risk is still considered essentially from an industrial perspective. The Solution Offered by Minsky (1975) Minsky's solution provides a further step for a possible synthesis, referring to an explicit treatment of economic behaviors. Banks are first to be delineated in the scheme (Minsky, 1975, chapter 6). Their objective is to maximize their own profits. Therefore, they have to formulate expectations relative to the profitability performances of firms. Expectations are uncertain, as they were in Keynes. During the expansion phase of a business cycle, bankers' expectations are prevalently optimistic, such that the problem of the solvability of firms is not essential; that is why they agree to increase the ratio between commitments and reserve assets and to lend liquid resources to firms. During the recession phase, it is quite the reverse. Bankers' expectations become pessimistic and the solvability of firms becomes a fundamental problem; the ratio between commitments and reserves then decreases. Money supply therefore depends on the level of general economic activity. In other terms, it is clearly endogenous (ibid.: 76). Moreover, Minsky uses Kalecki's principle of increasing risk and Keynes's notion of lender's risk to characterize banks' behaviors. Now, the lender's risk increases when the relation between a firm's external finance and production becomes wider. Financial intermediaries are also considered (ibid.: 123). Their behaviors reinforce the weight of those of the bankers. For instance during the expansion phase of the cycle, they change the composition of their portfolios, preferring long-term assets which are more profitable to money or liquid money assets. Households and firms also have to decide the composition of their own port folios. Their choices concern four assets: money, liquid assets, bonds and capital goods. In compliance with traditional portfolio analysis, they will depend on the respective risk and profitability of assets. The cash-flow of a given asset is equal to q - c+ J. where q is the flow of future expected returns of the asset; c is the volume of liabilities necessary to purchase the planned volume of assets; and I is the liquidity premium (ibid.: 81). This portfolio choice concerns firms as well as households. Minsky explicitly considers the possibility of firms reallocating capital goods and liquid resources (ibid.: 88-9). 426 Circuit and Post Keynesian Investment Decisions In this framework, it is clear that the prices of various assets depend not only on the state of confidence but also on the volume of the money supply, since quantities and prices of assets are inversely related. Moreover, this volume is a function of bankers' expectations of firms' profitability, these expectations providing the basis of an investment supply price. The higher the level of investment, the higher is the recourse of firms to external finance. We have already noted that a high level of external finance increases banks' or lenders' risks. Therefore, to compensate this risk, banks lay higher costs on firms, for instance, through an increase in interest rates. The lender's risk thus implies an increasing capital-good supply price, as soon as the level of investment is superior to the self-financing possibilities of the firm (ibid., chapter 5). On the other hand, according to the principle of increasing risk applied to firms, the borrower's risk also increases with the level of investment and external finance. That is why, for levels of investment superior to the self-financing kind, the capital-good demand price decreases. It is then easy to draw Minsky's famous diagram (Figure 15.1) where the demand price of capital goods Pk is superior to the supply one PI, where IA is the self-financing level of investment, and t the final one (ibid.: 108). The respective slopes of the curves reflect the state of the economy and of agents' expectations. For instance, in the expansion phase of a business cycle, entrepreneurs and bankers are optimistic and do not focus on risk. Banks lend liquid resources easily and firms do not hesitate to use external finance. In this case, the two curves are more elastic, which means that a change in expectations always implies a variation in the level of investment (ibid.: 110-13). This approach has some similarities with Tobin's Q-theory, although it differs from it because of the treatment of risk and uncertainty. It might also be combined with the Eichner-Harcourt-Wood way of connecting investment decisions and long-term pricing strategies. Therefore, firms could use their market power to affect future streams of returns instead of foreseeing uncertainty in this respect. Taking the precisions into account, Minsky's approach provides a possible synthesis between Kalecki's and Keynes's views of investment decisions. On one side, variables such as the importance of entrepreneurs' capital and the capacity for self-financing are taken into account according to the principle of increasing PK PI Supply price curve PKr-~~----==~~~£----- Plr-------~~----~----~~ I i IA 1* Figure /5./ Demand price curve Internal funds curve I Richard Arena 427 risk. On the other side, the notions of borrowers' and lenders' risks and the influence of the state of expectations are also introduced into the scheme. Finally, Minsky's solution stresses the independence of the factors which determine the prices of capital and consumption goods, emphasizing the influence of money and financial conditions on the former. However, if Minsky's solution provides a possible synthesis for Post Keynesian authors, it is not certain whether it fits in with Circuit theory. We will now try to understand why. 3 CIRCUIT AND POST KEYNESIAN APPROACHES Minsky's view of investment decisions is acceptable to the Post Keynesian approach because it captures one of the essential features of the latter, that is, the integration of real and monetary determinants of investment. It was also noticed elsewhere (Kregel, 1987: 15-16) that Minsky's scheme is different from but perfectly compatible with Davidson's, based on a stock flow (Davidson, 1972; 1986), or with Kregel's scheme, inspired by chapters XVI and XVII of the General Theory (Kregel, 1982, 1988). We may be more sceptical about the possibility of Circuit theorists welcoming Minsky's scheme, as it stands, within their analytical framework. This hesitation might be considered as rather strange. According to Kregel: the circuit approach has [indeed] done much to reawaken interest in Keynes' monetary theory of production and to extend it in new directions. Yet, it appears to remain confined within the limits of the industrial circulation of Keynes' Treatise; it thus does not deal with the determination of investment goods prices nor can it deal with the differential determination to the prices of assets and liabilities introduced in the General Theory in terms of the marginal efficiency of capital and liquidity preference. Since the importance of money in Keynes' theory is intimately linked with these concepts it leads us to conclude that circuit theory needs further development in order to capture Keynes' views on the dominance of money over the real sector in a capitalist economy. (Kregel, 1987, summary of the paper: 11) A similar view seems to be defended by Lavoie in his very informative'synthesis' of the circuit theory: the main critique we may address to circuit theorists is that they were unable to formalize a presentation of the monetary economy they pretend to describe. What I mean is that in a monetary production economy, we might find that prices and employment are influenced by the monetary conditions, the rate of interest for instance, or its level as regards the general rate of profit of the economy, or by the level of indebtness of firms. (Lavoie, 1987: 92) 428 Circuit and Post Keynesian Investment Decisions In other words, circuit theory would not be misleading but incomplete. However, Minsky's scheme seems to provide exactly what is missing in this theory, according to Kregel and Lavoie. A first analysis of most circuit theory contributions seems to corroborate this view. Many of them do not consider financial circulation as such - giving only the conditions of logical compatibility between an achieved theory of industrial circulation and a theory of financial circulation - and thus still needs to be done. A good example of this tendency is provided by an interesting paper written by a French author, Vandevelde (1987). This offers one of the most complex versions of circuit theory, going as far to include financial markets and intermediaries and allowing for the possible appearance of hoarding. However, the paper does not show the presence of any theory of the determination of the rate of interest, of the prices of bonds or capital goods, or of equities. What is described is the macroeconomic network of money goods or assets flows among agents through markets or financial intermediaries. Different degrees of liquidity preference are considered but they are not analyzed directly, through their influence on prices or on the rate of interest. They are viewed indirectly, through different typical distributions of flows between money and bonds. The consequence of this theoretical prospect is that investment decisions as such are not considered at all. A more careful investigation of circuit theory literature may however change this initial impression. Circuit theory does not consider investment decisions or price formation processes, not so much because it is incomplete but, above all, because it has some logical reason for that. Some reasons are methodological. Thus, as Lavoie noticed, one of the basic ideas of circuit theory is'the emphasis put on the existence of pure macroeconomic laws, of structural relations independent from microeconomic agents behaviors' (Lavoie, 1987: 73). A good example of this type of'pure macroeconomic law' is given by the circuit interpretation of the macroeconomic relation between saving and investment, which is described as an identity independent of any level of national income or of any rate of interest (Friboulet, 1985;. Parguez, 1980). Within this framework, it is not really conceivable to describe portfolio or investment decisions, such as those retained by Minsky. Secondly, circuit theory - at least in Schmitt's version - carefully distinguishes between'virtual' and'realized magnitudes' (Schmitt, 1971). The former are only defined in agents' minds and have no social existence. Therefore, the theory cannot take hold of them. The latter are socially observable by anybody. They are expressed in book-keeping figures; and so an objective discourse could be developed round them. This conception is obviously related to circuit conceptions of economic measurement (Schmitt, 1975) and time (ibid., 1984; Cencini, 1984). If we accept this, we are led to exclude expectations or decision formation processes from economic analysis and to focus on realized magnitudes. These expectations or decisions are indeed formed in the world of'virtual magnitudes' and the economist only takes into account'realized' ones. Richard Arena 429 We come now to the third reason which might explain some of the scepticism concerning Minsky's view of investment decisions. This last reason is more important than both the previous ones because of its theoretical contents. We can sum up the main argument, stating that, for circuit theory, neither the liquidity preference nor the financial rate of interest can affect investment decisions. Let us develop this argument, the merit of which is due to Graziani. In circuit theory, transactions which occur among firms are not analyzed as transactions occurring between firms and consumers. What is assumed is that firms, as a whole, always produce what they need. Therefore, as far as they are able to obtain money from the banks, firms can perform transactions among themselves, with no risk of over or underproduction (Graziani, 1988a:. xxi-xxii). In the same way, in as much as they obtain cash for their capital goods sales, firms pay money back to the banks so that they never hold liquid balances during the circuit. Therefore, the only possible final use of firms' income is the purchase of consumers or of capital goods. In other words, for firms, the sole form of saving is capital investment (ibid.: xxi). The sole agents who can hold liquid balances are therefore the wage earners. But this is not the only possible use of their income. They can also buy consumers goods or financial assets. If wage earners spend their entire income on consumer goods and/or on assets, they reveal no liquidity preference. Moreover, firms, considered as a whole, are likely to receive the totality of the incomes they distributed to wage earners, in counterpart either of consumer goods or of bonds' sales. They are therefore able to repay the banks entirely. In this case, the rate of interest plays no role for firms, whatever its level may be. We recognize here a Kaleckian conception: To put it in a more concrete fashion: if some capitalists increase their investment by using .... bank credit, the spending of the amounts in question will cause equal amounts of saved profits to accumulate as bank deposits. The investing capitalists will thus find it possible to float bonds to the same extent and thus to repay the bank credits. (Kalecki, 1971: 84) Kregel, commenting on this same passage, notices: Since the deposits accumulate within the'period', the bonds could also be floated, within the period without effect on bonds prices, and thus on interest rates. (Kregel, 1989: 204) A second possibility may also occur if wage earners do not spend their whole income, but hold liquid balances. In this case, firms will be led either to decrease their prices in a Kaleckian way or to face corresponding unsold inventories. Both decisions cannot prevent them, however, from being unable to pay back the banks. Two solutions may then need to be faced. The first is to demand money on 430 Circuit and Post Keynesian Investment Decisions the financial markets and run into medium- or long-term debt in order to pay back to the banks the short-term finance. However, this new money demand might increase the rate of interest, especially if, as in Keynes's General Theory, the money supply is assumed to be constant. If firms do not want to pay a higher rate of interest, they must face the second solution and borrow money from banks, taking on medium- or long-term bank loans. Such a scheme implies that, in circuit theory, investment decisions are independent from the conditions of the financial market. This market exerts only an influence on the relations between banks and firms and on the possibility for firms to avoid an increasing indebtness as regards banks. Therefore, the financial market is not a prevailing factor of capital accumulation. However, determining the firms indebtness, it governs net industrial profit on one side and the existing money supply on the other side. (Graziani, 1988a: xxxviii) Indeed, even if the rate of interest increases because of an extra money demand from firms, firms can always increase their prices in order to cover this extra cost. The result might be inflationary but would not exert any influence upon profits. However, if the rate of interest does not exert any influence on investment decisions, what are their prevailing determinants? Do we have to go back to'animal spirits'? Some circuit theory papers might suggest this type of direction. Parguez, for instance, uses the expression'bet' to characterize investment decisions (Parguez, 1987: 778-80). However, as well as Graziani, he also accepts the idea that factors like the state of competition or the technological conditions of production provide constraints for investment decisions (Graziani, 1988a: xix, xxv, xxxiii). But the most important economic constraint is obviously the nature of the contractual relations between banks and firms (ibid.: xxxvii; Messori, 1985). These relations are crucial because they express the relative strengths of banks and firms, as they appear through the level of the bank rate of interest. Now, what about the notion of liquidity preference and the role of the financial rate of interest, the importance of which is central to Keynes's General Theory? They still have a part to play. However, this part is perfectly independent from firms' investment. It is only related to wage earners' saving preferences, when they have to choose between hoarding and assets-holding. The choice is not negligible at all. Its main consequence might be a disequilibrium in industrial circulation and lead to the necessity for firms to borrow long- or medium-term loans, as mentioned earlier. This conception of investment decisions is not so different from that of Asimakopulos, who strongly criticized the notion of marginal efficiency of capital (Asimakopulos, 1971), replaced it by the double-sided relationship between investment and profits (ibid.), and considered the rate of interest as a second-order factor in investment decisions (ibid. 1990: 81 ).3 Richard Arena 431 The differences between Post Keynesian and circuit approaches concerning investment decisions are thus more clearly distinguished. If we focus on the theoretical ones, we may easily consider that circuit theory is more Kaleckian than Post Keynesian. To go further, the theme of investment decision provides us with a good opportunity to stress the general differences between Post Keynesian and circuit approaches. First, both Kalecki's theories and circuit approaches consider the social framework of the economy as a whole, including complementary social groups, while the Post Keynesian view is not so classically oriented. According to the former tradition, the crucial distinction is between bankers and entrepreneurs, the latter being identified with capitalists: In circuit theory, the analysis is founded on a net distinction between banks and firms sectors. According to such a distinction, banks are characterized as institutions which produce liquidity but do not use it, while firms use liquidity without being able to produce it. This is a typical distinction of classical economic analysis, which was fully used by Marx. (Graziani, 1988b: 98) Contrary to this, the Post Keynesian approach does not consider banks as a specific group of agents. Banks belong to the group of firms, even if they are firms of a specific type. Banks and firms have an analogous purpose: the maximization of their incomes. Secondly, in Kalecki as well as in circuit theory, financial markets do not exert any noticeable influence on investment. Investment decisions are formed independently of the Stock Exchange expectations or from the variations of the financial rate of interest. This view is in clear contrast to the Post Keynesian one, according to which investment decisions and changes in the level of this rate of interest are clearly interdependent. Thirdly, the'short-term' rate of interest plays a major part in Kalecki's as well as in circuit theory. The level of this rate is a consequence of the bargaining between bankers and firms. It governs the distribution of income between banks and industrial capital. Conversely, we have already seen that, in Keynes as well as in Post Keynesian theory, this rate is secondary. These analogies between circuit and Kaleckian theories clearly lead to a common vision of capitalist investment. This vision entails a predominance of industrial capital over the financial kind and attributes a major role to the banks, to the monetary market and to the endogenity of money, in compliance with the continental tradition. Opposed to this, Post Keynesian theory stresses the danger of the interference of speculation on productive investment, due to the growing part attributed to the modem development of financial markets, in compliance with the Anglo-Saxon tradition. These remarks suggest that both circuit and Post Keynesian approaches may be more complementary than analogous. In this respect, and paradoxically, the crux of the debate between Post Keynesian and Circuit and Post Keynesian Investment Decisions 432 circuit theorists is perhaps more institutional than theoretical. It would not, however, be inherent in the establishment of a more consistent synthesis but rather in the building of a more general approach, able to explain a larger range of 'stylized facts'. Notes * I. 2. 3. I wish to thank Augusto Graziani, Edward Nell and the participants to the conference on Monetary Theory: Post Keynesian and Circulation Approaches for their valuable criticisms, comments and suggestions on the first draft of this chapter. Remaining mistakes and confusion are mine alone. The adjectives 'industrial' and 'financial' have the same meanings as in Keynes's Treatise on Money, when the author defines industrial and financial circulations (Keynes, vol, I, 1973). Steindl (1990: 15-19) also analyzes the more complex cases where C is permitted to change, rand e are dependent of the size of investment. One of the last paper of Asimakopulos (1990) exhibited definite sympathy with circuit theory. References Arena, R. (1988),'Keynes apres Lucas: quelques enseignements recents de la macro economie moneta ire' , Economies et Societes, I. Asimakopulos, A. (1971),'The Determination of Investment in Keynes's Model', Canadian Journal of Economics, IV (3), August. Asimakopulos, A. (I983),'Kalecki and Keynes on Finance, Investment and Saving', Cambridge Journal of Economics, 7, September. Asimakopulos, A. (l986),'Finance, Liquidity, Saving and Investment', Journal of PostKeynesian Economics, 9 (I). Asimakopulos, A. (l990),'Investment, Finance, Saving and Profits: A Kaleckian Approach to the Dynamic Circuit', Economies et Societes, Series Monnaie et Production, 6, February. Ccncini, A. (1984), Time and the Macroeconomic Analysis of Income (London: Frances Pinter). Davidson, P. (1972), Money and the Real World (London: Macmillan). Davidson, P. (l986),'Finance, Funding, Saving and Investment', Journal of Post Keynesian Economics, 9 (I). Feiwell, G. (1989), 'The Legacies of Kalecki and Keynes', in M. Sebastiani (ed.), Kalecki's Relevance Today (London: Macmillan). Friboulet, J.J. (1985), 'Le Traiti de la monnaie et l'inflation d'equiIibre', in F. Poulon (ed.), Les Ecrits de Keynes (Paris: Dunod). Graziani, A. (l988a),'11 Circuito Monetario', in M. Messori (ed.), Moneta e Produzione (Milan: Einaudi). Graziani, A. (I988b),'La Teoria del Circuito et la Teoria Generale di Keynes', in M. Messori (ed.), Moneta e Produzione (Milan: Einaudi). Kalecki, M. (1933), Proba teor;; koniun-ktury (Warsaw), partial Italian translation of A. Chilosi (1979) (ed.), Kalecki (Bologne: II Mulino). Kalecki, M. (I936),'Ka\ecki's Review of Keynes' General Theory', translation by F. Targetti and B. Kinda-Mass, in Australian Economic Papers, 21, December. Richard Arena 433 Kalecki, M. (1937),'The Principle oflncreasing Risk', Economica, 3. Kalecki, M. (1954), Theory of Economic Dynamics (London: Al\en & Unwin). Kalecki, M. (1971), Selected Essays on the Dynamics of the Capitalist Economy, 1933-1970 (Cambridge: Cambridge University Press). Keynes, J.M. (1936), The General Theory of Employment Interest and Money (New York: Harcourt, Brace). Keynes, J.M. (1973), The Treatise on Money, vol. I, in Collected Writings of John Maynard Keynes, vol. V (London: Macmillan). Kregel, J.A. (1982),'Monnaie, anticipations et prix relatifs', Economie Appliquee, 35 (3). Kregel, lA. (1987),'Shylock and Hamlet or are there Bulls and Bears in the Circuit?', Economies et Socieres, 9. Kregel, lA. (1988),'Rate of Return Over Cost and Over Rates oflnterest: Fisher, Sraffa and the Development of the General Theory', Cahiers d'Economie Politique, 13-14. Kregel, J.A. (I 989),'Savings, Investment and Finance in Kalecki's theory', in M. Sebastiani (ed.), Kalecki's Relevance Today (London: Macmillan). Lavoie, M. (l987),'Monnaie et production: une synth~se de la th~orie du circuit', Economies et Societes, 21(a), 65-101. Malinvaud, E. (1981), Theorie macroeconomique, vol. I: Comportements, croissance (Paris: Dunod). Meacci, F. (1989),'The principle oflncreasing Risk Versus the Marginal Efficiency', in M. Sebastiani (ed.), Kalecki's Relevance Today (London: Macmillan). Messori, M. (1985),'Le circuit de la monnaie: acquis et probl~mes non r~solus', in R. Arena and A. Graziani (eds), Production, monnaie et circulation (Paris: Presses Universitaires de France). Minsky, H. (1975), John Maynard Keynes (London: Macmillan). Parguez, A. (l980),'Profit, ~pargne, investissement: ~I~ments pour une th~orie mon6taire du profit', Economie Appliquee, 2. Parguez, A. (1987),'Introduction 1I1'6conomie de rentiers', Economies et Societes, 9. Patinkin, D. (1989),'Michael Kalecki and the General Theory', in M. Sebastiani (ed.), Kalecki's Relevance Today (London: Macmillan). Schmitt, B. (1971), L'analyse macroeconomique des revenus (Paris: Dal\oz). Schmitt, B. (1975), Theorie unitaire de la monnaie nationale et internationale (Albeuve: Castel\a). Schmitt, B. (1984), Inflation, chOmage et malformations du capital (Paris/Albeuve: Economica, Castel\a). Steindl, J. (1945),'Capitalist Enterprise and Risk', Oxford Economic Papers, 7 (March); republished and revised in J. Steindl (1990), Economic Papers, 1941-1988 (London: Macmillan). Steindl, J. (1952), Maturity and Stagnation in American Capitalism (New York and London: Monthly Review Press). Vandevelde, F. (1987),'La th~saurisation dans Ie circuit de I'~conomie mon~taire de production', Economies et Societes, 9. Part IV Endogenous~oney Introduction to Part IV The idea that the supply of money adapts to 'the needs of trade', and therefore is in some respects at least, beyond the control of authorities, is central to both the PK and the CA visions. But there are a number of different ways of conceptualizing these ideas, and it may also be the case that the monetary system functions differently in different historical periods. A. THE RELEVANCE OF THE INSTITUTIONAL FRAMEWORK Wray in Chapter 16 begins with an analysis of the nature and functions of money, arguing that money is a form of debt inherently enmeshed in uncertainty. Monetary contracts are forward contracts with interest built in. Credit money is more fundamental than commodity money. Five stages in the development of monetary institutions can usefully be distinguished. The circuit in a fully developed capitalist system can be usefully analyzed in a two-sector model. The shortterm credit will equal the total wage bill, which will circulate the consumption goods. Either profits are realized in physical terms or investment goods purchasers will sell long-term debt to holders of short-term debt. Interest creates complications, as does the fact that production is continuous. Adding a banking sector allows for a simple model to exhibit the relationships between the rates of interest on loans and on deposits, and the rates of growth of loans and deposits. This provides a basis for an analysis of the way credit money is related to a lender of last resort. The chapter concludes with a comparison of PK and CA, suggesting that there is room in circuit analysis for liquidity preference, while PK needs to think in circuit terms. Renversez in Chapter 17 focuses on the role of the Central Bank in a financial markets' economy and in an overdraft economy. In the former case, the open monetary market, articulated with the capital markets, is the preferred site of intervention for the Central Bank's regulating activities, which are discretionary. In the latter case, by contrast, the financing of the economy is basically provided through credit, and the banks continually tum, as a last resort, to the definitive refinancing by the Central Bank. As a. consequence, the intervention of the Central Bank is compulsory; the contribution of the liquidity of the monetary system to the economy, resulting from the primary debt of the productive sector, has an endogenous origin. 437 438 Introduction to Part IV B. EXPLANATIONS OF ENDOGENEITY: ACCOMMODATION OR STRUCTURE? Pollin in Chapter 18 begins by noting the extent to which the financial system has changed in recent years. These changes have not been reflected in mainstream theory. Yet they would fit well with one tradition of monetary thought - the tradition of credit money associated with Tooke. Recently Post Keynesians have revived this approach in two different but related theories of endogenous money, 'accommodationist' and ·structuralist'. Chapter 18 analyzes these and identifies significant differences. Recent empirical work is discussed; while not decisive, it appears that trends favor the structuralist approach. It might be that in a certain period the system appeared to work according to the accommodationist model, while later it is better described by the structuralist approach. But the structuralist model operates at a deeper level of insight into the system. Palley in Chapter 19 seeks to extend the endogenous money approach to cover finance more generally. The theory of endogenous money focusses too closely on the banking sector, ignoring other financial activities. The contrast between the accommodationist and the structuralist approaches suggests that the latter can be developed to explore the ways in which the entire financial system responds to changes in the level of activity. In periods of liquidity shortage credit can replace money as the medium of exchange. The purpose of Lavoie's Chapter 20 is to identify a few general guidelines around which Post Keynesians could agree when it comes to monetary policy recommendations. Lavoie distinguishes between cost-inflation, due to conflicting claims on the surplus, and demand-inflation. In the former case, which is considered by Post Keynesians as the major type of inflation, monetary policy should aim at fixing a fair rate of interest, consistent with the prevailing rate of growth of productivity and with the distributive shares of rentiers and workers. Fighting inflation through increases in interest rates may only lead to higher prices if one accepts the full-cost pricing model. In the exceptional case of demand-inflation, monetary policy should take the shape of direct credit controls. A. The Relevance of the Institutional Framework 16 Money in the Circular Flow L. Randall Wray'" INTRODUCTION Money plays an important role in both the Post Keynesian and circuit theory approaches. In the circuit approach. macroeconomic identities and the logic of circular flows are emphasized. In this method. all values are in nominal terms; thus. money is analyzed according to the role it plays in these nominally valued circular flows. In contrast, most Post Keynesians have emphasized money as a stock and have paid particular attention to the impact rising liquidity preference has on circular flows. Some Post Keynesians. however. have also recognized that as all spending must be financed. the money supply must expand endogenously to finance a growing circular flow. In addition. Post Keynesians have tended to pay more attention to individual decision-making than have those who adopt the circuit approach. The most important difference between the two approaches. however. is probably the treatment of uncertainty. Because Post Keynesians are particularly concerned with ex ante decision-making. uncertainty plays a major role in their view. On the other hand. circuitistes tend to focus on ex post aggregate identities. where uncertainty plays no role. Thus. the different treatments of money in the two approaches can be attributed at least in part to a difference in emphasis over where analysis should begin. This chapter will address several issues which have not been adequately treated by either approach. First. we will examine the nature of money: what is money. why did money originate. and why is money used? Unfortunately. Post Keynesians have generally taken current financial institutions as given and have failed systematically to examine the nature of money. Next we will compare and contrast credit money. commodity money and fiat money. Both Post Keynesians and circuitistes have emphasized that credit money is the most important form of money in capitalist economies. but have not provided an analysis of the reasons for the coexistence of credit money with other forms of money. In our analysis. it will be useful to categorize the development of monetary systems according to five stages. While others have also used a 'stages' approach to analyses of the evolution of the financial system. the approach taken here will be quite different. I We will also discuss why all monetary circuits must include interest. and why inclusion of interest necessarily leads to a logic of accumulation. Thus. because money is included in a circular flow, the nominal value of the flow must grow: accumulation must be in money 440 L. Randall Wray 441 form. However, uncertainty provides a role for expectations, which means that accumulation will not be steady; therefore, crises are possible. This analysis will help to synthesize the Post Keynesian and circuit approaches. 2 WHAT IS MONEY? Money is frequently identified with the physical objects which sometime function as money. Thus, it is sometimes said that 'gold is money', or that 'demand deposits are money'. This then leads to a focus on several of the uses to which these forms of money are put: 'medium of exchange'; 'store of value'; and so on. This approach, however, cannot uncover the fundamental characteristics of money and the necessary conditions underlying the development of money, as well as those conditions which induce economic agents to use money. As Schmitt and Greppi (this volume) rightly argue, money plays no important role in orthodox analysis - precisely because orthodoxy has ignored those factors which give rise to the use of money. Money arises out of a social relation; it exists only in societies which take a specific social form; and it is used only in a world which operates in real, historical time. First, money is debt; it represents one type of contractual liability of one party to another.2 Money is created as one independent agent delivers something of value today to another independent agent who promises to deliver a welldefined value tomorrow. As such, monetary contracts always exist in historical time and indicate a future obligation in return for something delivered today.3 Keynes (1982: 252) recognized that money is most importantly a unit of account in which debts are written: 'Now for most important social and economic purposes what matters is the money of account; for it is the money of account which is the subject of contract'. Later, as will be discussed in a moment, money functions as the means of payment, or means of contractual settlement. As Keynes (1971b, p. 3) argued, 'Money itself, namely that by delivery of which debt contracts and price contracts are discharged . .. derives its character from its relationship to the money of account, since the debts and prices must first have been expressed in terms of the latter'. This also gives rise to the 'store of value' function of money, for money reserves ensure that one has a liquid position, defined as the ability to meet contmctual obligations as they come due (Davidson, this volume). Still later, money evolves its 'medium of exchange' function as markets are developed. Such monetary contracts cannot exist unless private property is recognized. As Heinsohn and Steiger (1989: 193) argue, 'With the establishment of private property, we at once have the elements of a money economy'. In the absence of private property, one may seize objects of value without entering into a contractual obligation to deliver value later. 4 Similarly, independence of agents is essential so that they may enter into such contracts.s 442 Money in the Circular Flow Independence of agents also generates existential uncertainty of a form which does not exist in communal societies. Of course, tribal societies and the individual members of these societies face great uncertainty: wiU nature provide sufficient means of subsistence? WiU the neighboring tribe attack this year? Will I be kiUed by a bear? However, this type of uncertainty is fundamentaHy different from that faced by the independent agent operating in a private property society. By existential uncertainty, I mean the kind of uncertainty referred to by Keynes: By 'uncertain' knowledge, let me explain, I do not mean merely to distinguish what is known for certain from what is only probable. The game of roulette is not subject, in this sense, to uncertainty; nor is the prospect of a Victory bond being drawn. Or, again, the expectation of life is only slightly uncertain. Even the weather is only moderately uncertain. The sense in which I am using the term is that in which the prospect of a European war is uncertain, or the price of copper and the rate of interest twenty years hence, or the obsolescence of a new invention, or the position of private wealth owners in the social system in 1970. About these matters there is no scientific basis on which to form any calculable probability whatever. We simply do not know. (Keynes, 1987: 113-14) In a communal society, many of the risks are shared, and, in some sense, are calculable. Even the individual risks (will the bear eat me?) are calculable in a probability sense. On the other hand, in a private property economy, the risks are individually borne and are fundamentally uncertain for reasons to be discussed below. For this reason, monetary contracts, which exist only in private property economies, always involve existential uncertainty.6 The agent who enters into a forward contract that delivers something of value today and accepts a promise of another party to deliver something of value tomorrow cannot know whether the obligation wiU be met. Neither can he know the hardships he will face by surrendering use of the object during the period of the contract - even if the contractual obligation is met. 1 Similarly, the agent who promises to deliver something of value 'tomorrow' is uncertain about the hardship he wiU face in attempting to meet this obligation. Because such contracts entail uncertainty, they always stipulate 'value today for more value later'. That is, interest is always built into forward contracts. The interest rate will depend on the perception of uncertainty faced by the agent accepting the IOU (Heinsohn and Steiger, 1989). It may be useful to break the history of the evolution of financial systems into five 'stages'. This is, of course, merely meant to ease exposition and is not meant to suggest that divisions between actual, historical stages are clear-cut.8 In the first stage, forward contracts include interest, but no universal standard of measure has arisen. In the second stage, a universal equivalent or debt numeraire has been developed, with money emerging as a unit of account. Thus, loans of physical objects may be repaid with other objects valued in money terms. In the third stage, loans may be made directly in the universal equivalent. At this stage, money begins to operate directly as a means of payment and as a medium of exchange. Through the' third stage, money is exclusively credit money, created L. Randall Wray 443 through private forward contracts. In the fourth stage, commodity money is finally developed, while fiat money is developed during the fifth stage. Thus, our modem financial system in which private credit money is made convertible into commodity money or government fiat money, and in which government guarantees and reserves of commodity money or government fiat money back up private credit money, is the result of a long period of evolution. At first, forward contracts took the form of 'two bushels of wheat today, for three bushels of wheat at harvest', or of 'one cow today, for a calf at the end of each calving season and return of the cow in three years'. In such cases, the value delivered 'today' took the form of a physical object and 'interest' was often generated by the natural, physical productivity of the object loaned. 9 The owner alienated his property and expected to receive interest to compensate for the uncertainty he now faced regarding loss of use and probability of its retum. IO At this first stage, a money of account had not been developed. Furthermore, forward contracts essentially consisted of a two party, inalienable IOU. Greater flexibility and lower uncertainty were allowed when contracts came to be written in terms of a universal equivalent during stage two. For example, the first universal equivalent used in monetary contracts appears to have been wheat, and later barley, grain. I I The reasons for this development are obvious: grain is divisible into convenient units; it is reasonably storable; and it is of fairly uniform quality and size (barley supplanted wheat because its grains were more uniform). Thus, the loan of a cow could be repaid in its wheat equivalent. At this stage, money has developed as a social relation between 'borrower' and 'lender' measured in terms of a universally recognized measure of social value. Of course, neither cows nor wheat are money - rather, money is the contractual obligation, or IOU, of the borrower who promises to deliver so many units of the universal social measure of value. The physical form of payment may be cows or wheat - but these clearly are not money, for the quantity of these to be delivered will depend upon their monetary value. 12 In stage three, repayment of debt occurs directly in the form of money functioning as a means of payment. At this stage, monetary IOUs become alienable: they may now circulate among third parties, acting as stores of value and means of payment for individuals unconnected to the forward contract in which the money was created. Finally, as we shall see, loans may be made directly in the form of money - in which one alienates an IOU through a loan to a third party such that the 'borrower' accepts an IOU and creates his own IOU. Of the functions of money, it is the unit of account function - or the terms in which debts are measured - which was initially of primary importance. This, of course, runs counter to the orthodox story in which money was first used to replace barter. That is, orthodoxy imagines a market economy based on barter into which money was introduced to facilitate exchange (Levine, 1983). Actually, money existed before the development of markets, and barter economies have never existed (except as emergency or temporary behaviour).13 Indeed, it was the development of the money of account which then led to the development of markets and to the use of money as a medium of exchange. 444 Money in the Circular Flow The explanation is fairly simple: as money became the universal measure of value, possession of money (and objects of nominal value) became the socially recognized form of wealth. Accumulation of money - that is, of promises to pay reduced existential uncertainty and reduced the probability that one would have to borrow from others. Since forward contracts always involve interest, those who issued IOUs would find that their contractual obligations continually grew, while those who held IOUs would find their nominal wealth expanding. If borrowers had issued IOUs to obtain use of a naturally productive object (say, a cow), interest commitments could be met due to nature's fecundity as long as those who accepted these IOUs wished payment in the form nature provided (say, calves). For three reasons, however, the development of monetary contracts tended to orient production toward sale on markets: monetary contracts in the form of the universal equivalent provided greater flexibility but also meant that the loan of an object would have to be paid not with the 'natural' product, but with the universal equivalent (a cow loan must be repaid in idealized wheat terms); the loan of nonproductive objects (say, beefsteak) would not naturally generate interest, so the borrower was forced to engage in production to meet contractual commitments which could be met, for example, by producing for market; and, most importantly, those who lost their private property (in some cases, because they were unable to meet debt commitments) could work for money wages to earn the universal equivalent which would purchase necessities on the market. Thus, the development of monetary debts oriented production toward the market, and the development of a class of property-less workers generated market demand. During this stage, stage three, money came to be used widely as a medium of exchange. With the development of large-scale markets, loans could finally take the form of direct loans of money. That is, rather than loaning an object with use value in exchange for a monetary IOU, loans could take the form of 'money now for more money later', in which a third party IOU is loaned as part of a forward contract between two parties. Alternatively, bank 'intermediation' becomes possible: a financial institution may make a loan (accept an IOU) by issuing an IOU, used by the 'borrower' as a medium of exchange, accepted as payment by a seller. It also means that accumulation of money becomes the supreme goal, with money becoming the ultimate store of value. It is at this stage - with the development of markets and loans of money - that money circulates widely among third parties (those unrelated to the original contract in which the money was created) as the medium of exchange and the means of payment. As we shall explore in the next section, commodity money was developed in the fourth stage, while fiat money was developed in the fifth stage. This allowed the development of a 'pyramidal' financial structure in which forward contracts written in terms of the money of account are settled through the use of money which is issued by parties higher in the pyramid (Foley, 1989). Thus, non-banking firms settle accounts through the use of bank money acting as a means of payment, while banks settle acounts through the use of government fiat money. This does not mean L. Randall Wray 445 that non-bank finns cannot create money, or that non-bank money cannot be used as a medium of exchange, means of payment, or store of value. Indeed, bank money and the pyramidal structure based on fiat money are relatively recent developments, occurring many dozens of centuries after the development of money as a unit of account, store of value, medium of exchange, and means of payment. 14 It is important to emphasize that money is created as part of a forward contract between two parties, and that it is destroyed when this contract is retired through the use of a third party liability or through the use of commodity and fiat money. The development of the pyramidal structure with the central bank at the apex came much later as a means of dealing with two chronic problems: lack of state purchasing power and instability of a financial system based on private credit money. 3 CREDIT MONEY, COMMODITY MONEY AND FIAT MONEY In orthodox theory, commodity money is the 'natural' fonn of money. Economic agents discovered they could use shells or other physical objects to reduce transactions costs in exchange. Gradually, for reasons which are not always made clear, commodity money took the universal fonn of gold and silver coins. IS The value of this commodity money is supposedly detennined by supply and demand; thus, an excess supply of money will cause its value to fall and the money prices of all other commodities to rise. Governments discovered that they could somehow fool the public into accepting debased coin (that is, coins with little gold or silver) or even purely nominal paper money. At this point, governments obtained the ability to issue fiat money - whose nominal value was detennined by decree. However, if governments issued 'too much' fiat money, inflation would increase. In other words, while governments could control the nominal quantity of commodity or fiat money, they could not control the real quantity of commodity or fiat money. Finally, credit money is either ignored in orthodox theory, or is treated as credit, which is supposedly fundamentally different from money, and is treated as if its quantity were closely regulated by the quantities of commodity and fiat money. This leads to the fonnulation of the 'deposit multiplier' found in every money and banking text. 16 In reality, credit money is the 'natural' fonn of money: money as a privately issued IOU. With the development of the money of account, these IOUs were at first denominated in terms of wheat, and later in terms of a precious metal wheat equivalent - but neither wheat nor gold was (or is) money:7 Money represents a social relation between borrower and lender, the tenns of which gradually became standardized according to some universally accepted measure. Much later, during stage four, precious metals were coined to provide a somewhat exogenous source of money in addition to the endogenously generated credit money. Purely nominal coins were an even later development, coming in stage five, long after the development of credit money. 446 Money in the Circular Flow As I will discuss below, commodity money was developed, at least in part, to deal with the problem of the possibility of default on credit money. The commodity money would be denominated in the universal unit of account, which evolved out of privately created credit money. Even after the development of commodity money, credit money is still the most important form of money in private property economies, created as loans are made. In such economies, credit money is used as the medium of exchange and as the means of payment (in which third party IOUs are used to settle debts); even where commodity money exists, credit money circulates alongside it. For example, McIntosh finds that in early (1300-1600) England, Any two people might build up a number of outstanding debts to each other. As long as goodwill between the individuals remained firm, the balances could go uncollected for years. When the parties chose to settle on an amicable basis, they normally named auditors who totalled all current debts or deliveries and determined the sum which had to be paid to clear the slate. (McIntosh, 1988: 561) Indeed, virtually all sales were made on the basis of credit money acting as the medium of exchange (McIntosh, 1988: 560). Consumers would later settle accounts with coin, but merchants would use coins only to settle net debits. The evidence shows that debts might be carried for months (and even for years) on the accounts of merchants before being finally settled. Thus, in monetary economies, money is 'naturally' credit money, while commodity money represents a 'special' form of money. 18 It is worth noting that in all cases (Babylonia, Italian city states, Western Europe), 'deposit banking' develops after the public has become accustomed to using credit money. For example, banks in Western Europe operated primarily on the basis of banknotes until the nineteenth century, rather than as deposit banks. Knapp (1924) argues that deposit banking cannot evolve until the public has developed the 'banking habit', that is, the habit of accepting banknotes. Thus, rather than acting as intermediaries from 'depositors' to 'borrowers', early banks made 'loans' by issuing banknotes; that is, they financed their position in assets by creating liabilities. As I will discuss below, banks would then hold reserves (notes of other banks, assets that could be liquidated quickly, and commodity money) so their notes could be redeemed. Only later did deposit banking develop, which casts further doubt upon the orthodox story of the development of banking, in which the goldsmith discovers a deposit multiplier based on reserves of commodity money (Wray, 1990). Government fiat money is in all cases a relatively recent development coming in stage five, long after the public had become used to credit money. Indeed, throughout history most governments were very limited in their ability to gain purchasing power by issuing fiat money (or, more generally, issuing IOUs). Where government attempts to gain purchasing power by issuing debased coin were met by rising prices, this is evidence that the government had not attained the power to issue fiat money. Neither an increase in the supply of credit money nor an increase in the supply of fiat money leads directly to an increase in prices. L. Randall Wray 447 When governments are not able to issue fiat money, the value of government money will be determined by the quantity of commodity money embodied within it. This was recognized by Keynes, who argued: When ... a coin is no more than a quantity of bullion, of which the stamp may certify the quality and indicate the quantity, [it] ... will not circulate except for its bullion value. In this elementary stage the expedient of debasement is not available. It cannot appear, until with the development of contract the conception of a money of account has emerged, and the coins issued by a state have acquired the character of legal tender and enjoy a cours force as the legal discharge of obligations calculated in this money of account. It is at this stage that money, in the sense in which we understand it, makes its entry into human institutions. (Keynes, 1982: 226) Thus, when government money is merely commodity money, debasement cannot generate additional purchasing power as those who sell to the government become quite adept at detecting debasement and merely raise prices in terms of government money. 19 I will postpone further discussion of commodity money and of government money until later. The next several sections will focus only on credit money, as this is the important form of money in private property economies. 2o 4 MONEY, INTEREST AND ACCUMULATION In orthodox theory, money is traditionally taken to be barren. This is the fundamental reason for its insignificant role in orthodoxy. Attempts to remedy this have generated rather fantastic stories, about the 'real' services provided by money in reducing transactions costs, in order to generate ad hoc explanations for its use. Keynes rightly argued that no one in the neo-Classical world would hold money, and correctly identified one of the reasons for the use of money: uncertainty. Thus, in Keynesian theory, money yields liquidity and can never be barren. Indeed, we may go further. Because there is uncertainty, all money contracts must include interest - thus, money is not barren and it yields interest precisely because uncertainty generates liquidity preference. Because of uncertainty, loans in a private property economy must incorporate interest. This means that accumulation of monetary values is ensured in the aggregate as long as debt contracts are not abrogated. In other words, the existence of monetary contracts ensures nominal economic growth (so long as defaults do not occur) since every loan requires repayment of a larger nominal value. There are always two ways to meet interest commitments: obtain money or expand the nominal value of IOUs held by creditors. Use of either of these methods will require an increase of the total volume of money outstanding. 448 Money in the Circular Flow Over time, specialized institutions evolved to take up private liabilities and government debt and to issue their own liabilities. These institutions would act as 'expert judges' in deciding which liabilities were acceptable; they could reduce risk by specializing in this activity, by diversifying, and by seeking government protection (deposit insurance, lender of last resort intervention, monopoly charters, and so on). In this way, they were able to issue liabilities that paid lower interest than did the liabilities they held as assets. Thus, they would live on the differential between the 'loan' rate of interest and the 'deposit' rate of interestas long as this differential is positive, their assets would grow faster than their liabilities, so that their net worth would rise. If money is nothing more than the aggregate volume of outstanding promises to pay, then the rate of growth of the aggregate money supply must be closely related to the interest rate if interest commitments are always met by issuing new IOUs. In a moment, we will clarify this relationship. Even if interest commitments are met by payment of money (third party IOUs) rather than by issuing new IOUs, the rate of growth of the money supply will be related to the interest rate. The reason is quite simple. In order for one economic agent to make payments on outstanding debt, this agent must run a surplus on current account. This means that some other agent must be running a deficit on current account - that is, must be issuing monetary IOUs to finance deficit spending. In order for the first agent to meet interest and principal commitments, the second must issue an equivalent quantity of IOUs which means that the outstanding money supply is growing at the rate of interest. (We will use a simple model to demonstrate this relationship below.) The existence of uncertainty leads to interest, which generates a drive to accumulate. Accumulation takes the form of instruments of further accumulation. 21 Because money earns interest, it is one means to further accumulation. However, as markets develop, alternative means to further accumulation develop with an expected return which exceeds that on money. Those who issue IOUs, creating money on which interest is earned, do so increasingly to purchase a variety of assets which are expected to generate net money receipts. That is, one 'borrows' to purchase 'capital' to be used to produce commodities to be sold for money on the market. One would not do so unless the rate of return expected to be generated on 'capital' were greater than the interest to be paid on IOUs to compensate for 'borrower's risk' (the uncertainty faced concerning ability to meet contractual obligations).22 Thus, the expected return on 'capital' must exceed the contractual return on money in order to induce 'borrowers' to issue the money held by 'lenders'. In Keynes's terminology, due to liquidity preference the expected return on illiquid assets must exceed that on liquid assets. Since money is the most liquid asset, its return sets the standard which must be achieved by the return on all other assets. Over time, the whole spectrum of interest rates has tended to fall. In prehistoric societies with underdeveloped markets, the rate of interest was closely related to the natural productivity of the objects loaned. However, as markets developed and as loans took the form of 'money now for more money later', the rate of L. Randall Wray 449 interest was released from the fecundity of nature. By stage two, the rate of interest came to represent a premium to compensate for uncertainty. This premium has tended to fall over the centuries as various financial innovations have reduced lender's risk. 23 As the premium has fallen, the required 'marginal efficiency of investment' has fallen as investment is pushed to the point where the 'marginal efficiency of money' equals the 'marginal efficiency of capital'. 5 MONEY IN THE CIRCUIT Post Keynesians and circuitistes alike emphasize the importance of circular flow analysis. Schmitt and Greppi (this volume) argue that neutral accounting reveals the 'prior existence' of circular flows - that is, such flows are inherent in monetary economies. Circular flow analysis always has to do with nominal flows, thus, with absolute prices and with sums of money, rather than with relative prices as in general eqUilibrium analysis. Finally, circular flow analysis is concerned with the creation (not allocation) of macroeconomic values. Even after all financial assets and liabilities are canceled, real quantities of goods and services remain and these quantities tend to grow over time because circular flows must grow over time, inducing accumulation of means of further production. 24 An analysis of monetary expenditure and income flows lends itself naturally to a circuit approach. The essence of this approach is the recognition that all market activities involve interconnected balance sheets. It may at first seem paradoxical that money arises only where economic agents attain a level of independence so that each is able to enter into contracts, but that money inexorably links balance sheets so that individual fortunes are always tied to the actions of others. Thus, the seeming independence of individuals in market economies is always illusory: pecuniary activities are always interdependent. In a non-monetary, non-market society, balance sheets need not interlock. Thus, Robinson Crusoe may pick his coconuts and decide whether to consume or save them with no impact on the balance sheet of Friday. If, however, Crusoe lends four coconuts to Friday today in return for an IOU which promises five coconuts tomorrow, then their fortunes are tied: Crusoe's surplus equals Friday's deficit; Friday will not be able to retire his monetary debt unless he is able to generate a surplus tomorrow; and Crusoe's wealth depends on the fortunes of Friday. In this very simple example, the circuit is 'closed' and the money is destroyed if Friday is able to retire his debt and pay the contracted interest. At this point, Crusoe receives the four coconuts originally lent plus a 'profit' of one coconut. Thus, when money is destroyed, 'profits' are accumulated in non-monetary form. If Friday is not able to do so, his outstanding commitment (which represents the total money supply) will grow at the rate of interest. In this case, the circuit is not closed, but grows instead. Crusoe accumulates wealth in the form of money (Friday's coconut IOUs) rather than in the form of coconuts. 450 Money in the Circular Flow Once we leave our hypothetical Crusoe society and examine a private property, monetary, capitalist society, the nature of the circuit becomes more complex. It is useful to first examine two subcircuits: the consumption goods circuit and the investment goods circuit. 2S As a first approximation, we may begin with the assumption that all expenditures at the beginning of the period are financed by issues of short-term obligations - that is, by creating money. These issues allow the capitalist to hire labour (and pay for raw materials - which will be ignored since these complicate but do not change the argument). We will also assume that society consists only of these two classes, and that the marginal propensity to consume is one for workers but zero for capitalists. Finally, workers do not own means of production so that their only source of income is wages, while capitalists do not work for wages. The short-term credit issued by capitalists is equal to the wage bill in the two sectors and equals total spending on the output of the consumption goods sector. Revenues received by capitalists are initially in the form of their short-term credits, and their profits equal the wage bill in the investment goods sector. Receipt of these short-term credits allows the capitalist class as a whole to cancel the debt incurred to produce consumption goods. At this point, the consumption goods subcircuit is 'closed'. Remaining short-term credits received as gross profits are equal to the cost incurred in producing investment goods. If these goods are sold to those capitalists holding short-term credits, then the investment goods subcircuit is also closed as the remaining short-term debt is canceled. In this case, all capitalist profits are realized in the form of physical goods. Alternatively, investment goods purchasers might sell long-term debt to holders of the short-term debt - again, the investment goods subcircuit is closed in the sense that all short-term debt is retired, although outstanding long-term debt equals the value of the investment goods. Again, capitalist profits equal the value of outstanding (long-term) debt, which equals the value of the physical investment goods. Finally, it is possible that those who hold short-term debt refuse to purchase either the investment goods or the long-term debt used to fund positions in physical goods. That is, those who have received profits prefer to hold them in liquid fornl. In this case, intermediaries may spring up to issue short-term liabilities held by those with high liquidity preference, and to accept the long-term debt issued by those who wish to hold the physical investment goods. In this case, the outstanding short-term debt equals aggregate profits, and the long-term debt, and the value of investment goods.26 There are two complications which must be considered: we have ignored interest rates, and we have neglected to consider that production is an ongoing process. The short-term debt which is issued to pay the wage bill is not retired at the end of each production-consumption period. Instead, this short-term debt is continually renewed to allow production to proceed anew. In other words, the stock of outstanding credit is not retired at the end of each period, but is rolled over as a revolving fund of finance. In this way, the aggregate profits which are distributed among capitalists (equal to the wage bill in the investment sector) are L. Randall Wray 451 realized in money form (the short term credits issued to cover the wage bill in the investment sector). As Marx, Keynes and many others have emphasized, the object of production in a capitalist economy is 'to end up with more money than it started with' (Keynes, 1979, p. 89). In other words, the object is to realize profits in money form; accumulation must take the form of monetary accumulation. Monetary accumulation occurs where the circuits are not closed and occurs as the 'revolving fund' is continually renewed and as the volume of short-term credits grows over time. This volume grows for two primary reasons: growth of the wage bill and growth of debt due to interest. The wage bill grows either because nominal wage rates rise or because the volume of employment increases. In the discussion of the circuit above, we ignored interest charges on outstanding monetary liabilities. As the monetary IOUs issued to finance the wage bill must carry interest, it is apparent that closure of the circuit is not possible unless firms issue enough short-term debt to include interest commitments, and that all received interest income is always spent (for example, on consumption goods). In this way, the circuit can be closed. Alternatively, interest commitments may be 'carried on the books' so that the circuit grows at the rate of interest applied to the 'revolving fund of finance'. This can be shown very easily in the following model. We will expand the discussion slightly to add a third sector: banks. 27 Banks accept the short-term liabilities of firms (non-bank money) and issue their own short-term liabilities (bank money). As the wage bill grows, bank assets and liabilities will grow. Given a uniform production period (that is, all production processes begin and end on the same dates), a marginal propensity to save out of wages equal to zero, a marginal propensity to buy long-term bonds out of profits equal to one, and lOOper cent financing of the wage bill through the use of bank money, then the rates of growth of bank money and of non-bank liabilities are relatively easy to calculate. We will assume that banks provide deposits D on which they pay an interest rate of RD' Firms create non-bank liabilities L (loans) to finance the wage bill Wand interest payments at an interest rate of R L • At the end of period one, the outstanding stocks of loans and deposits are as shown in equations (I) and (2) respectively. 4= Wo+ RLLo (1) Do= Wo+RoDo (2) As long as the interest rate on loans exceeds that on deposits, the net worth of the banking system will increase. The indebtedness of firms has increased by the amount of the wage bill plus interest on the total quantity of loans. Similarly, bank indebtedness has increased by the amount of the wage bill plus interest on bank money. In the second production period, the indebtedness of firms will grow by any increase in the wage bill plus interest on any increase of indebtedness, and a similar argument applies to bank indebtedness. We are assuming, of course, that Money in the Circular Flow 452 interest is accumulated in the aggregate as debt. If firms used a portion of receipts from each production period to meet interest commitments, they would find that their 'revolving fund of finance' would shrink so that additional loans would have to be issued just to maintain the wage bill, unless bank receipts of interest are spent on consumption goods. 28 It is common in circuitiste models to follow the example of Wicksell, in which the loan rate of interest equals the deposit rate of interest; if all interest receipts are spent on consumption goods, then firms receive sufficient income to pay interest. I do not follow this example. First, this ignores profit-seeking behavior of banks. Bank net worth exists (and grows) because the loan rate normally exceeds the deposit rate of interest. Second, assuming that all interest receipts are spent on consumption goods (whether the interest is received by banks or depositors) ignores the desire to accumulate wealth in monetary form (which is the object of all accumulation in a monetary economy). Third, this approach focuses on the medium of exchange function of money, rather than on the unit of account function. Thus, I prefer to assume that non-bank interest obligations are carried on the books of the banks, and that bank interest obligations are carried on the books of depositors. Since all money (in this model) is debt, growth of nominal values also implies growth of indebtedness - as interest must be paid on debt, one undertakes debt now only if one believes even more debt will be undertaken later. That is, one is willing to become indebted only if one expects the stream of returns to increase at a rate that exceeds the loan rate of interest. This is possible in the aggregate only if the aggregate level of indebtedness grows over time at a rate that exceeds the loan rate of interest. Furthermore, if the loan rate exceeds the deposit rate, even if all interest receipts were spent on consumption goods, firms cannot cover a portion of interest expenditures unless banks buy a portion of the output of firms so that the entire net worth of banks is realized in real terms. This is not consistent with one of the functions of banks, which is to maintain a substantial portion of net worth in liquid form. For these reasons, I prefer to carry interest on the books of 'borrowers' and 'depositors'; thus, firms (as borrowers and depositors) are always net debtors to the banking system, with their net indebtedness equalling the net worth of the banking system. In this case, the outstanding quantities of loans and deposits at the end of period two are: L, =Lo + (W,- Wo) + RL(L, - D, 1.0) + RLLo =Do + (W, - Wo) + RD(D, - Do) + RoDo (3) (4) By subtracting these equations from those presented above, we may obtain the growth of loans and deposits over the two periods: AL =(_I_)(AW + RLLo) l-RL (5) L. Randall Wray 1 ilD= (--)(ilW + RoDo) l-Ro 453 (6) Thus, the growth of firm indebtedness is determined by the interest rate paid on loans, on the growth of the wage bill, and on the initial quantity of loans issued. Similar arguments apply to bank money. The rate of growth of each is: ilL - Lo 1 ilW = (--)(R L + - ) l-RL Lo ilD 1 ilW = (--)(Ro + - ) Do l-Ro Do - (7) (8) This means that the interest rates paid and the rate of growth of the wage bill set a minimum rate of growth on loans and deposits. Even if the wage bill does not grow, the minimum rates of growth are: ilL -= Lo RL (1-Rd ilD Ro -= Do (1- Ro) (9) (10) If the rate of growth of non-bank liabilities exceeds that of bank money, bank equity rises. Equation (9) shows the minimum rate of growth of non-bank liabilities which must be taken up by the banking system to maintain the revolving fund of finance and to permit fulfilment of interest obligations by firms. Equation (10) shows the minimum rate of growth of bank money consistent with maintenance of the revolving fund of finance and bank fulfilment of interest obligations on bank money. Equation (7) shows the rate of gross, nominal accumulation of the economy. The rate of nominal accumulation in the consumption sector is determined by the rate of growth of the wage bill in the investment sector, less the differential between equations (9) and (10). As we have assumed that the marginal propensity to buy bonds out of profits is one, the rate of growth of long-term bonds will be determined by the rate of growth of the wage bill in the investment sector, plus the long-term interest rate on bonds. This allows the investment sector to realize produced goods in money form. Thus, uncertainty leads to the development of money and interest, while interest ensures that the circuit must grow. This means that a capitalist system, which is fundamentally a monetary economy, must accumulate 'capital': in the case of firms, 'capital' includes those real and financial assets expected to generate net money receipts at a rate which exceeds the interest rate on the combination of short-term and long-term liabilities issued to fund positions in this capital. For banks, 'capital' 454 Money in the Circular Flow primarily consists of the short-term and long-term debts of firms, and positions in this capital are financed by issuing short-term liabilities - or bank money. Due to uncertainty and interest rates, the outstanding stocks of liabilities - including bank money - must grow to allow accumulation in money form. In the case of non-bank firms, nominal accumulation is as shown in equation (7); in a credit money economy, the monetary wealth of non-banks equals their liabilities. When we net out financial assets and liabilities of the non-bank sector, we are left with the real assets created by the labour in the investment sector. The creation of these real assets occurs only where expected money receipts at the individual level outweigh nominal commitments. In the aggregate, as mentioned above, this is made possible by nominal accumulation - that is, by growth of the revolving fund of finance (or the wage bill and interest) over time. The interest rate on bank money must be less than that on non-bank money in order to er.sure accumulation of net wealth by banks. This is normally ensured because bank money is more liquid due to the special status of banks - as will be discussed in the following section. 6 THE RISE OF FIAT MONEY AND CENTRAL BANKS, AND THE SPECIAL STATUS OF BANKS As discussed above, the first and most characteristic form of money in a monetary economy is credit money. Later, however, commodity money in the form of coined precious metals was developed and circulated side-by-side with credit money. Still, in a monetary economy, the vast majority of money takes the form of credit money. There are two reasons for this. First, a volume of credit money is built up during the normal production process, which is forward looking through an uncertain future, and which necessarily operates in historical time. Thus, production always involves 'money now for more money later' and a system of credits is necessarily created. This has long been recognized, as indicated by Marx's affirmati ve quote of Coquelin: In every country the majority of the credit transactions takes place in the circle of the industrial relations themselves ... the producer of the raw material advances it to the capitalist, who works it up, and receives from him a promise to pay on a certain day. The manufacturer, having completed his share of the work, in his tum advances his product on similar conditions to another manufacturer, who has to manipulate it further, and in this way credit extends more and more, from one to the other, down to the consumer .... All borrow with one hand and lend with the other, sometimes money, but more frequently products. In this manner an incessant exchange of credits, combining and crossing in all directions, takes place in the industrial relations. The development of credit consists precisely in the multiplication and growth of these mutual credits, and here is the real seat of its power. (Coquelin, from Du Credit et des Banques dans L'lndustrie, 1842, quoted in Marx, 1909: 472) L. Randall Wray 455 Second, a monetary economy cannot function on the basis of commodity money. As discussed above, monetary economies must grow if for no other reason than that loans always carry interest - thus, accumulation is necessary. There is no guarantee, however, that the supply of commodity money will grow at a sufficient pace to allow fulfilment of interest commitments. There is one significant problem with credit money, however: its value is only as good as the creditworthiness of its issuer. If the issuer defaults, credit money loses its value. 29 Commodity money, on the other hand, is a representative money whose value is not determined by the creditworthiness of any particular issuer - in a sense, its value is established by the value of money issued by the most creditworthy debtors. 3o As such, there is no fear of default on commodity money. This generates two special characteristics of commodity money: it need not pay interest, and privately issued credit money will naturally be made convertible into commodity money: in order to enhance the circulation of a particular private money, it can be made convertible into commodity money on demand (or under certain specified circumstances), allaying the fears of holders. While the credit money will pay interest (the rate of which will depend on the perception of the ability of the issuer to convert these liabilities into commodity money), commodity money need not. Thus, we observe that credit money comes to be issued on the basis of reserves of commodity money, leading to the orthodox deposit multiplier story. However, it is important to note that this arrangement occurs after the development of sophisticated credit arrangements, and not as in the orthodox story in which goldsmiths discover they need keep only fractional reserves. Only recently did governments become sufficiently creditworthy that their liabilities would circulate as money. Previously, governments had to rely on commodity money, which could be obtained through seigniorage, taxes or borrowing. Ability to obtain purchasing power through seigniorage was very limited because the value of commodity money is based on the quantity of precious metals contained in the coins. Attempts by the government to debase coin only led to rising prices in terms of the commodity money (although not necessarily in terms of privately issued credit money).3l Before the development of representative government, governments were also limited in their ability to collect taxes (relying on force) or to borrow. Typically, a monarch could only borrow against a specific revenue source ('anticipate' specific taxes) or issue debt on the basis of a guarantee of creditworthiness provided by prominent individuals or institutions. That is, unlike the situation today, government liabilities were backed by private guarantees of creditworthiness. The first central banks were created specifically to provide purchasing power to the government. 32 In return for taking up government debt, these central banks were given favourable treatment and various monopoly rights. Most importantly, various restrictions were placed on private note issue by other banks. As notes were the principal liability of early banks, sucb restrictions gave the central bank great advantages. Gradually, central banknotes replaced private notes and came to be used as the reserves and as the means of payment for settling interbank liabilities 456 Money in the Circular Flow (which gradually took the predominant form of checkable deposits). Furthermore, private bank liabilities became convertible into central bank liabilities, rather than commodity money. The central bank came to recognize that as its liabilities operated as bank reserves, this not only gave the government a very large source of purchasing power in the form of central bank fiat money, but it also gave the central bank some control over private banks through the discount mechanism. By the late nineteenth century, lender of last resort operations were discovered - which essentially gave a privileged status to bank money over all other forms of money since the central bank could guarantee that bank money bore no default risk. 33 Thus, a pyramidal structure has gradually evolved in which non-bank money is guaranteed by banks, is made convertible into bank money, and is retired using bank money; while bank money is guaranteed by the central bank, is made convertible into central bank money, and is retired using central bank money. Money which has no default risk need not pay interest. Thus, as discussed above, commodity money does not pay interest. With the development of the pyramidal structure, government fiat money (normally central bank liabilities) will not pay interest, either. Finally, as long as bank money is fully guaranteed by the central bank, it need not pay interest. However, for reasons which are beyond the scope of this analysis, regulations, institutional arrangements, and competition can generate a positive rate of interest even on bank liabilities which have no default risk. All this means that a capitalist economy cannot function on the basis of commodity money or fiat money of fixed supply (due to the necessity of interest). Rather, it must rely on credit money. Since all monetary economies inexorably link balance sheets, individual fortunes are tied to macroeconomic performance. This means that individual ability to meet contractual obligations depends on macroeconomic functioning of the system as a whole. If one defaults on commitments, others may follow. Attempts to convert credit money into commodity money lead to a liquidation crisis since it is impossible to do so in the aggregate. Nominal values must shrink, circular flows diminish in value, and a general depression results. One way out of this is to tie credit money to fiat money, which may be issued without limit to avoid a liquidation crisis. This, however, is not possible until a pyramidal financial system has evolved so that fiat money replaces commodity money at the apex. At this point, the worst consequences associated with depreciation of credit are eliminated through lender of last resort operations - but this intervention might itself create other sorts of crises (the analysis of which is beyond the scope of this chapter). 7 POST KEYNESIAN AND CIRCUIT APPROACHES TO MONEY: A SYNTHESIS? As Arena (this volume) discusses, the circuit approach emphasizes the existence of macro laws and structural relations which are independent of micro-level behaviour. In this approach, ex post circular flows are analyzed as logical necessities, L. Randall Wray 457 and closure of circuits at the end of each period is sometimes taken as a logical conclusion to the analysis. Thus, ex post equalities, such as expenditure equals income, or saving equals investment, arise as identities which must always hold over any period of any length. This had led some (such as Schmitt and Greppi in this volume) to argue that Say's Law is an identity which must always hold in monetary economies because spending equals income, thus, all incomes must be spent. Unfortunately, as emphasized by Arena (this volume), the circuit approach has not dealt adequately with liquidity preference, with the determination of asset prices, or with financial markets, in general. The problem really involves inadequate treatment of expectation formation under uncertainty. On the other hand, much of the Post Keynesian work deals explicitly with such matters. Thus, for example, Post Keynesians argue that while ex post identities do hold, ex ante divergence between expected spending and expected income can influence the levels actually achieved ex post. Say's Law does not hold, according to Post Keynesians, because those who receive income may prefer to hold liquid positions. As liquidity preference rises, asset prices fall, causing production of physical assets to decline and ex post spending and income flows to diminish. As a first approximation, one might say that circuitistes have focused on the flow aspects of credit money, while Post Keynesians have emphasized money as a stock which is held because the future is uncertain. This dichotomous treatment has, however, become less clear cut. Davidson (1978) emphasized two aspects of money: the income generating-finance process and the portfolio-change process. In the income generating-finance process, money enters the economy endogenously as banks provide the finance which allows the circular flow to expand; while in the portfolio-change process, the money supply is increased exogenously as the central bank buys government bonds. The first of these is clearly a flow-supply approach to money and is quite consistent with the circuit approach. Similarly, Moore (1988) has developed a 'horizontalist' approach to money which is also consistent in this respect with the circuit approach. Finally, Rousseas (this volume) defines money demand as a demand for financing of investment - again, this is similar to the role played by money in the circuit approach. Thus, the Post Keynesian focus on money has shifted from an early emphasis on a stock approach (as in the textbook Keynesian liquidity preference theory) to include a flow-supply of credit money. Does this mean that Post Keynesians will abandon liquidity preference theory and return to Say's Law? They need not. First, we must continue to recognize that the future is uncertain, that production takes time, and that time is irreversible. This gives rise to the use of money contracts in private property economies and to the incorporation of interest into such contracts. As Keynes argued, an own rate of interest (or marginal efficiency) can be calculated for each commodity, which is the return in terms of the commodity on a loan in that commodity. Each of these own rates can also be quoted in money terms. Keynes then argues that: '[T]he rate of interest on money plays a peculiar part in setting a limit to the level of employment, since it sets a standard to which the marginal efficiency of a capital- 458 Money in the Circular Flow asset must attain if it is to be newly produced' (Keynes, 1964: 222-3). If the preference for liquidity rises, all asset prices are affected, and those of physical assets, in particular, will fall relative to the prices of liquid assets. This means that capitalists will reduce orders of physical assets, reducing the size of the industrial circuit as the wage bill falls - initially in the investment goods subcircuit, but later in the consumption goods subcircuit as well (through the 'multiplier'). Reduced spending also causes income to fall - for example, capitalist profits fall by the decline of the wage bill in the investment goods sector - and this has further effects on capitalist expectations. Because income is always tied to spending, and because forward contracts always involve 'money today for more money tomorrow', declining spending makes it impossible to meet contractual commitments. Thus, there is a role for liquidity preference in the circuit. Even if the money supply is comprised solely of privately created credit, and even if the flow-supply were to expand to meet planned spending and expansion of the circuit, liquidity preference (a 'stock-demand') plays a role precisely because of its impact on demand prices of capital assets. As Keynes argued, rising liquidity preference does not mean that money hoards rise, but that asset prices adjust so that wealth-holders are content with the existing quantity of money. This does not mean, as is often assumed, that liquidity preference theory is inconsistent with an endogenous money approach, or with a circuit approach. It merely means that when the preference for liquidity rises, it is unlikely that this will be met by an increase in the flow-supply of money. That is, as capitalists become pessimistic about the future, they reduce production (or, the industrial circulation declines) and create fewer short term IOUs to finance the circuit. At the same time, the demand for liquid assets rises, raising prices of certain assets involved in the 'financial circulation'. Of course, previous commitments undertaken to finance a growing circuit will continue to grow at the rate of interest. As the industrial circulation shrinks, it will become increasingly difficult to meet these commitments. As discussed above, a debt deflation is a likely outcome. Although spending always equals income, Say's Law does not hold. Spending determines income, and at the aggregate level, spending is identically equal to income, but individuals need not spend all their income. If workers choose to hold some of their income in liquid form (for example, in the form of short-term liabilities issued by firms or banks), then expected sales of produced goods may be lower than anticipated. Firms find that outstanding short-term liabilities (held by workers as savings or by banks which have intermediated from firms to workers) equal the value of unintended inventory accumulation. Thus, capitalist income is partially 'realized' in non-money form. This does not negate any aggregate identities, but it is likely to reduce the expectations firms have about future levels of effective aggregate demand. Since the goal of capitalist production is always to 'end up with more money than it started with', production which cannot be realized in money form will affect expectations and decisions regarding the level of production to be taken in the following period. In this way, current outcomes influence the size of the circuit as we move through time. L. Randall Wray 459 In summary, capitalist decisions to spend on the wage bill determine worker income, while capitalist decisions to spend on the wage bill in the investment goods sector determine capitalist income. Worker decisions to consume or to save determine whether capitalists can realize income in money form or in the form of involuntary inventory accumulation. That is, if workers save part of their wages, then capitalists cannot sell all produced consumer goods at prices sufficient to cover the total wage bill. In this case, profits can only be realized in the form of physical consumption goods. Worker and capitalist decisions over the form in which savings will be held will determine how positions in investment goods will be funded. For example, if savings are held in liquid form, then long-term bonds cannot be sold to savers to finance purchase of the produced investment goods. However, financial intermediaries might hold long-term IOUs and issue the sorts of short-term IOUs preferred by savers. Thus, while the saving decision cannot directly affect investment in the same period (since the income which is saved is created simultaneously with the production of the investment goods), saving decisions do affect the form in which profits are realized (money or physical goods) and the method used to fund long-term positions in real assets. In this way, expectations about the future may be impacted so that the ex post results achieved this period can affect production decisions to be made concerning the following period. In private property economies, money is a social relation between debtors and creditors. In capitalist economies, money is representative of this general social relation, but it also represents a more specific social relation: the private creation of money allows capitalists to direct labour power toward the production of greater nominal values. This production, however, occurs in an uncertain world in which forward contracts and individual responsibility for one's own welfare create a preference for liquid positions. As apprehension about the future waxes and wanes, the industrial circulation tends to cycle round a path which only coincidentally corresponds with that which society, as a whole, might desire. The financial circulation relies little upon labour power for expansion of nominal values, however. As liquidity preference rises, nominal prices of liquid assets rise, raising the 'marginal efficiencies' of these above the marginal efficiency of capital. Greater nominal profits can then be realized in the financial circulation than in the industrial circulation. 34 Unemployment results precisely because labour is not required to produce nominal values in the financial circulation (Davidson, 1978). Thus, there is room for liquidity preference in the circuit. Furthermore, there is some independence of the financial and industrial circulations and liquidity preference helps to determine where capitalist efforts will be directed. When liquidity preference is high and efforts are directed toward the financial circulation, then 'the position is serious when enterprise becomes the bubble on a whirlpool of speculation. When the capital development of a country becomes a by-product of the activities of a casino, the job is likely to be ill-done' (Keynes, 1964: 159). By combining the circuit and Post Keynesian approaches to money, we may come closer to understanding one of the major features of a capitalist economy: 460 Money in the Circular Flow cycles. While steady accumulation of nominal values is required due to the workings of compound interest, uncertainty in a non-ergodic world means that accumulation will not be steady. As expectations fall and accumulation is interrupted, payment commitments cannot be met so that defaults occur. A crisis ensues when a rush to liquidity cannot be met through the normal economic processes. Lender of last resort intervention helps to constrain the crisis, but may not eliminate the processes which generate crises. Notes ... The author would like to thank John Henry, David Levine and Edward Nell for l. 2. 3. 4. 5. 6. 7. 8. 9. valuable comments. See Chick (1986), Chick and Dow (1989), Moore (this volume), and Niggle (1990) for uses of 'stages' approach to the study of the evolution of the financial system. Of course, money is not the only type of contractual liability: the indentured servant must also fulfill certain contractual obligations. Davidson (this volume, and 1978) also emphasizes that real world contracts are written in money terms, and defines money as the means of contractual settlement. Thus, tribal societies do not create or use money. While it is true that members of tribal societies are always under various social obligations, these are not in the form of well-specified contracts which enumerate individual responsibility. See Wray (1990) for a criticism of the conventional view that tribal societies use money. Private property is a necessary, but not sufficient, condition for the existence of money. For example, a private property economy based on slavery does not require use of money. The relation between the slave and master is not based on a contract, thus does not involve a unit of account. On the other hand, a capitalist society does require money as the simultaneous existence of private property, independence of agents, and individual responsibility leads to the use of money contracts. As the domain of private property shrinks - as in feudal society - the sphere within which money contracts are used also shrinks. Thus, feudal society, in which private property plays a diminished role, has little use for money (Dobb, 1963). In tribal society, such risks are shared communally. Individuals in a tribal society are not normally held to be individually responsible for return of a loaned item if its return would result in individual hardship. See Heinsohn and Steiger (1983). In the 'stages approach' presented here, I rely primarily upon a logical reconstruction of the evolution to the modern financial system. However, I believe that this reconstruction is consistent with the theory of money presented here, and with the few historical facts we have about the origins of money and the evolution to the modern system. I will refer readers to more detailed, historical treatments. It must be emphasized, however, that historical facts do not speak for themselves; all historical analyses present an interpretation of these facts. The economist presenting an economic interpretation that relies to some extent on analyses provided by historians faces a formidable barrier: historians frequently have only a rudimentary understanding of economics. Indeed, most of the analyses of money written by historians appear to adopt a simple monetarist theory in which the supply of commodity money determines prices - even where this interpretation clearly conflicts with the 'facts' presented. To some extent, economists are forced to rely on logic, historical specUlation, and theories about what money is and what money does; the analysis that follows is in this vein. See Homer and Sylla (1991) for a discussion of such interest payments without money in primitive societies. L. Randall Wray 10. II. 12. 13. 14. 15. 16. 461 Of course, contracted interest payments allowed the 'lender' to appropriate a portion of the surplus generated by the 'borrower'. As individualistic, profit-seeking behavior became the norm, this share of the surplus came to be seen as a 'fair' reward. The orthodox economist does not even distinguish between interest and profit, seeing both as a 'return to waiting'. Clearly, when a property owner alienates property, it is to reap a share of the surplus generated by the 'borrower'. Interest income represents a prior allocation of a portion of the gross profits to be generated, and the interest rate is determined by uncertainty. Keynes showed that the early monetary units were based on a specific number of grains of wheat or barley. See Keynes (1982: 233-6) and Wray (1990, ch. I). Once a universal measure of social value is recognized, it becomes a purely notional construct whose value is freed from the physical objects (such as wheat) with which it was formerly identified. Thus, the money of account denominated in idealized wheat units is separated from actual wheat. At this point, an ideal money is created and prices are expressed in terms of this hypothetical money of account. See Polanyi (1971), Heinsohn and Steiger (1989), Levine (1983), and Wray (1990) for critiques of the orthodox notion that money developed as a medium of exchange after the development of market economies based on barter. The modern pyramidal structure with central bank fiat money at the apex was developed during the eighteenth and nineteenth centuries. In contrast, credit money is at least as old as writing. Of course, the 'historical stages' occur at different times in different societies. The ancient Near East developed a variety of monetary institutions and financial instruments between 3000 and 500 BC. Most of the societies with which we are familiar that used money had already reached my stage four. For example, both Greece and Rome used credit money and commodity money. Heinsohn and Steiger (1987) argue that the temples of ancient Babylonia played a role in the transition to a uniform unit of account (the transition to what I have called stage two); to reduce carrying costs, the temples substituted a metal money of account for the barley unit of account (my stage four); to eliminate counterfeiting, the temples finally switched to silver money. In every case, however, the metal monies were valued according to the barley grain weight equivalents. Heinsohn and Steiger (1983) argue that private property, and with it, credit money, developed in Mycenae in the eighth century BC; while the first purely nominal coin did not appear there until the seventh century BC (my stage five). Medieval Islamic society had developed a range of private financial instruments (paper notes, bills of exchange) by the eighth century even though deposit banking and usury were prohibited; commodity money was used in conjunction with credit money (thUS, this society had entered stage four). The Italian city states entered stage five with monetization of city debt as early as the mid-fifteenth century. However, in most of Western Europe, fiat money does not appear until the eighteenth or nineteenth centuries (See Wray, 1990). 'Metalists' argued that gold was chosen as money because it is intrinsically valuable, thus, the value of money is determined by the intrinsic value of gold. On the other hand, 'cartelists' argued that the value of money has nothing to do with the intrinsic value of gold. This position is more consistent with the view presented here. See Cesarano (1990) and Knapp (1924) for discussions of these views. Thus, in the orthodox approach, a deposit of commodity money (gold) or fiat money (high-powered money) leads to an expansion of loans and deposits. It is interesting to note that orthodoxy ignores the case of Islamic society, in which harsh penalties ensure that deposits cannot be loaned, and cannot be used to meet 'withdrawals' or redemption of liabilities. This has not prevented such societies, however, from developing sophisticated financial systems based on credit money which could not have arisen through the deposit expansion process. If orthodoxy took account of 462 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. Money in the Circular Flow this, it might have recognized that there is something wrong with the story of the goldsmiths. See Wray (1990, chapter 2). As mentioned above, Keynes showed that early monetary units were based on a specific number of grains of wheat or barley, while use of metals as money was a much later development. The value of 'metal money' would be determined by weight in terms of the number of grains of wheat or barley it represented. This is the reverse of the 'metalist' position that the value of commodity money is determined by the inherent value of the gold contained in the objects used as money. Rather than being chosen as money due to an intrinsic worth, precious metals were chosen for purely technical reasons: limited supply, density, and impervious nature. According to Keynes, coins were a later development, which came some 2000 years after the development of money (Keynes, 1982: 233-6, 255). Initially, the metal used for coinage would merely be stamped to indicate fineness, while the value of the coins would stilI be determined by weight. The first purely nominal coins were produced much later, under Pheidon of Argos in the seventh century Be. See note 14 above and see Heinsohn and Steiger (1984) and Wray (1990). I will return later to a discussion of the reasons for the creation and use of commodity money. See Wray (1990, chapter 2) for a discussion of government attempts to increase purchasing power and of the development of fiat money. As discussed in Wray (1990), once money has first developed as credit money, credit money may disappear if private property disappears. That is, when private property is abolished (or diminishes in significance as an institution) as in Roman society, or feudal society, or even Soviet society, credit money loses its importance. To the extent that money is used in such societies, it takes the form of commodity money or fiat money, except in the restricted sphere in which private property remains. There is no satiation point in the logic of accumulation because the existence of interest necessitates never-ending accumulation. See Keynes (1964: 144-5) and Arena (this volume) for discussions of lender's and borrower's risk. See Homer and Sylla (1991) for an examination of the historical trend of interest rates. Until the last couple of decades, the interest rate had fallen. Recently, interest rates have reached levels not seen since the development of the modern financial system (beginning in the seventeenth century in Holland or the eighteenth century in England). See Schmitt and Greppi, this volume. See Wray (1991) for more detail. Graziani (1990) presents an alternative treatment. The initial value of the investment good equals its cost. However, upon accumulation of physical investment goods, their value is not determined by cost but by the gross profits they are expected to generate. Note that I have ignored profits in the investment sector. The model could be extended to take account of investment sector profits by alIowing for additional deficit spending (that is, the consumption sector would spend more than its receipts to buy the capital goods) financed by additional money creation; or by adding a third sector (government or export) so that its wage biII would generate profits for the investment sector (although the third sector would not receive profits). In any case, however, I am interested in presenting the minimum rate of growth of the money supply required to 'finance' a circuit; adding profits for the investment sector would merely increase the quantity of finance required above that shown in my model. This section provides a brief summary of a more detailed treatment which can be found in Wray (1991). As we have assumed that the marginal propensity to save out of wages is zero, alI interest received by workers must be spent on consumption goods - thus, bank inter- L. Randall Wray 29. 30. 31. 32. 33. 34. 463 est obligations are carried on the books of capitalists. Since we assume workers don't borrow, all nonbank interest obligations are incurred by capitalists and are carried on the books of banks. Note also that I assume capitalist interest receipts are not used to reduce interest obligations; as Marc Lavoie has pointed out to me, this would - at the limit - reduce the required money growth rate to the differential between the loan and deposit rates. Of course, this can set off a 'debt deflation' as described by Irving Fisher (1933) and, later, by Hyman Minsky (1978). Marx also recognized that 'It is precisely the enormous development of the credit system during a period of prosperity, hence also the enormous development of the demand for loan capital and the readiness with which the supply meets it in such periods, which brings about a shortage of credit during the period of depression' when there is a run to liquidity (Marx, 1909, p. 532). Kregel (this volume) argues that monetary economies must have some exogenous money to use as a final means of payment. However, credit money can be used as final means of payment by those lower in the pyramid. See Wray (1990) for a detailed discussion of government attempts to obtain purchasing power and of the rise of fiat money. See Wray (1990) and Goodhart (1989) for analyses of the rise of central banking. Kregel (this volume) examines the creation of the Bank of England as a means to provide finance for a war with France. Tooke argued that the central bank should provide liquidity as necessary by discounting good paper in order to stop runs, but, as Bagehot described, the Bank of England generally operated in precisely the opposite manner until late in the nineteenth century because it did not see itself as a central bank. See Tooke and Newmarch (1848), Bagehot (1927), and Wray (1990). As Aglietta (this volume) argues, lender of last resort operations were initially prescribed as a way to save healthy banks, and not as protection for insolvent banks. However, if there is a run to liquidity so that asset prices fall, banks will become insolvent. Thus, the distinction between use of lender of last resort operations to provide liquidity rather than to rectify problems of insolvency is not so clear. See Rousseas (this volume) for a discussion of the distinction between financial and industrial circulations. References Aglietta, M. (1995), 'Systems Risks, Financial Innovations and the Financial Safety Net'. this volume. Arena, R. (1995),'Industrial and Financial Factors in Investment Decisions: A Significant Difference between Circuit and Post-Keynesian Approaches'. this volume. Bagehot, W. (1927), Lombard Street: A Description of the Money Market (London: John Murray). Cesarano. F. (1990). 'Law and Galiani on Money and Monetary Systems', History of Political Economy. 22 (2). p. 321. Chick, V. (1986). 'The Evolution of the Banking System and the Theory of Saving. Investment and Interest', Department of Economics, University Co\lege London, Discussion Paper No. 86-01. Chick. V. and S. Dow (1988). 'A Post-Keynesian Perspective on the Relation Between Banking and Regional Development'. Thames Papers in Political Economy. Spring. Davidson, P. (1978). Money and the Real World (London: Macmillan). Davidson, P. (1995), 'What Are the Essential Elements in Post Keynesian Monetary Theory?,. this volume. Dobb. M. (1963). Studies in the Development of Capitalism (New York: International Publishers). 464 Money in the Circular Flow Fisher, I. (1933), 'The Debt-Deflation Theory of Great Depressions', Econometrica, I, p.337. Foley, D. (1989), 'Money in Economic Activity', in J. Eatwell et al. (eds), The New Palgrave: Money (New York and London: W.W. Norton), p. 248. Goodhart, C. (1989), 'Central Banking', in 1. Eatwell et al. (eds), The New Palgrave: Money (New York and London: W.W. Norton), p. 88. Graziani, A. (1990), 'The Theory of the Monetary Circuit', Economies et Societes, 7, p. 7 Schmitt, B. and S. Greppi (1995), 'The National Economy Studied as a Whole: Aspects of Circular Flow Analysis in the German Language', this volume. Heinsohn, G. and O. Steiger (1983), 'Private Property, Debts and Interest Or: The Origin of Money and the Rise and Fall of Monetary Economies', Studi Economici, 21, p. 3. Heinsohn, G. and O. Steiger (1984), 'Marx and Keynes: Private Property and Money', Monnaie et Production, ISMEA, I, p. 37. Heinsohn, G. and O. Steiger (1987), 'Private Ownership and the Foundations of Monetary Theory', Economies et Societes, 9, p. 229. Heinsohn, G. and O. Steiger (1989), 'The Veil of Barter: The Solution to "The Task of Obtaining Representations of an Economy in which Money is Essential"', in J.A. Kregel (ed), Inflation and Income Distribution in Capitalist Crisis: Essays in Memory of Sidney Weintraub (New York: New York University Press). Homer, S. and R. Sylla (1991), A History of Interest Rates (New Brunswick and London: Rutgers University Press). Keynes, J.M. (1964), The General Theory of Employment,lnterest, and Money (New York and London: Harcourt Brace Jovanovich). Keynes, J.M. (1971), The Collected Writings of John Maynard Keynes, vol. V, ed. by D. Moggridge (London: Macmillan). Keynes, J.M. (1979), The Collected Writings, vol. XXIX (London: Macmillan). Keynes, J.M. (1982), The Collected Writings, vol. XXVII (London: Macmillan). Keynes, J.M. (1987), The Collected Writings, vol. XIII (London: Macmillan). Knapp, G.F. (1924), The State Theory of Money (London: Macmillan). Kregel, J.A. (1995), 'The Policy Implications of the Current Bank Crisis; or "Is Free Market Capitalism Compatible with Endogenous Money?"" this volume. Levine, D. (1983), 'Two Options for the Theory of Money', Social Concept, I (I), p. 20. Marx, K. (1909), Capital, vol. III (Chicago: Charles H. Kerr). McIntosh, M.K. (1988), 'Money Lending on the Periphery of London, 1300-1600', Albion, 20 (4), p. 557. Minsky, H.P. (1978), 'The Financial Instability Hypothesis: A Restatement', Thames Papers in Political Economy, Thames Polytechnic. Moore, B.J. (1988), Horizontalists and Verticalists: The Macroeconomics of Credit Money (Cambridge: Cambridge University Press). Moore, B.J. (1995), 'The Money Supply Process: A Historical Reinterpretation', this volume. Niggle, C.J. (1990), 'The Evolution of Money, Financial Institutions, and Monetary Economics', Journal of Economic Issues, 24 (2), p. 443. Polanyi, K. (1971), 'Aristotle Discovers the Economy', in K. Polanyi et al. (eds), Trade and Market in the Early Empires (Chicago: Regnery), p. 64. Rousseas, S. (1995), 'The Spheres of Industrial and Financial Circulation Revisited and their Implications for Post Keynesian Economic Policy', this volume. Tooke, T. and W. Newmarch (1848), History of Prices and of the State of the Circulations, From 1792 to 1856, vols III and IV (New York: Adelphi Press). Wray, L.R. (1990), Money and Credit in Capitalist Economies: The Endogenous Money Approach (Aldershot: Edward Elgar). Wray, L.R. (1991), 'The Inconsistency of Monetarist Theory and Policy', Economies et Societes, 11-12, p. 259. 17 Monetary Circulation and Overdraft Economy* Fran~oise Renversez If the international economic literature deals extensively with problems of indebtedness, on the level of businesses as well that of governments, the concept of the overdraft economy is rarely invoked as an analytical tool outside French research. There is one study dealing with the Finnish economy I that is often cited early on by authors concerned with the particular form of operation of the economies concerned. The concept has been developed to permit the theoretical specification of the operation of the French economy up to 1980, and it has served as a basis for the financial equations of models such as Metric2 and, more generally, for the studies of the research divisions at the Banque de France and the National Institute for Statistics (INSEE), which is responsible for national accounting. The distinction first posited by Vivien Levy-Garboua3 and Gerard Maarek4 between the overdraft economy and the traditional system, now designated as a financial markets economy, continues to be used. In the latter case, the financing of businesses is essentially provided by the financial markets, and the open monetary market, articulated with the capital markets, is the preferred site of intervention for the central bank's regulatory actions. The concept of the overdraft economy, by contrast, refers to those systems where financing is basically provided through credit, the volume of which presumes that the central bank itself extends credit to the ordinary banks through a variety of refinancing measures. Thus the contribution of liquidity from the monetary system to the economy, resulting from the primary indebtedness of the productive sector, has an endogenous origin. The concept of the overdraft economy should not be confused with that of the debt economy. It is entirely possible that a financial markets economy achieves equilibrium (the adjustment of needs to financing capacities) through foreign indebtedness. This is the case for the United States, which was a creditor vis-a-vis the rest of the world until 1984 but which is now a debt economy, with foreign indebtedness representing 15 per cent of the gross national product. Unlike overdraft economies, however, where the· financing process can only be balanced through intervention of the central bank, in a financial markets economy, the eqUilibria of the different markets are reached in such a way that the intervention of the central bank remains discretionary. The interest rate retains its role as indicator, and the system remains that of the financial markets economy and can be 465 466 Monetary Circulation and Overdraft Economy regulated as such. In the overdraft economy, by contrast, the equilibrium of the monetary system presupposes specific regulatory techniques such as credit restriction or the 'corset'. The passage to an open economy decreases its efficiency considerably. In France, one response to this problem has been the effort to modify financing conditions in favor of the securities markets. Such efforts are faced with difficulties of regulating a system in transition, but it remains to be seen if the endogeneity of money which is proper to the overdraft economy is really called into question. The concepts of the overdraft economy and the financial markets economy have served well to clarify the modes by which the eqUilibrium of the financial systems is determined and the appropriate instruments of monetary policy for their regulation. The articulation (through credit) of the financing of the productive sector in the overdraft economy gives an endogenous character to monetary circulation (1). This leads in tum to a specific mode of determining the equilibrium of the monetary system, the closure of which is guaranteed by the central bank (2). The modes of regulation particular to the overdraft economy system are not easily adaptable to a period of transition (3). THE ENDOGENEITY OF MONEY IN THE OVERDRAFT ECONOMY The financial system proper to the overdraft economy develops on a real base in the sense that in this case, the productive system arrives at a rate of investment that can only be financed through ongoing recourse to indebtedness. The latter is predominantly realized in the form of credit. The Real Bases of the Overdraft Economy System The constraints that transform an overdraft economy into an ongoing financial system emerge on the level of the financing of businesses. This indeed where J.R. Hicks,S who is often cited in this connection, situates the overdraft economy: he bases the distinction between 'overdraft economy' and 'auto-economy' on the modes of financing of non-financial enterprises. If Hicks insists on the external origin of business resources in the 'overdraft economy' , in contrast to the 'autoeconomy', where business is financed by its own funds, the main distinction between enterprises in the two systems arises from the volume of resources withdrawn from their flow or stock of savings which the firms can allocate to the financing of their investments. 6 These investments are determined first of all by the value added realized and secondly by the deductions levied on it (wages, financial charges, taxation). In terms of national accounting, the gross savings are calculated after a double deduction, financial charges and taxes, taken from the firm's gross returns. 7 Companies' savings serve as an indicator of their ability to finance themselves. The question obviously arises as to how a developed Franroise Renversez 467 economy can show a real structural gap between the rate of savings achieved and the level of investment that is not only anticipated but actually realized at the end of the period. One of the explanations is to be sought on the level of the division of value added between wages and profits, although in an open economy the value added increases with productivity. A certain social compromise may work to stabilize the relative shares of wages and profits at the expense of firms' savings. But on the macroeconomic level, it seems more significant to focus on the rate of self-financing. In terms of national accounting, the self-financing rate is equal to gross savings divided by the gross formation of fixed capital (with or without stock). This aggregate effectively permits the comparison of firms' financing capacity with the investments that they consider justified by their expectation of forthcoming demand. In the medium or long term, this can be taken as the level of investment corresponding to the operative growth trend of the economy at hand. Thus the question is why this investment does not subsequently lead to a sufficient volume of savings to guarantee its own financing. An initial response can be found on the level of the growth rate. It may be noted that Japan and France, which have functioned with an overdraft economy system as described above, have maintained rapid industrial renewal and, in the case of France, significant reconstruction for a period of thirty years by depending on credit financing. The second reason why the markets have difficulty providing the volume of external savings necessary for the businesses to ensure their financing in an overdraft economy depends on this growth-rate level and the high rate of investment that it presupposes. For the period from 1974 to 1984, the selffinancing rates of companies in France and Japan are around 60 per cent, thus much lower than the averages for the US and West Germany, which are over 90 per cent. Even with household savings rates representing a good international performance (averaging over 17 per cent for the period, with the current shift beginning only in 1984), the recourse to credit is inevitable. This is mainly due to the weakness of the self-financing rate in the overdraft economy, which in the long run gives rise to the creation of financing institutions and practices adapted to the financing needs of the productive sector. If the theoretical case presupposes the tightness of the financial markets, they can well be imagined to develop without calling this financing structure into question. The Preponderance of Credit in the Overdraft Economy In the overdraft economy, the appeal to indirect financing, to the intervention of financial intermediaries, is more precisely an appeal to bank credit. Thus what is sought is not intermediation as a transformation of settlement dates or of the nature of the resources, but the creation of means of financing that anticipate the growth of the product, It is reached through an operation - credit - which presupposes a personalized bilateral relationship between borrower and lender and not a supply 468 Monetary Circulation and Overdraft Economy destined to the undifferentiated demand of the market. It is clear that relative to the borrower, the position of a bank that buys an issue of securities or acquires securities on the secondary market is different from that of one which grants credit. In the overdraft economy, firms are the main beneficiaries of credit. This characteristic is apparent in Japan, as it has been in France. Conversely, household credit is the first to be affected by possible restrictive measures, concerning either consumption or housing investment. Considered as a source of pressure on demand, an inflationary factor, such credit is viewed less positively by economic and monetary authorities. The importance of credit in the financing of overdraft economies lies in the structure of banks' balance sheets. In Japan, for example, credit constitutes 60 per cent of the assets of the main banks, with the security portfolio/credit ratio running between 25 and 30 per cent. It may be noted that collected deposits constitute the banks' main resource. In France, as in Japan, certain financial institutions appear to be specialized in collected resources and others in the distribution of credit. The big banks which are able to offer the most diversified credit are subject to 'structural deficit', in the sense that they have to complement their resources through refinancing operations bearing on certain elements of their assets on the money market. In Japan, the urban banks that are among the country's main banks are dependent on refinancing for 10 per cent of their liabilities. These are, according to the Japanese terminology, 'overborrowing' banks. In France, likewise, it is the big banks that show 'structural deficits' , and since the nationalizations of 1982. it can be said that the mutualist sector shows a surplus and the nationalized sector shows a deficit, but neither has anything to do with the efficiency of management or the profits of the sectors in question. This situation became particularly evident when certain tax advantages for the mutualist banks were suppressed without any basic changes in the positions. If the French and Japanese economies fall into the overdraft category, it is at the level of financing, because the supplies of financial institutions with structural surpluses cannot meet the demand of institutions with structural deficits on the interbank market. Thus the obligation to refinance falls on the central bank. The volume of investment to be financed in order to renew the growth rate of the activity in the productive sector leads to excess demand for bank credit, which implies a structural gap between the supplies and demands of the money market and makes the intervention of the central bank inevitable. Thus credit appears to determine the liquidity of the national economy in the overdraft economy. Operating Analysis Approaches and the Endogeneity of Money in the Overdraft Economy The monetary statistics of the National Credit Councils developed by the Banque de France suggest an analysis of the sources of issuing money that brings out the endogenous character whereby money is put into circulation in France. Franroise Renversez 469 If the nomenclature9 relative to the definition of the components of the money supply have varied in France, notably with regard to the significant monetary aggregate,1O the method of presenting the sources of issuing money in terms of the counterparts of the money supply remains constant. According to this approach, the counterparts of the money supply consist of the claims whose acquisition by the monetary system leads to the injecting of money into circulation. The definition of the monetary system itself has varied, II particularly with regard to the inclusion of the largest French financial institution, the Caisse des Dep6ts et Consignations, in the bank category; likewise, the presentation of claims as counterparts of the money supply has also varied. But the analytical principle remains the same: it is by being entered as an asset in the monetary syste,m balance sheet that a claim becomes a source of the issuing of money. Within the three categories of claims - foreign, Treasury, and domestic l2Treasury claims were preponderant in France until the end of the 1950s. But since that time, they have yielded to domestic claims, whether in the form of securities or, most often, in the form of credits that might or might not be drawn against the Banque de France. In the long run, the relative share of counterparts of the money supply has evolved. In the beginning of the 1950s claims on the Public Treasury represented over 40 per cent of the counterparts of the money supply, but by 1980 they were down to 8.5 per cent. Their decline favored domestic claims, which in 1980 constituted 86 per cent of the counterparts of the money supply. Given the very limited volume of foreign claims, it is in effect the different forms of credit to the economy that have constituted the essential counterparts of the money supply in France over the past twenty years. Thus the importance of foreign money can be said to have decreased considerably, and the relative stability of the two other cate~ories of counterparts make credit the determining source of the creation of money. In France, budgetary deficits were contained, and the amount of the public debt showed a decrease in the period where the financial system most resembled the theoretical model of the overdraft economy. Until the success of the financial innovations, the French specificity, represented by the Treasury circuit with its network of correspondents, ensured a significant liquidity for the yearly management of the Public Treasury and thus reinforced this mode of operation, without, however, as it sometimes thought, determining it. The analysis in terms of the counterparts of the money supply thus permits an approach to the endogeneity of the creation of money in France. The preeminence of anti-inflationist objectives has favored this perspective. However, macroeconomic modeling, less current since the decline of indicative planning procedures but still used in forecasting, was based on data from the French national accounting system, where the financing of the economy is monitored in terms of the F1ow-of-Funds balance sheet (Tableau des operations financieres, or 470 Monetary Circulation and Overdraft Economy TOF). The structure of the TOFs l3 makes it clear that money is a bank debt and brings out the articulation of its issuance to the primary indebtedness; from these tables it is possible to arrive at a synthetic approach to monetary circulation. The TOFs are double-entry tables analyzing by agent the respective flows of credits and debts, with the agents presented in the columns and the operations in the rows. Thus they establish for each agent the variations in assets and liabilities, so that what is provided is an accounting of the flows rather than the cumulated amounts. The operations are presented in order of decreasing liquidity, which clearly brings out the narrow substitutability of the fonns of asset and liability from money to long-term securities. The theoretical inspiration behind the structure of the TOF is obvious: it offers the novelty of integrating national accounts into Keynesian macroeconomics through the studies of Denizet,14 the creator of French financial accounts, whose structure was in tum adopted by the financial accounts of the OECD, the analyses of Gurley and Shawls on financial intennediation, and those of J. Tobin l6 on the substitutability of financial assets in agents' investment demands. The sectors used are in fact those of the national accounting system: businesses, households, public administration and foreign. This sets up a more detailed classification than the theoretical distinction between deficit and surplus financing agents, and the financial institutions sector comes under close analysis. In the current presentation, however, insurance is considered as an autonomous category and not a financial intennediary, as its activity on the financial markets would justify. There are two large categories within the financial institutions: banking and related establishments, the Organismes de Placement Collectifs en Valeurs Mobilieres (OPCVM, the organisms for collective investment in secur ities). The first results from the changing classification of certain financial institutions such as the Caisse des Dep6ts. The current version recognizes the evident role that this institution plays in determining the equilibrium of the monetary market and thus in establishing the equilibrium of the banking system. The columns of the TOF thus show the flow of assets and liabilities for the Banque de France and the Exchange Stabilization Funds, the banks (including those of the mutualist system), the Caisse de Dep6ts and the savings funds, the finance companies, specialized financial institutions such as the Comptoir des Entrepreneurs and the Credit Foncier de France, and other financial institutions including the Caisse Nationale de I'Industrie but also the brokerage houses. The OPCVM are considered as financial institutions. The creation of securities portfolios, the size of which is a factor in the diversification of risks, permitting the securities offered to the clientele to combine return and diminished risk, and thus better liquidity, results in fact from their activity as financial intennediary, even though they were not included as such in the statistics of the National Credit Countil until 1991. 17 Following the presentation of the national accounting system, the TOF does not set up a specific category for the Treasury; its operations are compiled in the government operations category under the account of the public administration agent. Franroise Renversez 471 The combined reading of rows and columns for the variation of credits and debts shows, on the one hand, the ways that the surplus resources of agents offering a means of financing has been placed at the disposal of agents in need of financing. On the other hand, they allow the identification of the share of financing of primary indebtedness that has been provided by bank credit adjusted to the banks' liabilities through deposit reserves and thus show that primary indebtedness generates the creation of money. Thus, money can be defined as a claim on the banking system, like the banks' circulating debt. It appears as an element of the banking liability accepted as a means of payment. V. Levy-Garboua 18 has pointed out the ambiguity of the idea of debt attached to a claim that cannot really be reimbursed. On the microeconomic level, any agent can obtain the reimbursement of his deposits and their payment in another currency, in banknotes from the central bank, for example, or, since the lifting of exchange controls, in foreign currency. But in macroeconomic terms, the payment system would cease to exist if all deposit-holders demanded to be paid. This hypothesis is incompatible with the current operations of the payment system. The TOFs thus provide an accounting approach to the banks' function of issuingicreating money that can only be conceived in the exchange with nonfinancial agents. This takes place within the framework of the circulation of the flow of annually produced revenue. The accounting entry of monetary circulation proposed by the TOF is thus that of a money representing a claim on the wealth produced within a national economy. This explains the difficulty of situating extra-national currencies, in so far as the closure of the monetary system is itself also guaranteed at the national level. 2 THE EQUILIBRIUM OF THE MONETARY SYSTEM IN THE OVERDRAFT ECONOMY In the financial system of the overdraft economy, where credit constitutes the principal means of financing productive activity, the central bank is obliged to ensure the closure of the system. If the characteristic conditions of the theoretical case of the overdraft economy are met, the central bank is subject to an obligation to refinance that reverses the causal relationship between base and money supply. The Central Bank's Refinancing Constraint In the overdraft economy, the banks are not limited to transforming the terms of the resources borrowed from the public: they create the means of financing through credit. The obligation to maintain required reserves and the need to ensure payments in banknotes force them to turn to refinancing,I9 even if the annual short-term flows of foreign operations feed their liquidity, which is also affected by the variations in the Treasury debt. Thus they seek to obtain complementary 472 Monetary Circulation and Overdraft Economy Treasury reserves, either from other banks in the interbank division of the monetary market or from the central bank, in the central bank division. Both the procedures used and the securities that back them up are specific. The techniques employed are based on the negotiation of short-term securities before maturity date (discount), or purchase and resale (buying with the promise to resell in the short run). The securities are either public or private with specific guarantees. France has seen the emergence of the 'blank' advance without the back-up of securities. In a growth period, refinancing on the monetary market is inadequate; thus the banks tum, as a last resort and in a permanent way, to refinancingfrom the central bank. The problem posed is not that of refinancing from the banks but that of the obligation to refinance that weighs on the central bank. The analysis of the simplified balance sheets of the financial institutions placed in this operating situation makes the point clearly. On the basis of the French case, it may be hypothesized that, for historical and institutional reasons, certain intermediaries have a structural advantage in collected resources and employ their surpluses on the monetary market, while others, and in particular the ordinary banks specialized in the distribution of credit, borrow the means to ensure their liquidity in the central money there. In other terms, they tum to the monetary market for the amounts of central money necessary for financing the balance of their operations of compensation with other banks, and for establishing their required reserves and their payments in banknotes. The monetary market, and more precisely the interbank division, thus permits the operation, at a certain cost, of the transfer of financing resources among the financial intermediaries. Contrary to what is sometimes argued, the draining off of a not inconsiderable portion of the economy's liquidity, in France, by para-public financial intermediaries and mutualist banks does not reduce the possibilities of the banks' activities, but does increase those of the monetary market and possibly the cost of credit. The structure of household investments thus cannot be placed among the causes of the overdraft economy. Lags and leads operations, through which the banks ensure the profitability of their available funds, can, in the short run, hide the structural positions (Tables 17.1 and 17.2). To simplify, let us assume that RFE = O. If RFE = RFD the interbank division of the monetary market is in equilibrium, and the potential preponderance of bank credit falls within a system of balanced financing where the central bank can intervene to achieve certain goals of monetary policy. It can take the classic step of selling securities to 'sponge up' the banks' liquidity or buy them to increase it. Through these mechanisms it can also seek to bring about interest rate movements or, if the institutional structure permits, use variations in the interest rate to provoke the contraction or expansion of the distribution of bank credit. This situation corresponds to the case of the financial markets economy regardless of the development of bank credit. Franc;oise Renversez Table 17.1 Banks and financial institutions with structural surpluses Liabilities Assets Money Required reserves Securities guaranteeing refinancing Credit Other securities Table 17.2 Deposits Securi ties issued RFE CRE BES Banks with structural deficits Liabilities Assets MD Money Required reserves ROD Securities guaranteeing refinancing RFe Credit Other securities 473 Deposits Refinancing Securities issued BD CR D BDS If in general RFD > RFE, there is a disequilibrium on the central money supply side, and a Keynesian-type mechanism of credit contraction through the restriction of bank liquidity. The difference between these two modes of equilibrium of the monetary system balance sheet does not stem from differences in the techniques of providing bank liquidity. The central bank restores the equilibrium of the banks' balance sheets by accepting the request for refinancing from the ordinary banks. Even when this closure is carried out after inventorying the demand before adjudication (intervention upon invitation to tender), it does not correspond to a market practice. It depends on specific techniques - purchase and resale or discount - which seem to be procedures of voluntarist adjustment of supply to demand and not the Walrasian tdtonnement that allows supply and demand to adjust themselves to a price that brings them into equilibrium. Here the price - the refinancing rate - is given: in price theory, its determination would be called 'administrative'. This is to say that it is fixed according to the categories of operations at hand in function of either the goals of general economic policy (export credit) or monetary policy (adjudication upon invitation to tender), and notably in function of the outlook for the exchange of the national currency. 474 Monetary Circulation and Overdraft Economy Table 17.3 Central Bank assets and liabilities Assets Foreign debts Treasury debts: Advances Securities Claims on the economy Liabilities TRE RFBr RFBp Required reserves ROE + RO v =RO Banknotes RO Mb The banking system is indebted to the central bank (Table 17.3), which thus appears to function as a lender of last resort under constraint. In certain institutional systems, the central bank automatically refinances certain credits at more favorable rates than those of the interbank market; the amount of (RFB r + RFBp) - (RFD - RFE) depends on the scale of these practices (e.g. the representative effects of export credit in France). Thus the central bank ensures the closure of the system, but this inevitably entails a compromise of both the operation of the financial machinery and the overall equilibria of the economy. The functioning of the system depends on the contribution of the central bank. The equilibrium of the monetary system is thus radically different from that achieved under the control of the central bank in the financial market economy. In all organized monetary systems there is a possibility of recourse to the issuing institution as lender of last resort, as with the facilities offered by the Discount Window of the Federal Reserve System in the United States, for example. But this recourse remains exceptional. The central bank's interventions in the monetary market to regularize the rates or the volume of bank liquidity should not be confused with the virtually automatic practices of refinancing. If regulatory interventions are to remain effective, they must be able to be carried out in both directions: expansion of liquidity, contraction of liquidity, braking of tensions at the rise or fall of the interest rate. In other terms, the intervention of the central bank is discretionary in the financial markets economy, but it is obligatory in the overdraft economy. A good quantitative approximation of this differing position of the central bank in the two financial systems is provided by the IMF's international financial statistics, which posts central bank credits to money-creating institutions (line 12e). France, Japan, Italy, and West Germany show heavy, or at least significant amounts of credit; Great Britain shows small amount, while no figures are posted for the US because of their negligible amount. Franfoise Renversez 475 The overdraft economy is thus defined by a double level of indebtedness: that of the firms to the banks and that of the banks to the central bank. In this system the creation of money is endogenous in the sense that the essential counterpart of the issuing of money is the credit demanded by the agents to achieve a certain level of activity in the real sector. It is the credit, then, that finance a determining share of investment, and the central bank does not have to room to manoeuver that would really permit it to modulate its refinancing. For this reason the link between central currency and the money supply should be interpreted in terms of the credit divisor. The Credit Divisor The analysis of the relationships between money supply and central currency in the overdraft economy has given rise to an alternative concept to that of the base multiplier: the credit divisor. Given that the central bank cannot avoid meeting the demand of the ordinary banks for central currency, either to make up their required reserves or to meet the demand for banknotes from non-financial agents, then, as V. and L. LevyGarboua have shown,20 it is necessary to modify the description of the mechanisms linking the central bank and the banks. This relationship is traditionally expressed in simple terms reflecting the accounting equilibria of the monetary macroeconomy. M=Bp+D (1) H=Bp+Bb+Mb (2) where M is the money supply; Bp is banknotes held by the public; D is bank deposits; H is monetary base; Bb is banknotes held by banks; Mb is sum of creditor accounts of banks at Central Bank. It is further postulated that Bb=eD (3) Mb=fD (4) Bp=gD (5) where e is banknote ratio of banks; g is banknote ratio of public; f is ratio of creditor accounts of banks at central bank to total checking deposits. The ratio bank reserves that can be held in the form of e.ither banknotes or cash balance at the central bank is e +f According to the way the monetary authorities regulate the banks, either norms or a choice of bank portfolios is produced. 476 Monetary Circulation and Overdraft Economy In this framework: Bp+D Bb+Bp+Mb (6) M gD+D -= H eD+gD+fD (7) M (1 + g)D -= H (e+g+j)D (8) L= D- (Bb + Mb) (9) M -= H where L is the total volume of bank loans and other remunerated assets. From this it follows that the determinants of the issuing of money are the volume of the monetary base H and the coefficients e, f, g. They permit the establishment of three multipliers: the monetary multiplier M/= l+g I::.H e+g+J (to) the bank deposits multiplier I::.D= e+g+j I::.H (11) the bank credit multiplier I::.L=l-e-j I::.H e+g+j (12) In such a model, the supply of money is under the control of the central bank if e +j + g are constant or predictable. 21 In the overdraft economy, the banks do not lend the multiple of the central money they possess. The central bank is a lender of last resort 'under constraint' (either to feed the monetary market or to accept the rediscount or the nearly automatic buying and resale of certain securities~ and particularly public securities). Franfoise Renversez 477 As a result, the banks will be able to distribute the credit, and only later will they procure the sum necessary for their operations in central currency. As C. de Boissieu indicates, the direction of the causal relationship between base and money supply is reversed. 22 The bank's needs in central currency are the three kinds: payment of the balance of the compensation of their payments to and from the other banks, payment in banknotes, and establishment of required reserves. Such a mode of operation implies that the banks only hold the amount of reserves that they strictly need, in particular to meet the norm of required reserves. The relationship between increased resources and increased credit remains the same, but it is based on the increase in credit: in other words, the financial system of the overdraft economy presupposes the complete endogeneity of the monetary base. The banks' demand for credit to the central bank, i.e. the demand for refinancing becomes (Bb + Mb) = (e +/) D (13) The volume of bank deposits given by the accounting equation (9) becomes D= L+ (Bb + Mb) (14) The deposits can then be expressed in terms of credits (L) D= L 1-(e+f) (15) According to equations, the public demand for banknotes is given by Bp=gD Equation (1), which expresses the supply of Money M = Bp + D, becomes: M = (l+g) 1-(e+f) L (16) Two interpretations of the credit divisor have been formulated to express the relationships between the monetary base and the banks' credit deposits. The strong form of the credit divisor posits totally automatic refinancing. In such a system, the supply of credit is perfectly elastic to demand at a given rate of interest because it permits banks to realize a maximum profit by increasing their turnover. Furthermore, since the demand for credit is only slightly elastic to the interest rate, at least below certain thresholds, the supply of credit appears to be perfec1ty elastic to demand. This hypothesis has given rise to the interpretation of bank behavior in the interpretative models of the growth period (metric is a good example). The weakform of credit divisor is more broadly based on bank behaviors in the face of risk and the resulting diversification of the balance-sheet structure. The money supply is a function of the volume of bank credit that the banks deem profitable to distribute at a given rate of interest. Good bank management implies that the volume of bank commitments be determined by the maximization of 478 Monetary Circulation and Overdraft Economy bank profit compatible with the demand for credit. The weak form appears to be better suited to taking recent developments into account than the strong form and also allows for less mechanistic conclusions to be drawn from the banks' purchase of securities. 23 For these reasons, in a period of growth, the overdraft economy system shows clear inelasticity to variations in the interest rate. 24 As a result, although the practice of refinancing entails the administration of the monetary market interest rate, since what is involved is no longer the eqUilibrium price but a determining rate fixed by the issuing institution, the first segment of a segmented interest rate structure25 regulation by the central bank when monetary policy is restrictive is carried out on the quantity of credit. In France this is known as credit squeeze. This system of financing comes under the bank principle system defined by Tooke, where the issuing of money is determined by the volume of bills accepted for discount by the central bank. But it no longer has limits set by a metal currency standard. This is replaced by the differential inflation rates of the main partners in the exchange. In a financial system thus articulated to the activity of the real sector and in particular the productive sector, the other sources of increasing the assets of the central bank seem to be at the origin of the inflationary tensions since they do not have a counterpart in the increase in production. In this context, government indebtedness, in so far as it does not have recourse to prior savings but to the creation of money, is considered inflationary. The trend is thus to limit its growth. It has been noted that budgetary deficits have been contained, and that the amount of the public debt in France has declined during the period where the financial system closely resembled the overdraft economy model. 26 The French specificity of the Treasury circuit with its network of correspondents guaranteeing considerable liquidity for annual management of the Treasury has been only one element reinforceing the French overdraft economy system, but it has not been determinant in its formation. In fact, neither the low level of firms' self-financing rates nor the preferences of households for real investments, especially housing (which explains in large part the tightness of the financial market over the last decades) are related to it. None the less, it must be noted that deposits resulting from credit operations can operate in the Treasury circuit and diminish bank resources accordingly, thus reinforcing the need for their refinancing. The presence of financial institutions specialized in particular forms of credit results more directly from the operating logic of an overdraft economy articulated on credit and where the interest rates are administered. The 'Lagayette' report on the financial outlook of the French economy defined it as having been an 'economy of administered financing' until 1986. 27 If the banking sector cannot meet the inadequately lucrative demand, the government uses the channel of financial institutions with special legal status to generate the form of credit leading to an overdraft economy. Until 1985, the practice of improved loans was quite developed in France. This involves loans benefiting from preferential interest rates, with the government picking up the difference from the market rates. In 1985, these represented. Fran~oise Renversez 479 nearly half the economy credits, which means that only half of the credits were determined by the market rates. Here one can perceive the expansion of the •administered financing economy'. However meaningful this formula may be, it does not establish the same link as the overdraft economy, namely that between the economy's need for borrowing and the closure of the monetary system's equilibrium by the central bank. The concept of the overdraft economy, as already noted, takes into account a two-level structure of indebtedness, that of businesses to the banks and that of the banks to the central bank. This articulation of the indebtedness of the real and financial sectors functions as the formative element of the financial system in the very sense of systems theory. Only the modification of this structure, anchored in the real variables of the economy and in particular the division of value added between wages and profit, can bring about a veritable evolution of the system and notably that of its mode of regulation. 3 MONETARY REGULATION IN THE OVERDRAFf ECONOMY In the theoretical case of the overdraft economy, as it was largely realized in France until the 1980s, with credits financing a determinant share of investment, the central bank does not have the room to manoeuver that would allow it to really modulate its refinancing. More recently, tempered forms of the overdraft economy have permitted the establishment of a new configuration of monetary control. Regulation in a Pure Overdraft Economy System Essentially focused on refinancing by the central bank, studies on the overdraft economy28 have tended to characterize it as the economy where credit supply, at a given rate of interest, is perfectly elastic to demand. This is 'the hypothesis of the Metric model, where the credit market is defined as one where, for a regulated debit rate, the banks have an infinitely elastic supply.29 The model includes a financial sector where the eqUilibrium level determines the conditions of realizing real eqUilibria. The presentation here will be limited to overall equations, with more detailed treatment of credit and demand for money. The following symbols are used: demand for credit short-term interest rate PCM household consumer price I all the magnitudes of the real sector (in prices) leading to demand for credit (productive investment, housing investment, stock fluctuations) CRBA supply of banking credit CREDEX net supply of non-banking credit Cd TXCT Monetary Circulation and Overdraft Economy 480 BMEX PARB TXAJJ REFIN DEPO M4 TXE RESOB TORES TXO Y o exogenous monetary base (cumulative budget deficit and net exchange resources at the Banque de France) bank's share in collected liquidities daily cash rate on monetary market refinancing of banks total deposits in banks money in M4 sense (total liquidities) savings fund interest rate required reserves deposited by banks rate of required reserves interest rate on bonds revenue of nonfinancial agents bonds held by the Caisse des dep6ts The demand for credit is expressed by Cd = Cd (TXCT, PCM, I) (17) The total demand for bank credit is CRBA = Cd - CREDEX (18) The determination of the short-term interest rate by the banks TXTC = TPXCT(l1CRBA, l1BMEX, PARB, TXAJJ, REFIN / CRBA (19) Since all the demand for credit is met by the banks at the debit interest rate that they fix, they meet their required reserves by refinancing themselves on the monetary market according to their share of the market in the collected deposits, as shown by the analysis of the central bank's obligation to refinance. This is expressed in the following form: Banks' share in total liquidities PARB= DEPO PARB(CRBA,TXE) M4 (20) Amount of required reserves RESOB = TORES x DEPO (21) Determination of banks' demand for refinancing REFIN = CRBA + RESOB - DEPO (22) The total demand for liquidity is then M4 = M4 (TXO, PCM, y) (23) Franc;oise Renversez 481 The Metric model thus takes into account the specific role of the Caisse des Dep6ts in establishing the equilibrium of the French monetary system. This financial institution, which manages the deposits of the savings fund network, intervenes in a regulatory capacity on the money market. The hypothesis of the model, which corresponds to a prior phase, is that the Caisse des Dep6ts automatically corrects any disequilibrium that could arise between the supply and the demand for money. Equation (24) expresses the eqUilibrium between supply and demand for money in accounting terms M 4 = CRBA + CREDEX + BMEX + 0 (24; The structure of the interest rates is given by: TXO = TXO (TXCn (25) The modeling of the French case shows the importance of practices that guarantee the security of refinancing 30 and signals the strong endogeneity of money in a financial system of overdraf economy. Is this to say that the investment rate realized in the overdraft economy is reached because every investment project compatible with the interest rate finds its means offinancing through the banking system? Such an analysis must be qualified on two levels: the behavior of the banks and that of the monetary authorities. Whatever the elasticity of credit supply to demand, the banks often opt for prudent risk management, as verified by econometric studies of the overdraft economy.31 At a given rate of interest, they differentiate the 'non-prices conditions' of credit, as it is termed by Moctigliani and Jaffee following Harris.J2 Furthermore, they impose loan conditions notably by taking the contractors' personal goods as security and by arranging the dates of payment, which, according to the risk classification that they place on the operation, modify the modalities of obtaining credit. Credit rationing 33 ensues, as has been shown for France by a series of studies in the context of credit supply whose management on the macroeconomic level made it appear perfectly elastic. 34 Since they cannot refuse to 'close the system', the monetary authorities try to act on the level of closure through credit restriction. This mode of credit regulation takes into account the necessity for an increasing volume of credit to guarantee the financing of the investment, but it tries to limit it by determining a rate of increase. This mechanism was implemented in France between July 1957 and February 1959, February 1963 and January 1967, and November 1968 and October 1970 and remained continuously in place between December 1972 and December 1984. The indicator of its effectiveness, set up by the Banque de France, is known as the 'rate of credit-restriction bite' .35 When it is significant, this bite rate corresponds to an organized rationing of credit. In effect, in this situation the banks have no reason to abandon their behavior in face of risk; rather, they are put in the position of intensifying it. The bankruptcy of the SMEs following the bite period of 1974 shows that the rationing behavior affecting the 482 Monetary Circulation and Overdraft Economy riskiest categories of bank management is reinforced. In a period of credit restriction and biting of the restriction. the credit supply is thus rationed for a given rate of interest. Here we find ourselves faced with the paradox of monetary policy in the overdraft economy: the financial system is based on the necessity of boosting the economy through credit, but the preferred means of controlling the increase of the money supply is likewise credit. Indeed, the tools of monetary regulation that are· appropriate for the functioning of the financial markets economy cannot produce their effects in the overdraft economy. More precisely, neither regulation by means of the interest rate nor regulation from the base can be applied without correctives. This is not simply a question of institutional specificities, but rather of institutional practices that have only been able to be implemented and maintained through mechanisms proper to the systems of the overdraft economy. What analysis may help to establish is a distinction between the characteristics of circumstantial operations and those that are intrinsic to the overdraft economy and which must be taken into account in the definition of monetary policy. Regulation in a Weak Form of Overdraft Economy The progressive establishment of a unified capital market in response to the desire for modernizing the means of financing the economy has modified the circumstances surrounding French monetary policy. With the adoption of new regulatory instruments, the question of their limited effectiveness raises that of the system's risk. In an open financial system, portfolio anticipations and arbitrages greatly strain the links between credit and monetary aggregates, as Brunner and Meltzer showed many years ago for the financial markets economy. 36 Thus regulation must take into account all the sources of financing and not only the credit distributed. In the French case this implied a reappraisal of the instruments of monetary policy in favor of the interest rate, in so far as the significant recovery of the firm's self-financing rate presented certain obstacles to its use. The interest rate is considered as an appropriate regulatory tool in a financial system where the different segments of the capital market are linked to each other. By controlling the rate of the interbank market, the central bank maintains its hold over all of the interest rates, on the hypothesis, sometimes invalidated in recent years, of the relative stability of the interest rate structure. This intervention on the short-term rates is classically complemented by an action on long-term rates through the bonds market. Here the size of the economy plays an essential role. To allow the interest rate to fluctuate, as did the Federal Reserve System under Paul Volcker, does not have the same impact in the United States and in a mediumsized open economy. The latter is very vuJnerable to interest rate fluctuations and to international flows of capital. For the central banks, the interest rate is the tool that permits these two elements of external constraint to be contained.31 Franroise Renversez 483 The Banque de France implements regulation through the interest rate essentially by means of interventions on the interbank market. Two techniques are used: invitations to tender and seven-day purchase and resale. The rate of invitations to tender establishes the bottom price of operations on the interbank market, while the rate of seven-day purchase and resale, which higher, should establish the ceiling price, although the bank's anticipation may temporarily place the interbank market rate above this level. These techniques for regulating banking liquidity are intended to keep the increase of the money supply within margins permitting the containment or reduction of inflation differentials with the main economic partners. Additional measures are intended to regularize the money rate from day to day. These consist of intervention on the interbank market either in the form of purchase and resale transactions to increase the banks' liquidity and thus lower the interest rate, or conversely, in the form of sale and repurchase transactions to draw on bank liquidity and thus sustain the interest rate. Since December 1986, the Banque de France has intervened more broadly through open-market techniques, which consist of buying or selling Treasury note sat the market price in order to make the interest rates go down or up accordingly. The procedures used include purchase and sale as well as sale and repurchase. The limited volume of notes issued by the French Treasury does not allow this technique to assume the place it holds on the US market. None the less, it marks a step toward regulation by means of market techniques. The required reserves are used to 'reinforce' the action of the interest rate, to borrow the expression of the National Credit Council.38 These had already been used in the earlier period; the regulation concerning them has been revised. It is primarily intended to make variations in the rate of required reserves function as a signal of the Banque de France's intentions. The elimination of the quantitative control of credit on the first of January J987 led monetary authorities to modify the bases for calculating reserves. Credit was no longer the variable to be controlled; reserves on assets were eliminated in favor of an increase in the reserve rates on the elements payable from liabilities. The reserve rate on checking deposits was set at 5 per cent, and that on fixed-term deposits at 1 per cent. These rates place the required reserves of French banks at a level below those of Germany and the US. The control of bank liquidity is thus expected to regulate domestic demand through variations in the interest rate. The internationalization of the capital markets that underlies the establishment of this mode of regulation is likewise at the origin, if not of an international determination of interest rates, then at least of a gesture toward the meaning of often uncontrollable variations. The possibilities of effective monetary policies appear limited in the current economic climate. Monetary regulation today occurs in a context where the central banks' field of action seems to be diminishing.39 484 Monetary Circulation and Overdraft Economy If the central banks' regulatory action carried out in the context of an open economy is subject to the effects of external constraint,40 it also intervenes in a financial system marked by the evolution of banking activities toward riskier assets. Thus the central banks are faced with systemic risk. But the action of the central bank, rather than basing itself primarily on the costs of refinancing, as was the case prior to 1987, must take into account portfolio behaviors and the sensitivity of financial intermediaries, and especially the banks, to the interest rates, which not only alter the cost of their resources but also, through the price of securities, the value of their assets. In other words, there is a cost effect of capital. The banks' behavior in fact tends to be based less broadly on a margin of exploitation calculated in relation to the interest rate of the interbank market than to be progressively articulated to the taking into account of appreciation or depreciation. Thus the central bank exerts its regulatory function over behaviors whose determination is much more complex than in a pure overdraft economy. There is a pronounced transition toward the financial markets economy, yet bank refinancing from the central bank remains important. It is this last feature relative to the closure of the monetary system's eqUilibrium that gives the impression that the French economy is still in an overdraft economy system, while the reappearance of a substantial rate of self-financing (over 80 per cent since 1986) could suggest an accelerated transition toward the financial markets economy. This evolution of the determinants of bank behavior underlies the marked concern for elaborating a prudential regulation among both the central banks and the institutions confronted by difficulties of international monetary regulation such as the International Payments Bank and the European Economic Community. The latter in particular has now implemented a harmonizing procedure for financial systems with differing notions of prudential rules. In France, studies conducted at the initiative of the Banking Commission have shown the differences existing between the French and Anglo-Saxon positions relative to the assessment of banking assets. The traditional French assessment is carried out in a patrimonial perspective based on the sale value of the assets, while the Anglo-Saxon assessment is built on the economic value of the operations, which is to say, on the balance sheet value at the time the commitments are fulfilled. The risk resulting from the growing importance of securitization of the banking balance sheets in France is greater than it would seem from the emerging share of securities portfolios in bank assets (12 per cent in 1989 as compared to 2.5 per cent in 1980). In fact, through the channel of the OPCVM and the Common Investment Funds (Fonds Communs de Placement, or FCP), which they manage, the banks take risks analogous to those of the financial intermediaries but which do not appear on their balance sheet. Instead, there is a unit of the bank as enterprise which bears the consequences of losses incurred by its subsidiaries OPCVM or FCP. Even though the prudential ratios are intended precisely to oblige banking Franfoise Renversez 485 establishments to ensure the necessary counterparts to their commitments, the volume of transactions off the balance sheet considerably increases banking risk. The question may then arise, as it did in the 1988 report of the National Credit Council, of the possibility that the microeconomic risks may spread and be transformed into macro financial risks. This explains the recent efforts of the central banks and their consultative bodies to come up with prudential rules guaranteeing greater security for the functioning of the system. A significant example is the central banks' acceptance of the Cooke ratio requiring that the banks have own funds equal to 8 per cent of their weighted credit commitments. The size of the French banks, which places them in good stead on the world level, is certainly an element in their favor. It should be noted that until now the base rate of the French banks has remained relatively stable and less affected by market fluctuations than that of the American banks.41 This can be seen as a manifestation of the effects of the credit market's oligopolistic structure. But, as indicated above, the banking base rate in France has lost the preponderant place that it held until 1984 as the reference rate for credits. Whatever the efforts of economic policy to encourage the financing of French enterprises with their own funds by facilitating their creation of gross savings, to the extent that these businesses depend on external financing and in large part credit financing, it is not desirable for the interest rate, which is the basic cost of this determining resource, to show great variation. In such a financial system, the task of the central bank thus consists of limiting such variations, which transmit the strong impulses resulting from the volatility of the exchange rates. The concept of the overdraft economy allows us to specify on the theoretical level the characteristics of a financial system marked by the endogeneity of monetary circulation. This circulation is strongly articulated, through credit, to the primary indebtedness of the productive system. Thus the central bank ensures the closure of the monetary system by refinancing the banks. In this respect its role is very different from that of the discretionary lender of last resort as attributed to the central banks in the general case called the financial markets economy. The opening of the economies has led to an evolution of this rather hardy financial system, which permitted the financing of Japan's growth together with that of France. The international transmission of the rise in interest rates and financial innovations has, in France at least, allowed for a considerable place to be given to the financing of businesses by the securities markets. Monetary circulation still remains quite endogenous, however, with public securities taking up only a small share of productive sector claims in the banks' assets. The endogeneity of money proper to the overdraft economy system continues, as does the paradox of its monetary policy: the new variable of intervention by the monetary authorities, the interest rate, can only act effectively in an antiinflationist perspective by affecting credit, which is the necessary condition for stimulating growth. 486 Monetary Circulation and Overdraft Economy Notes 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 'This text was completed at the end of 1990. R. Airikkala and T.R.G. Bingham (1978), 'The Formulation and Conduct of Monetary Policy in an Open Overdraft Economy: the Case of Finland', Banque de France, Cahiers Economiques et Monetaires, 6. H. Sterdyniak and P. Villa (1977), 'Du cllte de I'offre de monnaie', Annales de l'INSEE, 24; P. Artus, J. Bournay, A. Pacaud, C. Peyroux, H. Sterdyniak and R. Tessier (1981), METRIC: Une modelisation de /'economiejranfaise (Paris: INSEE) V. Levy-Garboua (1978), 'Le taux de change et la politique monetaire dans une economie d' endettement', Annales de l'INSEE, 32. G. Maarek (1978), 'Monnaie et inflation dans une economie d'endettement', Revue d'Economie Politique, 1. J.R. Hicks (1974), The Crisis in Keynesian Economics (London: Basil Blackwell pp.50--1. This is also the thesis of V. Levy-Garboua and G. Maarek (1985), lA dette, Ie boom, la crise (Paris: Atlas-Economica). G. Maarek (1984), 'Le partage de la valeur ajoutCc dans I'economie fran~aise. Une analyse retrospective 1970--1982', Revue de l'lPECODE, 4, pp. 1-24; G. Maarek (1984), 'Rentabilite et endettement des entreprises fran~aises. Une analyse retrospective, 1963-1982,' Revue de l'lPECODE, 6, pp. 1-31. Annual reports of the National Credit Council. A new nomenclature adopted in October 1990 has modified the 1987 definitions. See D. Bruneel (1985), 'La redefinition des agregats monetaires', Bulletin Trimestriel de la Banque de France, 57; D. Bruneel (1987), 'La nomenclature des institutions financieres', Bulletin trimestriel de la Banque de France, 62. J.P. Patat (1987), 'Les nouveaux agregats monetaires en France', Cahiers Economiques et Monetaires de la Banque de France, 27. Banque de France, Direction Generale des Etudes (1987-8), 'Presentation de la methodologie des statistiques monetaires', Bulletin Trimestriel de la Banque de France, 65. Government credits and economy credits are now combined under the heading 'net internal credit'. Systeme elargi de Comptabilite Nationale (1987), Collection de I'INSEE, series C, nos 140-1. J. Denizet (1982), Monnaie et jinancemenl. Essai de IMorie dans un cadre de comptabilite economique (Paris: Dunod, J. Denizet (1982), Monnaie etjinancement dans les annees quatre-vingl (Paris: Dunod). J. Gurley and E. Shaw (1960), Money in a Theory oj Finance (Washington: Brookings Institution). J. Tobin (1958), 'Liquidity Preference as Behaviour toward Risk', Review oj Economic Studies. This has been the case since January 1 1991. V. Levy-Garboua and B. Weymuller (1982), Macro-economie contemporaine (Paris: Economica). C. Toullec (1979), 'Economie de marche, economie d'endettement et politique monetaire', Banque (March), ·1979 V. and L. Levy-Garboua, (1972) 'Le comportement bancaire,le diviseur de credit et I'efficacite du controle monetaire', Revue economique, February. For a modeling that takes into account the financing of the public deficit and the diversification behavior of the banks' portfolio, see D.T. Llewellyn (1983), 'The Money Supply Process: A Generalised Approach', in D.T. Llewellyn, The Fran~oise 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. Renversez 487 Framework of UK Monetary Policy (London: Heinemann), and F. Renversez (1988), 'L'incidence mon~taire du financement des soldes budg~taires', in P. Llau and F. Renversez (eds) Stratlgies de financement des soldes budgltaires. Une comparaison internationale, (Paris: Economica). C. de Boissieu (1985), 'Economie d'endettement, economie de marcMs financiers et taux d'int~ret', in A. Barr~re (ed.), Keynes aujourd'hui: Thlories et politiques (Paris: Economica) 309-328. For a modeling, see Renversez in Stratlgies definancement, pp. 261-9. Denizet, Monnaie etfinancement (1987). C. de Boissieu, 'Economie d'endettement, ~onomie des marcMs financiers et taux d'int~r~t'. Llau and Renversez, Stratlgies definancement. Commissariat G~n~ral du Plan (1987). Perspectives definancement de l'lconomie franfaise, vol. 1 Rapport du groupe de trarail pr~sid~ par M. Ph. Lagayette (Paris: La Documentation Fran~aise), 1987. V. Levy-Garboua and G. Maarek, La dette,le boom,la crise. 'Mod~le METRIC', (1977), Annales de I'INSEE, 26-7. Ch. Joffre and S. Roubine, 'Le fonctionnement du marcM mon~taire', and 'La demande de refinancement aupr~s de I'Institut d'~mission: une ~tude empirique', Banque de France, Cahiers Economiques et Monltaires, 8, pp. 25-39,41-74. J. Laudy and N. Lombard, 'Comportement des entreprises en p~riode de morsure de I'encadrement du cr~dit', Banque de France, Cahiers Economiques et Monltaires, 8, pp. 167-95. D.M. Jaffee and F. Modigliani (1969), 'A Theory and Test of Credit Rationing', American Economic Review (December). A. Coutiere (1977), 'Le rationnement du cr~it en France'. Economie Appliqule. M. Chazelas, J.F. Dauvisis, and G. Maarek, 'L'ex¢rience fran~aise d'encadrement du cr~it', Banque de France, Cahiers lconomiques et Monltaires. B. Enfrun, J. Pecha, and L. Griffoul (1983), 'La nouvelle m~thode d'~laboration de I'indicateur de morsure de I'encadrement du cr~dit', Bulletin trimestriel de la Banque de France, 47. K. Brunner and A.H. Meltzer (1968), 'Liquidity Traps for Money, Bank Credit and Interest Rates', Journal of Political Economy; K. Brunner and A.H. Meltzer 'Money, Debt and Economic Activity', Journal of Political Economy 1972 80(5) 951-978 P. Artus (1990), 'La politique mon~taire dans les ann~es 1980', Document de Travail, Caisse des D~pots et Consignations, 1990-201E. Conseil National du Cr~it, Rapport Annuel (1987) p. 33. C. Bordes and M.O. Strauss-Kahn (1987), 1977-1986, Dix ans de politique d'objectifs en France, ou Ie 'targeting' la francaise (Tilburg: Suerf 1987). P. Artus, 'La politique mon~taires' and 'MarcMs finan~iers: internationalisation, mecanismes, risques et politiques' ibid M.A. Levy (1990) 'La rentabilit~ bancaire de 1978 ~ 1988 in Economice et statistique, 234, July-August. a References Artus, P. et al. (1981) METRIC. Une modllisatio1l de I'lconomiefranfaise (Paris: INSEE). Banque de France. Direction g~n~rale des etudes (1987-8), 'Pr~sentation de la m~thodologie des statistiques mon~taires', Bulletin trimestriel de la Banque de France, 65, December-January . 488 Monetary Circulation and Overdraft Economy Boissieu, C. de (1985), 'Economie d'endettement, economie de marches financiers et taux d'interet', in A. Barrere (ed.), Keynes aujourd' hui: Theories et politiques (Paris: Economica). Bruneel, D. (1985), 'La redefinition des agregats monetaires', Bulletin trimestriel de la Banque de France, 57, December. Bruneel, D. Commissariat General du Plan (1987), Perspectives de jinancement de I' economie fran~aise, vol. 1. Rapport du groupe de travail preside par M. Ph. Lagayette (Paris: La Documentation fran~aise). Conseil National du Credit. Rapports Annuels. Davidson, P. (1988), 'Endogenous Money, the Production Process and Inflation Analysis', Economie Appliquee. 41 (1), 151-169. Denizet, J. (1967), Monnaie etfinancement: Essai de theorie dans un cadre de comptabilite economique (Paris: Dunod). Denizet, J. (1982), Monnaie etjinancement dans les annees 80 (Paris: Dunod). Goux, J.F. (l990),'Les fondements de I'economie de decouvert: Apropos de la tMorie de la liquidite de Hicks', Revue Economique, 41 (4). Gurley, J. and E. Shaw (l960), Money in a Theory of Finance (Washington: Brookings Institution). Hicks, J.R. (1974), The Crisis in Keynesian Economics (London: Blackwell). Lavoie, M. (1985), 'Credit and Money: The Dynamic Circuit, Overdraft Economy, and Post-Keynesian Economics', in M. Jarsulic (ed.), Money and Macro Policy (Boston: Kluwer-Nijoff). Levy-Garboua, V. (1978), 'Le taux de change et la politique monetaire dans une economie d' endettement', Annales de I'INSEE, 32. Levy-Garboua, V. and G. Maarek (1985), La dette, Ie boom, la crise (Paris: AtlasEconomica). Llewellyn, D.T. (l983), The Framework of UK Monetary Policy (London: Heinemann). Livre blanc sur la reforme du jinancement de I'economie fran~aise (Paris: La Documentation fran~aise). Llau, P. and F. Renversez, (eds) (1988), Strategies de jinancement des soldes budgetaires: Une comparaison internationale (Paris: Economica). Maarek, G. (1978), 'Monnaie et inflation dans une economie d'endettement', Revue d'Economie Politique, 1. Metais, J. and P. Szymczack (1986), 'Les mutations du systeme financier fran~ais, innovations et dereglementation' , La Documentation fran~aise, Notes et etudes documentaires, Paris. Minsky, H.P. (1985), 'La structure financiere, endettement et credit', in A. Barrere (ed.), Keynes aujourd'hui: Theories et politiques (Paris: Economica). Moore, B. (1985), 'Wages, Bank Lending and the Endogeneity of Credit Money', in M. Jarsulic (ed.), Money and Macro Policy (Boston: Kluwer-Nijoff). Patat, J.P. (1982), Monnaie, institutions jinancieres et politique monetaire (Paris: Economica). Patat, J.P. (1987), 'Les nouveaux agregats monetaires en France', Cahiers Economiques et Monetaires de la Banque de France, 27. Rapport sur les comptes de la Nation (Paris: INSEE). Renversez, F. (1986), 'La France: Une economie d'endettement', in 'Les systemes financiers', La Documentation fran~aise, Cahiers Franfais, 226. Steindl, J. (1985), 'Epargne et endettement', in A. Barrere (ed.) Keynes aujourd' hui: Theories et Politiques (Paris: Economica). Tobin, J. (1958), 'Liquidity Preference as Behaviour Toward Risk', Review of Economic Studies, February. Toullec, C. (1979), 'Economie de marcM, economie d'endettement et politique monetaire', Banque, March. B.Explanations of Endogeneity: Accommodation or Structure? 18 Money Supply Endogeneity: What are the Questions and Why do They Matter?* Robert Pollin SPHERES OF THEORY AND REALITY The defining characteristic of the monetary and financial systems of advanced capitalist economies over the past thirty years has been one of persistent and fundamental change. Focusing on the US system, financial markets were highly regulated in the early 1960s. and the regulations created sharply segmented markets among various intermediaries. By the early 1990s, the regulatory system had been almost entirely abandoned, either through formal repeal of laws or the circumvention of remaining rules through innovation. The distinctions between banks. thrifts, insurance companies and brokerage houses have greatly diminished and promise to become even weaker, perhaps non-existent, in the near future. Interest rates were relatively stable and low in real terms in the early 1960s, partially because of regulations establishing ceiling rates, but also because of the general stability and constraints on competition in the market. By the 1990s, interest rates had long been freed from regulation. Rates became much more volatile from the 1970s onward, and, over the 1980s, rose to and maintained a level that was unprecedented at least since the 1920s. 1 The dollar was still the official reserve currency for international transactions through the 1960s, and its exchange rate was fixed relative to other major currencies through Bretton Woods. At the same time, US credit markets were largely self-sufficient in the 1960s, as, on average, only 1 per cent of credit market funds came from foreign sources. Currency markets became highly volatile after the demise of Bretton Woods, and exchange rate fluctuations became a major factor in setting US monetary policy, especially since the contribution of foreign sources to US credit markets rose to an average of 6.5 per cent over the 1980s. No serious financial crisis had occurred from the Depression years to the mid1960s, and indeed, in the 1960s high noon of macroeconomic fine-tuning, such 490 Robert Pollin 491 phenomena were considered as relics of a bygone era. By the early 1990s, financial crises have become a regular feature of the financial landscape, from the credit crunch of 1966, to the Wall Street crash in 1987 and the collapse of the Savings and Loan industry in 1990. The reality of a fundamentally changing financial structure is also evident from more formal measures of macrofinancial activity. This can be seen in Figure 18.1, which plots a measure of the velocity of money (Ml velocity) and the ratio of aggregate net borrowing by domestic non-financial sectors to GNP. We see first that Ml velocity has risen sharply. In 1960, each dollar of currency and transaction accounts purchased $3.67 of GNP, while in 1989 it bought $6.64 (a figure which, as the figure shows, is below the 1981 peak of $7.18 but still high by historical standards). The ratio of borrowing to GNP shows an equally dramatic increase. In 1960, we can say roughly that 6.6 cents of every dollar of GNP were purchased with borrowed funds. By 1989 over 13 cents of each GNP dollar were financed through borrowing (again, down from the 1986 peak year of nearly 20 cents). Remarkably, amidst such profound changes in the sphere of money and finance, the sphere of theorizing about money and finance has largely neglected these developments. Indeed, as the forces of change in the financial system gathered momentum from the mid-1960s onward, the economics profession came increasingly to embrace theoretical approaches in which financial markets and private institutions play essentially no role in explaining economic outcomes. I refer first, of course, to the dramatic ascendence of monetarist theory advanced by Milton Friedman and associates. Regarded initially in the 1950s as a negligible rehash of pre-Keynesian Quantity Theory, by the 1980s it represented a pre-eminent orthodox position throughout the world. According to monetarism, the only channel through which financial forces affect the aggregate economy is through the money supply; and significant changes in the money supply are generated not by market forces but by central banks - that is, they are exogenous to the private economy. Money demand, according to monetarism, is a stable function of permanent income and, as such, the velocity of money is also stable within any short-term period (Friedman argued that long-term velocity should fall because money is a luxury good). Fluctuations in the supply and demand for credit are merely a reflection of the more fundamental changes in money growth - they are 'supporting players', according to Friedman (1969: 189), in a drama in which the money supply is the headliner. It therefore follows that financial institutions, and the role of historical change in the financial system, carry almost no importance in the monetarist system: this, despite the massive researches into the history of UK and especially US monetary institutions conducted by Friedman and Schwartz (1963a, b; 1970; 1982). J ~ 60 -I, '" /' , 62 , 64 66 ~ , /' 68 70 '--' 72 , /" " /' \ ,," 74 \ '" Ml Velocity / / , " 76 / " " '" 78 ~ .- ,,', , " 82 ~ 80 .- / \ \ 84 , 86 ~ " "' ..... \ / r6 l7 88 .M! Scale 3 1-4 -- r5 '" / Figure 18.1 Measures of change in US financial structure: Ml velocity (GNPIMl) and net domestic borrowing/GNP Sources: Citicorp &onomic Database and Flow of Funds Accounts of The Federal Reserve System. 0.05 0.10{- 0.15 BmrowmW GNP 0.20 ; 0.25 8 ~ tv \0 Robert Pollin 493 But monetarists are not the only post-war school which cast aside institutional considerations in their theoretical models. Orthodox Keynesians, following the ISLM framework initiated by Hicks, accepted that financial institutions and market forces could be adequately characterized within a model which focused only on money supply and demand rather than a broader array of institutional variables. The Keynesians primary dispute with monetarists here was over the degree of interest elasticity of the money demand function, and this only weakly established any independent influence for private market forces in determining aggregate activity. Beyond this, orthodox Keynesians accepted Modigliani and Miller's conclusion that financial structures were 'irrelevant' for the valuation of non-financial firms, and, by implication, for broader macroeconomic outcomes as well. 2 It is not surprising, of course, that these theoretical approaches have encountered severe difficulties explaining the radical changes over the past thirty years. Thus, orthodox monetary policy is predicated on the central bank's ability to define, measure and control the growth of 'money'. Yet, as the market created interest-yielding demand deposits and other instruments possessing high degrees of liquidity, it became increasingly difficult to distinguish monetary from nonmonetary assets. The definition of what constituted 'money' has thus gone through numerous revisions over the past twenty years. Seen from the demand side of the market, these same effects of innovation led to the breakdown of what had been understood as an empirically established stable money demand function. Subsequent efforts to 'search for the missing money' proved fruitless, since researchers continued to neglect the central point, that institutional innovation was continuing to create an array of assets possessing more finely distinguished degrees of liquidity, and as such, there could be no stable demand function with any fixed set of 'money' assets. 3 Observing these and related problems, even as prominent a mainstream analyst as Benjamin Friedman (1988) acknowledged by the late 1980s that institutional change had so overtaken financial markets that mainstream theory no longer offered useful guidance for economic policy. 2 PATHSNOTTAKEN This yawning gap between the profound institutional changes in money and finance over the past thirty years and the indifference of mainstream theory to such considerations has brought an obvious intellectual challenge to the fore: the development of a monetary theory firmly grounded in institutional and historical realities. But of course, the foundations of such a theoretical approach already exist, and indeed possess a long and honorable heritage tracing back at least to the work of Thomas Tooke in the 1840s. Tooke developed a 'credit theory of money' in debates with the proponents of the 'currency school', the predecessors of contemporary orthodoxy.4 Against this approach, Tooke advanced two primary arguments that carry relevance today: 494 Money Supply Endogeneity there is a fundamental identity between different financial instruments, so that theory needs to focus on the array of instruments rather than on any single one; and the creation of credit by intermediaries takes place only because the nonbank public demands its creation. Thus, from Tooke, we begin to develop a theoretical framework in which, contrary to modem orthodoxy, broad credit conditions rather than narrow monetary aggregates are the focus of concern; and that demand forces, relative to supply, are given at least as important, if not a more prominent role in determining financial market behavior. In short, Tooke opened the door to the very type of institutional analysis which appears so appropriate for the present. This approach resurfaced in many subsequent incarnations - through Volume III of Marx's Capital; Wicksell, in his discussions on institutional forces determining velocity change; and of course Keynes in his Treatise on Money, though not in the General Theory. More contemporary efforts included the work in the 1950s and early 1960s of Gurley and Shaw, Tobin, Minsky, and the Radcliffe Commission in Britain. While there were significant differences in these various approaches, commonalities were evident: the focus, as with Tooke, was on credit rather than 'money'; and on private financial institutions, not simply central banks, as sources of aggregate increases in credit. Or to frame these points somewhat differently, to recognize the wide and perhaps even unlimited variability of money velocity in a financial system capable of innovation. As the Radcliffe Commission (1959: 133) started, 'the more efficient the financial structure, the more can the velocity of circulation be stretched without serious inconvenience being caused'. One of the great puzzles of contemporary economics is why this approach, which appeared so promising in the early 1960s, was completely disregarded thereafter, during precisely the period when the reality of a rapidly changing financial system emerged before our eyes. S 3 POST KEYNESIAN INTERVENTIONS One of the major contributions of Post Keynesian economics has been to keep this tradition alive and develop it in several important directions. In particular, one of the major advances of Post Keynesians has been the development of a modem theory of money supply endogeneity. In opposing the mainstream notion that the money supply grows strictly through central bank initiatives, Post Keynesians have advanced the Tookean argument that pressures emerging endogenously within financial markets are the basic determinant both of fluctuations in money supply growth and, more broadly, of credit availability. As such, Post Keynesian theory has again centered monetary theory on the behavior of private financial institutions as well as central banks; and also on how demand forces from nonintermediaries shape the behavior of the intermediaries and central bank. As this Post Keynesian literature has developed, what has also become clear is. that two distinct theories of money supply endogeneity have emerged within this Robert Pollin 495 tradition. To my knowledge, Rousseas (1986) was the first to present an extended discussion demonstrating the sharp distinctions between the two approaches. In a recent paper (pollin, 1991) which I delivered at the 1990 Levy Institute conference, I sought to elucidate these differences further, by isolating some key propositions associated with the alternative views, specifying them in a manner amenable to empirical testing, and conducting tests to measure the relative strength of the two approaches. The two approaches share the basic Tookean insight that the rate of money supply growth and, more importantly, broad credit availability are fundamentally determined by demand-side pressures within the financial markets. More specifically, both approaches accept the assessment by the former New York Federal Reserve Bank senior vice president Alan Holmes (1969: 73) that 'in the real world banks extend credit, creating deposits in the process, and look for the reserves later'. Where the two approaches diverge is in moving beyond Holmes's point to explaining the process of 'looking for the reserves later'. That is, how and where do the banks (and other intermediaries) obtain the needed additional reserves once they have 'extended more credit, creating deposits in the process?' One perspective argues that when banks and other intermediaries hold insufficient reserves, central banks must necessarily accommodate their needs. To act otherwise would threaten the viability of the financial structure, and hence the overall economy. Central banks can choose the means through which they will accommodate: either by increasing the availability of non-borrowed reserves through expansionary open market operations, or by forcing the banks to obtain borrowed reserves through the discount window. This decision will affect the cost to the banks of obtaining their needed reserves. But because central banks are obligated to accommodate the demand for reserves at the discount window, no effective quantity constraint exists on banks' reserve needs. I have therefore termed this approach as a theory of accommodative money supply endogeneity. According to the other perspective, central bank efforts to control the growth of non-borrowed reserves through open market restrictiveness exert significant quantity constraints on reserve availability. Discount window borrowing, in this view, is not a perfect substitute for open market operations. But what this view also stresses is that, when central banks do choose to restrict the growth of nonborrowed reserves, then additional reserves, though not necessarily a fully adequate supply, are generated within the financial structure itself - through innovative liability management practices such as borrowing in the federal funds, Eurodollar and certificate of deposit markets. We may thus refer to this second Post Keynesian approach as a theory of structural endogeneity.6 4 ALTERNATIVE CREDIT MONEY CURVES Commentators on my earlier paper have correctly pointed out that a basic way of distinguishing the different approaches is through their treatment of the loan 496 Money Supply Endogeneity supply curve (Palley, 1991) or credit money curve (Moore, 1991). To pursue this point, let us first define these terms within a T-account framework. A simple balance sheet for a modem depository institution would be one in which Assets = L + RR + E, and (1) Liabilities =D + N + F, (2) where L is loans, RR is required reserves, E is excess reserves, D is transaction deposits, N is non-transaction deposit funds, such as large Certificates of Deposit, and F is non-deposit liabilities, such as money market mutual accounts. Since assets equal liabilities, we can define the outstanding supply of loans as L= D+ N + F-RR-E (3) The 'credit money' supply is a more ambiguous concept, reflecting the difficulties inherent in any attempt to specify 'money' through a given set of assets rather than the functions that money performs. As discussed in Moore (1988) and Rodgers (1989), the term credit money is used by way of contrast with commodity money. Rodgers, for example, defines commodity money as 'one or the other of the precious metals such as gold or silver', and credit money as 'a generic term which refers to unbacked fiat money and credit in all its forms' (ibid., p. 171). Following Rodgers, we can define credit money broadly as M=C+D+N+F (4) where C is total currency held by the non-intermediary public. 7 According to these measures, loan supply and credit money are closely related categories, since M=L+ C+RR+E (5) Without as yet inferring assumptions about causality, it will be helpful to establish some relationships within a simple money multiplier model. We first measure the change of various components of M relative to the growth of transaction deposits D: aM= a(l + n +/+ c)D, (6) where n = NID,/= FlD, and c = CID. We define the monetary base as aB =aBOR + aNBOR and (7a) aB= aRR + aE+ ac, (7b) where BOR is borrowed reserves, NBOR is non-borrowed reserves, RR is required reserves and E is excess reserves. Robert Pollin 497 Since RR = rdD + r"N, where rd and rn are the reserve requirements on transaction and non-transaction deposits respectively, we have llB = llrp + llT"N + llE + llC (8) Substituting, we obtain llB = llD (rd +rnn + e + c), or llM - ll( rd + rn:+ e + (9) c) (10: The full expression for the credit money multiplier is therefore llM = II (1 + n + f + c) B rd +rnn+c+c (10: Figure 18.2 attempts to depict the money creation/lending process in a way which clarifies causal relationships within both accommodationist and structuralist approaches. We define the 'market period' as that between t\ and t4. According to both approaches, the credit money growth process begins in t \ with the increase in loan demand, llLd. This then leads to the increase in loan supply, llL' in t2. In t3' the impact of llL' is felt in both the private financial market and the central bank. Those who hold llL' may hold these assets as D, N, F, or C. The choices they make will then establish parameters for the central bank's decision as to how much it wishes to increase non-borrowed reserves, llNBOR. It is at this point that the accommodationist and structuralist analyses diverge. Two versions of the accommodationist position can be derived. The most prominent is that once the central bank has decided on a value for llNBOR, private intermediaries will then be able to determine how much they will need to borrow from the discount window. The central bank will be forced to supply the -fl -~ 6D ~ LS flRR I + .,. I +6Ni;DF + + flc I D.E - --~ \M fl NBOR + fl BOR Figure 18.2 The money supply process: structuralist and accommodationist transmission paths 498 Money Supply Endogeneity amount of borrowed reserves demanded, regardless of the amount: this is what establishes the infinitely elastic money supply. Within this argument, the values for n,!. c and e are implicitly held constant. At the same time, the central bank independently determines the interest rate through two mechanisms: by fixing the discount rate and by setting the federal funds rate through its control over tlNBOR. This is how a horizontal credit money supply curve is generated at the administered interest rate. A variation on this approach is suggested in some of Moore's (1991) recent writings, including his essay in this volume. Once tlNBOR is established, private intermediaries may engage in liability (and asset) management, which in our model means seeking to shift the values of n,!. c and e, specifically to raise nand ! and lower c and e. Moore especially discusses the significance of raising funds through issuing Certificates of Deposit, that is, increasing nand f. If nand! are successfully raised, this will of course reduce ARR, in which case, assuming ANBOR is constant, the amount of borrowed funds demanded (ABOR) will fall. Here, the infinitely elastic credit money supply can thus be established through a variable combination of tlNBOR + ABOR + A(n +fl. Nevertheless, in the accommodationist approach, this latter strategy will still have no effect on interest rate determination: the interest rate is still exogenously administered by the central bank through its capacity to set the federal funds and discount rate. The alternative approach advanced by structuralists becomes most clear if we assume, non-controversially, that I1NBOR O. V"> 0 =(iv + cv)/(l - k) (26) (27) (28) (29) (30) v + S + kD' =D' + BR + 1" (31) (32) DS =J)d (33) 'P (34) TS= where iv is the interest rate on deposits; iB is the interest rate on bonds; iT is the interest rate on time deposits; NBR is non-borrowed reserves; BR is borrowed reserves; id is the discount rate; A 1 is the expansionary open market operation variable; A2 is the expansionary loan demand shift variable; MRj is the marginal revenue for banks from asset j; MCj is the marginal cost to banks of 522 Beyond Endogenous Money liability j; p is the illiquidity discount on loans relative to bonds; CL is the constant marginal cost per dollar loaned of monitoring loans, including provision for expected defaults per dollar loaned;? CD is the constant marginal cost per dollar deposited of administering deposit accounts; Cr is the constant marginal cost per dollar deposited of administering time deposits; Y is nominal income; S is bank holdings of secondary reserves. Signs above all functional arguments represent assumed signs of partial derivatives. Equations (16), (17) and (18) represent the demands for currency, demand deposits and time deposits. These demands are assumed to be characterized by gross substitutability and depend positively on own returns and negatively on other asset returns. Equation (19) is the demand for reserves, and it is assumed for simplicity that time deposits have no reserve requirement. Equation (20) describes the supply of reserves which consists of a non-borrowed and borrowed component. The shift factor AI captures one off expansionary open market operations, while the slope of the NBR function with respect to iF captures the feedback behavior of the monetary authority. Equation (22) describes the demand for loans which depends negatively on the loan rate, and positively on the bond rate. If loan rates rise relative to bond rates some borrowers switch to bond financing. Equation (29) determines nominal income which is a positive function of the level of loan demand. Finally, equation (31) is the aggregate balance sheet identity. 8 The model has a number of significant features. First, banks hold bonds as secondary reserves and issue both checkable (demand) and non-checkable (time) deposits. This has important implications for the way in which banks accommodate increases in loan demand. Second, the model incorporates optimizing behavior on the part of banks through equations (23)-(29), which represent the first-order conditions for a representative mUlti-input (liabilities) multi-output (assets) banking firm. These conditions serve to incorporate asset and liability management by banks, a feature which is central to the dispute between accommodationists and structuralists, and bears importantly on the issue of endogenous finance. Additionally, the first-order conditions show how the activities of banking, regarding acquiring and applying financial funds, link interest rates across markets. By a process of substitution, equations (16)-(34) can be reduced to a two equation system given by ;-+ .-+ .++ .+. C(IF' A z) + kD(IF' A 2) = NBR(IF' AI) + BR(IF - 'd) (35) L(IF' A2) = (I - k)D(IF' A2) + T(IF' A2) + BR(IF -Id) - S (36) .-+ .-+;t+ .+. The endogenous variables are iF and S: the exogenous variables are AI' A 2, k, and id. The signing of the demand for checkable deposits with respect to iF assumes that increases in the general level of interest rates reduce the demand for checkable deposits. This occurs despite the fact that the rate on deposits also rises, and it reflects the fact that the interest differential in favor of time deposits rises. Thomas I. Palley 523 Additionally, it is assumed that the shift factor A2 increases loan demand and income. Thus the direct impact of increased loan demand outweighs any subsequent interest rate crowding out effect. Totally differentiating equations (35) and (36) enables solution for the comparative statics. The jacobian is unambiguously negative. The comparative statics are then given by di~dA, <0 di~dA2 >0 dSldA, <0 ifL; -(1-k)D; -T; -BR; <0 dSldA 2 < 0 if [(C; + kD; - NBR; - BR; )] [(1 - k)DA + TA - LA ]- F F F F F F F F 222 [(CA + kDA )(L; - (1 - k)D; - T; - BR; )] > 0 2 2 F F F F The signing of the change in iF with respect to the exogenous variables is standard. Note, however, that interest rates rise in response to an increase in loan demand, as the induced expansion in demand deposits creates an incipient aggregate scarcity of reserves. This induces partial accommodation by the monetary authority, increased borrowing from the discount window, and asset and liability management transactions (discussed below) that end up economizing on uses of reserves. The effect of A2 on S is ambiguous because of offsetting interest and income effects. For instance, an increase in A2 causes an initial tightening of the loan market which induces banks to sell secondary reserves to fund increased loan demand, but the subsequent increase in income raises loan rates which in tum raises the demand for non-checkable deposits while reducing loan demand. The combination of these adjustments may ultimately make for enlarged bank holdings of secondary reserves. So much for the technicalities of the model: what about differences from the accommodationist model? The critical difference is that the banking system undertakes active asset and liability management in response to a tightening of the federal funds market. Whereas in the accommodationist model the sole source of reserves is the monetary authority, in the structuralist model not only do banks have recourse to the monetary authority through the discount window, but they can also obtain reserves from the non-bank public by undertaking asset swaps and inducing liability transformations. These latter methods of raising reserves include: 1. portfolio substitutions by banks between secondary reserves and loans. These substitutions are accomplished by selling secondary reserves to the non-bank public, which extinguishes deposits and releases reserves to support the new deposits created by bank lending. This process reveals an important buffer-stock role for secondary reserves. If there are unexpected deposit outflows, banks sell secondary reserves to recapture liquidity: if there is an increase in loan demand, banks sell secondary reserves to finance 524 2. Beyond Endogenous Money the increase in lending. In a sense banks therefore perform their own openmarket operations with the non-bank public, so that although the total stock of reserves remains unchanged, the banking system is enabled to finance more loans; raising interest rates on time and other non-checkable deposits, which encourages a substitution out of currency and demand deposits into these liabilities. This process of substitution releases reserves, which the banking system can use to back deposits newly created by increased lending. Considered together, the above mechanisms illustrate the nature of the structuralist argument: the private asset and liability management initiatives of banks represent an integral part of the banking system's response to changes in the economy's financing requirements. Not only does this process apply to the financing of deposits created by expansions of bank lending, but it also applies to conditions of monetary policy tightening. Thus, whereas in an accommodationist framework banks passively accept a tightening of monetary policy, in a structuralist framework they take active steps to circumvent the tightening. 3 TOWARD ENDOGENOUS FINANCE The previous section highlighted the essential difference between the accommodationist and structuralist perspectives regarding the role of bank asset and liability management in the accommodation of loan demand. However, it is the argument of this chapter that both approaches are flawed because of their exclusive attention to the banking sector. In a monetary economy banks represent one element in the financial system, and in principle there is room for substitution in the manner in which payments are arranged. This means that it is necessary to move the analysis beyond the confines of the banking system and include other forms of financing. It is this feature that prompts the notion of moving beyond endogenous money toward endogenous finance. This section begins this process, and develops a model in which agents use trade credit as a medium of exchange, and money as the medium of settlement. Imagine an economy in which credit is the exclusive medium of exchange, and money is the exclusive medium of settlement. If transactions per period for the representative agent are Y;, and each period consists of T days, then the average daily money balance of the representative agent is M;= Y; IT (37) The actual pattern of agent i's money balances is shown in Figure 19.3. During the course of the period the agent has no need for money since all purchases are paid for with credit, but at the end of the period the agent needs to settle her outstanding debt, and this causes a spike in money balances. Thomas l. Palley 525 Money y.I - - - - - - - T Time Figure 19.3 Shows the pattern of agent i's money holdings in a pure credit card economy Now consider an economy in which both money and credit are used as the medium of exchange. If Y is the total level of expenditures, and b is the proportion of expenditures paid for with credit, we have O E~T which requires that average money balances for money transactions exceed daily money expenditures. Thomas I. Palley 531 References Bar-lIan, A. (1990), 'Overdrafts and the Demand for Money', American Economic Review, 80, pp. 1201-16. Baumol, W. (1952), 'The Transactions Demand for Cash: An Inventory Theoretic Approach', Quarterly Journal of Economics, November. Minsky, H.P. (1982), Can /t Happen Again? (Armonk, NY: M.E. Sharpe). Moore, B.1. (1989), 'A Simple Model of Bank Intermediation', Journal of Post Keynesian Economics, 12, pp. 10-29. Palley, T.I.(1987-8), 'Bank Lending, Discount Window Borrowing, and the Endogenous Money Supply: A Theoretical Framework', Journal of Post Keynesian Economics, 10, pp. 282-303. Palley, T.1. (1991), 'The Endogenous Money Supply: Consensus and Dissent', Journal of Post Keynesian Economics, 13(3), pp. 397-403. Pollin, R. (1991), 'Two Theories of Money Supply Endogeneity: Some Empirical Evidence', Journal of Post Keynesian Economics, 13, pp. 366-96. Rousseas, S. (1985), 'A Mark-up Theory of Bank Loan Rates', Journal of Post Keynesian Economics, 8, pp. 135-144. Tobin, J. (1956), 'The Interest Elasticity of the Transactions Demand for Cash', Review of Economics and Statistics, August. 20 Monetary Policy in an Economy with Endogenous Credit Money* Marc Lavoie Although quite a lot has now been written about money and credit by Post Keynesian authors, the literature on the consequent monetary policy has been rather sparse. The objectives of this chapter are to gather the little that has been written on Post Keynesian monetary policy, and to identify a few general guidelines around which Post Keynesians could agree when it comes to non-specific monetary policy recommendations. In the following, the term Post Keynesian is meant to cover contributions on money and inflation from the so-called American Post Keynesians, as well as those of the Cambridge (England) Keynesians, and those of the neo-Ricardians. It will be obvious that the views of the radical school are quite similar. I shall start by establishing how I view the links which exist between circuit theory and Post Keynesian economics. CIRCUIT THEORY AND THE POST KEYNESIANS Augusto Graziani (1990) has recently presented and analyzed the main characteristics of circuit theory, initially developed by Bernard Schmitt and Alain Parguez. I There is thus no need to duplicate his work. I shall focus here on what I consider to be the main contribution of circuit theory, which is to make us realize that no production in a monetary economy is possible unless the banks make advances to the firms. In a closed economy without government deficits, firms must necessarily borrow from the banks, and banks must necessarily borrow from the central bank if the latter has a monopoly on some sort of money required for transactions. Firms are, by logical necessity, perpetually indebted towards the banking system, and the latter is necessarily indebted towards the central bank. The structure of any monetized production economy is thus one of the overdraft type, as Hicks (1974) would have it, in contrast to the autoeconomy or the market economies. What circuit theory asserts is that all economies are overdraft economies: some are pure overdraft economies; others, the so-called market 532 Marc Lavoie 533 economies, are impure overdraft economies. In the latter, firms hold liquidities and banks detain liquid assets which they can dispose of. However, the existence of the accompanying financial markets does not in any way distort the causation which exerts itself so obviously in overdraft economies. The presence of financial markets, open market operations and static portfolio effects confuses the analysis but it does not modify the underlying structure. Much attention has been devoted recently to the issue of the endogeneity of the money supply. Its strongest exponent has claimed, and rightly so, that this endogeneity is not a matter of institutions but rather one of logical necessity (Moore, 1988: xi). By analyzing how credit money is integrated to production economies, circuit theory demonstrates why indeed endogenous money is a logical necessity, and it provides proofs to other standard Post Keynesian claims, such as the causal predominance of investment over savings. Several economists recognize to some extent the endogeneity of the money supply. There is some disagreement, however, as to the domain of validity of the concept of endogenous money. The monetarists, for instance, believe that money is endogenous in an open economy with fixed exchange rates. Their view of endogeneity is different from that held by Post Keynesians or the circuitistes. It is still based on a supply-determined stock of money, rather than a demanddetermined one. There are also economists who believe that overdraft economies are an anomaly rather than a logical necessity. Endogenous money is a phenomenon limited in space. Some countries are under an endogenous money regime whereas others are not. In the former, if only enough savings could be generated, credit would not be needed. The supply of money could become exogenous and under the control of the central bank. The views of these overdraft economists, as I have called them elsewhere (Lavoie, 1985), are represented in this volume in the chapter of Fran~oise Renversez. There are also economists who believe that endogenous money is a phenomenon limited in time. There would be historical periods where money is endogenous and others where it is not. In particular, it has been argued by Chick (1986) and Niggle (1991) that money is not demand-determined any more, since commercial banks impose supplies of money through their practices of liability management. The other, related argument, is that money is endogenous when central banks accommodate, whereas money is exogenous when banks refuse to accommodate. A tuned-down version of the above is to argue that money is only partially exogenous when monetary authorities refuse to accommodate, because of the presence of financial innovation, this being the position held by Pollin (1991). I believe that the chapter by Moore in this volume attempts to respond to this sort of objection based on historical regimes. In my own view, accommodation or the lack of it, liability management or the lack of it, and financial innovations or the lack of it are second-order phenomena compared to the crucial causal story that goes from debt creation to the supply of means of payment. For instance, liability management has been a pervasive and 534 An Economy with Endogenous Credit Money penuanent phenomenon in all financial systems of the pure overdraft type. It does not change the causality that goes from credit demand to the supply of money in any monetized production economy. Circuit theory provides the monetary foundations of a proper analysis of monetized production. As it has recently been admitted by a neo-Walrasian author, circuit theory is a particularly acute attempt at describing the reality of circulating money flows, something which cannot be found in models of the mainstream sort, except in Clower's (d'Autume, 1989: 27). The importance of bank advances has also recently been reasserted by Davidson. It is ironic to note that Davidson (1987: 151) also refers to Clower's admission that some authority must provide advances for the whole (neo-Classical exchange) process to get started. There is thus evidence that those researchers who desire to go beyond the helicopter money story have had to conf:ont the problems that circuit theory tackles. My own view of circuit theory is that it constitutes the proper foundations to a non-orthodox monetary theory, which itself must be part of a larger non-orthodox research programme encompassing effective demand as well as value theories. With respect to the policy issues that are to be discussed in the next sections, the crucial contribution of circuit theory is the importance attributed to bank credits. According to mainstream theory, money is considered to be the major short-run cause of price increases, and it is the only long-run cause of inflation. Circuit theory emphasizes the predominance of credit granting decisions by bankers, and the marginal role played by cash balances which are residual to the whole macroeconomic production process. Under that approach, it is clear that targeting money rates of growth is a very inefficient way to go about monetary policy since it deals with the residual consequences of price or output growth. Instead, if restrictive monetary policy must be pursued, it is the causal credit aggregates that must be constrained. One may have the impression that circuit theory is far top unorthodox to be taken seriously by practicians who are in positions of power. Consider, however, the following quote from a former Governor of the Banque de France, which summarizes neatly some of the major themes of the circuitistes. Recall that banks grant loans mainly to finus and borrow funds foremost from households. The ... mechanism of money creation may then be summed up in the following manner: banks lend to firms money which they simultaneously create (money creation). The finus hand it over for instance to their employees to pay wages, and the households keep it until they decide to spend it; then they return the money to the finus as they make purchases (money circulation). Finally the finus use the money received from the households to pay back the banking credits (monetary destruction). This scheme may be complicated but its essential structure cannot be changed. (translation of de La Geniere, 1981: 271) Marc Lavoie 535 No one, then, will be surprised to learn that at the time when monetarism was at the peak of its influence on bank officials, de la Geniere was objecting to the control of monetary aggregates through attempts to restrain the creation of highpowered money or manipulating interest rates (ibid.: 278). Because he believes that such a policy gives very uncertain results, and because of the causality implicit in the process of money creation and destruction quoted above, he favours direct quantitative controls on credit. This, we shall see, is also the policy view of many Post Keynesians. 2 STANDARD MONETARY POLICY The major role of monetary policy in standard economics is to check inflation. Of course, central banks face other tasks such as the management of foreign exchange rates, or the control of the intensity of domestic activity or the level of employment, but these tasks are related and often subordinated to the control of the rate of inflation. In mainstream economics, inflation is mainly an excess demand phenomenon, induced by an excess supply of money. In recent years, at least in North America, the burden of the fight against inflation has fallen upon the shoulders of the central bank, budgetary policy playing a subsidiary role. The fight against inflation in mainstream economics has taken two paths. There are those who, like the monetarists, would like the central bank to set an appropriate rate of growth of the money supply, enforcing that rate by directly controlling the high-powered money base; and there are those, the neo-Classical synthesis Keynesians, who recommend the use of contra-cyclical monetary policies, by raising or decreasing the rate of interest. More recently, this has taken the form of equating the real rate of interest to the long-term trend of the marginal physical product of capital. Several Post Keynesians have made policy recommendations which deviate little from these obvious and apparently reasonable mainstream policies. For instance, in one of the rare Post Keynesian chapter explicitly devoted to monetary policy, Dow and Earl (1982) recommend the use of both strands of mainstream monetary policy. The rate of growth of money assets should not be very different from that of long-run output. However, because of unstable expectations, this rate of expansion 'should be such as to produce the correct interest rate and expectations combination to encourage investment in downturns and discourage it if it runs ahead of capacity in upturns' (ibid.: 233). The latter statement is the neoClassical synthesis Keynesian proposition. As recently pointed out by Wray (1989: 1188), there are also paragraphs in Moore's recent book, the most complete statement of Post Keynesian monetary theory, which also sound quite mainstream. In numerous instances, Moore (1988: p. 264) indicates that interest rates could be too low and lead to inflation, or that when interest rates are not at their appropriate level, savings could constrain 536 An Economy with Endogenous Credit Money investment over the long run (ibid.: 348). As we shall see, there are however other parts where Moore goes beyond the case of demand inflation and this is precisely what we should do. 3 COST INFLATION The Interest Rate: A distributive Variable Although demand inflation cannot be rejected out of hand, Post Keynesian authors generally consider inflation to be a cost-push phenomenon. The simplest of these Post Keynesian views considers the unit wage rate to be the prime mover of the rate of inflation, as it has been forcefully argued in the past by Sidney Weintraub (1978) through his mark-up equation. 2 Cost-push inflation can, however, be seen in a more complex manner, as resulting from the conflicting claims over aggregate income by the three main classes of society. 3 It can be considered that entrepreneurs (firms), rentiers (financial institutions), and workers (labor unions) all attempt to get hold of a larger share of domestic product. 4 Rentiers try to appropriate a larger real rate of interest (i); entrepreneurs want a larger net real rate of profit (r - i); workers desire higher real wages and purchasing power (w/p). Under these circumstances, the rate of interest, in monetary or real terms, is not so much a variable designed to control the level of investment and that of aggregate demand, but rather it is a distributive variable, as it has been argued emphatically by some neo-Ricardians (Pivetti 1985; Roncaglia 1988) and also by some post-Keynesians (Eichner 1987: 860). Making such an admission does not necessarily imply that the rate of interest is the main determinant of the macroeconomic rate of profit, as the same neo-Ricardians would have it; it simply means that the rate of interest is an important determinant of the distribution of income between social classes and presumably between individuals. Let us then take for granted that inflation is not a monetary phenomenon, but basically the consequence of conflicting claims over income shares, and let us accept the statement that the monetary interest rate is fundamentally a distributive variable. It then becomes clear that monetary policy should not so much be designed to control the level of activity, but rather to find the level of interest rates that will be proper for the economy from a distribution point of view. The aim of such a policy should be to minimize conflict over income shares, in the hope of simultaneously keeping inflation low and activity high. Two points of view may be proposed at this juncture. One may be of the opinion that rentiers are parasites and that, as a consequence, interest rates should be kept as low as possible, even at negative real rates, eventually to get rid of this useless economic class. This was the view put forth by Keynes (1936), in his call for the euthanasia of the rentier class. Although they are not necessarily related to Marc Lavoie 537 this strategy, one should note that various Post Keynesian authors have repeatedly called for a monetary policy of low interest rates, either to keep the spot prices of capital assets high so as to induce accumulation (Davidson, 1972), or because it was felt that the behavior of interest rates was asymmetrical and that any increase would make it very difficult for interest rates to go down again (Lavoie and Seccareccia, 1988: 149). Another approach is to view rentiers as a necessary evil, a class the ranks of which are constantly replenished by the arrival of new retiring generations. Under these circumstances, the elimination of the rentier class is a utopia, the more so if that class is associated with the vested interests of financial institutions which, like industrial firms, are permanent institutions. The well-known warnings of Kalecki (1971, ch. 12) on the politics of full employment must then be taken into account. This amounts to finding a fair share for those earning interest income, in order to appease them while economic activity is high and inflation is looming. Such a fair share can be found in Pasinetti's (1981, ch. 8) concept of the natural rate of interest. Pasinetti shows, grosso modo, that when the real rate of interest is equal to the rate of increase of total productivity, there is no transfer of resources from the industrial to the rentier sector, or vice-versa. This means that the rentiers are holding on to their purchasing power in terms of labor units. Normally the rate of interest should be equal to the rate of inflation plus the rate of growth of labor productivity. Under those conditions, an amount of money equivalent to one hour of labor time, if lent at that normal rate of interest, will still be worth one hour of labor time when recovered with its interest payments. The purchasing power of the rentiers will increase if the productivity of the overall economy has increased. The relative situation of the rentiers in the social hierarchy stays the same, whatever economic conditions. Such a social contract, set by the monetary authorities, could convince the rentiers that they should not seek a slowdown of the economy, and could persuade labor unions to accept some form of permanent incomes policy. Such a policy, in one form or another (Tip, for instance), as is well known, has always been advocated by Post Keynesians as a way out of the inflation/unemployment trade-off dilemma (Davidson, 1972: 362; Eichner, 1976, ch. 8; Moore 1988: 389). However, the feasibility of such a scheme has just as often been questioned, in part because these schemes could not take care of the peculiar situation of the rentiers. The Interest Rate: An Exogenous Variable and a Cost Some may object that it is not very reasonable to suppose that interest rates can be set at their fair levels by the monetary authorities. Various objections spring to mind: what about the market forces determining rates of interest; what about Keynes's liquidity preference; what about the international constraints set by the interest rates abroad? One of the main theme of Post Keynesian monetary theory 538 An Economy with Endogenous Credit Money is that the rate of interest is a convention. It is the monetary authorities who decide which convention will prevail. The base rate of interest set by the monetary authorities - discount rate, minimum lending rate, federal funds rate, Treasury bill rate, or what have you - is the rate upon which all other fixed rates of return gravitate. The base rate of interest is exogenous. If the monetary authorities are sufficiently persistent, the base rate will prevail. Market forces and liquidity preferences can only influence the structure of the rates of the various assets, not the base rate itself, unless the central bank relents to market opinion (Lavoie, 1993). As to the real interest rate, since it is assumed that monetary policy has no direct bearing on the inflation rate, it is just as exogenous as the nominal rate. Indeed, this has been' verified in empirical studies (Mishkin, 1984). The most often heard objection to the presumed exogeneity of the rate of interest is that interest rates are set on international markets. Americans usually complain that high interest rates in Europe preclude the Fed from decreasing interest rate levels, while Europeans whine about high interest rates in the States. In Canada, when faced with complaining citizens, government officials would pretend that market forces are such that international investors ask for a traditional premium of 1 per cent over American interest rates to be induced in lending their money to Canadians. During the 1990-1991 period the short-term interest rate differential (nominal and real), has been between 4 and 5 per cent, an obvious proof that the so-called international markets are a myth. While the Bank of Canada was creating the first ever Canadian home-made recession, it was demonstrating to all that, even in an open economy with the United States as its free-trade partner, the rate of interest is indeed an exogenous variable, under the control of the central bank. Viewing the interest rate as an exogenous variable which is a component of the distribution mechanism, and also as a cost of producing goods and services, brings to the fore an impact of rising interest costs which is usually not taken into consideration when traditional restrictive monetary policies are advocated. NeoRicardians have recently emphasized this distributive role of the interest rate. Their view is that in the long run higher interest rates lead to higher profit margins over unit labor costs, that is, higher interest rates are conducive to higher prices (Pivetti, 1985: 95). This full-cost approach to pricing, which would include interest costs (explicit or implicit), is also sometimes endorsed by Post Keynesians, Kaldor (1982: 63) for instance, when he writes that 'there is evidence for believing that interest costs are passed on in higher prices in much the same way as wage costs'.s There is thus a further reason not to use high rates of interest as a monetary tool to fight against inflation. High interest rates initially induce more inflation since their cost might be transposed into higher prices. Furthermore, higher interest rates, if they are initially ineffective in restraining investment and employment, might simply increase the level of aggregate demand, since consumers will wind up with more income to spend, thanks to the increased revenues distributed Marc Lavoie 539 by loaned-up firms and by governments with large public debts. High interest costs might thus lead to higher inflation, either directly through the full-cost effect on prices, or indirectly as a consequence of higher effective demand, thus justifying the higher prices. One can build long-run models of effective demand yielding such results, with excess capacity and cost-plus pricing procedures incorporating the impact on prices of interest costs (a La neo-Ricardian) and of Eichnerian growth costs (Lavoie, 1993). Indeed, models where higher rates of interest induce higher rates of inflation have already been built (Dutt and Amadeo, 1993). 4 DEMAND INFLATION We have seen that in the most likely case of inflation, that is, cost inflation induced by conflicting claims over income, Post Keynesian theory called for an incomes policy accompanied by monetary measures to keep the rate of interest at a low level, compatible with some neutrality of income distribution between social classes. What if an economy is subjected to overheating or demandinflation? On standard Keynesian grounds, contra-cyclical budgetary policy should be followed in that case. However, it may be that, for political or bureaucratic reasons, budgetary policies take too much time to be put in place at the appropriate moment. Monetary policies would then have to be activated to fight against existing inflationary pressures. What then would a Post Keynesian anti-inflation monetary policy look like? At this stage, we have to remember that the Post Keynesian theory of money claims that central banks have little direct control over monetary aggregates but that they set the base interest rate, that is, the rate upon which the whole structure of interest rates is based. Furthermore, as the circuitistes would demonstrate, monetary aggregates are a consequence of the loans that have been granted in the past by (mainly) the banks. As Moore (1989: 483) points out, 'monetary change is determined by changes in the demand for bank credit, not for bank money'. An appropriate Post Keynesian monetary policy should thus focus either on the base interest rate, which the central bank can control, or on the flow of credit, which is the prima causa of economic activity. We have seen, however, that high interest rates can themselves be a cause of inflation. Furthermore, there has always been a belief among Post Keynesians that substantial movements in interest rates are necessary to have any impact on investment and thus on economic activity (Kaldor, 1964: 132). Therefore, any attempt at restraining inflation through changes in interest rates would require drastic steps to be effective. A major drawback would then be the fact that once interest rates have risen, the central bank needs to be very resolute and persistent to get the whole spectrum of interest rates down again. It follows that a policy of high interest rates should 540 An Economy with Endogenous Credit Money generally be rejected and that some alternative must be found. This alternative is direct credit control. 6 There is a tradition in Post Keynesian writings in favour of credit controls which goes back to Richard Kahn. 7 As he says, a monetary policy is best expressed 'in terms of rates of interest on loans of various maturities and of the amount of bank advances' (Kahn 1972: 128). In his evidence to the Radcliffe Committee, Kahn argues that if a restraining monetary policy must be put into place 'there is much to commend a direct approach, and to proceed by way of edict' (ibid.: p. 151) to limit the total of banks' advances. This approach of direct regulation of credit was partly endorsed by the Radcliffe Committee, which advocated quantitative ceilings on bank advances in case of emergencies (Rousseas, 1986: 103).8 Although initially he was opposed to credit control, when Kaldor (1982: 107) adopted reverse causation and the endogenous credit money view, he was also led to favor its use. Direct credit control, through the use of quantitative ceilings on loans, was also advocated by Le Bourva (1962: 54), the original modem exponent of the theory of endogenous money. At this stage, a reader brought up in demand and supply analysis might wonder how limitations on loans would not simultaneously affect the level of interest rates. The answer lies in the fact that borrowers are not rationed by the price mechanism but rather by procedural rules. These are Keynes's fringe of unsatisfied borrowers. Collaterals, liquidity ratios and the like are used by banks to select those customers who can have access to credit. The demand for credit money which would appear on a graph is thus the 'effective' demand for credit money once these rules have been applied. This explains both why credit is not totally unconstrained, although the money supply is said to be endogenous and its schedule horizontal, and why 'the behaviour of bank advances exercises an influence on investment which is separate from that exercised by rates of interest' (Kahn, 1972: 146). For the sake of deregulation, competition and fashionable monetarism, and also because limiting access to credit without disrupting the economy is not an easy task, most central banks in the 1970s reverted to trying to influence economic activity through the indirect means of interest rate variations and money supply growth. European central banks, which used to function under a system of discounting bank loans, have attempted to embrace an open market type of financial system, where the central bank does not appear to intervene directly. Central bankers are now much more pragmatic. Monetarism has failed as a theory, but it was politically successful in allowing central banks to pursue restrictive policies blindly while imputing to the impersonal market forces the drastic hikes in interest rates that they were imposed to the economy. As a reSUlt, central bankers are still reluctant to go back to the more direct and more obvious credit controls. However, in the little that has been written recently about monetary policy in Post Keynesian circles, one can see some trend in favor of the re-emergence of Marc Lavoie 541 some credit control when .necessary, or of the reappearance of a system of central bank discount. Rousseas (1986, ch. 6) has advanced a series of proposals involving various restrictions or requirements on the asset side of the accounts of banking institutions, calling these selective control proposals. Although Rousseas does not seem to favor direct controls as such, he is aware of the necessity to control the creation of money at its source, that is, when banks grant loans to businesses. Rousseas has thus designed schemes to modify the behavior of banks and their borrowers at the critical junction where a flow of credit is being created. This is where the inflationary process becomes activated, since firms are then allowed effectively to bid for the physical resources they require. Minsky (1986: 325) also favors selective credit controls based on inducements. He believes that these kinds of controls, which can even avert the tendency of capitalist economies to move towards more fragile financial structures, can be more efficient within a system of bank assets discounting by the central bank. Whereas in Rousseas' case inducements take the form of differentiated assets reserve requirements, for Minsky inducements are shaped by penalty discounting rates. In his book, Moore (1988: 132-6) also discusses credit controls. Although he recognizes that direct quantitative controls on bank credit are far more efficient than standard reserve requirements, he does not favor their use. His main objection is the possibility of disintermediation. When direct controls on bank advances are effective, businesses borrow from non-banking financial institutions or set up credit arrangements between themselves. This fear of disintermediation should be linked to Moore's proposition, which we have already cited, that rates of interest could be 'too low'. With credit ceilings and interest control, banks obviously have to weed out legitimate borrowers. But this is precisely what banks do anyway. Banks always ration credit, lending only to those that they consider to be credit worthy borrowers. In that sense, credit control is a natural function of banks, just as one may say that price control is a natural function of manufacturing oligopolies. They do it anyway. If credit control is of a permanent nature, the definition of banks must be more encompassing to avoid evasion through disintermediation. If direct credit controls are to be imposed in emergency situations only, we know, as it has been argued by Wojnilower (1980: 307), that they have 'startling potency'. 5 CONCLUSION Monetary policy in a world where endogenous credit money prevails and where interest rates are a conventional distributive variable cannot be identical to that of standard monetary policy. It has been presumed that cost-inflation, due to conflicting claims on the surplus, was the major type of inflation, while demand- An Economy with Endogenous Credit Money 542 inflation would occur in exceptional cases. Within that framework, three propositions have stood out: 1. 2. 3. Monetary policy, in normal times, should aim at fixing a fair rate of interest, consistent with the prevailing rate of growth of productivity so as to keep equal the relative potential purchasing power of rentiers and workers. Unless the rate of growth of productivity is negative, that rate would be above the rate of inflation. Commodity inflation, and conflicts generated by rentiers' inappropriate share of income, could thus be avoided. Fighting inflation through increases in interest rates needs excessively drastic steps, incompatible with the previous objectives, and leads paradoxically to higher prices if one accepts the full-cost pricing model. In times of demand-inflation, monetary policy should take the shape of direct credit controls, which should be as encompassing as possible, since bank loans are the fuel of inflationary expenditures. 9 Of course, monetary policy is not exclusive of other measures that could be taken in line with monetary measures. As we are all now aware, a successful capitalist economy which is close to full employment is prone to generate wage-led inflation, and as a consequence incomes policies could be permanently imposed (Cornwall, 1983). Furthermore, swings in economic activity should be compensated by budgetary policies, in particular investment planning by the state. Also, direct credit control is quite compatible with the advent of an industrial policy. In fact credit control may well be an absolute requirement for industrial policy. Those who believe that such recommendations are too radical and unrealistic in the present political climate should remember how fast the Fed in the United States has changed its monetary stance. Economic policies that are not fashionable today may well be tomorrow. Naturally, money and the real world are quite complex. The three propositions just advocated need to be developed and thought about. My intention in suggesting them was hopefully to trace a path upon which all Post Keynesians concerned with monetary policy could engage their reflection. Notes >I< \. I thank Ghislain Deleplace and my colleague Mario Seccareccia for their comments. Naturally, they cannot be held responsible for the remaining errors and omissions. I have myself discussed the similarities between Post Keynesian monetary theory and the strictly monetary aspects of circuit theory (Lavoie, 1985), as well as the more general relationships between the circuitistes and the American Post Keynesians (Lavoie, 1990). In the US, closest to the circuitistes are papers by Nell (1986), Parker Foster (1987), Terzi (l98~7) and Wray (1988, 1990, ch. 9). Incidentally, when Moore (1988, p. 314) denies all validity to the standard multi- Marc Lavoie 2. 3. 4. 5. 6. 7. 8. 9. 543 plier analysis, considering that the multiplier is identically equal to one, he is rediscovering a result previously demonstrated by Schmitt (1972). Authors such as Kahn (1972: 142) and Wood (1978) have put a lot of emphasis on wage inflation caused by relative wage conflict among the workers themselves. Kotz (1987) has pointed out that the conflict theory and mark-up theory of inflation may appear to be different but have a similar structure. A recent attempt to formalize a conflict theory of inflation can be found in Dutt (1990, ch. 4). The role of the rentier class has been underlined frequently by Parguez (1987). As Pivetti points out, various Classical authors endorsed this view. Non-orthodox authors have regularly suggested it, for instance Hotson (1976: 103). In their critique of monetary policy, and one could say of monetarist policy, Dow and Saville (1988) come to the same conclusions: monetary policy should generally be one of low interest rates; in emergencies, high interest rates and credit controls should be briefly imposed. One can even say that a tradition of credit control in Post Keynesian theory goes back to Thomas Tooke and the banking school. As Arnon (1984: 324) recalls, while Tooke was opposed to the control of money, on the Kaldorian grounds that any excess money would be extinguished by the repayment of loans, the mature Tooke admitted that excess credit could be inflationary. Arno-n concludes that Tooke must have implied that if there had to be any monetary policy, it is credit rather than money that had to be controlled. See Wulwick (1987) for a description of the similarities between the views of the central bankers at the time of the Radcliffe Committee and those of modern day Post Keynesians. A fourth proposition, designed to deal with the inherent instability of financial institutions, has also been put forward by various authors. Financial constraints and regulation should be reinstated: non-discretionary down payment and maturity limits on mortgage and installment credit (Wojnilower, 1980: 326), prudential controls on the structure of assets (Dow and Earl, 1982: 235). These constraints could help fight against speculative (instead of productive) lending activity. References Arnon, A. (1984), 'The Transformation in Thomas Tooke's Monetary Theory Reconsidered', History of Political Economy, 16(2), pp. 311-26. Chick, V. (1986), 'The Evolution of the Banking System and the Theory of Saving, Investment and Interest', Economies et Societes, 20, pp. 111-26. Cornwall, J. (1983), The Conditions for Economic Recovery (Oxford: Martin Robertson). D'Autume, A. (1989), 'La macroeconomie contemporaine', in La science economique en France (Paris; La Decouverte), pp. 16-30. Davidson, P. (1972), Money and the Real World (London: Macmillan). Davidson, P. (1987), 'Sensible Expectations and the Long-run Non-Neutrality of Money', Journal of Post Keynesian Economics, 10, pp. 146-53. De La Geni~re, R. (1981), 'Les I~ons de la politique monetaire', Banque, 404, pp. 269-79. Dow, I.C.R. and I.D. Saville (1988), A Critique of Monetary Policy: Theory and British Experience (Oxford: Oxford University Press). Dow, S.C. and P. Earl (1982), Money Matters (Oxford: Martin Robertson). Dutt, A.K. (1990), Growth. Distribution and Uneven Development (Cambridge: Cambridge University Press). 544 An Economy with Endogenous Credit Money Dutt, A.K. and E.J. Amadeo (1993), 'A Post Keynesian Theory of Growth, Interest and Money', working paper. Eichner, A.S. (1976), The Megacorp and Oligopoly (Cambridge: Cambridge University Press). Eichner, A.S. (1987), The Macrodynamics of Advanced Market Economies (Armonk, NY: M.E. Sharpe). Graziani, A. (1990), 'The Theory of the Monetary Circuit', Economies et Societes, 24, pp.7-36. Hicks, J. (1974), The Crisis of Keynesian Economics (Oxford: Basil Blackwell). Hotson, J.H. (1976), Stagflation and the Bastard Keynesians (Waterloo: University of Waterloo Press). Kahn, R. (1972), Selected Essays on Employment and Growth (Cambridge: Cambridge University Press). Kaldor, N. (1964), Essays on Economic Policy, vol. I (London: Duckworth). Kaldor, N. (1982), The Scourge of Monetarism (Oxford: Oxford University Press). Kalecki, M. (1943), Selected Essays on the Dynamics of the Capitalist Economy (Cambridge: Cambridge University Press). Kleeks, M. (1971), Selected Essays on the Dynamics of the Capitalist Economy 1933-1970 (Cambridge: Cambridge University Press). Keynes, J.M. (1936), The General Theory of Employment, Interest and Money (London: Macmillan). Kotz, D.M. (1987), 'Radical Theories of Inflation', in The Imperiled Economy, book I (New York: Union for Radical Political Economics), pp. 83-92. Lavoie, M. (1985), 'Credit Money: The Dynamic Circuit, Overdraft Economics, and Post Keynesian Economics', in M. Jarsulic (ed.), Money and Macro Policy (Hingham, MA: Kluwer Academic Publishers). Lavoie, M. (1990), 'Le circuit dans la pensee post-keynesienne americaine', Economie, 6, pp. 105-18. Lavoie, M. (1993) 'A Post-Classical View of Money, Interest, Growth and Distribution', in G. Mongioui and C. RUhl (eds), Macroeconomic Theory: Diversity and Convergence (Aldershot: Edward Elgar), pp. 3-21. Lavoie, M. and M. Seccareccia (1988), 'Money, Interest and Rentiers: The Twilight of Rentier Capitalism', in O.F. Hamouda and J. Smithin (eds), Keynes and Public Policy after Fifty Years, vol. 2 (Aldershot: Edward Elgar), pp. 145-58. Le Bourva, J. (1962), 'Creation de la monnaie et multiplicateur de credit', Revue Economique, 13, (I), pp. 29-56. Minsky, H.P. (1986), Stabilizing an Unstable Economy (New Haven: Yale University Press). Mishkin, F.S. (1984). 'The Real Interest Rate: A Multi-Country Empirical Study', Canadian Journal of Economics, 17 (2), pp. 283-311. Moore, B.J. (1988), Horizontalists and Verticalists (Cambridge: Cambridge University Press). Moore, B.J. (1989), 'On the Endogeneity of Money Once More', Journal of Post Keynesian Economics, 11 (3), pp. 479-87. Moore, B.J. (1995), 'The Money Supply Process: A Histological Reinterpretation', this volume. Nell, E.J. (1986), 'On Monetary Circulation and the Rate of Exploitation', Thames Papers in Political Economy, Summer, pp. 1-35. Niggle, C.J. (1991), 'The Endogenous Money Supply Theory: An Institutionalist Appraisal', Journal of Economic Issues, 25, pp. 137-51. Parguez, A. (1987), 'Introduction 11 I'economie des rentiers', Economies et Soc;etes, 23 (4), pp. 103-20. Marc Lavoie 545 Parker Foster, G. (1987), 'Financing Investment', Journal of Economic Issues, 21 (1), pp.101-12. Pasinetti, L.L. (1981), Structural Change and Economic Growth (Cambridge: Cambridge University Press). Pivetti, M. (1985), 'On the Monetary Explanation of Distribution', Political Economy, 1 (2), pp. 73-103. PoIlin, R. (1991), 'Two Theories of Money Supply Endogeneity: Some Empirical Evidence', Journal of Post Keynesian Economics, 13 (3), pp. 366-96. Renversez, F. (1995), 'Monetary Circulation and Monetary Economy', this volume. Roncaglia, A. (1985), 'The Neo-Ricardian Approach and the Distribution of Income', in A. Asimakopulos (ed.), Theories of Income Distribution (Boston: Kluwer-Nijhoff, pp. 159-80). Rousseas, S. (1986), Post-Keynesian Monetary Economics (Armonk, NY: M.E. Sharpe). Schmitt, B. (1972), Macroeconomic Theory (Albeuve: CasteIJa). Terzi, A. (1986-7), 'The Independence of Finance from Saving: A Flow-of-funds Interpretation', Journal of Post Keynesian Economics, 9 (2), pp. 188-97. Weintraub, S. (1978), Capitalism's Inflation and Unemployment Crisis (Reading, MA: Addison Wesley). Wojnilower, A.M. (1980), 'The Central Role of Credit Crunches in Recent Financial History', Brookings Papers on Economic Activity, 2, pp. 277-326. Wood, A. (1978), A Theory of Pay (Cambridge: Cambridge University Press). Wray, L.R. (1988), 'Profit Expectations and the Investment-saving Relation', Journal of Post Keynesian Economics, 11 (1), pp. 131-47. Wray, L.R. (1989), 'Two Reviews of Basil Moore', Journal of Economic Issues, 23 (4), pp. 1185-9. Wray, L.R. (1990), Money and Credit in Capitalist Economies: The Endogenous Money Approach (Aldershot: Edward Elgar). Wulwick, N. (1987), 'The Radcliffe Central Bankers', Journal of Economic Studies, 14 (4), pp.36-50. Part V Innovative Finance Introduction to Part V The financial system has developed both new instruments and new relationships, on a global scale, in recent years. This both opens opportunities and raises new questions about stability, as well as posing the danger that distant events may have local repercussions. A. FINANCE, RISKS AND CRISES Aglietta's Chapter 21 first analyzes the microfoundations of systemic risk in a monetary economy under uncertainty. It emphasizes the role of asymmetrical information in the credit system and of dynamic externalities in the payments system. It dismisses as unviable the ultra-liberal view of a moneyless economy eradicating systemic risk. On the contrary, it asserts the functional uniqueness of banking and provides a rationale for the Lender of Last Resort and for prudential regulation. The chapter also applies this theoretical model to the present financial globalization, examines the emergence of an international financial safety net, and states the basic principles needed to build the required institutional framework. Wolfson in Chapter 22 begins by sketching the outline of a typical business cycle in conditions in which the fragility of the financial system increases during the upswing, leading to a financial crisis. This can be defined as the emergence of a sudden, intense demand for money in the context of a sharp interruption in the supply of credit. The long, slow expansion of the 1980s exhibits the pattern of growing financial fragility. Interest coverage falls, debt-equity ratios rise, profit rates first rise then fall, but remain lower than in the 1960s. Nominal interest rates are lower than their peak period 1979-81, but are still high. Bank loan losses rise. In general by the end of the 1980s the signs point to serious financial problems. B. FINANCE AND THE BUSINESS CYCLE Semmler and Franke in Chapter 23 focus on financial-real interactions, spelling out and providing theoretical rationales for empirical hypotheses regarding such relationships. They introduce measures to help study the impact of both financial 549 550 Introduction and real forces on investment. Regression results using detrended variables provide strong support for lagged investment functions based on a broadly defined accelerator. Variables representing financial conditions significantly contribute to explaining cyclical variation in investment expenditures. The best contribution to explaining the cyclical variation of investment was provided by the interest rate spread. There is evidence as well of financial-real interaction. Jarsulic in Chapter 24 is concerned with financial dynamics, in particular with the dynamic complications introduced by the mere existence of debt. Given that investment demand is influenced by net cash flow, fixed payment commitments create the possibility of instability. The point is explored in an aggregate dynamic model with two endogenous variables, the rate of capital stock growth and the debt--capital ratio, and three functional relationships-growth is determined by lagged net cash flow, profits by capacity utilization, and the debt--capital ratio by a dynamic flow of funds constraint. In general the conclusion is that, for plausible values of the parameters, almost any dynamic outcome is possible. There can be no presumption of stability. A. Finance, Risk and Crises 21 Systemic Risk, Financial Innovations, and the Financial Safety Net Michel Aglietta 1 INTRODUCTION: THE LENDER OF LAST RESORT: FROM BAGEHOT TO THE PRESENT TIME Money does not manage itself. This was Bagehot's contention, backed up by his large experience with the London money market. He drew operational conclusions and stated what is known to be the Classical theory of the lender-of-Iastresort (LLR). It was the most articulate set of principles to guide LLR interventions in the nineteenth century. The lender-of-Iast-resort shall be devoted to the overall stability of the financial system and shall not conflict with the long-run objectives of monetary policy. To achieve this goal, the LLR shall provide liquidity assistance quickly but temporarily in case of liquidity stringency, to stifle any panic before it gathers momentum. The LLR has responsibility towards the entire financial system. But, in order to avoid moral hazard, he shall not bail out individual failures. The role of LLR function is to preclude potentially adverse consequences of global financial stocks on the real economy and to shackle contagious effects of individual failures upon sound financial firms. To apply these principles, Bagehot gave his most famous recommendation to central bankers: lend freely during a financial crisis to all sound borrowers against good collateral. The potential borrowers might be any economic agent without discrimination, banks or non-banks, financial or non-financial institutions. The acceptable collaterals are any kind of securities which usually back loans and are good standing paper on normal time; the eligible paper shall be accepted by the Bank at its pre-crisis market value, so that debtors do not suffer from the externality of the general slump in the prices of all financial and property assets during the crisis. However loans shall be delivered at a penalty rate, carrying a premium over the Bank discount rate. There are two reasons for this premium: first it deters borrowers from resorting to the emergency loans while other sources of liquidity are still available; second, it compensates the Bank for the risk of potential losses she takes while accepting collaterals at their pre-crisis values. Finally, to sustain public confidence and discourage excessive risk taking that could stem 552 Michel Aglietta 553 from the rescue of insolvent financial institutions, the ultimate lender shall announce the guiding principles and keep them firmly. Reviewing Bagehot's big book from the perspective of twentieth-century developments both in monetary theory and policy, one cannot help being struck by the topical interest of the debate he raised, by the large deviations of present practice from his principles, and by the insulation of his ideas from mainstream economic theory in his time and ours. Let us take these three points in turn, starting with the last one. Bagehot provided the first comprehensive statement of central banking history, but it was completely severed from economic theory. Acknowledging systemic risk means dealing with an economy where perfect competition does not work. Financial crises can arise in the shape of global disruptions occurring in the process of allocating capital and which have lasting negative effects on the real economy, if not impairing it. From one crisis to another, a pragmatic body of knowledge evolved in path-dependent fashion, assigning to a reluctant central bank the responsibility of thwarting financial disaster. Bagehot codified the pragmatic knowledge. However, it was completely separate from the basic concepts of a competitive economy which were founded at the time of Bagehot's writings. For, in a competitive economy there can be imbalances and diversifiable risks but there cannot be financial crises and systemic risks. Therefore, the role of the lender-of-Iast-resort was deemed empirically necessary long before its necessity was theoretically established. It was only after Knight and Keynes that the proper analytical tools were made available. It could then be demonstrated that systemic risk is the product of competition itself, the endogenous outcome of competitive forces under conditions of uncertainty. In the twentieth century, LLR interventions have operated in quite different contexts from those envisaged by Bagehot. Last resort loans were the only interference with an otherwise unregulated financial system in the pre-1914 liberal economic order. They impeded the free interplay of economic forces in extreme circumstances. But they were made for the sake of orderly market adjustments. This is why Bagehot insisted upon lending to solvent firms and not to insolvent institutions, and upon providing liquidity assistance to the economy at large and not to banks as specific entities different from all others. In the Classical view, the lender-of-Iast-resort was an extension of monetary policy. It was not a component of what was later called banking policy and which might conflict with monetary policy. On rebuilding the devastated financial system, LLR was amalgamated into the much more ambitious conception of a safety net. This concept singled out banks as entities requiring special regulations: reserve and (or) capital requirements, deposit insurance (publicly or privately run according to the countries concerned), supervision, compartmentalization of financial activities (some exclusively run by banks, some that banks were forbidden to enter). The safety net was a set of rules, specific to each country. Such rules had to be coordinated 554 Systemic Risk, Financial Innovations by one or several regulatory authorities. This coordinating process is the banking policy. Within the structure of the safety net, the role of the lender-of-Iast-resort changed drastically. It moved away from Bagehot's Classical principles and its rationale became blurred. Since banks make up the hub of the financial system. it became common wisdom within the regulatory authorities that large banks should not fail. Therefore, lender-of-Iast-resort might be obliged to support insolvent institutions, with possible adverse consequences on money supply and on the incentives to take risks. The rationale of LLR interventions cannot be easily derived from general and transparent principles. The criteria of their effectiveness and equity depend on the structure of the safety net. They must be assessed by a comparative analysis of how the safety net works in different countries. For nearly two decades after World War II, the safety net worked efficiently. Bank failures were very limited; stable financial conditions made LLR interventions rare and modest in size; deposit insurance schemes were not seriously tested. Meanwhile, the underlying financial structure began to change, especially in the 1960s for the US and in the early 1970s for the UK. The stimulus of competition made banks expand their assets tremendously, enter new business (consumer finance) and recover former ones which had contracted drastically in the segmented financial systems following the Great Depression (international finance). To fund their greatly expanding credit, banks began to circumvent existing regulations and traditional banking practices, seeking more and more borrowed funds. Aggressive liability management became common practice. As long as their core represented a declining share of their assets, banks were exposed to interest rate and exchange risks on top of their higher credit risks. Banks had become much more vulnerable to international shocks and to unpredictable changes in monetary policy, which itself responded to more unstable financial conditions. In the 1980s, competition picked up a qualitatively new momentum. The barriers between banks and non-banks, built into the safety net for reasons of prudence. crumbled because of a wave of deregulation. The securitization of the liabilities of non-financial entities (businesses and individuals) radically changed the bearing of credit risk. On the one hand, the direct issue of secl,lrities induces a cohort of contingent liabilities for banks in complex financing schemes that make credit risks much more difficult to assess and allocate. On the other hand. the securitization of credit itself. that is the repackaging by banks and other financial institutions of conventional loans into securities. completely fragments the credit process. Furthermore. the globalization of finance has increased international financial interdependence in more than one way: the size and speed of financial transactions puts alarming tensions into the electronic wholesale payments systems; the international interbank money market is a network of potentially volatile deposits, vulnerable to runs induced by a flight to quality, and is as strong as its weakest link; the securities markets are directly linked by the contagion of expectations, creating the possibility for generalized and highly exaggerated fluctuations in the prices of financial assets all over the world. Michel Aglietta 555 The radical changes in domestic and international finance have renewed the debate on the relevance and scope of lender-of-Iast-resort activity, and have begun with delays in making the authorities aware of the need to rebuild the safety net in order to meet the challenge of globalization. The present time is a time for asking basic questions and receiving confused or highly conflicting answers. Many experts in finance ask themselves what is special in banking to justify a particular regulation. Some radical liberals even consider that free market institutions should get rid of money itself. Some proposals want to push the trend of innovation to its ultimate consequences by breaking up banks and reshaping banking functions separately. Such sweeping structural reorganizations would have far reaching consequences for two pillars of the safety net: lender-oflast-resort and deposit insurance. We have lived with a safety net based upon the differentiation of institutions. What would a safety net remodeled along functionallines look like? Does it mean that the umbrella of the lender-of-Iast-resort should be extended to the entire financial system? Or, on the contrary, does it mean that it should concentrate on the protection of the payments system and on the prevention of runs on deposits more narrowly defined than they are now? Nobody is yet able to provide definite answers for all these questions. However, it is possible to sketch the analytical framework necessary to clear up some theoretical puzzles and political dilemmas. First, I will try to provide a synthesis upon the theoretical underpinnings of systemic risk and the disruptive processes that have to be held at bay (section I). Then I will examine some of the dilemmas of the lender-of-Iast-resort in the changing financial structure (section 2). Finally, I will outline a framework for a possible safety net compatible with the globalization of finance. THE RATIONALE OF BANK RISK Keynesian theory never cast any doubt upon the rationality of individuals. But the informational contents of the competitive interplay of individual behaviour is all important in determining the outcome. A market economy where money matters is plagued with uncertainty. It is not an assumption a theorist is free to make or not. Uncertainty is the basic fact of human life: human beings have limited rationality simply because they are not gods; markets are incomplete simply because human societies live in historical time. If time is an irreversible arrow directed from the past to the future, the body of knowledge people possess and draw upon to represent their relationships and make their decisions is unable to forecast the knowledge which will generate future knowledge. This is radical uncertainty. In such an environment, individual rationality is expressed with strategic decisions, with expectations upon expectations of others with mimetism. In such an environment, relationships are built upon asymmetrical information. There can be systems risk, that is, potential dynamics launched by competition whose outcome 556 Systemic Risk, Financial Innovations is global disruption of the market economy. If potential systems risk becomes effective, it is a financial crisis. To keep it latent and preserve financial security, a market economy needs a collective organization which is not only a set of rules to guide private behaviour, but also a set of institutions empowered with the public ability to act. This is the foundation of central banking. The analysis of competitive behavior under uncertainty brings into playa combination of the sources of uncertainty, the individual rationality under uncertainty, the germinal processes of a crisis, and the dynamics of global disruption. The four anaIyticallevels of a behavioral theory of uncertainty are observed in three interrelated sectors of finance: securities markets, intemediated credit, and payments systems. Crossing the behavioral items and the financial sectors, one can get a synopsis of systems risk (fable 21.1). The phenomena involved have been studied in numerous works. I am only concerned here with the connection to systems risk. This will legitimize for a financial safety net and a lender-of-Iast-resort in a free market economy. Bursts of Speculative Bubbles in Financial and Property Markets were often Triggers of Crises in the Past They no longer are. To cause a severe disruption, the slumps in assets' prices must spread from one market to another. Only a general slump would depress the wealth of non-financial entities to the extent that they cut real expenditure. Only the attempt of financial firms to sell securities forcefully and try altogether to enhance their liquid balances would increase the interest rate to the extent that businesses have to unload their inventories of primary commodities. The loss taken by firms in commodities and capital goods' sectors can impair the financial health of banks which lent to them. Finally, depositors of those banks can become anxious and can trigger runs on deposits. However, the chain reaction from a crisis originating in a particularly speculative market can take place only if the interest rate increases in response to the decline of those particular asset prices. It can happen only if monetary policy is passive before the substitution to safer assets and the shift of the liquidity preference schedule. The money supply can be rigid because of convertibility rules on a metallic standard. But, in a modem monetary system, it does not have to happen. In principle, the financial crisis can be thwarted not by a LLR intervention, but by a flexible monetary policy. Besides, the aftermath of the 1987 crash in equities markets clearly demonstrates this case. All the central banks increased the money supply, lowering short-term interest rates and making it known that liquidity would be available in the next few months. Market opinion expected future interest rates to remain at their lower level, inducing long rates to fall. People could safely shift from shares to bonds, fulfilling their expectations of a lower bond rate. Therefore, the value of their bonds in portfolio increased, offsetting much of the decline in equity values. Finally, there was no negative wealth effect upon consumption. The decline in interest rates was all-important in stimulating consumer and business loans, giving a renewed impetus to real expenditure. Sources of uncertainty Fundamental values unknown or misunderstood Time inconsistency. moral hazard and externalities in the creditor- 0 with r the rate of return on capital, (i - II) the real interest rate (nominal interest rate minus the rate of inflation), and p the deviation of the prospective rate of return from the current rate of return. The excess of the expected rate of return, r + p, over the real interest rate, i-II, is the measure of risk of investment projects. In this context, the cost of capital funds consists of the (real) interest rate and the cost of risk stemming from borrowing outside funds. The latter is considered to be al) increasing function of the ratio at which investment is debt financed. 22 Thus, given default risk and a real interest rate, investment should vary with the expected rate of return or, for a given expected rate of return and real interest rate, investment is expected to vary (inversely) with the risk arising from the investment decision. 23 Financial risk has also been made central in the theory of imperfect capital markets. 24 Invoking the theory of asymmetric information, it is assumed that lenders and borrowers of funds have different knowledge about the possible success of the investment project. Given this information structure lenders of funds will be unable to screen borrowers perfectly. High-quality firms, when competing with low-quality firms for funds, have to pay a premium to obtain external funds. If the spread between high- and low-quality firms increases in a downswing, agency and borrowing cost of external funds rise and this exerts a negative impact on investment. Net worth of low-quality firms is predicted to move cyclically and borrowing cost of external funds countercyclically, amplifying real economic disturbances. Similarly to the Kaleckian view, the transmission mechanism from default risk to the real sector works through the financial 614 Finance. Profit Expectations. and Investment fragility of firms. Now, for an empirical test, the aforementioned theoretical variables have to be translated into empirical proxies. This is not too difficult for the actual rate of return and the (real) interest rate. Although (some) macrotheories nowadays point in the direction of a lesser importance of the interest rate for investment,25 we nevertheless prefer not to exclude the real interest rate a priori as an independent variable. Since there are strong comovements the actual rate of return can most easily be captured by the utilization of capacity. The latter variable is, in particular, used as the basic reference variable against which the contributions of other variables are measured. Later the actual rate of return will also be employed directly. The risk of bankruptcy, on the other hand, is in principle an unobservable variable at the macro level. Traditionally, empirical studies have employed variables such as credit flow, the debt-asset ratio and the interest coverage ratio26 as appropriate proxies to measure default risk of firms. Other studies have proposed liquidity variables as proxies for risk21 rather than debt variables. In another type of studies default risk is proposed to be measured by interest rates spread. If lowquality firms - financially fragile firms - face a higher bankruptcy risk their net worth would be lower and external financing cost would rise with a decline in economic prospects. One thus expects, in a downswing for example, an increasing spread between low- and high-quality bonds. If lenders can accurately assess the default risk of individual firms or industries, the changes of risk will be reflected in interest rate spreads. As an aggregate measure of the spread to be used as proxy for risk we take the difference between the short-term commercial paper rate and the interest rate of Treasury bonds. 28 The monetarist tradition would suggest to include the stock of M2 money as an additional factor in investment decisions. More specifically, we employ the velocity of M2 money among the independent variables. If there is a strong endogenous money supply via banks in the business cycle then one might expect a countercyclical movement in M2 velocity.29 On the other hand, the problem usually arises as to what extent the money supply is also affected by monetary policy. If a restrictive (anti-inflationary) money supply is pursued by the FED at the late period of the boom and continued at the beginning of recessionary periods, this, too, might contribute to a countercyclical movement in M2 velocity and a (possibly lagged) negative correlation with investment.3o A more difficult problem is to measure prospective profits. If the stock market were a good predictor for firms' prospective profits one could rely on Tobin's q as an important factor in investment decisions of firms. 31 As a first proxy for an expectational variable we tried capital gains (computed in Fair, 1988) as substitute for the market evaluation of firm's expected rate of return. A corresponding regression, however, did not yield significant results. In the subsequent regressions only Tobin's q is used among the independent variables. On the other hand, one might argue that investment decisions depend more directly on business prospects. This suggests to employ variables such as, for Willi Semmler and Reiner Franke 615 example, the leading indicators for estimating firms' expected returns. In the subsequent part we therefore add to the independent variables an aggregate form of the leading indicators. Also the arithmetic average of Tobin's q and that aggregate measure is invoked. 32 We can now summarize the empirical measures that are employed in the following regressions. The trend deviations of the growth rate of capital stock is taken as the dependent variable. As financial variables, from the balance sheets of firms, in a preliminary step we explore credit market debt (flows and stock), liquid assets, quick assets, working capital,33 Moreover, M2 money stock,34 or, alternatively, the velocity of money, is introduced. All variables are used as ratios over capital stock, in 1982 dollars (cf. Fair, 1988). As the real interest rate the six months commercial paper rate minus the actual inflation rate (measured by the consumer price index) is employed. As said above, the interest rate spread is measured by the difference between the six months commercial paper rate and the six months Treasury bill rate (Citibase Data 1989). Tobin's q35 and an aggregate of the leading indicator are added to the independent variable to measure expected returns. 36 Before discussing the regression results we want briefly to present results on possible loops between the growth rate of capital stock and some financial variables. A view often expressed in the literature is that causal relation between real and (some) financial variables may be ambiguous due to some simultaneity problems between the financial and real variables. 37 Co movements between the growth rate of capital stock and some financial variables may thus be generated by the fact that financial variables themselves are affected by real variables. Candidates to be checked are M2 money (or its velocity), liquidity, and credit flow. The interaction of the above variables with investment was explored for each of the three cycles 1960:4-1970:4, 1970:4-1975: I, 1975: 1-1982:2. A representative example of a loop between velocity and investment is given in Figure 23.2 which plots the trend deviations of gk and velocm. 38 The straight line is the regression line. Its negative slope holds true for all three cycles. Moreover, the interaction always moves clockwise. The apparent structure gives rise to strong (first-order) autocorrelated errors in a regression estimate. Loops are equally obtained for liquidity and investment. Here the regression line is positively sloped, while the loops move countercyclically (again with presence of autocorrelated errors); cf. Figure 23.3. A similar pattern results for the other cycles. Less distinctive loops show up for the interaction of credit flow and investment. In all three cycles significantly positive slopes can be observed, but the looping behavior is only weakly similar to the one depicted in Figure 23.4. Depicting loops for the other financial variables above displayed that if there are cyclical patterns at all, the turning points of the those variables have much shorter loops compared to those of investment, an observation that might suggest more causal relationships. Regressions to test for such a hypothesis are presented next. Finance. Profit Expectations. and Investment 616 2~------------------------~ fD 0 v -1 e -2 -3 -15 -10 -5 0 5 10 velocmDev Figure 23.2 Loop ofve!ocity and investment. 1975:4-1982:4 3 2 f D e v 0 -I -2 -2 -1 0 2 3 liquDev Figure 23.3 Loop of liquidity and investment. 1960:4-1970:4 Willi Semmler and Reiner Franke 617 2~------------------------~ ~ D e v o -I -2 -l-!:E~:::".---.--.-----.---J 2 4 -2 3 -1 o gscmdDev Figure 23.4 Loop of credit flow and investment, 1970:4-1975: 1 5 FINANCE AND INVESTMENT: REGRESSION RESULTS In the regressions to follow, two alternative real variables are employed, namely either capacity utilization, uDev (Table 23.2a, b) or the rate of return on capital, pratDev (Table 23.3a, b), both lagged by three-quarters. In some respect, the variable pratDev leads to more clear-cut results. Proxies for financial risk and profit expectations are as discussed above. Tables 23.2a and 23.3a concentrate on traditional measures of financial risk of firms while Tables 23.2b and 23.3b show results when the recently proposed measure of risk, interest rate spread, is added to them. To explore the extent of serial correlation also regressions with first-order autocorrelation are run. The t-statistics are given in parentheses and the R2 is adjusted. We leave out the report of the constant terms, since they were always insignificant as they should be in view of the detrending procedure. From the set of a large number of regressions the following specifications of variables and lags seem to us most important. Turning to Table 23.2a let us repeat that the terms uDev, liquDev, velocmDev, gscmdDev, ireaLDev, icovrDev, qsumDev, dleadDev and confDev stand for the utilization of capacity, liquidity, velocity of M2 money, growth rate of credit market debt, the real interest rate, the interest coverage ratio, Tobin's q, the aggregate leading indicator, and the constructed leading indicator (the mean of dleadDev and qsumDev). 618 Finance, Profit Expectations, and Investment Table 23.2a OLS-regression of the capital stock growth rate on capacity utilization, expected profits and financial risk (exclusive of interest rate spread), 1960:4-1982:4, percentage trend deviations (gkDev as dependent variable) Ind. var. Real I Financial risk 2 4 3 uDev_3 .305 (16.2) .236 (9.3) .087 liquDev_l .342 (2.9) (1.1) -.136 -.038 (6.7) (2.2) .169 (3.0) .348 (4.7) velocmDev-2 gscmdDev_l gscmdDev_2 ireaLDev_2 ircovrDev-2 qsumDev_2 5 .58 .77 .64 .79 .54 .78 8 Autocorrel. 9 .138 (3.6) -0.25 (.8) .004 (2.2) -.053 -.034 (1.8) (1.0) -.0006 (.3) -.006 (.3) .024 .031 (1.8) (2.0) .041 (2.4) .87 (13.1) .89 .38 1.70 .004 (.5) .068 (4.1) rho .76 .58 .62 7 .234 .261 .256 .248 (8.7) (13.7) (12.0) (8.2) .049 (.5) -.032 (1.6) .150 .028 (2.5) (.5) confDev_2 S.E. DW 6 .243 (9.7) .011 (.2) -.009 (.5) .122 (2.1) dleadDev_2 R2 adjusted Profit expectations .80 .54 .85 .79 .54 .60 .035 (2.6) .77 .56 .60 .79 .54 .80 .78 .56 .60 .023 (.8) Though in Table 23.2a, as compared to Table 23.1, a number of different variables is added in the regression equations, it is remarkable that the prominent role of (lagged) capacity utilization is maintained throughout. The variable uDev_3 is always highly significant and can still explain a substantial portion of the variation in investment. Its coefficient, however, tends to decrease somewhat, thus pointing to the influence of the other factors. All these other variables also exhibit the expected sign, even if they are not always significant. The three previously discussed variables liquDev, velocmDev, and gscmdDev are significant when regressed alone on investment (column 2)39 with suitable short lags, and the latter two remain so when jointly regressed with utilization (column 3). Of course, including uDev brings about a considerable decrease in the size of their coefficients. 4o Willi Semmler and Reiner Franke 619 So far the result that investment is positively correlated with liquidity appears to give some support to Foley's approach of a liquidity growth dynamics (Foley, 1987; cf. also Semmler and Sieveking, 1993). Furthermore, the result of an important role of credit flows for investment confirms the study undertaken by Friedman (1986) on the existence of 'credit cycles' in the US economy.41 The negative sign of the variable velocmDev is also of theoretical interest. The negative sign of the coefficients for velocmDev appears to support the view - as one might derive from, for example, Romer and Romer (1990) - that a restrictive monetary policy, starting before recessionary periods, was quite effective in limiting money supply. Adding further traditional variables to measure finance conditions and risk (irealDev and icovrDev) leads to multicollinearity, so that not all of the above variables remain significant (column 4). In all regressions including irealDev, the contribution of the real interest rate is only of minor significance (cf. column 7). It always has, however, the correct sign.42 For this reason, and since it is in many approaches referred to as an important variable in the determination of investment, we have not discarded it. Besides, it may perform better if, in the definition of the real interest rate, the current inflation rate is replaced by some appropriate measure for expected inflation. As regards the profit expectations variables, qsumDev, dleadDev and confDev lagged by two-quarters, the leading indicator displays a high t-value (column 5) quite in contrast to Tobin's q (which was only significant if regressed alone, or jointly with uDev; results not reported here).43 Though the variable dleadDev is significantly correlated with investment, the coefficient, compared to the one of capacity utilization, is always much smaller. Moreover, both variables, dleadDev and qsumDev, contribute less to an overall R2 when uDev is already present. On the other hand, when regressed alone on gkDev they brought about an R2 = .56 with qsumDev being insignificant. Numerous other studies have demonstrated that there may be strong comovements of the utilization of capacity and expectation variables, so that, for example, an expectational variable does not come out as distinctly as expected. 44 It is nevertheless worth reiterating that the coefficient of utilization and its level of significance is always the least affected. With regard to the proxies for profit expectations qsumDev and dleadDev, we finally want to point out that qsumDev and dleadDev are variables capturing the expectations formation for periods ahead of time. On the other hand, the variable uDev, or pratDev, and their lags are to be interpreted as indicating gestation lags in the production of investment goods. Though both the expectations as well as real variables exhibit lags in our regressions above, they should be interpreted differently.45 Table 23.2b reports regression results when the spread of interest rates, in deviation form, of course, and with a three-quarter lag is added to the independent variables. This is the only change with respect to Table 23.2a, and the columns are numbered correspondingly 620 Finance, Profit Expectations, and Investment Table 23.2b OLS-regression of the capital stock growth rate on capacity utilization, expected profits and financial risk (including interest rate spread), 1960:4-1982:4, percentage trend deviations (gkDev as dependent variable) Ind. var. Real I uDev_3 .237 (16.2) liquDev_1 velocmDev_2 gscmdDev_1 gscmdDev_2 ireaLDev_2 icovrDev_2 sprdDev_3 gsumDev_2 Financial risk 3 4 2 (.275 (2.3) -.109 (4.6) .347 (4.7) .253 (11) -.003 (0.5) .007 (.4) .147 (2.9) 5 .237 (9.3) -.027 (.35) .017 (.87) .167 (3.2) .253 .264 .246 .237 (11.0) (14.4) (12.8) (9.2) -.025 (.3) .018 (.9) .140 .170 (2.7) (2.9) .152 (4.4) .006 .001 .007 (.05) (.2) -.003 (1.3) -.100 -.110 -.110 -.110 (3.8) (5.1) (4.8) (4.5) .005 .007 (.4) (.22) .002 (1.4) -.118 -.079 -.115 -.114 (5.8) (2.1) (4.8) (4.5) .022 (1.2) confDev_2 rho OW .041 (1.4) -.070 (3.6) (.7) dleadDev_2 R2 adjusted S.E. Profit expectations Autocorrel. 7 8 9 6 .83 .50 .75 .60 .76 .73 .84 .48 .84 .84 .48 .85 .83 .49 .68 .015 (1.3) .005 (.5) .002 (.2) .82 .49 .83 .48 .80 .83 .48 .86 .72 0.30 (1.8) .83 (11.6) .90 .36 1.88 As can be observed. sprdDev_3 is of the expected sign and always highly significant. It generally brings out a considerable decrease in the standard error. The explanatory power of uDev-3 roughly remains the same (its coefficient partly decreases or increases slightly). Furthermore. sprdDev_3 appears to interact weakly with the other (traditional) measures of financial risk. and makes some of them less significant (see velocmDev_z. incovrDev_z). For others, though remaining insignificant, the sign of the coefficients changes (i.e. liqudev_1 and irealDev_2). The significance of the variable gscmdDev partly increases or decreases. On the other hand. the proxies for profit expectations now turn out to be insignificant. indicating some multicollinearity with the financial risk variable sjJrdDev_3. Overall. sprdDev-3 adds visibly to the explanatory power of the independent variables. 46 Willi Semmler and Reiner Franke 621 As noticeable in Tables 23.2a, b, in the first set of tests (columns 1-8) the possible presence of serial correlation is completely neglected. The consequences are low DW-statistics. In order to reduce serial correlation we employ the often applied method and included an AR(I)-term with coefficient rho.47 The outcome with the much improved DW is reported in column 9 in Tables 23.2a, b. Yet, the reader might be reminded, as noted in the introduction, that this method does not satisfactorily resolve all the problems that might be involved in the regression specifications. In the econometric literature (for example, Granger and Newbold, 1986, ch. 6), it is suggested to employ regressions with first differences to obtain stationary time-series data and to attain a higher DW. This was undertaken for the original level variables as well as for the series representing the percentage trend deviations. We pursued this for sample tests, namely by running the same regressions as reported in Table 23.2a but using, however, first differences of the dependent as well independent variables (for level and for trend deviations). The result was indeed that the DW moved up to a level between 1.6 (with only few independent variables as in columns 1,2, 5-6) to a number close to 2 (with a larger number of independent variables as used in the other columns). The R2s were much lower, namely between .14 (for a small number of independent variables as in columns I, 2, 5-6) and .36 (for a large number of independent variables). The signs of the regression coefficients were the same as obtained for the detrended variables. Some coefficients lost significance whereas some others became significant. Overall, the results were quite supportive of our above regressions reported in Table 23.2a, although the coefficient for the utilization of capacity was mostly slightly reduced. Before discussing this matter on the DW-statistics further some additional results will be reported. A second set of tests was run in which the lagged trend deviations of the rate of return, pratDev, replaces the utilization as real variable. 48 Here the results are more distinct. Again distinguishing between traditional variables of financial risk and interest rate spread, we report the essential results in Tables 23.3a, b, respectively. (Again the constant terms are left aside since they were as expected all insignificant.) In Table 23.3ajointly with the variable pratDev the proxies for financial conditions and default risk of firms almost always come out significantly (columns 2, 3 and 5). The real interest rate ireaLDev remains insignificant, though the negative sign is preserved. The explanatory power of the variable qsumDev for gkDev is only slightly improved and when dleadDev is included its effect becomes much weaker (not reported here). The average financial (qsumDev) and business prospects (dleadDev), called confDev, are always significant. Those results can be read from columns 4-6. Again, the gain from the regressions including the financial fragility variables and an expectational variable, compared to the regressions with pratDev only, appears important. As in the regressions reported in Table 23.2a, here also the presence of a real variable, now pratDev, already explains a substantial part of the variation of investment, which is an expected result compared to previous 622 Finance, Profit Expectations, and Investment Table 23.3a OLS-regression of the capital stock growth rate on the rate of return, profit expectations and financial risk (exclusive of interest rate spread), 1960:4-1982:4, percentage trend deviations (gkDev as dependent. variable) Ind. var. pratDev_3 liquDev_l Real 1 .410 (12.8) velocmDev-2 gscmdDev_2 icovrDev_l Financial risk 2 3 .277 (7.3) .254 (2.6) -0.67 (3.6) .167 (2.7) ireaLDev_2 conjDev_2 .248 (6.0) .175 (1.8) -0.45 (2.1) .209 (3.2) .005 (1.9) .0001 (0.04) rho R2 adjusted S.E. DW .66 .65 .51 .74 .61 .56 .74 .60 .59 Profit expectations Autocorrel. 4 5 6 .326 (9.2) .145 (3.0) .125 (1.9) .269 (7.1) .173 (1.7) -0.59 (3.0) .144. (2.3) -.051 (1.5) .048 (3.1) -.013 (.4) .028 (1.8) .71 .64 .58 .74 .60 .52 -.008 (.4) .049 (3.0) .910 (15.7) .89 .39 1.50 .030 (1.2) work on the effect of Tobin's q on investment. 49 Employing utilization uDev, however, usually yields a higher R2 and a smaller S.E. As visible in Table 23.3b, including the interest rate spread, sprdDev_3 partly increases or decreases the coefficient for the rate of return on capital pratDev_3' The coefficient for the spread has as expected a negative sign and is always significant. It everywhere reduces the standard error and increases the DW. Here too, as before in Table 23.2b compared to Table 23.2a, the traditional measures of financial fragility of firms, liqueDev, velocmDev and icovrDev, lose significance when the interest rate spread is included in the regressions. Interestingly, the interest rate irealdev_2 has the wrong sign but it remains insignificant. Credit market debt gscmdDev_2 keeps its significance, though its coefficient is also slightly decreased. The importance of the expectations variables confDev_2 partly declines. Overall, as for the regressions with utilization of capacity uDev_3' the spread between risky and safe interest rates turns out to be a quite important additional variable, measuring financial risk of firms and its effect on investment. We now want to come back to the problem of the low DW-statistics in Tables 23.2a-23.3b. For the purpose of eliminating first-order serial correlation, we have Willi Semmler and Reiner Franke 623 Table 23.3b OLS-regression of the capital stock growth rate on the rate of return, profit expectations and financial risk (including interest rate spread), 1960:4-1982:4 (gkDev as dependent variable) Ind. var. pratDev_3 liquDev_t Real 1 Financial risk 2 -.140 (6.5) -.120 (8.7) .280 (7.4) .097 (1.0) -.011 (.51) .176 (2.9) .003 (1.4) .030 (.9) -.119 (4.2) .76 .57 .73 .78 .55 .82 .79 .55 .81 .360 (13.4) velocmDev_2 gscmdDev_2 icovrDev_t .304 (8.7) .148 (1.7) -.019 (1.0) .147 (2.6) ireaLDev_2 sprdDev_3 confDev_2 3 rho R2 adjusted S.E. DW Profit expectations 4 5 .317 (10.3) .296 (8.4) .Ill Autocorrel. 6 .144 (3.1) (1.1) -.021 (1.1) .122 (2.1) .136 . (2.3) .011 .019 (.6) -.137 0.020 .021 (1.4) .024 (.7) -.119 (4.1) .013 (.9) .79 .55 .86 .78 .55 .79 (504) (A) (.1) -.060 (2.8) .044 (2.7) .88 (15.1) .90 .38 1.65 included, for the regression with uDev_3 and sprdDev_3 on gkDev (see column I, Table 23.2b) an AR (I) term and can compare the numerical results as well as graphs of fitted values. The effect of including an AR(I) term in the regression reported in column I, Table 23.2b is that the DW indeed substantially increases (from DW=.75 to DW=1.81).sO A measure of the goodness-of-fit is usually the mean squared prediction error (MSPE). For the regression without the AR(I) term we have an MSPE=.52 whereas for the regression inclusive of the AR(I) term we get MSPE=.37. For our purpose to track cyclical components of business cycle variables, a comparison of the actual and fitted series is prudent. Even without the AR( 1) term the goodness-of-fit, as depicted in Figure 23.5 for the actual and fitted line for gkDev is noteworthy. The fitted line traces peaks and troughs well and generally the deviations are quite small. This result holds except for the period shortly before 1966 and the 1975 trough. These exceptions might be due to non-Iinearities in the time-series data when they reach extreme values. We want to note that the succession of positive and Finance. Profit Expectations. and Investment 624 3,---------------------------~ 2 o -1 -2 -3~--------------------------~ 62 64 66 68 70 72 74 76 78 80 82 - - Actual gkDev - - - - - - Fitted gkDev Figure 23.5 Actual gkDev and fitted gkDev for the regression of column 1. Table 23.2b (excluding AR(l» negative deviations in Figure 23.5 also shows the source of the low DW-statistics. The graphic depicted in Figure 23.5 compared to the one in Figure 23.6 nevertheless suggests that the discrepancies compared to the fit obtained by employing the AR(l) term are not too serious. Thus the low DW-statistics emerging from our original regressions. without the AR(l) term. do not necessarily generate an inadequate fitted line. The mechanical way of correcting for possible (first-order) autocorrelations through the use of an AR(I) term appears to be of quite limited use (in particular, in the light of the fact that other problems, as noted in section 2. may be disguised). A further question concerns the simultaneity of financial and real variables. This has been explored initially in section 4 by means of the techniques of loops. The problem can also be further investigated by employing instrument variables (IV) and Two-Stage-Least Square regressions (TSLS). Our previously discussed variables liquDev. velocmDev and gscmdDev are proper candidates for such a test. These three variables may be endogenous, and thus may be proxying other possibly real - factors affecting investment. Indeed, various test results revealed that these three variables especially turned out to respond more sensitively to instrument variables. The results of the IV-tests for the three variables are reported in Table 23.4 where we also include an AR(l) term. As instruments we chose a (lagged) endogenous variable and dleadDev. These two instruments approximately fulfilled the requirement for IV-tests. namely, that the instrument variables are uncorrelated with the residual term resulting from the OLS-equation and reveal a high R2 with the endogenous regressor for which our IV-variables act as instruments. The last variable in the three different types of regressions reported in Table 23.4 is always Willi Semmler and Reiner Franke 625 3,-----------------------------, 2 o -1 -2 -3~--------------------------~ 62 64 66 68 70 72 74 76 78 80 82 - - - Actual gkDev ------- Fitted gkDev Figure 23.6 Actual gkDev and fitted gkDev for the regression of column t, Table 23.2b (including the AR(l» the regressor to which the instruments were applied, however, with one additional lag. We again leave aside the constant terms since they were insignificant. As observable from Table 23.4 all three variables (liquDev, velocmDev and gscmDev) lost significance when instruments were applied to them but the overall R2s and the signs of the coefficients remained the same. The loss of significance points to the fact that those variables may indeed. to some extent, be proxying other (real) factors. This somewhat reconfirms our statements in section 4 where we have already discussed the conjecture that the financial-real interaction may be particularly distinct for variables such as liquDev, velocmDev and gscmdDev. 51 6 CONCLUSIONS In order to test some hypotheses pertaining to the financial-real interaction and determinants of investment in the business cycle, we have detrended the relevant variables to separate the trend from the cyclical component by employing a segmented trend procedure. The output-capital ratio served as the basic measure of the business cycle. This was justified by its strong correlation with other measures such as GNP and indicators of capacity utilization. When, in our regression studies, investment was expressed as the percentage deviation of the growth rate of capital stock from its trend value, the results concerning real variables were quite clear cut. With respect to macrodynamic modeling they represent a strong case for the use of (delayed) investment functions that include, as their core argument, a strong accelerator effect (in the wide sense). That is, to a large 626 Table 23.4 Finance, Profit Expectations, and Investment IV-regressions (with lagged endogenous regressor and dleadDev as instrument variables) (gkDev as dependent variable) OLS-regressions rho uDev (-3) liquDev (-I) R2 f3 t 14.1 3.9 1.8 .891 .153 .115 IV-regressions f3 .89 rho uDev (-3) velocmDev (-2) R2 .845 .\34 -.084 rho uDev (-3) gscmdDev (-1) R2 .882 .150 .039 .89 ·89 t 6.2 6.2 .1 .817 .lIS .029 .89 12.5 3.5 2.7 .906 .084 -.131 13.4 3.7 1.8 .890 .121 0.48 .89 13.6 3.9 .85 26.4 5.6 1.0 .89 extent capital growth can be seen as being dependent on the output-capital ratio, or on some rate of return on capital (which in reality is weakly correlated with the former) although, as demonstrated, some non-linear relationships may be present. As also shown, the inclusion of variables for finance conditions and financial risk of firms in the investment function, as well as proxies for profit expectations, significantly contributes to explaining the cyclical variation in investment expenditures. This holds regardless of whether capacity utilization or, alternatively, the rate of return on capital are used as real variables. The traditional measures of financial fragility such as liquidity, velocity of M2 money, flow of credit, interest coverage ratio (to a minor extent the real interest rate), as well as the profit expectations variables (our aggregate leading indicator and the composite measure) contribute to the variation in investment, if the variables are properly lagged and are measured as deviations from their respective trends. 52 Yet with respect to traditional finance variables, we obtained some indication that there may also be a strong interaction effect of some of the finance variables with real variables. The best contribution to explaining the variation of investment in the business cycle, beside the utilization of capacity, was provided by the interest rate spread which indeed appears to serve as a good summary index for the financial conditions of firms. This variable when regressed with other variables on investment proved always to have the correct sign and the coefficient was highly significant. In part, it replaced the effects of the traditional financial variables and the proxies for profit expectations and improves the goodness-of-fit. Willi Semmler and Reiner Franke 627 Let us conclude with some suggestions for further research. An attempt could be made to distinguish more sharply between the phase of expansion and contraction. With the latter procedures, non-linear relationships could be explored further (asymmetries and regime switching). This is undertaken in a non-linear bivariate model in Kockesen and Semmler (1995). On the other hand, the analysis could be pursued with a more disaggregate data set. One may argue that the differences of financial risk or profit expectations among firms and their effects on investment are partly washed out through the use of aggregate data. The direct impact of firms' balance sheet variables on financing and investment is usually more noticeable when disaggregate data sets are used. 53 For our purpose, however, of highlighting the financial-real interaction in the business cycle - and pointing out their significance for small-scale macrodynamic models - the above results are already sufficiently encouraging. Noles * 1. 2. 3. 4. 5. 6. 7. 8. The authors would like to thank Raul Zambrano for valuable research assistance. We also want to thank Richard Arena and Tracy Mott for helpful comments and David Gordon for making available to us a large data set employed in the econometric part of the paper. The work on this paper has been started while Semmler was visiting scholar at the Economics Department of Stanford University. The hospitality of Stanford University is gratefully acknowledged. Cf. the contributions by Minsky, Davidson, Tobin and Moore in Nell and Semmler (1992). Most of these contributions, however, do not go as far as Kaldor concerning the endogeneity of money and thus posit some interaction of money with real activity. In the tradition of the credit view it is then also maintained that financial markets (intermediaries) affect the real side of the economy mainly through lending operations. Accordingly, monetary policy exerts its influence on real activity not through the liability but the asset side of financial intermediaries (bank lending). Cf. Bernanke (1983), Bernanke and Gertler (1989), Greenwald and Stiglitz (1989). See, for example, Akashi and Asada (1986), Asada (1989), Asada and Semmler (1995), Foley (1987), Franke and Semmler (1989), Taylor and O'Connell (1989) and Woodford (1989). The impact of the inflation rate on the financial structure of firms in form of a debt-asset ratio is studied in Franke and Semmler (l99Ia). Recent macroeconomic work mostly follows this road of generating stationary time series; see, for example, Blanchard and Fischer (1989, ch. 1). The procedures we tried included: (i) log-linear trends, (ii) segmented trends, (iii) polynomial trends, (iv) trends generated by repeated moving averages, (v) trends generated by repeated arithmetic averages, and (vi) the HodrickPrescott filter. In many cases, but not in all, the results were quite similar, but the segmented trend proved hardly ever inferior. The Hodrick-Prescott filter has recently been used widely. In our context it did prove to be largely equivalent to the segmented trend procedure. For a more systematic evaluation of the usefulness of the different detrending procedures for a business cycle analysis, cf. Canova (1992). An early discussion on this problem can be found in Blatt (1983, ch. 16) who lists the following specification errors, which can never be ruled out completely (even if 628 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. Finance, Profit Expectations, and Investment there were no structural change): truncation error, lag structure error, aggregation error, omitted and unknown variable error. For a thorough survey on econometric tests when non-linear relationships between the variables hold; cf. Potter (1994) and Sayers (1994). For the insensitivity of the Durbin-Watson test to important types of specification errors, cf. the reset test by Ramsey (1969). A well-known example in the empirical Iiterature where inclusion of autoregressive errors fails to remove the serial correlation are models of overlapping annual wage changes, which adopt the assumption that the same percentage of workers negotiate their wage increase in each quarter. It is easily seen that this statistical device transforms the error terms in the underlying behavioral relationships into moving averages. The same type of argument, however, applies to any quarterly data when a fraction of its constituent components is to change in a longer interval. Of course, this is the case for investment decisions: they are made infrequently by a single firm, but different firms decide at different points in time. Even the MSPE is not a sufficient criterion of a goodness-of-fit. A statistic as to the significance of the MSPE when different forecasts are compared should be added, cf. Mizrach (1991). The prominent role of the output-capital ratio in these models derives also from the fact that the other dynamic variables are scaled down by the capital stock. This may, however, be different for unemployment data (cf. Neftci, 1984). To be precise, investment is a function of the level of profits in Kalecki or the level of income in Kaldor, not of their rate of change, which would be an accelerator in the strict sense. In Taylor (1983, ch. 2), the utilization of capacity directly enters the investment function, whereas in other works of Taylor the investment function could be reduced to this argument if the income distribution is kept constant (cf. Taylor, 1985; Taylor and O'Connell 1989); see also Franke and Semmler (1989, 1991a, b). Different versions have been tested where the independent variables are the rate of change of income (or sales), the level of sales, the sales-capital stock ratio, the rate of change of cash Hows, the level of cash Hows or different measures for the utilization of capacity. Mostly, a more or less complicated lag structure is included. For empirical tests and discussions on the different versions of the accelerator principle, cf. for example, Blanchard and Fischer (1989, ch. I), Clark (1979), Eisner (1967), Fazzari et al. (1988) and Kopcke (1985), who make a strong case for the empirical relevance of the accelerator principle. This result corresponds, at least in spirit, to investment functions used in Goodwin (1948, 1951) and Kalecki (1937a). It is also in line with the investment function used in Kaldor (1940) and Kaldorian cycle models, where (in the framework of a stationary economy) investment is a (non-linear) function of the level of GNP. Notable exceptions in earlier times are the work of Anderson (1964), Gurley and Shaw (1955) and Meyer and Kuh (1957). Particularly, Anderson (1964) constructs variables for financial risk which are in his view, next to cash Hows and cost of capital, most relevant for investment decisions. Theoretical models of this type, though from different perspectives, can be found in Bernanke and Gertler (1989, 1990), Foley (1987), Franke and Semmler (1989, 1991a, b), Greenwald and Stiglitz (I986a, b, 1988a, b), Taylor and O'Connell (1989) and Woodford (1989). Empirically, the relevance of finance for investment has mostly been reduced to the question of whether cash Hows and self-financing is a strong explanatory variable in investment behavior of firms. This"type of investment study, initiated by Meyer and Kuh (1957) has been recently revived by Fazzari et al (1988). Other financial variables measuring financial fragility of firms have been included in some other studies. Willi Semmler and Reiner Franke 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 629 For a specification of an investment function with those components in a macrodynamic model, cf. Taylor and O'Connell (1989), and Franke and Semmler (1991 b). Kalecki (l937a) argues that investment will be undertaken up to the point were the excess of prospective profits over the interest rate is equated to the bankruptcy risk arising from debt financing of the investment project. Therefore, the cost of funds consists of two components: the interest rate and the 'increasing risk' due to debt finance. References to Kalecki's work start with Klein (1948) Anderson (1964) and continue in the recent work on imperfect capital markets. More specifically, one may posit that it is the marginal risk of bankruptcy that affects investment. A theory pertaining to the relation of imperfect capital markets and investment is usually presented on the basis of a one-period model; cf. Bernanke and Gertler (1990). Similar results can, however, also be obtained on the basis of an infinite horizon model, see Asada and Semmler (1992). Cf. Greenwald and Stiglitz (l986a, b), Stiglitz and Weiss (1981). The debt-asset ratio and interest coverage ratio may be considered as important variables in Minsky type models, cf. Fazzari and Mott (1986/87). Theoretical models in whichHhe ease and tightness of liquidity impact investment, possibly channeled through the interest rate, have been developed by Day (1989), Day and Shafer (1985), Foley (1987). Bernanke (1983), Gertler et al. (1991), Friedman and Kuttner (1992) and Mishkin (1991) have already employed interest rate spreads as measures for financial fragility. There, however, other proxies for financial risk are left aside. Interest rate spread has also been proposed as an additional leading indicator, cf. Stock and Watson (1989). The (detrended) time series of money supply revealed a clear procyclical movement of M2 money. A strong endogenous (procyclical) money supply is empirically shown in King and Plosser (1984). If the variability in output is weaker than that of M2 money then a countercyclical movement in M2 velocity will occur. Hall (1987), by referring to the financial instability hypothesis, has conjectured that the velocity of M2 money will be positively correlated with the growth rate of nominal GNP. This entails that there might also be a covariation of velocity and investment. As will be seen below, for our time period and our variables measured relative to the trend, this does not seem to hold true. Tobin (1990) maintains that such a policy was pursued since the 1960s. This view is also supported by Romer and Romer (1990) who particularly point out that an antiinflationary monetary policy became effective in December 1968, April 1974, August 1978 and October 1979 (Romer and Romer 1990: 161). There is, of course, a large body of literature which contests that the stock market fulfills its role as efficient forecaster of firms prospective profits, cf. Shleifer and Summers (1990). We want to point out here that Keynes, for example, never thought of a variable solely reflecting the financial evaluation of the firm, as being the most important determinant of investment. He more accurately referred to the 'state of confidence of investors' and business prospects when discussing the role of expectations (Keynes, 1936, chs. 5 and 12). This includes general business conditions, consumption behavior, credit conditions and financial market prospects. It is on these grounds that we will refer to financial as well as to business prospects of firms in our regressions. The above measures were used in real terms (in 1982 dollars) with the following definitions: (I) flow and stock of credit market debt net flows of corporate debt instruments; (2) liquid asset =stock of liquid assets; (3) working capital = stock of working capital; (4) interest coverage ratio cash flows over net interest paid by = = 630 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. Finance, Profit Expectations, and Investment non-financial corporations (Fair, 1988). The data for those variables are taken from the Flow of Funds Accounts (1989). M2 money stock was obtained from Citibase Data (1989), in 1982-dollars. The time series data on Tobin's q were kindly made available to us from David Gordon. In order to construct an aggregate predictor for expected returns we aggregated with equal weights the four leading indicators of Business Conditions Digest (Citibase Data, 1989). The leading indicators are dleac (composite index, capital investment), dlead (composite index, inventory investment and purchase), dleap (composite index, profitability) and dleaj(composite index, money and financial flows). In section 5, real variables are employed as instruments for the above financial variables to study simultaneity further. The loops not printed in the chapter can be obtained from the authors upon request. In addition to the liquidity-asset ratio, the quick ratio and the ratio of working capital to assets were also tried. When liquidity was replaced by the latter two variables, essentially the same results as reported in column 2 were obtained. Note, however, that there may still be a possible simultaneity between those three variables and real variables; a problem to be further explored below. Friedman uses a vector autoregression to test for the interaction of credit and the business cycle. Also a simple regression between the credit flow as dependent variable and M2 money as independent variable (both in deviation form) revealed a high positive correlation. When ireaLDev was included in the regression of column 4, but other variables such as icovrDev and qsumDev excluded, the sign of ireaLDev came out correct but the coefficient was not significant either. Studies on earlier periods, starting with the 1860s, have shown that the short-term interest rate moves procyclically (cf. Zarnowitz and Moore, 1986, who studied the nominal commercial paper rate and the Treasury bill rate). At least for the real rate the recent period since the 1960s does not seem to support such a strong position. When the levels of gsumDev and dleadDev, with two lags, were regressed joinly with the level of capacity utilization on the growth rate of capital stock, qsumDev was significant. Cf. for example, von Fuerstenberg (1977), Abel and Blanchard (1984). Cf. also Abel and Blanchard (1984). We want to note here, however, that there might be an interaction between the money supply and the interest rate spread. As Bernanke (1990) points out, periods of contractionary money supply may result in higher interest rate spread, since firms will - with tight money and restricted loans from banks - look for substitutes and employ more credit market instruments to finance their investment, which may also give rise to an increase in the commercial paper rate. Cf. Pindyck and Rubinfeld (1981, ch. 6.2), or Abel and Blanchard (1984). An additional AR (2) term, as used in Abel and Blanchard (1984), is not necessary for the elimination of serial correlation. Below a further discussion on the error term will be provided. The r~te of return on capital is computed from Fair (1988). It is the non-financial sector firms' cash flow over the (net) capital stock, deflated by the price deflator for plant and equipment. Previous results, mostly using level variables for Tobin's q and investment, usually demonstrate a high colIinearity of the utilization of capacity and q; cf. von Fuerstenberg (1977), Abel and Blanchard (1984) and Kopcke (1985). Through the introduction of the autoregressive term the size of coefficient for uDev and sprdDev, however, alters from .27 and .12 to .15 and .07, respectively. The Willi Semmler and Reiner Franke 51. 52. 53. 631 coefficient for the autoregressive term is as high as .85. All coefficients are significant and the standard error falls to .38. In addition to the above IV-approach we employed the Granger-causality test for each of those variables and uDev, gkDev and dleadDev respectively, using a lag structure of four lags. Whereas the results for the variables velocmDev and gscmdDev were less distinct when their causal relation with uDev, gkDev and dleadDev was tested, very clear causal relations were found for liquDev and the three real variables. It was uncovered that liquDev is Granger-caused by uDev, gkDev, dleadDev respectively but does not Granger-cause those three real variables. A likewise positive interaction of liquidity and a real variable (inventory change) is observed in Kashyap et al. (1992). It is, however, worth reiterating that owing to the multicollinearity some of the variables lose significance when combined with others. Cf. Fazzari et al. (1988), and Bernanke and Campbell (1988). References Abel, A.B. and 0.1. Blanchard (1984), 'The Present Value of Profits and Cyclical Movements in Investment', Econometrica, 54 (2), 249-73. Akashi, S. and T. Asada (1986), 'Money in a Kaldorian Cycle Theory', Economic Review, 37, 169-77. Anderson, W.H.L. (1964), Corporate Finance and Fixed Investment (Cambridge, MA: Harvard University Press). Asada, T. (1989), 'Monetary Stabilization Policy in a Keynes-Goodwin Model of Business Cycles', in W. Semmler (ed.), Financial Dynamics and Business Cycles: New Perspectives (Armonk/London: M.E. Sharpe). Asada, T. and W. Semmler (1995), 'Growth and Finance: An Intertemporal Model', forthcoming, Journal of Macroeconomics, September. Bernanke, B.S. (1983), 'Non-Monetary Effects of the Financial Crisis in the Propagation of the Great Depression', American Economic Review, 73, 257-76. Bernanke, B.S. (1990), 'On the Predictive Power of Interest Rates and Interest Rate Spread', NBER-working paper, no. 3486, Cambridge, Mass. Bernanke, B.S. and J.Y. Campbell (1988), 'Is there a Corporate Debt Crisis?', Brookings Papers on Economic Activity, no. I, 83-139. Bernanke, B. and M. Gertler (1989) 'Agency Cost, Net Worth, and Business Fluctuations', America Economic Review. 79(1),14-31 Bernanke, B.S. and M. Gertler (1990), 'Financial Fragility and Economic Performance', Quarterly Journal of Economics, 55, 87-114. Blanchard, O. and S. Fischer (1989), Lectures on Macroeconomics (Cambridge, MA: MIT Press). Blatt, G.M. (1983) 'Economic Policy and Endogenous Cycles', Journal of Post Keynesian Economic, 635-47. Canova, F. (1992), 'Detrending and Business Cycle Facts', Dept. of Economics, Brown University, mim60. Citibase Data (1989), Data on diskettes. Clark, P.K. (1979), 'Investment in the 1970s: Theory, Performance and Predictions', Brookings Papers on Economic Activities, 1,73-124. Day, Richard H. 1989 'Comparative Monetary and Fiscal Policy Dynamics', in W. Semmler, (ed.), Financial Dynamics and Business Cycles: New Perspectives (Armonk, NY: M.E. Sharpe). 632 Finance, Profit Expectations, and Investment Day, R.H. and W. Shafer (1987), 'Keynesian Chaos', Journal of Macroeconomics, 7 (3), 277-95. Eckstein, O. and A., Sinai (1986), 'The Mechanism of the Business Cycle in the Postwar Era', in RJ. Gordon (ed.), The American Business Cycle: Continuity and Change, NBER (Chicago: University of Chicago Press), 39-105. Economic Report of the President (1984, 1988) (Washington, DC: United States Government Printing Office). Eisner, R (1967), 'A Permanent Income Theory for Investment: Some Empirical Explorations', American Economic Review, LVII (3), 363-89. Fair, RC. (1984), Specification, Estimation, and Analysis of Macroeconometric Models (Cambridge, MA : Harvard University Press). Fair, RC. (1988) 'Fairmodel, Fairmodel Plus', Fairmodel Monthly Service, Macro, Inc. Fair, R.C. and W. Parke (1984), The Tape of the Fair-Parke Program and the Fair US Model (New Haven: Yale University). Fazzari, S. and T. Mott (1986/87), 'The Investment Theories of Kalecki and Keynes: An Empirical Study of Firm Data, 1970-1982', Journal of Post Keynesian Economics, 9, 171-87. Fazzari, S., R.G. Hubbard and B.C. Petersen (1988), 'Financing Constraints and Corporate Investment', Brookings Papers of Economic Activity, no. I, 142-95. Fisher, I. (1933), 'Debt Deflation Theory of Great Depressions', Econometrica, 1, 337-57. Flow of Funds Accounts (1989), Balance Sheets for the US Economy 1948-87, Board of Governors of the Federal Reserve System, Washington, DC. Foley D. (1986), 'Stabilization Policy in a Nonlinear Business Cycle Model', in W. Semmler (ed.), Competition, Instability, and Nonlinear Cycles: Lecture Notes in Economics and Mathematical Systems (New York/Heidelberg: Springer Verlag), 200-11. Foley, D. (1987), 'Liquidity-Profit Rate Cycles in a Capitalist Economy', Journal of Economic Behavior and Organization, 8, 363-77. Franke, R. and W. Semmler (1989), 'Debt Financing of Firms, Stability, and Cycles in a Macroeconomic Growth Model', in W. Semmler (ed.), Financial Dynamics and Business Cycles: New Perspectives (Armonk, NY; London: M.E. Sharpe). Franke, R. and W. Semmler (1991a), 'A Dynamical Macroeconomic Growth Model with External Financing of Firms: A Numerical Stability Analysis', in E. Nell and W. Semmler (eds), Nicholas Kaldorand Mainstream Economics (London/New York: Macmillan). Franke, R. and W. Semmler (1991 b), 'Expectation Dynamics, Financing of Investment and Business Cycles', in D. Papadimitriou (ed.), Profits, Deficits and Instability (London/New York: Macmillan). Franke. Rand W. Semmler (1991c), 'Trend and Cycles in the Corporate Debt-Asset Ratio: A Theoretical Note', University of BielefeldlNew School for Social Research, mimoo. Franke, Rand W. Semmler (1992), 'Empirical Evidence on Some Macroeconomic Relations over the Business Cycle', University of BremenlNew School for Social Research, mimoo. Friedman, B.M. (1986), 'Money. Credit, and Interest Rate', in RJ. Gordon (ed.), The American Business Cycle (Chicago: University of Chicago Press). Friedman, B.M. and K.N. Kuttner (1992), 'Money, Income, Prices and Interest Rate" American Economic Review, June. 472-92. Friedman, M. and AJ. Schwartz (1963). A Monetary History of the United States, 1867-1960 (Princeton, NJ: Princeton University Press). Fuerstenberg, G.M. (1977), 'Corporate Investment: Does Market Evaluation Matter in the Aggregate?', Brookings Papers on Economic Activity, 2, 347-408. Gertler, M., R.G. Hubbard and A. Kashyap (1991), 'Interest Rate Spreads, Credit Constraints, and Investment: An Empirical Investigation', in R.G. Hubbard (ed.). Financial Markets and Financial Crises (Chicago: University of Chicago Press). Willi Semmler and Reiner Franke 633 Goodwin, R.M. (1948), 'Secular and Cyclical Aspect of the Multiplier and Accelerator" in Employment, Income and Public Policy: Essays in Honor of A.H. Hansen (New York: Norton), \08-32. Goodwin, R.M. (1951), 'The Nonlinear Accelerator and the Persistence of Business Cycles', Econometrica, 19 (1),1-17. Granger, C.W.J. and P. Newbold (1986), Forecasting Economic Time Series (New York: Academic Press). Greenwald, B.C. and J.E. Stiglitz (1986a), 'Imperfect Information, Finance Constraints, and Business Fluctuations', in M. Kohn and S.C. Tsiang (eds), Finance Constraints, Expectations, and Macroeconomics (Oxford: Clarendon Press), 103-40. Greenwald, B.C. and J.E. Stiglitz (l986b), 'Money, Imperfect Information, and Economic Fluctuations', ibid., 141-65. Greenwald, B. and J.E. Stiglitz (1987), 'Keynesian, New Keynesian and New Classical Economics', Oxford Economic Papers, 39,119-32. Greenwald B. and J.E. Stiglitz (1988a), 'Examining Alternative Macroeconomic Theories', Brookings Papers on Economic Activity, 207-70. Greenwald, B. and J.E. Stiglitz (l988b), 'Financial Markets Imperfections and Business Cycles', NBER Working Papers, no. 2494, Cambridge, Mass. Gurley, J.G. and E.S. Shaw (1955), 'Financial Aspects of Economic Development', American Economic Review, 4. Hall, R. (1987), 'Monetary Policy under Financial Innovation and Deregulation', in Y. Suzuki and H. Yomo (eds), Financial Innovation and Monetary Policy (Tokyo: University of Tokyo Press). Kaldor, N. (1940), 'A Model of the Trade Cycle', Economic Journal, 50, 78-92. Kalecki, M. (1937a), 'A Theory of the Business Cycle', Review of Economic Studies, February, 77-97. Kalecki, M. (193 7b), 'The Principle ofIncreasing Risk' , Economica, November, 441-7. Kashyap, A.K., O.A. Lamont and J.C. Stein (1992), 'Credit Conditions and Cyclical Behavior of Inventories: A Case Study of the 1981-82 Recession', University of Chicago, mimeo. Keynes, J.M. (1936), The General Theory of Employment, Interest and Money (London: Macmillan). King, R.G. and C.1. Plosser (1984), 'Money, Credit and Prices in a Real Business Cycle', American Economic Review, 74, 363-80. Klein, L.R. (1948), 'Notes on the Theory of Investment', Kyklos, 2 (I), 97-117. Kockesen, L. and W. Semmler (1994), 'An STR Model to Test for Nonlinear FinancialReal Interaction', Department of Economics, New School for Social Research, mimtSo. Kopcke, R. W. (1985), 'The Determinants of Investment Spending', New England Economic Review, July/August, 19-35. Leijonhuvud, A. (1989), 'Hicks on Time and Money', in J.C. Wood and R.N. Woods (eds), Sir John R. Hicks: Critical Assessments (New York: Routledge). Meyer, J.R. and E. Kuh (1957), The Investment Decision (Cambridge, MA: Harvard University Press). Minsky, H.P. (1975), John Maynard Keynes (New York: Columbia University Press). Minsky, H.P. (1982), Can 'It' Happen Again? Essays on Instability and Finance (New York: M.E. Sharpe). Mishkin, F.S. (1991), 'Asymmetric Information and Financial Crisis, A Historical Perspective', in R.O. Hubbard (ed.), Financial Markets and Financial Crises (Chicago: University of Chicago Press). Mizrach, B. (1992), 'Forecast Comparison in U', Department of Finance, Wharton School, mimeo. 634 Finance, Profit Expectations, and Investment Neftci, S.N. (1984), 'Are Econometric Time Series Asymmetric around the Business Cycle?', Journal Of Political Economy, 92(2), 307-328. Nell, E. and W. Semmler (1992) (eds), Kaldor and Mainstream Economics (LondonlNew York: Macmillan). Pindyck, R. and D. Rubinfeld (1981), Econometric Models, Econometric Forecasts (New York: McGraw-Hill) Potter, S.M. (1994), 'Asymmetric Propagation Mechanisms', in W. Semmler (ed.), Business Cycles: Theory and Emperical Methods (Boston: Kluwer). Ramsey, J. (1969), 'Tests for Specification Errors in Classical Linear Least Squares Regression Analysis', Journal of Royal Statistical Society, Series B, 31, 350-71. Rappoport, P., and L. Reichlin (1987), 'Segmented Trends and Nonstationary Time Series', Economic Journal, 99,168-77. Romer, C.D. and D.H. Romer (1990), 'New Evidence on the Monetary Transmission Mechanism', Brookings Papers on Economic Activity, I, 150-213. Sayers, C.L. (1994), 'Testing for Chaos and Nonlinearities in Macroeconomic Time Series', in W. Semmler (ed.), Business Cycles: Theory and Empirical Methods (Boston: Kluwer). Semmler, W. (1986), 'On Nonlinear Theories of Economic Cycles and the Persistence of Business Cycles', Mathematical Social Sciences, 12 (I), 47-76. Semmler, W. (1989) (ed), Financial Dynamics and Business Cycles: New Perspectives (ArmonklLondon: M.E. Sharpe). Semmler, W. and M. Sieveking (1993), 'Nonlinear Liquidity-Growth Dynamics with Corridor-Stability', Journal of Economic Behavior and Organization, 22, 189-208. Shleifer, A. and H. Summers, 'The Noise Trader Approach to Finance', Journal of Economic Perspectives, vol.4, no. 2, Spring 1990, 19-34. Stiglitz, G.E. and G.B. Greenwald (1989), 'Towards a Theory of Rigidition', American Economic Review, 79 (2) May, 364-69. Stiglitz, J. and A. Weiss (1981), 'Credit Rationing in Markets with Incomplete Information', American Economic Review, June, 353-76. Stock, J.H. and M.W. Watson (1989), 'New Indexes of Coincident and Leading Economic Indicators', in 0.1. Blanchard and S. Fischer (eds), NEBER Macroeconomics Annual 1989 (Cambridge, MA: MIT Press). Summers, L.H. (1981), 'Taxation and Corporate Investment: A q-Theory Approach' Brookings Papers on Economic Activity, 1,67-127. Taylor, L. (1983), Structuralist Macroeconomics (New York: Basic Books), Taylor, L. (1985), 'A Stagnationist Model of Economic Growth', Cambridge Journal of Economics, 9, 383-403. Taylor, L. (1988), 'Real and Money Wages, Output, and Inflation in the SemiIndustrialized World', MIT, mimeo. Taylor, L. and S.A. O'Connell (1989), 'A Minsky Crisis', in W. Semmler (ed.), Financial Dynamics and Business Cycles: New Perspectives (ArmonklLondon: M.E. Sharpe). Tobin, J. (1965), 'Money and Economic Growth', Econometrica, 33,671-84. Tobin, J. (1975), 'A Keynesian Model of Recession and Depression', American Economic Review, 65, 195-202. Tobin, J. (1982), 'Money and Finance in the Macroeconomic Process', Journal of Money, Credit, and Banking, XIV (2), 171-203. Tobin, J. (1990), 'Endogenous Money', in E. Nell and W. Semmler (eds), Nicholas Kaldor and Mainstream Economics (London/New York: Macmillan). Woodford, M. (1989), 'Finance, Instability and Cycles', in W. Semmler (ed.), Financial Dynamics and Business Cycles: New Perspectives (ArmonklLondon: M.E.Sharpe). Zarnowitz, V. and Moore, B. (1986), 'Stylized Facts of the Business Cycle', in R.I. Gordon (ed.), The American Business Cycle: Continuity and Change (Chicago: University of Chicago Press). 24 Aggregate Determinants of Financial Instability * Marc Jarsulic INTRODUCTION There is a growing body of theoretical literature which analyzes the connections between financial factors and macroeconomic stability. Following Minsky (1982), some authors (Franke and Semmler, 1989; Taylor and O'Connell, 1985) emphasize the importance of expectations. They connect expectations of future profitability to endogenously determined interest rates. For some configurations of parameters, the effects of expectations on aggregate .demand can create instability and self-reinforcing declines in aggregate activity. Other theorists have examined the interaction of multiplier-accelerator behavior and financial factors. Foley (1986, 1987) shows that when accelerator effects produce local instability, interest rate movements can determine turning points in cycles. Skott (1994) has modified Kaldor's (1940) cycle model to examine the dynamics of 'tranquillity' and 'fragility'. Local instability is produced by a Kaldorian investment function and sufficiently rapid adjustment of savings to investment. While the multiplier-accelerator and Minskian models have much to tell us about financial dynamics, they appear to underestimate the dynamical complications introduced by the mere existence of debt. Intuition suggests that if investment demand is influenced by net cash flow, and there are fixed payment commitments, the necessary ingredients for instability are already present. Shifts in expectations, interest rate shocks, or explosive demand may not be needed to produce aggregate instability and financial crises. This point of view will be explored by means of an aggregate dynamical model which has two endogenous variables, the rate of capital stock growth and the debt-capital ratio of firms. The behavior of these variables is determined by three functional relationships. The growth rate is determined by a lagged adjustment to net cash flow. Profits are determined by capacity utilization. The evolution of the debt-capital ratio is determined by a dynamic flow of funds constraint for firms. Some important elements of the model are left exogenous. The interest rate is assumed to be fixed. Long-period expectation formation, always problematic in any model, is left exogenous. Instead, it is assumed that long-period expectations are reflected in the stable financial rules governing capital accumulation. 635 636 Aggregate Determinants of Financial Instability Where possible, parameters and equilibrium values of endogenous variables will be calibrated to constrain model behavior. It will be shown that, even with these restrictions, a wide range of dynamical behaviors is plausible. Stability, instability, saddle point behavior and limit cycles all appear to be possible outcomes. The model is consistent with several ideas which are central to alternative economic theory. The investment function is consistent with Kalecki's(l97l) ideas on increasing risk and financial constraints. It reflects the general Post Keynesian view (Arena, Lavoie and Seccareccia, this venue) that credit provision is central to the process of capital accumulation. The interest rate assumption is also part of Post Keynesian (Kaldor, 1982; Moore, 1989) and circuit (Lavoie, and Parguez, this volume) theory. To the extent that Post Keynesians and circuit theorists are in accord with the model, it may be viewed as an exploration of the dynamical implications of their positions. 2 GOODS MARKET EQUILIBRIUM AND AGGREGATE DISTRIBUTION The model assumes a closed economy with a constant price level, in which there are representative firms and households. Although firms will be assumed to retain a constant portion of their net cash flow, it will be convenient to assume that the aggregate savings propensity for the economy is a constant 1 > s > 0, regardless of the distribution of income. Hence equilibrium in the goods market will be determined by the equality of savings and investment, which can be written in intensi ve form as YIK =IlsK =gls (1) where Y is real NNP, K is real capital stock, I is real net investment, and g = 11K. As the work of empirical business cycle researchers has shown (Boddy and Crotty, 1975; Bowles, Gordon and Weisskopf, 1989; Hahnel and Sherman, 1982; Weisskopf, 1979), corporate profitability in the United States, as measured by the average profit rate of the non-financial corporate business sector, is often moving in the direction of capacity utilization. For simplicity, the observed non-linearities in this relationship will be ignored and it will be assumed that profits are an increasing function of utilization. It will be seen that outcomes are complicated enough, even with this simplification. To incorporate this relationship into the model, YIK will be taken as a proxy for utilization and the rate of profit will be determined by 'IT = IlIK = 11(g), 'TTg > 0 where n is the flow of profits and (2) 'IT is the profit rate. Marc Jarsulic 3 637 DYNAMICS OF CAPITAL ACCUMULATION Starting with a Keynesian view of expectation formation, it is impossible to represent the economic future by specifying a probability distribution. However, as Lawson (1985) cogently argues, this does not meant that one needs to assume that decisions about the future are irrational and exogenous to the functioning of the economy. When there are social decisions to accept conventional assumptions, it is reasonable to depict the resulting behavior as rule governed. The rules are not based on explicit optimization, but they are rules none the less. In the case of investment demand, which is concerned with the future if anything is, we will assume that long-term expectations about matters such system viability and t~e possibility of reasonable profits are determined exogenously. It will be useful to depict shifts in these expectations without explaining their formation. Given this, however, the rules governing the determination of feasible investment, and the actual execution of investment plans, still need to be spelled out. In his discussion of the principle of increasing risk, Kalecki (1971: 105-9) provided an influential formulation of the idea that investment is financially constrained. Because of genuine, non-probabilistic uncertainty, firms will be discouraged from borrowing by large or increasing debt burdens. Fixed payment commitments put income and capital in danger if business conditions sour. Lenders, who are also unable to know the future, and who are somewhat less well informed about a firm's organization and behavior than its owners, have reason to restrict lending as firms become more highly leveraged. Kalecki used this depiction of limited knowledge on both sides of financial markets to provide part of his account of investment demand. In his work realized profits, along with the evaluations of financial riskiness by firms and financial markets, are major determining factors of capital accumulation. Recently, 'new Keynesian' macroeconomists have used the existence of finance constraints to generate explanations of market failure in an economy characterized by probabilistic uncertainty (e.g. Greenwald and Stiglitz, 1988; Mankiw, 1986; Stiglitz and Weiss, 1981). Asymmetric information and agency problems are shown to produce restrictions on finance, which affects capital accumulation. There is some empirical work consistent with these theoretical ideas. There is a history of including profitability in macoeconometric investment functions for the US economy (Abel and Blanchard, 1986; Kopcke, 1985), sometimes with multiyear lags. Fazzari and Mott (1986) and Fazzari et at. (1988) use micro data to show that, among other factors, two-year distributed lags on cash flow are statistically significant determinants of investment rates for some firms. Also, Bowles et at. (1989) have shown a strong correlation between the rate of capital accumulation and the lagged profit rate. On the basis of this empirical evidence, a lagged cash flow constraint can reasonably be considered a major determinant of investment. 638 Aggregate Determinants of Financial Instability Note that the finn level studies show a very weak explanatory power for q variables, which are meant to represent the effect of financial market conditions on investment decisions. This finding may be counted as evidence against the Minskian explanation of cycles, since asset prices are central to this explanation of behavior. To represent investment in light of these considerations, let us begin by defining net cash flow as retained earning less required financial payments. To simplify matters, it will be assumed that the only required financial payments are interest on the outstanding stock of debt. This is equivalent to assuming that all debt is short term and instantly rolled over; or that debt contracts require the constant payment of an instantaneous interest rate, with the principal repayable at will. Given this, net cash flow will be CF= P(TI - rD) (3) where 1 > P > 0 is the corporate retention rate, assumed constant; r is the rate of interest, also assumed constant; and D is the stock of debt. The idea that there is a roughly constant p for corporations is supported by data on the NFCB sector for the US economy (Council of Economic Advisers, 1988, p. 262). If CF is normalized by K, it becomes cf = p( 7T - rd) (4) where d = DIK. Using (3) and (4), the feasible rate of accumulation will be given by gl = H(cf) (5) The cyclical behavior of the ratio g!cf(Wolfson, 1986: 143) indicates that He1can be positive or negative, since the ratio increases in the early cycle and decreases in the late cycle. The investment studies previously discussed, as well as business cycle research (Wolfson, 1986: 142-50; Zarnowitz, 1975), suggest significant adjustment lags between feasible and actual growth rates. Therefore actual accumulation will be determined in the model by F(g, d) = g = a(gf - g), a> 0 (6) This specification allows for a lagged detennination of g, with the lag time between gf and g being reduced as a increases. 4 FIRM FLOW-OF-FUNDS AND THE EVOLUTION OF DEBT The evolution of the debt4:apital ratio is derived by beginning with the identity ~=~-~ m Marc Jarsulic 639 which gives d= D/K -gd (8) If we assume that firms will use any net retained revenue to payoff the stock of outstanding debt or accumulate financial assets, then a dynamical cash flow constraint for the firm is given by D =interest payments + investment expenditure net cash flow. Negative values of Dwould correspond to net financial accumulation. Using (4) this constraint is D = rD +1- CF (9) When (8) is substituted into (7) we obtain d = G(g, d) =(1 + p)rd + g - p7l'(g) - gd (10) which describes the evolution of the debt-capital ratio. 5 LOCAL STABILITY ANALYSIS The local stability of the system (6), (10) is easily stated in terms of the Jacobian (II) evaluated at (g*, d*), where gt' and d* are eqUilibrium values. When Det (J) < 0, (g*, d*) is a saddle point. When Det(J) > 0, stability will be determined by Tr(J). When Tr(J) < 0, the point is stable; when Tr(J) > 0, unstable. To understand the dynamics of the model, it is necessary to consider the factors which determine the elements of the matrix J, Det(J) and Tr(J). The specific expressions are aA = a[Hc/.p7Tg ) - l] aB=-aHcJPr C= l-d- P7Tg E=(J +p)r-g Det(J) a[HcJP [r(l - d) - (r - g)p7Tg } TrW aA +E = = + {g - (l + p)r}] On the basis of casual empiricism, the entries in J can be restricted somewhat. The average value of g in the postwar period is .03, which will be taken as an estimate for g*. The real interest rate varies over cycles, and cyclical averages trend upward in the postwar period. Therefore, we assume that r will be in the interval [.03,.08], which is the observed range of cyclical averages after 1958. The value of p will be set at .6, its average value over the period. Given these 640 Aggregate Determinants of Financial Instability Table 24.1 Stylized facts about the US non-financial corporate business sector A verage growth rate of capital I .03 A verage retention rate 2 .6 Average investment/cash flow rati0 3 1.2 A verage real baa bond Debt capital ratios I 2 3 4 S rate4 [.03, .08] [.5, .9] Author's calculation from data supplied by Thomas Michl, for 1948-88. Economic Report of the President, 1988 (Washington, DC: Goverment Printing Office), p. 262. Period is 1948-87. Colin Mayer, 'Financial Systems, Corporate Finance and Economic Development' , in R. Hubbard (ed.), Asymmetric Infonnation, Corporate Finance and Investment (Chicago: University of Chicago Press, 1990), p. 310. Period is 1970-85. Computed from nine postwar cyclical averages (1947-82). Data in R.I. Gordon and 1. Veitch, 'Fixed Investment in the American Business Cycle, 1919-83, in R.I. Gordon (ed.), The American Business Cycle (Chicago: University of Chicago Press, 1989), p. 294. Cycle averages range from .01 to .08. Post-1958 values are .03 or higher. Board of Governors of the Federal Reserve System. Balance Sheets for the US Economy, September 1991, pp. 31-6. Period is 1945-90. Observations range over the given interval. increasing from a low of.5 in 1945 to a high of.9 in 1991. values, it can be assumed that E > O. It has already been noted the Hefcan be of any sign. If we limit our considerations to the case Hef > 0, then B < O. Local stability is not a necessary consequence of this configuration. The equilibrium (g*, d*) could be a saddle point. Setting r = g* = .03, d* = .9, and Her = 1.2, all empirically possible outcomes as can be seen from Table 24.1, then Det(J) = 0:[.72(.003) - 0.18] = -.016. If Det(J) > 0, the eqUilibrium is not a saddle point. However, the local stability of the eqUilibrium will depend on TrW = a:A + E. If A > 0, then the equilibrium point is unstable. If A < 0, changes in the parameter 0: will affect the local stability, but not the eqUilibrium values. As ex ~ 0, TrW becomes positive. Hence in this case local stability is affected by lags in adjustment in investment. Longer lags make local instability more likely in this case. I Changes in assumption about Hef and 7Tg will of course produce different outcomes, but there seems little reason to suppose that local stability will ever be assured. 6 BIFURCATION ANALYSIS OF LOCAL BEHAVIOR Let us assume that, A, B < 0, and C, E > 0 and Det(J) > O. Although the equilibrium will not be a saddle point under these conditions, Tr(J) can have negative, Mare Jarsulie 641 positive and zero values, depending on the size of parameter O. These eigenvalues mean the local behavior of the equilibrium will be respectively stable, neutrally stable, and unstable. When the change of a parameter can cause such an exchange of local stability, while leaving equilibrium values unchanged, then it is possible to deduce additional information about the local behavior of the dynamical system. The tool for doing this is the Hopf bifurcation theorem. The Hopf theorem (Guckenheimer and Holmes, 1986: 151-2; Wiggins, 1990: 270-8) says that if, as a result of a parameter change, there is an exchange of stability at the equilibrium point, then there may be a limit cycle surrounding the equilibrium. Furthermore, the cycle becomes longer as the value of the parameter increases. 2 Thus, although system (6), (10) can have a stable fixed point when Det(J) > 0 and TrW < 0, sufficient reductions in O. This means that the real parts of the complex eigenvalues at the equilibrium become positive. And it also means that as the lag with which actual accumulation adjusts to feasible, the equilibrium is destablized and a limit cycle appears. This is depicted in the phase diagram of Figure 24.1. The stability of the limit cycle can also be determined. The Hopf theorem states that the sign of an index constructed from the derivatives of (6), (10), evaluated at (g*, d*) will do this. As can be seen in the appendix, even in a relatively simple model such as this, the sign of this index may be difficult to determine. However, in some simple cases clear results fall out. If H ef, 'lTg > 0 and H(ef) and 'IT(g) are locally linear, then the index is a = -2 0 gd (6) (10) Marc Jarsulic 643 write the Jacobian as J= [ Fg Gg (l2) evaluated at (g*. d*; ao), i.e. at an equilibrium point, and with a value for the parameter a that sets TR(J) =o. Liu et al. (1986) derive the following expression for the index used in the Hopf theorem to determine the stability of the limit cycle: a = [(L(Fggg + Gggd) + 2M(Fggd + Ggdd ) + C(Fgdd + Gddd» Det(J) + (MFgg + CFgd)(LFgg + 2MFgd + CFdd ) - (MGdd + LGgd)(LGgg + 2MGgd + CGdd ) - ML(FglJgg + FggG gd} + MC(FglJdd + FdlJgd) - UFggGgg + CZFdPddJlDet(J)C where all derivatives are evaluated at (g·.d·; ao). Now it is assumed that second and higher derivatives of H(cf, B) and 1T(g} are equal to zero. This means that there will be no third order derivatives for (6), (10). Moreover, Fgg =Fdd =Fdg =Fgd =Ggg =Gdd =0, and Gdf =Ggd =-1. Hence the index reduces to a = -[2LM]IDet(J)C. In terms of (II), a =-[ 2 a BEJIDet(J)C. Notes 1. 2. I would like to thank Basil Moore, Tracy Mot!, Peter Skot! for helpful comments an earlier versions of this paper. All errors remain my responsibility. The effect of time lags on local stability and global dynamics is explored in a different case in Jarsulic (1993). There it is shown that when discrete or distributed lags are incorporated into a continuous-time Keynesian model of accumulation. a wide variety of behaviors are possible. In general. longer discrete lags and narrower distributed lags contribute to local instability and global complexity. The Hopf bifurcation theorem (Wiggins. 1990: 270-8) can be stated as follows: Suppose the parameterized system x= F(x. y. f.L) j= G(x. y. f.L} which is CS. has a fixed point at the origin for all values of the real parameter f.L. Furthermore. suppose that at the origin there is a f.L = f.Lo for which the system has a pair of imaginary eigenvalues. Further, suppose that the real parts of these eigenvalues are functions of the parameter f.L. Then define = d oRe(A(f.L»/0f.L. evaluated at f.Lo (i) where A and Xare the eigenvalues of the vector field linearized about the origin. It is then possible to show that there is an index. a, such that: if d < 0 and a > O. then the origin is an unstable fixed point for IL < f.Lo and an asymptotically fixed point for IL> f.Lo. with an unstable periodic orbit for f.L < f.Lo. 644 (ij) Aggregate Determinants of Financial Instability if d < 0, a < 0, then the origin is an unstable fixed point for I-" < 1-4l, and an asymptotically stable fixed point for I-" > 1-"0' with a stable periodic orbit for I-" < 1-4l. The calculation of the index a is discussed in the appendix. References Abel, A. and O. Blanchard (1986), 'The Present Value of Profits and Cyclical Movements in Investment', Econometrica, 54, 249-73. Board of Governors of the Federal Reserve System (1991), Balance Sheets for the US Economy, September 1991, Washington. Boddy, R. and J. Crotty (1975), 'Class Conflict and Macro Policy: The Political Business Cycle', Review of Radical Political Economics, Spring. Bowles, S., D. Gordon and T. Weisskopf (1989), 'Business Ascendancy and Economic Impasse', Journal of Economic Perspectives, 3, 277-95. Council of Economic Advisers (1988), Economic Report of the President (Washington, DC: US Government Printing Office). Fazzari, S. and T. Mott (1986), 'The Investment Theories of Kalecki and Keynes: An Empirical Study of Firm Data, 1970-82', Journal of Post Keynesian Economics, 9, 141-206. Fazzari, S., R. Hubbard, and B. Petersen (1988), 'Financing Constraints and Corporate Investment' , Brookings Papers on Economic Activity, no. 1. Foley, D. (1986), 'Stabilization Policy in a Nonlinear Business Cycle Model', in W. Semmler (ed.), Competition, 1nstability and Nonlinear Cycles (New York: Springer-Verlag). Foley, D. (1987), 'Liquidity-Profit Rate Cycles in a Capitalist Economy', Journal of Economic Behavior and Organization, 8, 449-68. Franke, R. and W. Semmler (1989), 'Debt-financing of Firms, Stability, and Cycles in a Dynamical Macroeconomic Growth Model', in W. Semmler (ed.), Financial Dynamics and Business Cycles (Armonk, NY: M.E. Sharpe). Greenwald, B. and J. Stiglitz (1988), 'Imperfect Information, Finance Constraints, and Business Fluctuations', in M. Kohn and S.C. Tsiang (eds), Finance Constraints, Expectations, and Macroeconomics (Oxford: Oxford University Press). Guckenheimer, J. and P. Holmes (1986), Nonlinear Oscillations, Dynamical Systems, and Bifurcations of Vector Fields (New York: Springer-Verlag). Hahnel, R. and H. Sherman (1982), 'The Rate of Profit over the Business Cycle', Cambridge Journal of Economics, 6, 185-94. Jarsulic, M. (1990), 'Debt and Macro Stability', Eastern Economic Journal, 16,91-100. Jarsulic, M. (1993), 'Complex Dynamics in a Keynesian Growth Cycle Model', Metroeconomica, 44, 43-64. Kaldor, N. (1940), 'A Model of the Trade Cycle', Economic Journal, 50, 78-92. Kaldor, N. (1982), The Scourge of Monetarism (Oxford: Oxford University Press). Kalecki, M. (1971), Selected Essays on the Dynamics of the Capitalist Economy (Cambridge: Cambridge University Press). Kopcke, R. (1985), 'The Determinants of Investment Spending', New England Economic Review, July-August, 19-35. Lawson, T. (1985), 'Uncertainty and Economic Analysis', Economic Journal, December, 909-27. Liu, W., S. Levin and I. Yoh (1986), 'Influence of Nonlinear Incidence Rates upon the Behavior of SIRS Epidemiological Models', Journal of Mathematical Biology, 23, 187-204. Marc Jarsulic 645 Mankiw, G. (1986), 'The Allocation of Credit and Financial Collapse', Quarterly Journal of Economics, 101,455-70. Minsky. H. (1982). Can 'It' Happen Again? (Armonk, NY: M.E. Sharpe). Moore, B. (1989), Horizontalists versus Verticalists (Cambridge: Cambridge University Press). Skot!, P. (1994), 'On the Modeling of Systemic Financial Fragility', in A.K. Dutt (ed.), New Directions in Analytical Political Economy (Aldershot, UK: Edward Elgar). Stiglitz, J.E and A. Weiss (1981) 'Credit Rationing in Markets with Imperfect Information', in American Economic Review, 71(3) June, 393-410. Taylor, L. and S. O'Connell (1985), 'A Minsky Crisis', Quarterly Journal of Economics, 100, 871-86. Weisskopf, T. (1979), 'Marxian Crisis Theory and the Rate of Profit in the Post-War US Economy', Cambridge Journal of Economics, December, 3(4), 341-378. Wiggins. S. (1990), Applied Nonlinear Dynamics and Chaos (New York: SpringerVerlag). Wolfson, M. (1986), Financial Crises (Armonk, NY: M.E. Sharpe). Zarnowitz, V. (1975), 'Business Cycle Theory in Historical Perspective', Journal of Economic Literature, 23, 523-80. Part VI Policy Introduction to Part VI Both PK and CA tend to support an interventionist position, and adopt a critical stance towards much contemporary policy, which they see as weakening the defenses against instability. Kregel in Chapter 25 first observes that the monetary theory of the circuit criticizes economic theory for failing to recognize the existence and importance of endogenous money creation through bank financing of productive activity. Yet economists fro,m Hume and Smith to the Chicago school have clearly identified the importance of endogenous money and recommended that the efficient operation of the free market system requires that it be actively controlled, if not eliminated. This suggests that the important question is not the endogeneity of money, but rather the framework in which both money and credit are regulated. Kregel concludes on a paradox within circuit theory: endogenous money created by commercial banks is necessary to capitalism, but it also creates financial instability which threatens its survival. Rousseas in Chapter 26 returns to the ideas of Keynes' Treatise, to develop the distinction between the industrial and financial circulations, and to explore the relations between them. Two important conclusions emerge: the demand for cash deposits as a means of payment in the industrial sphere is relatively stable, but the demand for cash in the financial sphere is quite volatile. The total demand for money, therefore, is not linked in any stable way to nominal GNP. Secondly, in a speculative boom, the financial sphere can draw resources away from the industrial sector, leading the latter to slow down and even contract. This is a serious source of danger, but current political developments are eroding the instruments needed to control such speculative excesses. Epstein in Chapter 27 turns to international issues and argues that a profit squeeze is less likely under a system of flexible exchange rates than under a fixed rate regime. With flexible exchange rates domestic firms are better able to pass along increased costs in higher prices without being hurt by foreign competition, since the depreciating exchange rate will maintain the competitiveness of the domestic industry. Industrial capital will therefore tend to support expansionist measures under flexible exchange regimes. But financial capital tends to be wary of expansion under both regimes, since expansion tends to bring inflation. The argument is developed with reference to two classic analyses of macropolicy-Kalecki and Boddy and Crotty. The argument of Boddy and Crotty is shown to apply to a fixed exchange rate regime, that of Kalecki to flexible rates. De Brunhoff in Chapter 28 examines the European plan for the creation of a single currency, using official texts from the European authorities. The project 649 650 Introduction to Part VI recommends a stable and single currency, called BCU, which would be issued and managed by an independent, credit-worthy Central Bank. Whatever the advantages of a single currency, the determining of its unit of account and exchange rate poses difficult problems, as does the discrepancy between the European monetary and budget policies, the former being favored in the context of a market economy. 25 The Policy Implications of the Current Bank Crisis, or, 'Is Free Market Capitalism Compatible with Endogenous Money?' Jan A. Kregel 1 CIRCUIT THEORY AND THE CRISIS IN THE US BANKING SYSTEM What I am going to say is extremely simple, indeed so simple that I am sure that it will be misunderstood. To avoid that possibility I shall make the argument as complex as possible to try to circumvent that natural instinct to react negatively to any proposition which challenges the status quo. The proposition is not, however, new; indeed many may be surprised by the historical precedents. I also hope to be provocative, at least enough to start people thinking about a series of problems that once exercised the best monetary theorists, but seems to have disappeared from modem discourse. It is clear to any informed observer that the structure of the money and banking system which has been in place in the US since the New Deal is about to (perhaps it already has) collapse, or it might be more correct to say, self-destruct.! While most experts are concerned to formulate proposals to save it, I would like to suggest that since the arguments for its existence were never very strong, the opportunity should be seized to replace it. There are two issues which have attracted particular attention. The first is the moral hazard problem linked to federal guarantees for deposit insurance in a competitive financial system, and the second is the range of activities that insured deposit-takers should be allowed: the repeal of the Glass Act of 1933. Neither of these issues can be resolved by piecemeal adjustment to the system; discussion should be better directed to radical revision. To forestall the instinctive rejection of the proposal I am going to put forward? I would like to point out that it comes with the best of free-market credentials and is based on analysis by economists going back to David Hume, Adam Smith and includes Irving Fisher plus the members of the pre-Friedman and Schwartz Chicago school of economics. It is based on a principle common to all these 651 652 Policy Implications of the Current Bank Crisis writers, that in matters of the creation of money, the free competitive market cannot be left to regulate itself; that control over the creation of money is one of the basic responsibilities of representative government. I realize that in recent times the belief in the benefits of the unrestricted operation of free competitive markets has come to dominate every aspect of public and private affairs; I just want to make it clear that if one's instinct is to reject the proposal because it conflicts with this modern current of opinion, then most proponents of free markets from Adam Smith to Henry Simons must also be consigned to the same category. In fact, the extension of free market principles to the money markets, which has generated the wave of financial deregulation and innovation experienced in the United States since the 1970s, runs counter to a long history of both liberal and conservative economic thinking. The results of this extension of market competition, first gradually, and then increasingly rapidly, to the operations of deposit-taking commercial banks and savings and loans institutions in order to offer them a 'level playing field' in which to compete with other financial intermediaries, by giving them unrestricted access to asset markets and freedom to compete for funds, confirms this traditional position that money and financial markets are different from other competitive markets. The failure of the US government in the 1980s to effectively discharge its responsibility to determine the framework in which financial institutions operate has made it, paradoxically, one of the largest bank holding companies in the world. Although the US government 'takeover' of these institutions have been justified as preserving the 'stability' of the financial system, it has only served to preserve the deposit liabilities of private banks as the central component of the supply of means of payment. In the event, the market is imposing the radical free market position of Smith, Simons, Mints and others: that 'the money market cannot be left in private hands'.3 The implications of the analysis of these free market supporters is that commercial banks should either be abolished or regulation should prevent them from issuing means of payment which should be solely the responsibility of government. This approach solves the problem of refinancing and redesigning the deposit insurance system by making it unnecessary. The aim of this chapter is to argue that this traditional position is the correct response to the problem of deposit insurance,4 and to assess the position of the 'circuit theory' approach with respect to this proposal. 2 A SIMPLE EXAMPLE It is often easier to see a point by putting it in the international context; in the present case the argument may be more easily seen with reference to the traditional Fleming-Mundell resolution of the conflict between internal and external equilibrium, by means of foreign borrowing. If the trade account is in the balance at less than full employment, expansion leads to a trade deficit and pressure on the exchange rate requiring restrictive policy to reduce demand and imports. But, Jan A. Kregel 653 if the country can borrow in international capital markets, an expansionary fiscal policy accompanied by monetary restraint to raise interest rates sufficiently above international levels may produce capital inflows to offset the trade deficit and produce payment balance at full employment. Foreign borrowing thus eliminates the balance of payments constraint on full employment. It is now generally recognized, however, that this is at best a short-run solution, for in the absence of measures to reduce imports per unit of output interest on the foreign borrowing increases the current account deficit, requiring increasing levels of borrowing and what may be called 'Ponzi' financing of the trade deficit. Eventually foreigners lose confidence in the currency and restrictive policies will be necessary. Thus, although the solution of foreign borrowing is not generally valid for all countries (Kregel, 1991), it did apply in one case, the United States under Bretton Woods. The US had no foreign constraint as long as it could borrow without having to pay interest abroad, which it could do because foreign central banks and private agents were willing to hold dollars or dollar deposits (which then paid no interest) as reserves in the US; this is the international equivalent of the US using 'the printing press' (as De Gaulle pointed out) to finance the trade deficit associated with the desired level of employment at no interest cost.s The same reasoning can be applied to the closed economy. If the autonomous level of private investment produces a savings-investment balance at less than full employment potential, the government can increase expenditure above tax receipts to raise the level of output, financing the deficit by issuing government debt. In the absence of any change in private sector autonomous expenditures, the following year the government would have to return to the financial markets to finance the deficit plus the interest payments on the accumulated debt of prior years. Eventually the result is the same as in the international case: the government has to increase interest rates to continue borrowing, a given expenditure deficit requires an increasing amount of borrowing to cover accumulating interest. In the end the government loses the confidence of the capital markets and has to increase taxes and introduce an austerity policy to right the government finances. 6 But why does the position of the US under Bretton Woods not carryover to US government domestic finances? Is not the Federal Reserve (or any other central bank) in exactly the same position vis-a-vis US residents as it was to nonresidents under Bretton Woods? If demand is insufficient to produce full employment why should government expenditures be financed by borrowing from the financial markets at interest, when they could be financed by the issue of additional Federal Reserve Notes at no interest cost? 3 AN HISTORICAL EXAMPLE A historical example may also be helpful to understand how the current system developed: the creation of the Bank of England. In the seventeenth century the 654 Policy Implications o/the Current Bank Crisis Royal Mint in the Tower of London was used by merchants as a safekeep for their gold coin and plate. In 1640, Charles I, in urgent need of money committed an act of lese majeste and permanently 'borrowed' £ 120 000 worth of merchants' valuables held in 'his' Mint. After this incident another 'security' system became necessary and merchants entrusted their valuables to goldsmiths, who in turn placed them at interest in the Exchequer for safekeeping; goldsmiths thus became intermediaries and formal creditors of the Crown. But, Charles II, true to his namesake, in urgent need of funds to fight the Dutch in 1672, closed 'his' Exchequer and stopped payment of principal and interest on Exchequer debt, confiscating £1 328526 worth of goldsmiths' loans. This 'stop of the exchequer' generated widespread discontent (English understatement) and the risk of financial ruin for the goldsmiths, as well as their merchant clients. The King was eventually forced to promise repayment (a promise never made good, the 'Goldsmith Banker Debt' was eventually consolidated with the South Sea Annuities to initiate the British national debt). Not surprisingly, in 1694 William III, in urgent need of funds to continue his war with France, found he was unable to borrow due to the lack of creditworthiness of the Crown. The Bank of England was created as a 'financial innovation' which was intended to provide an intermediary to fund the Crown. A group of private businessmen was granted a Royal Charter, giving them the right to call themselves the Governor and Company of the Bank of England, in exchange for their promise to underwrite £1 200000 of government debt. The Bank of England in turn issued notes to the public to fund the government debt. The Bank of England started life as the lender of first resort to the Crown, intermediating with the public, which would not buy government debt because it was considered too risky. For its services to the Crown the Bank eventually acquired the complete monopoly of the note issue.? Note that this is precisely the opposite of current conditions in which government debt is considered as 'risk free', and banks can barely maintain investment grade ratings on their liabilities! The key to the subsequent development of the structure of Anglo-Saxon banking lies in the exchange of monopoly on the note issue for the obligation of being the lender of first resort to the Crown. But this meant that all other private banks lost one of their sources of revenue, the issue of their own notes. The banks thus responded with a financial innovation of their own, the creation of deposits subject to check which could serve as a source of income from lending to replace that lost with the prohibition of the private issue of notes. This is the source of the 'dual' definition of the money supply as composed of 'public' liabilities: the notes of the central bank, and 'private' liabilities: the demand deposits of the private banks.8 But, in modern conditions when the central bank has become a public agency (and is given formal responsibility for monetary policy via control of monetary aggregates), it is the private banks who have become the major lenders to the government through holdings of government securities and financing the holding Jan A. Kregel 655 of their clients. When it is no longer the government that is the high credit risk, but rather the banks, there seems to be little justification for compensating the banks for the 'services' of intermediating government debt by granting them the right to issue currency in the form of deposits. In effect, the government grants, without consideration from the private banks, the right to issue deposit liabilities which compete with its own central bank's notes; in addition it insures the full substitutability of the competitors' product, and finally acts as a major customer by borrowing those liabilities and paying interest to do SO.9 4 THE FREE MARKET CREDENTIALS That the commercial banks should be prevented from selling deposits which are used as means of payment and that the control and issue of all forms of money used for transactions purposes be left to the government is a proposition that dates back to the monetary discussion of David Hume and Adam Smith. Hume confessed to 'entertain a doubt concerning the benefit of banks and paper credit, which are so generally esteemed advantageous to every nation' (Hume, 1752: 35), while Smith, who generally approved of them, none the less recommended that they be regulated: To restrain private people, it may be said, from receiving in payment the promissory notes of a banker, for any sum whether great or small, when they themselves are willing to receive them; or, to restrain a banker from issuing such notes, when all his neighbours are willing to accept them, is a manifest violation of that natural liberty which it is the proper business of law, not to infringe, but to support. Such regulations may, no doubt, be considered in some respect a violation of natural liberty. But those exertions of the natural liberty of a few individuals, which might endanger the security of the whole society, are, and ought to be, restrained by the laws of all governments; of the most free, as well as the most despotical. The obligation of building party walls, in order to prevent the communication of fire, is a violation of natural liberty of the same kind with the regulation of the banking trade which are here proposed. (Smith 1976: 324) The regulation that Smith proposes is that notes issued by private banks be limited to sufficiently large denomination that they would in practice be used only for transactions between 'dealers' in the wholesale trades. In this way he hoped to prevent the 'frequent bankruptcies to which such beggarly bankers [who issue small denomination notes without full gold backing] must be liable'. The result would be a system in which 'paper money does not necessarily increase the quantity of the whole currency' so that notes simply substitute for the metaIlic currency. Smith can thus be considered as a member of the 'banking' school, for he 656 Policy Implications of the Current Bank Crisis enunciates the theory of 'reflux' as placing a natural limit on the issue of notes; since 'the quantity of gold and silver which is taken from the currency is always equal to the quantity of paper which is added to it' (ibid.). This is, of course, the same as 100 per cent reserve banking; it is achieved not by instituting reserve requirements, but by regulation to limit the use of paper currency to trade between wholesalers and dealers, and manufacturers. Note that Smith's theory is not based on the quantity theory: The proportion between the value of gold and silver and that of goods of any other kind depends in all cases ... upon the proportion between the quantity of labour which is necessary in order to bring a certain quantity of gold and silver to market, and that which is necessary in order to bring thither a certain quantity of any other sort of goods. (ibid.: 329). It is also important to note the way Smith approaches regulation,1O by designing the framework in which competition is presumed to function to produce acceptable results. Smith's regulations do not, for example, deal with the licensing of banks, but with regulating the areas in which banks can operate. 'If bankers are restrained from issuing any circulating bank notes, or notes payable to the bearer, for less than a certain sum; and if they are subjected to the obligation of an immediate and unconditional payment of such bank notes as soon as they are presented, their trade may, with safety to the public, be rendered in all other respects perfectly free. II Thus Smith can be cited as the first economist to suggest that banks be excluded from provision of public means of payment. 12 5 THE 'CHICAGO' CREDENTIALS These early discussions of the role of banks in the process of money and credit creation did not stop with the discussion between Smith and Hume. They are repeated in the late eighteenth-century bullion controversy and again in the nineteenth-century debates between the banking and currency schools, culminating in the passage of Peel's Banking Act of 1844 dividing the Bank of England into a private banking arm and a note issuing arm.13 The same argument resurfaced in the beginning of the twentieth century, but this time in the guise of monetary theories of business fluctuations of the trade cycle. In fact, monetary explanations of fluctuations were formulated as criticisms of the simple application of the Classical quantity theory of money, first by Wicksell, Schumpeter and Hawtrey; a tradition continued by Myrdal, Keynes, Robertson and Hayek in Europe and by Fisher, and then Simons, Douglas, Mints and Viner in the US. It is interesting to note that this latter group represents the 'early' Chicago tradition of monetary economics. The common element in the position of all these economists is similar to that first enunciated by Smith; that money could be a 'disturbing' element in Jan A. Kregel 657 the operation of the competitive forces in the economy, because in the absence of government policy or control there were no 'automatic' forces to regulate its creation in a manner compatible with the stable equilibrium of the rest of the system. The early Chicago school of Douglas (1935), Mints (1950) and Simons (1934), following in the tradition of Irving Fisher, took a rather different view from contemporary exponents of Chicago monetarism and started their analysis of monetary policy from the premise that money and credit markets were different from other private commodity markets because monetary stability was a prerequisite for the efficient operation of competition in a market system. 14 Institutions in the money market would thus have to be designed and controlled in such a way as to ensure the stability necessary to allow competition to function efficiently in the markets of the rest of the economy. One of the chief defects 15 of the US economy was considered to be its dependence on short-term bank debt to finance business operations, seen as the chief cause of business fluctuations and monetary instability.16 If it was the role of monetary policy to set the proper institutional framework for monetary stability, and if fractional reserve banking produced a system in which the production process was financed via short-term business lending, then the framework should be changed by eliminating commercial banks. Note that it was not endogeneity as such, but rather exogenous changes such as the instability of the public's desired currency coefficient (one could here substitute velocity or demand for money, or even liquidity preference) which created difficulties. It was endogeneity that could tum this instability inherent in a free market system into pathological crisis - debt deflation and generalised depression. Competition for reserves to meet changes in the public's portfolio preferences could cause distress sales which would produce a collapse of asset prices, and general bankruptcy in both banks and business. The policy response was not to provide temporary (or permanent, as in the New Deal proposals) price supports, but to change the framework. Private banks were to be eliminated or required to hold 100 per cent reserves.'7 This radical solution had clear implications for the role of the central bank. First, as a matter of logic, the elimination of discount operations - there would either no longer be any short-term paper to discount or no need to do so. At the same time, as already noted, the criticism was not of the elasticity of the currency, which was to be determined by action of the central bank through the preservation of open market operations - and which could be considered as exogenous. In particular it meant that growth of the money supply would be limited by the growth of government debt. In fact, Mints proposes that the tax rate be set such that, given government spending decisions taken on other (social, political or economic) grounds, the fully monetised increase in government indebtedness should exhibit a constant rate of expansion. Any variations in programmed expenditures or expected receipts would be offset by open market operations directly with the public or financial institutions. In Mints's view, only when monetary policy provides this sort of monetary stability would the forces of 658 Policy Implications of the Current Bank Crisis competition be capable of producing a natural tendency to overall economic stability. Government intervention is thus most important in the monetary field, and extends to active and precise controls over the scope and activity of financial institutions. Mints believed that this would make government intervention less necessary since the natural stabilisers of the private sector would be more efficient, but not exclude it altogether. 6 THE MODERN APPROACH TO MONETARY CREATION It is precisely this discussion of the role of endogenous money creation and economic instability of free market capitalism which has been lost in modem discussions of monetary policy which simply presume the stability of the private sector in competitive conditions, independently of its monetary framework; monetary policy has been reduced to a discussion of a choice of the rate of inflation as determined by the control of the money supply, given the presumption that any government intervention in the economy will distort the efficient operation of the free market system, implying a balanced government budget at the lowest possible level of expenditure. For examples, models of monetarist inspiration presume that the money supply is subject to exogenous determination, and proceeds to its policy recommendations on the basis of the statitical observation, rather than theoretical demonstration, of the stability of the demand for money. The textbook IS-LM approach differs only in its specification of of the demand for money; it also presumes that the money supply may be exogenously fixed and subject to control by the monetary authority. Monetary policy is more problematical in this approach only because of the inclusion of the Keynesian speculative demand in the money demand function. Finally, if one looks at the most evolved forms of this model, presented in terms of aggregate supply and demand, exogenous changes in the price level with a given nominal supply of money replace exogenous changes in the money supply, which is still defined as central bank money plus commercial bank deposits. Via the 'real balance effect', a lower price level leads to an increase in the real money supply and an increase in expenditure. It is not clear what kinds of assets commercial banks are holding in their portfolios, for if they include commercial loans, and business firms facing lower prices have difficulty meeting interest payments, banks may find themselves with non-performing loans and respond by changing their lending practices. This should cause a change in the amount of deposits outstanding and thus in the overall quantity of money, but this contradicts the assumption of a constant quantity of money. IS It would be natural to presume that some of the loans, contracted at higher price levels, could not be repaid as firms face lower output prices and go bankrupt, or that the squeeze on profits should lead to a fall off in investment, or to a reluctance by the Jan A. Kregel 659 banks to continue to lend at the same level. If any of these were to occur the quantity of money could not remain unchanged. It seems that the only way out of this contradictory treatment of the supply of money is to presume that there is only outside or central bank money, that is, there is no bank lending orthat prices of all contracts and goods in the economy change together, or there is only commodity money. This is the equivalent to 100 per cent reserves on commercial banks; modem theory thus seems to reason as if it had already been introduced! 7 THE CIRCUIT CRITICISM Against this modern background, the circuit approach may be considered as going back to the earlier tradition which recognises the direct and important influence of the endogenous process of money creation on the behaviour of the economy. The modem theory of the monetary circuit distinguishes itself from modem mainstream monetary analysis by separating the 'public' from the 'private' sector creation of money, and in doing so it joins up with the modem endogenous money approach which reaches an extreme form in the 'horizontalist' approach of Basil Moore (1988) arguing that the central bank cannot control even the 'public' creation of money. The 'public' creation of money involves the acquisition of government debt by the central bank in exchange for its own note or deposit liabilities, usually called 'base money' or high-powered money. Such a process clearly presupposes the existence not only of a budget deficit, but the existence of a central bank which intermediates between the government and the bank and non-bank public. But, for the circuit approach this is not the most relevant source of money creation: 'Money is created because firms need finance in order to pay wages and buy means of production.' In contrast to the 'typical macro-economic models' which 'assume the whole of the outstanding monetary base as coming from the government deficit, the theory of the 'circuit' assumes an 'equilibrium' position fully consistent with the existence of money created independently of any government deficit, and therefore with the permanent presence of a debt of firms towards banks, as well as of commercial banks towards the Central Bank.' It is on account of the assumption of only 'private' creation of money that the circuit approach 'considers the money stock to be a strictly endogenous variable' (Graziani, 1989: 20) There is no question that this is an important contribution with respect to the treatment of money in modem mainstream analysis. The question remains as to what circuit theory can contribute to the policy debate concerning the proper role of private banks in creating economic stability. As already noted, both the currency and banking schools, as well as most varieties of monetarists, agree that banks should be prevented from creating means of payment, although the reasoning motivating each position differs. The circuit approach's insistence that the endogenous creation of money by banks is crucial to the operation of the 660 Policy Implications of the Current Bank Crisis economy would appear to argue against the elimination of private money creation as a means ofincreasing stability. The position of Hotson and Hixson is, as I understand it, rather different and may be summarised as follows. The US Mt money stock is a little over $800 billion, composed of about $570 billion of demand deposits of commercial banks and $230 billion of notes in circulation. Federal debt held by private investors is a little over $2000 billion. Since notes and deposits are substitutes as means of payment, the public should be expected to continue to hold the existing money stock, irrespective of whether it is composed of 100 per cent Federal Reserve Notes or only one quarter. But, if it were composed of 100 per cent Federal Reserve Notes, the net Federal Debt would be about a quarter lower and current expenditures on interest payments would be about $50 billion lower (about the size of the reduction in the deficit currently .- November 1990 - being debated in Congress).19 The elimination of the outstanding debt would free budget policy from its current impasse in which economic conditions argue for stimulus, yet the absolute size of the debt is used as an argument for budget surpluses to reduce it. 8 THE CIRCUIT WITHOUT FRACTIONAL RESERVE BANKS: MONEY AND CREDIT AND INSTITUTIONAL STRUCTURE This leaves the question of how circuit theory, which considers the 'private' creation of money crucial, would assess the proposal to abolish the private banks which create endogenous money? The question can be approached in two ways. The first is by making a distinction between money and credit. This is a distinction that has bedeviled economists since the bullionist controversy, continued in the banking-currency school conflict, and is still alive in the dispute between monetarists and Keynesians. In formulating his position Mints discusses the distinction between public and private creation of money proposed by the circuit theory in a way which allows comparison with the distinction between "money and credit employed by both circuit theory and modem monetarists: Ideally all private lending operations, or any placing of funds which the possessor does not desire to use in the hands of one who will use them, would be merely a transference of money income and, consequently, of market control over a certain quantity of resources, from one person to another. Monetary policy, on the other hand, is a matter of controlling the total volume of circulating medium in accordance with some reasonable criterion. With a sensible financial structure, these two problems would have nothing in common ... To compound our difficulties, a whole theory of monetary control has been created upon the basis of commercial banks, which ar~ at best unnecessary and at worst a highly aggravating influence in business fluctuations. Bankers are ... creditors on short term. If business conditions are not good, the bankers are therefore in a Jan A. Kregel 661 position to curtail their lending; and this they are impelled to do, both because they fear outright losses on new loans and because of the danger of a demand for cash ... It is unlikely that deposit insurance has eliminated all of this unfortunate characteristic, since it cannot have eliminated other reasons for this perversity than the withdrawals of cash which the public is likely to make when doubts about the condition of the banks prevail. There is no force within the banking system which tends at any time to adjust the volume of bank loans spontaneously to the requirements of monetary stability ... more remarkable than the permission given the banks to develop is the apparent belief among many writers that they are beautifully designed to serve as instruments for the implementation of monetary policy. (Mints, 1950: 5-7; emphasis added) Graziani's (1989: 2) authoritative survey of circuit theory notes that both the 'definition of money and the analysis of the money supply process are basic issues in the theory of the economic circuit' and despite restricting its analysis to the private creation of money, the circuit theory maintains the distinction between money and credit. If 'goods are traded against promises of payment such as bills of exchange, an act of trade gives rise to a debt of the buyer and to a credit of the seller. A similar economy is not a monetary economy but a credit economy. If in a credit economy at the end of the period some agents still owe money to other ones, a final payment is needed, which means that money has not been used ... Money is therefore something different from a regular commodity and something more than a mere promise of payment'. Graziani then lists three conditions for money to exist; that money cannot be a commodity, but only a 'token', 'money has to be accepted as a means of final settlement of the transaction (otherwise it would be credit and not money)' and that no individual agent can issue money himself. He goes to note that 'the only way to satisfy those three conditions is to have payments made by means of promises of a third agent, the typical third agent being nowadays a bank' (ibid.: 3). On the basis of this definition,20 the first question that should be asked is whether the exclusion of public creation of money by government in purely endogenous money circuit models is consistent with a monetary economy, for this would be an economy similar to Wicksell's, a pure 'credit' economy. Such an economy would not be a 'monetary' economy because its means of payment do not satisfy the definition of money given above, or to be precise it would only do so if there were a single bank. In a multiple bank system no bank is obliged to accept the credit of another bank as a debt to a third party: items in process of collection are, under this definition, simply credits granted to a bank's clients until they are paid via the transfer of a 'final means of payment' issued by a central bank. This would seem to imply the paradox that an economy with only 'endogenous' money creation has no 'money', for the 'final means of payment' must always be 'exogenous' to some subsector of the financial system creating credit. 662 Policy Implications of the Current Bank Crisis The definition also seems to imply that 'money' as a final means of payment cannot be a 'liability' of anyone with exchange relations within the system, but must be a pure asset, or that a 'monetary' system, as opposed to a 'credit' system, requires an 'outside' money which, within this definition, cannot be government debt or a commodity such as gold. Money must be token money that does not grant privileges of seigniorage to any agent making a payment,21 This seems to lead back to the conclusion of the early Chicago school that 'money' is issued by a 'third party' the central bank (public liabilities), and serves as the means of settlement for the 'credit' which is issued by commercial banks in the form of deposits subject to check (private liabilities); it was an error to ever attempt to construct a monetary system in which money and credit were considered as similar. Every monetary system then must contain an exogenous element that it is the responsibility of the governmen~ to control. In the circuit approach and the Chicago approach there does not appear to be any difference in the definition of money and credit. Rather, I believe it will be found in the explanation of the behavior of prices by each. Milton Friedman puts it this way: Much of the misunderstanding about the relationship between money and interest rates comes from a failure to keep ... 'credit' distinct from 'quantity of money'. In discussing credit ... the interest rate is the price of credit. General price theory tells us that the price of anything will be lowered by an increase in supply '" Therefore ... an increase in credit will reduce the rate of interest ... The tendency to confuse credit with money leads to the further belief that an increase in the quantity of money will tend to reduce interest rates ... My main thesis is that this is wrong ... because the interest rate is not the price of money. The interest rate is the price of credit. The price level or the inverse of the price level is the price of money. What is to be expected from general price theory is what the quantity theory says, namely, that a rapid increase in the quantity of money means an increase in the price of goods and servic'es, and that a decrease in the quantity of money means a decrease in the price of goods and services. (Friedman, 1969: 362-3)22 In the circuit approach, on the other hand, equilibrium is independent of the price level, and thus of the 'price' of money, as well as independent of the price of credit, the interest rate. For example, Graziani (1985) shows that in a system with endogenous money creation, the rate of interest has no impact on firms' profits or investment and only interest paid by firms to banks creates a charge against profits: 23 the ability of firms to pass on higher interest costs on borrowing in terms of higher prices allows it to carryon investment plans independently of any impact of interest costs. Since this independence of the general price level in the circuit depends on the endogeneity of the 'money' s\Jpply created by the issue of bank credit, it appears Jan A. Kregel 663 to require a single-bank financial system identical to Mints's ideal solution of the central bank as sole, exogenous source of 'money'. If it does not, it will require a third party to mediate the debts and credits created among banks in a multiple bank system: a central bank creating 'money'. This leads naturally to the second way of approaching the problem of whether the system can survive without endogenous money: the institutional structure of the banking system. Circuit theory does not specify the institutional framework in which banks operate in great detail, for example, it appears to presume a single bank rather than a multibank system. But when Parguez (1987) notes, for instance, that in the 'dynamic circuit' banks have a potential for creating instability if they refuse to serve as residual purchasers of firm's long-term liabilities, he may be interpreted within the US context as arguing for the abolition of banking legislation introduced in the US during the New Deal to prevent banks from doing precisely that because it was thought to be destabilising! Indeed, the closest equivalent to the 'single bank' system which Schmitt's version of circuit theory definition of money seems to imply is a 'universal' banking system such as that found in Germany in which a small number of banks use their interest earnings - which Parguez suggests is necessary for stability - to recycle funds to firms as long-term investments. They are able to do this without creating the kind of instability feared by the Glass legislation because deposit drain is minimal in a four-bank multibranch banking system. Indeed, Germany manages to have an extremely stable financial system without a mandatory federal deposit insurance system and in which banks may operate in areas forbidden to US banks by the Banking Act of 1933. Thus an alternative to a 100 per cent reserve system proposed by the early Chicago school as a way of eliminating the need for deposit insurance would seem to be a system with a small number of extremely large, multibranch, multipurpose deposit-investment banks with an extremely strong central bank.24 The level playing field of competition would then imply the elimination of the unit bank system and a massive concentration of banking in the US. 9 IS A 'PUBLIC MONEY' ECONOMY A STABLE ECONOMY? The Chicago proposals would allow the elimination of deposit insurance, either through a 100 per cent reserve requirement or regulation to prevent the private provision of means of payment. At the same time, this would make it possible to suspend the provisions of the Glass Act separating deposit-taking from investment banking. Alternatively, the adoption of a 'universal' rather than a 'unit' bank structure would come close to the same result. 25 The Chicago proposals, however. go further and argue that the increased stability of the financial system, by dampening the process of 'debt deHation' associated with endogenous money creation, would make government stabilization 664 Policy Implications of the Current Bank Crisis policy less necessary. In this regard it should be noted that the essence of the role of banks in the circuit models (and in the Post Keynesian models which preceded them 26), as opposed to the Chicago position, is to allow some sectors to have financial deficits without requiring other sectors to have financial surpluses: factors can be employed in lengthy processes of production before they receive their return and prior saving does not have to precede expenditures on capital goods required for investment. The point applies equally to the prior appropri ation of resources via financing for current production as for financing investment expenditures. Of course, the Chicago proposals, on the other hand, are formulated precisely in order to prevent this from happening; under 100 per cent reserves the banks are forced to act as simple intermediaries between savers and investors, insuring that investment is always limited by available savings. Since all balance sheets in the economy must balance, if investment is to be financed independently of savings such that in the private sector deficit sectors may exist without offsetting surplus sectors, the banking system must play the role of the residual balance sheet. It does not have to be able to create 'endogenous money', as defined in circuit theory, nor does it have to create a liability which is a one for one substitute for Federal Reserve Notes, in order to be able to do this. This is a maxim of financial economics which dates back to William Petty, is the basis of Smith's proposals for control of the note issue and may be found in Kalecki's workY The Bank of Amsterdam functioned at the center of a system in which there was endogenous creation of means of payment, yet it did not itself engage in the issue of 'token' money, it did not lend or issue any means of exchange. 28 As suggested above, the fact that modem economies have banking systems based on a dual system of public and private liabilities results from historical accidents - most of which were financial crises. In a free, competitive market system, as Adam Smith noted, it is impossible to prevent individuals from accepting private liabilities in discharge of debt, or to prevent these liabilities from having values which vary with the conditions of the issuer and general economic conditions. They will produce the same kind of variability as the currency ratio under a fractional reserve system, and have exactly the same effect. As noted above, the real point at issue is the relation between saving and investment. Indeed, this was the point of departure of the cycle theories proposed at the beginning of the decade. Cycles were explained in terms of the excess of investment relative to real saving which was permitted by the fact that fractional reserves allowed banks to convert a given quantity of money, representing household savings, into a larger amount of financing for real investment through the money multiplier process. With prices determined by the balance between savings and investment. bank lending which was not linked to real savings led to an excess demand in commodity markets and price rises. As banks cut back on lending, the opposite occurred and produced cyclical instability. This was the basis of both Fisher and Wicksell's theory and most of those who followed in their footsteps, including the young Keynes. 29 It is in this tradition that, perhaps Jan A. Kregel 665 paradoxically, we find the 'financial fragility' hypothesis of Hyman Minsky. Minsky can refer at one and the same time to 'Chicago'-'Fisherian' heritage and to a Keynesian heritage, because both traditions place the origin of the cyclical behavior of the system in the variability of the prices of debts, although they differ in their theories of the determination of commodity prices. 3o An important difference with respect to circuit theory is that one of the major sources of instability in both the Keynes and the Chicago approaches is in the impact on asset prices of the natural instability of the demand for money and liquidity preferences;31 endogenous money creation only serves to magnify the impact of these variations on asset prices. In the Fisher-Chicago approach it is thus implicitly assumed that public and private money are complements rather than substitutes; the endogenous behavior of credit thus makes it more difficult for the government to satisfy its obligation to produce monetary stability. Keynes's position is the opposite, that money and credit are substitutes to the extent that neither changes in credit nor in money have any necessary or direct effect on asset prices which are determined by liquidity preferences relative to the supply of money. Both approaches thus arrive at the same conclusion - the free competitive market capitalist system can be made to work only if its monetary framework is designed and regulated so as to provide stability of asset prices. It was impossible to place controls on the freedom of individuals to allocate their wealth; policy proposals thus had to remedy excessive volatility in asset prices. For Chicago this meant curbing the elasticity caused by endogenous monetary creation and concerned a 100 per cent reserve requirement and money expansion determined by a price or quantity rule. This simply mirrors the position of those who had early argued in favor of 100 per cent gold backing for the central bank note issue. Implementation of these proposals would institutionalize the representation of the money supply process of traditional theory criticized by the circuit theorists, as well as eliminating the link between private sector financing needs and the money supply upon which the circuit theory bases its analysis. 32 For Keynes, on the other hand, the stability of asset prices required flexibility in the balance sheets of financial institutions which is now known as the 'lender of last resort' function. The elimination of private money creation does not eliminate the need for a balancing agent for financial balance sheets. In contrast to the Chicago position, here it is the open market policy which ceases to have major importance and, as Minsky (1986) has suggested the discount window, extended to a wider range of assets and institutions, will become the most important tool of stability. This also means an increased and more coordinated role for financial institution surveillance. The problem of the multiregulatory agencies in the US will have to be faced as well. But, whatever position one takes, the financial system cannot be left to be controlled by the forces of the market; the process of rational regulation of the system must be part of the agenda of all positions on the political spectrum. 666 Policy Implications of the Current Bank Crisis Notes I. 2. 3. 4. 5. 6. 7. 8. 9. 10. Here, by 'system' one must mean not only the financial institutions, but also the regulatory institutions: the regulatory agencies such as FSLlC, FDIC, the Federal Home Loan Banking Board and so on which have re-emerged in the Office of Thrift Supervision, etc. I have already put it forward in the context of the reconstruction of the former Eastern European command economies (Kregel, 1990b), but at that time was unaware of its intellectual heritage. Cf. Tonveronachi (1982: 183). In the recent Post Keynesian literature the 100 per cent reserve proposal of the early Chicago school is only mentioned briefly by Minsky (1986), p. 9 as being 'still worth considering' and by Rousseas (1986: III), as serving 'the purpose of focusing on an important issue'. More recently, Phillips (1988) deals more extensively with the proposal, noting the similarity between Veblen and Simon's approach, and advocating a reconsideration of the 'Chicago Plan' for current problems. Although I have arrived at it by a rather different route. Since this chapter was written initially as a paper for a conference assessing circuit theory, I started my research by trying to answer the question, what does circuit theory have to tell us about the crisis in the US financial system? This led to the distinction between money and credit and the difference between the Dijon 'circuit' theory and the 'dynamic' circuit (see Parguez, 1986); then to Irving Fisher's debt deflation and to the similar position of the early Chicago school, then back to Adam Smith, and finally a reconsideration of that position of the Chicago proposals which on reflection seem to have been even more prescient than the static circuit approach. Having completed this rather circuitous route I discovered that there were others who had already arrived at this conclusion (and whose arguments I had previously failed to take seriously, in particular John Hotson and W.F. Hixson). This chapter offers apologies to them. It has also been defined as 'seigniorage'. The example is purposely simple; it would be more sensible to finance investment expenditures which would then require a Domar condition concerning the relation between the rate of increase of expansion of income and the rate of interest, but this is not the main point which we want to emphasize here. The Bank quickly showed itself to be highly inept in its dealing with the private sector and had to suspend cash payment on its notes for the first time (there were others) within three years of its creation! The same response occurred in the United States when the attempt to sell bonds directly to the public to finance the Civil War failed; the National Banking Act of 1864 created National Banks which acquired the right to issue national bank notes on condition that they purchase government debt that the public had refused. The Act also placed a tax on the notes issued by State chartered banks, which quickly created the use of deposits subject to check by these banks. The Office of the Comptroller als9 resulted from this legislation. It is as if the Coca-Cola company, without consideration, granted Pepsi (and every other soft drink bottler) the exclusive rights to the Coke formula, told all its clients that Pepsi was just as good as coke, and then proceeded to buy Pepsi to sell under the Coke label, paying Pepsi not only cost of production but profit on the Coke patent to do so! Smith was also perhaps the first theorist of financial innovation. Since he considered money as part of circulating capital, he considered anything which reduced the circulating capital requirements of production to be technical progress which would Jan A. Kregel II. 12. 13. 14. 15. 16. 17. 18. 667 reduce costs and prices. Banknotes and bank credits which replaced metals a the means of circulation were thus technical innovations which reduced costs and were beneficial to the economy a whole. However, he also insisted that with 100 per cent backing for notes, prices would be the same as in their absence. This is a common approach with all classical economics and in direct contrast to the modern free market position. This distinction was first noted by Lionel Robbins (1952) and developed by Warren Samuels (1966). Cf. Kregel (l990a). And he goes on to recommend free entry into banking and a multiplication of banks which reads like a justification for the US no-branching, unit bank system: A 'multiplication of banking companies ... obliges all of them to be more circumspect in their conduct ... restrains the circulation of each particular company within a narrower circle, and reduces their circulating notes to a smaller number. By dividing the whole circulation into a greater number of parts, the failure of anyone company ... becomes of less consequence to the publick. This free competition too obliges all bankers to be more liberal in their dealings with their customers, lest their rivals should carry them away. In general, if any branch of trade ... be advantageous to the publick, the freer and more general the competition, it will always be the more so' (Smith, 1976, p. 329). The Act was meant to prevent financial crisis by regulating the note issue via 100 per cent gold backing of Bank of England notes above the base limit then issued against government debt of £14 million. By 1847 crisis had none the less broken out and the Act had to be suspended. Walter Bagehot's first published essay, on the appropriate response to the crisis, recommends abolishing the right of private banks to issue notes and granting 'a monopoly of the note issue in the hands of the state'. Cf. Buchan (1959: 45). This was considered to be of even greater importance than the elimination of government interference from economic activity by reducing the size of government. Cf. Mints (1950: 4). As Tonveronachi stresses, this position should not be taken out of context. Simons considered monopoly power - of both big business and big labor - to be the major impediment, with monetary factors exacerbating the difficulties caused by large size. The same views may be found in Mints and in a less pronounced form in Douglas. All were united against the policies of the NRA, for example, which they believed institutionalized such power and made it collusive. Note that this is not the same as criticism of elasticity of the reserve base practised by the central bank which was to have been remedied by the creation of the Federal Reserve System. 'The proposals with reference to banking contemplate displacement of existing deposit banks by at least two distinct types of institutions. First ... deposit banks ... maintaining 100 per cent reserves ... a good case could be made for extending the facilities of the postal savings system for the provision of something like checking accounts ... A second type of institution ... in the form of an investment trust, would perform the lending functions of existing banks' (Simons, 1934: 64). As Keynes pointed out, the assumption that the quantity of money is fixed is equivalent to the assumption that the level of demand is fixed, which means assuming what one is trying to prove. Keynes made the argument with respect to a reduction in the wage, but it applies equally to a reduction in the price level: 'In its crudest form, this is tantamount to assuming that the reduction in money-wages will leave demand unaffected. There may be some economists who would maintain that there is no reason why demand should be affected, arguing that aggregate demand depends on the quantity of money multiplied by the income-velocity of money and that there is no obvious reason why a reduction in money-wages would reduce either' (Keynes, 1936: 258). 668 19. 20. 21. 22. 23. 24. 25. 26. Policy implications of the Current Bank Crisis This is a static estimate. Had such a change been introduced say in 1980 instead of D1DMCA and Garn-St Germain in 1982, the size of the current deficit would have to be reduced by both the accumulated reduction and the reduced costs of bailing out the S&Ls and the commercial banks. Which Graziani bases on the work of Bernard Schmitt. But it was precisely this seigniorage that allowed the Bank of England to be created to resolve King William's poor credit rating. In the light of Mints's statement above, it seems plausible to interpret Friedman's statement as referring to an ideal condition in which commercial banks do not lend through deposit creation. 'Money' would then mean central bank money, while 'credit' would refer to financial market intermediation of saving and investment. The only difficulty is that in the existing financial system, pan of the transactions which properly belong to the financial market for loans are in fact entrusted to banks who use them to expand deposits which are created by means of lending based on fractional reserve deposit creation. In the real world this sharp distinction between money and prices and loanable funds and interest rates can only be maintained if the public's demand for currency relative to deposits is fixed or a predictable function of economic variables which are under the control of the monetary authority. In this way, Friedman is able to make the same proposals for monetary stability without having to take Mints's extreme position (which was also Hayek's) concerning the elimination of the commercial banking system. In this sense, all of Friedman's statistical sophistry is due to his unwillingness to state clearly that the application of monetary stability and the efficient operation of the free market system requires elimination of the free market as it applies to money and elimination of private deposit taking banks. The thrust of the early Chicago proposals is thus to separate the commercial banks from the determination of the price of money, while allowing them to participate in the formation of the price of credit by engaging in medium and long-term borrowing and lending activities. The difference between Friedman and the early Chicago position is that the latter did not accept the stability of the demand function for money and thus called for the elimination of the banks. This is what Parguez (1987: 5) calls 'Ie probl~me du profit', which is created when firms finance production by borrowing but can receive in receipts only what they have paid out, leaving them able to repay lending but not the interest on the loans. Parguez points out that by rejecting the 'classical' model and viewing the circuit in dynamic terms as encompassing the financing of fixed investment, then if the existence of net profit can be explained one can also explain how firms pay interest to banks. It is perhaps relevant to recall that the large German universal banks all started as private investment banks and only got into retail deposit-taking by rescuing failing deposit banks. Implying an increase in concentration in banking that would have been vigorously resisted by Simons. The circuit theory was initially viewed as being an elaboration of Keynes's ideas in the Treatise, but has subsequently also become critical of Post Keynesian theory. Graziani's analysis is meant to 'dispel a recurring mistake in the [Post Keynesian] literature, according to which finance supplied by banks is confused with the financing of fixed investment' (1989: 21), that is, the undue emphasis on the 'finance motive' to the exclusion of the necessity to finance the recurrent payments of wages and other costs of production, Parguez (1986), however, takes a diametricany opposed position and criticizes the extensive reliance of Post Keynesians on what he calls the 'wage postulate' which places emphasis only on the recurrent. Jan A. Kregel 27. 28. 29. 30. 31. 32. 669 financing of wage costs of current production to the exclusion of the financing of purchases of finished investment goods. I am not sure that these two criticisms are consistent. Roncaglia (1985: 30) quotes Petty: 'For if men were excellently Versed in accompts Money were not necessary at all.' According to Kalecki (1971: 13), 'Thus capitalists, as a whole, determine their own profits by the extent of their investment and personal consumption ... capitalists as a whole do not need money in order to achieve this.' The role of residual balance sheet is presumably what Graziani is trying to express with the idea of the 'promises of a third agent'. I disagree, however, with his extension of this proposition to what he calls the 'fundamental rule' of a monetary economy, 'that no agent may make a final payment by means of a simple promise of payment' because this is precisely how most merchants became merchant banks, long before clearing banks developed in England. I would rather use as the fundamental rule the idea that in a monetary economy anyone who wants to sell has to be willing to grant credit to the buyer: all real exchange creates debts which can only be validated if someone is willing to carry them over time. And the young Hayek; cf. his analysis of the US crisis of 1920 in Hayek (1984). For the Fisher-Chicago approach asset prices vary from equilibrium because of the amplifying effect of endogenous money creation by commercial banks resulting from changes in the public's asset preferences, while for Keynes assets prices are a direct reflection of variations in the liquidity preferences of the banks and the public. Commodity prices are determined directly by the expansion of the supply of money relative to the supply of goods for the former and by efficiency wage levels and the pressure of investment on resources for the latter. The relation between Minsky and the Chicago tradition is noted by Whalen (1988) and extended to Veblen in Phillips (1989). Minsky's theory makes this inherent instability endogenously determined. The circuit theory thus appears to go against both the banking and currency schools when it argues that monetary analysis should emphasize endogenous money creation, because even the 'public' creation by the central bank cannot be considered exogenous, yet it accepts the banking school proposition that private creation of money has no effect on the price level and thus on the conditions of equilibrium. To put it simply, it would seem to reject the linkage between the variability of asset prices and endogenous money creation which is the basis of the two major extensions of Fisher's debt deflation hypothesis: the early Chicago school and Minsky's financial fragility hypothesis. As I have argued elsewhere (Kregel, 1986), it is also inconsistent with Keynes's version of this linkage which argues that prices may change as a result of changes in liquidity preferences independently of any change in the quantity of money or its velocity of circulation. References Buchan, A. (1959), The Spare Chancellor: The Life of Waller Bagehol (London: Chatto and Windus). Douglas, P. (1935), Controlling Depressions (London: Unwin). Friedman, M. (1969), 'Factors Affecting the Level of Interest Rates'; repro in T. Havrilesky and J. Boorman (eds.), Current Issues in Monetary Theory and Policy (Arlington Heights, IL: AHM Publishing Co., 1976). 670 Policy Implications a/the Current Bank Crisis Graziani, A. (1985), 'Monnaie, interet et depenses publiques', Economies et Societes, Serie Monnaie et Production, no. 2. Graziani, A. (1989), 'The Theory of the Monetary Circuit', Thames Papers in Political Economy, Spring. Hayek, F. (1984), 'The Monetary Policy of the United States after the 1920 Crisis', in Money, Capital & Fluctuations: Early Essays (London: Routledge & Kegan Paul) Hixson, W.F. (1987), 'Marxism and Monetary Policy', Economies et Societes, Serie Monnaie et Production, no. 4. Hotson, J .H. (1987), 'The Keynesian Revolution and the Aborted Fisher-Simons Revolution', Economies et Soc;etes, Serie Monnaie et Production, no. 4. Hume, D. (1752), Writings on Economics, Eugene Rotwein (ed.) (London: Nelson, 1955). Kalecki, M. (1971), Selected Essays on the Dynamics of the Capitalist Economy (Cambridge: Cambridge University Press). Keynes, J.M. (1936), The General Theory of Employment. Interest and Money (London: Macmillan). Kregel, J.A. (1986), 'Hamlet and Shylock, or are there Bulls and Bears in the Circuit?', Economies et Societes, Serie, Monnaie et Production, no. 3. Kregel, J.A. (1990a), 'Market Design and Competition as Constraint to Self-interested Behaviour', in K. Groenveld, J.A.H. Maks and J. Muysken (eds), Economic Policy of the Market Process: Austrian and Mainstream Economics (Amsterdam: North Holland), 45-57. Kregel, J.A. (\990b), 'Plan, Market and Banking', in K. Dopfer and K.-F. Raible (eds), The Evolution of Systems (London: Macmillan), 133-40. Kregel, J.A. (1991), 'Is Keynesian Full Employment Policy for "Top Countries" Only?', in C. de Neubourg, The Art of Full Employment (Amsterdam: North Holland). Lavoie, M. (1987), 'Monnaie et production: une synth~se de la theorie du circuit', Economies et Socieres, Serie Monnaie et Production, no 4. Minsky, H.P. (1986), Stabilizing an Unstable Economy (New Haven, CO: Yale University Press). Mints, L.W. (1950), Monetary Policy for a Competitive Society (New York: McGrawHill). Moore, B. (1988), Horizontalists and Verticalists (Cambridge: Cambridge University Press). Parguez, A. (1986), 'Au coeur du circuit, quelques reponses', Economies et Societes, Serie Monnaie et Production, no. 3. Parguez, A. (1987), 'Avant-Propos' to 'Monnaie, Les Rentiers et la Crise', Economies et Societes, Serie Monnaie et Production, no. 4. Phillips, R.1. (1988), 'Veblen and Simons on Credit and Monetary Reform', Southern Economic Journal, July. Phillips, R.J. (\989), 'The Minsky-Simons-Veblen Connection: Comment', Journal of Economic Issues, 23. Robbins, L. (1952), The Theory of Economic Policy (London: Macmillan). Roncaglia, A. (1985), Petty: The Origins of Political Economy (Armonk, NY: M.E. Sharpe). Rousseas, S. (1986) 'The Finance Motive, Keynes, and Post-Keynesian', Economies et Societes, 20 (8-9) August-September, 189-20. Samuels, W. (1966), The Classical Theory of Economic Policy (Cleveland: World Publishing). Simons, H.C. (1934), 'A Positive Program for Laissez Faire', in Economic Policy for a Free Society (Chicago: Chicago University Press, 1948). Smith, A. (1976), An Inquiry into the Nature and Causes of the Wealth of Nations (Glasgow Edition: Oxford University Press). Jan A. Kregel 671 Spalding, W.F. (1933), The London Money Market (London: Pitman). Tonveronachi, M. (1982), 'Monetarism and Fixed Rules in H.C. Simons', Banca Nazionale del Lavoro, Quartely Review, 141. Whalen, C. (1988), 'The Minsky-Simons Connection: A Neglected Thread in the History of Economic Thought', Journal of Economic Issues, 22. 26 The Spheres of Industrial and Financial Circulation Revisited, and their Implications for Post Keynesian Economic Policy Stephen Rousseas I Economic theories are historically bound products of their time. Their purpose is to explain the phenomenological world around them as they see it. Different theories, however, entail different weltanschauungs and thus the 'realities' they see are different - as are the policy prescriptions that follow from them. Keynes's General Theory is a case in point. Keynes was not a revolutionary. He did not threaten to shake the system down to its foundations in some Messianic view of the future. His goal was to reform it and make it work at a minimum social cost. The reason the General Theory succeeded over other theories of its time was that it provided a more meaningful way of looking at and dealing with the problems of the 1930s by emphasizing those policy control variables that were politically feasible. But even this limited approach proved too much for many American economists who were quick to bowdlerize and transform the General Theory into a parody of itself and, in an act of reverse English, export it back to Britain in a new mechanical version which was then used to indoctrinate generations of postwar economics students on both sides of the Atlantic. Either Say's Law was artificially restored via fine-tuning Is-LM economists, or the notion of a natural unemployment rate was used to banish involuntary unemployment, once more, in an unabashed return to pre-Keynesian economics in retailored Keynesian attire. Despite the American taming and mechanization of the General Theory in the postwar period, the fact remains that the General Theory, compared to Keynes's earlier Treatise on Money, was a step back into a Marshallian world of equilibrium analysis. This is best seen in the Keynesian formulation of liquidity preference theory using a simplified money-bond formulation that was ultimately transformed into a more complex asset-portfolio model that came to rest solidly 672 Stephen Rousseas 673 on utility analysis. The two major examples are James Tobin's theory of risk aversion in a world of interest rate uncertainty, and Milton Friedman's reformulation of the Quantity Theory of money in the form of a stable velocity function. The use of asset-portfolio theory to derive the demand function for money and the obsession, by both 'bastard' Keynesian and Friedmanian monetarists, with the utility functions of maximizing individuals, or 'people', led Nicholas Kaldor to ask: But who are 'the people' ... ? Are they the wage-and salary-earners, who, between them, account for 70 percent of the national income, but hold, at any one time, a much less proportion, perhaps 10 to 20 per cent ... of the total money supply? Or are they the 'rentiers', whose 'portfolio selection' and 'portfolio shifts' are much influenced at any time by short-term expectations, as well as by the relative yields of various types of financial assets? Or are they businesses, for which holding money is just one of a number of ways of securing liquidity unexploited borrowing power, unused overdraft limits and so on being other ways - and for which the state of liquidity is only one of a number of factors that influence current expenditure decision?· The failure to question Keynes's liquidity preference theory and its reformulation ultimately into an asset-portfolio model missed the opportunity to emphasize that the demand for money is for financial credit primarily, under normal conditions, by the business sector and not, as overwhelmingly represented, for assets by utility-maximizing 'people'. Linking the demand for money to the financing of investment and the problem of capital accumulation would have yielded a far more promising prospect for the development of monetary theory - and would still do so. II For Keynes the supply of money was exogenously determined by the central bank, the attempt by some American Post Keynesians to prove the opposite notwithstanding. 2 The velocity of the exogenous money supply, given Keynes's flat rejection of Classical dichotomy, was seen to be highly variable and a function of the rate of interest. Within this theory, the volatility of velocity is explained in terms of the activation of idle balances and the economizing of transaction balances at progressively higher interest rates. Implicit in this argument is the stability of the underlying financial structure in the context of Keynes's preoccupation with the short run. Changes in velocity were therefore to be understood as movements along a given V(i)-curve in response to the central bank's refusal to accommodate fully increases in the demand for money by requisite adjustments in the exogenous money supply. In 674 Spheres of Industrial and Financial Circulation this context, the base period money supply multiplied by the percentage change in velocity yields the velocity equivalent of an increase in the absolute money supply. At some point, when all idle balances are exhausted and transactions balances pared to the bone, velocity would be at its maximum value and monetary policy would effectively hold nominal GNP at its corresponding maximum. Monetary policy would then be completely effective (or so it was thought), assuming: (i) a total absence of financial innovation, i.e. no change in the underlying financial structure of the economy, and (ii) a totally exogenous money supply The liquidity preference approach of the General Theory, however, is a collapsed and far less rich analysis of the demand for money than can be found in the Treatise of 1930. There, Keynes's concern with velocity was far more detailed and complex, and the demand for money, unlike its later treatment in the General Theory, was sorted out and assigned, for analytical purposes, to the two clear and distinct Spheres of Industrialllnd Financial Circulation. The recent turmoil in American financial markets, and the marked increase in the instability of the financial system as a whole, cannot be adequately understood exclusively in terms of the General Theory. A more insightful explanation of what is going on, and what it portends for the stability of the American economy, can be made using the ideas and formulations of the Treatise. The unambiguous separation of the industrial and financial sectors of the economy was for the purpose of better understanding the effects of changes in one sphere on the other. Modern economies cannot be dealt with exclusively in terms of the real sector, as is the case with growth theorists, neo-Ricardian followers of Sraffa, neoMarxists, and micro theorists in general. What we learn from the Treatise is that the financial structure must be brought explicitly into the analysis if the economy is to be properly seen as a functioning credit-money system. The interrelatedness and interaction of the two spheres are critical for coming to grips with the ongoing problems of a capitalist economy working under conditions of uncertainty. Although the causal arrow is bi-directional between the two spheres, more emphasis should be. placed on the negative impact of changes in the financial sector on the industrial (or real) sector of the economy, especially in view of the almost unbridled speculative binge of the 1980s in the United States. It is in this connection that Keynes's analysis in the Treatise takes on a special relevance. Two basic definitions were set out in the Treatise: By Industry we mean the business of maintaining the normal process of current output, distribution and exchange and paying the factors of production their incomes for the various duties they perform from the first beginning of production to the final satisfaction of the consumer. By Finance ... we mean the business of holding and exchanging titles to wealth ... including Stock Exchange and Money Market transactions, speculation and the process of conveying current savings and profits into the hands of entrepreneurs. 3 Stephen Rousseas 675 If we were to reverse and undo Keynes's movement from the analysis of the demand for money in the Treatise to the Lt- and L2 compressions of the General Theory, we would begin by distinguishing between the different kinds of bank deposits relevant to what Keynes called the spheres of Industrial and Financial Circulation. Transactions, or what Keynes in the Treatise described as 'cash deposits', are held by firms and households and are linked to the general level of economic activity in the industrial sphere. Business cash deposits, and hence their velocity, are more volatile than household cash deposits since the incoming cash flows of firms are more uncertain than those of households. Financial deposits, in contradistinction, consist of those business deposits used in the trading of financial instruments. These financial business deposits, however, are not dependent on the real level of economic activity, and their turnover rate, though large, is not highly variable because of existing clearing arrangements in the financial sphere. It is the savings deposits of 'bears' and 'bulls' operating over the entire spectrum of financial circulation that are the most volatile of all deposits. Given the highly variable circulation velocity of financial savings depostis, any attempt by the central bank to restrain the financial sector by decreasing the flow of money to this sector is bound to be of limited effectiveness. The financial velocity of money would (largely through financial innovations) rise as an offset to central bank policy, far more easily than would be the case in the Industrial Sphere. It should be obvious that uncertainty plays a far larger and more destabilizing role in the sphere of financial circulation. The demand for financial deposits is therefore highly unstable. The demand for cash deposits as a means of payment in the sphere of industrial circulation, on the other hand, is relatively more stable. The ratio of cash to financial deposits, as a result, is not only highly unstable. it is also erratic and unpredictable. The total demand for money (cash plus financial) is therefore not linked in any stable way to the level of nominal GNP nor, for that maUer, are cash deposits alone, given the unexpected and unpredictable divergences between expected and actual earnings of firms in the sphere of industrial circulation. In no way is a strict monetarist or 'bastard' Keynesian approach admissible within the framework of the Treatise, any more than they would be in Keynes's earlier 1923 Tract on Monetary Reform. III More important. for current purposes in the US and for Europe after the implementation of the Maastricht Treaty. is the potential conflict between the Industrial and Financial demands for money. Keynes was concerned that in a speculative boom, the financial sector could end up 'stealing resources from the industrial sector'4 with untoward effects on real output and employment, especially in the 676 Spheres of Industrial and Financial Circulation face of an uncooperative and non-accommodating central bank. Interest rates, in that case, would either be maintained on a high level or forced up leading to 'an immediately deflationary tendency'.5 Should the central bank reverse its position and increase the money supply, a runaway bull market would nevertheless cause the demand forjinanciai deposits to continue to soar, and there would be little the central bank could do to counteract it without raising interest rates - with the same negative results in the sphere of industrial circulation. Indeed, in the 1920s the Federal Reserve System went so far as to refuse to take any steps whatsoever to curb runaway land and stock speculation on the grounds that its mandate was restricted to operating exclusively in the sphere of industrial circulation. It had, in its view, no business meddling in the financial sector of the economy to curb excessive speculation. The rest of the story is well known. It would appear that we do not learn from history and are condemned to repeat the blunders of the past. The current debt-equity ratio in the United States has soared in the non-financial corporate sector at the same time as the quality of that debt has been badly eroded. The principal causes for this startling leveraging strategy (in which financial institutions have played such a prominent role at the expense of investment in real capital) can be attributed to a combination of new developments: (i) a changed attitude towards debt; (ii) the globalization of financial markets; (iii) the wave of mergers, leveraged buy-outs, and the massive issuance of high-Hying junk bonds in the 1980s; (iv) the mad dash to financial deregulation via the Depository Institutions Deregulation and Monetary Control Act (DIDMCA) of 1980, and the Garn-St Germain Act of 1982; (v) the phenomenal increase in financial innovations; and (vi) the transformation of non-marketable debt into marketable debt, or what has been called the 'securitization' of debt. With bank consortia having provided corporate raiders with the credit to buy stock at highly inflated prices from stockholders greedy for capital gains, and the loading up of banks, other financial institutions and pension funds with junk bonds and highly speculative real estate and energy loans, the makings of an economic disaster were well at hand - the collapse of the S&Ls being a case in point at a public cost of hundreds of billions of dollars. Here, the relation of Keynes's spheres of Industrial and Financial Circulation becomes critical. In a serious recession, the shortfall of income and the fall of asset prices could well compound the downturn into a mutually self-reinforcing inward spiral of the Financial and Industrial Sectors. Without sustained economic growth to generate the cash flows needed to service the bloated amount of private debt, developments in the financial sphere could send the economy into the violent contraction of a debt deflation. Defaults, bankruptcies and massive layoffs in both the industrial and financial spheres would then correct the debt imbalances at great cost in real output and employmel}t. Keynes's concern in the Treatise that the financial sector, under certain circumstances, could wind up 'stealing resources from the industrial sector' has Stephen Rousseas 677 been more than amply realized in the 1980s. The 'crowding out' of the industrial sphere is reflected in the failure, over the past twelve years from 1981 to 1993, of the increase in real assets to match the enormous increase in debt. In this context, monetary policy is no longer effective. It is no longer a matter of moving along a V(i)-curve toward some maximum velocity and GNP value, but the constant shifting of the Keynesian V(;)-curve is the principal problem. 6 In other words, the underlying financial structure of the economy is undergoing a series of violent and destabilizing changes that make a mockery of traditional Keynesian monetary and fiscal policy tools, not to mention the even greater absurdities of the monetarist and rational expectations schools. IV Out of the rubble of 'bastard' neo-Classical and monetarist Keynesians (Friedman is, in essence, a waynard Keynesian) there has emerged the American School of Post Keynesian economics, and in France the related 'circuit theory' approach. American Post Keynesians, following the lead of Michal Kalecki, reject the notion of 'equilibrium' and refuse to admit it into their analysis; nor do they allow a substantive distinction between the short run and the long run. There are, furthermore, no 'central tendencies' or 'centers of gravity, as there are in the Halo-Cantabrigian school and thus no 'law' to be discovered. The major emphasis is placed on the short run and its attendant problems. It leans less on the mysteries of growth theory and its steady state properties and more on Keynes as a monetary theorist. The Treatise on Money and the General Theory playa far more important role in their thinking. In a sense, they could be called neo-monetarists, with Keynes's concept of uncertainty the linch-pin of their system. American Post Keynesian economics is the exact reverse of 'bastard' and neoClassical Keynesianism: money wages are exogenously determined, the stock of money is passively endogenous, and prices are determined by the distribution of the social product (not the other way around as in neo-Classical marginal productivity theory). Labor, moreover, is not a function of real wages. Indeed, trade unions cannot and do not bargain for real wages; collective bargaining is over the level of money wages, with past rates of inflation a part of the bargaining process. With money wages the critical cost component of prices, prices are determined by a mark-up process where the degree of monopoly and the productivity of labor are the critical factors in a largely oligopolistic economy; that is, with the degree of monopoly and the productivity of labor both relatively stable over time, it follows that the general price level is mostly a function of the money-wage rate and that the level of the real wage is determined by the power of oligopolistis to set commodity prices. Of equal importance to this 'wage theorem' is the notion that the money supply plays, essentially, an accommodating role. The 'banking principle' applies and 678 Spheres of Industrial and Financial Circulation banks, like any other business, have customers and if customer demand for loan· able funds increases, so will the supply - willy-nilly. The causal arrow, therefore, is seen to run from the asset to the liabilities side of bank balance sheets, not from the liabilities to the asset side as conventional banking theory would have it. Central banks are seen, if they are to avoid imploding the economy, as having no choice but to ratify and validate the loans of banks by providing, ex post facto, whatever reserves are needed. Gunnar Mydral, long ago in the I920s, argued along these lines in his pioneering work, Monetary Equilibrium. Given all this and the current impasse, where do the American Post Keynesians come out on the issue of policy (the French circuit theorists will have to speak for themselves)? Although Post Keynesian economics has much to contribute to the analysis of current economic problems, it has, on the policy front, problems of its own. The general Post Keynesian response ilas been that the federal government will not allow a collapse to happen; that it will undertake whatever steps necessary to prevent it and that, furthermore, it will be able to do so in time and that such steps will prove to be effective. This combination of pessimism about the current state of the economy and an apparently unbridled optimism concerning the economic power of the state are reminiscent of Hyman Minsky's 'financial instability hypothesis'. Minsky has spent a professional lifetime denying that what he expects to happen will happen. 'It', according to Minsky can never happen because of the stabilizing presence of the government, and an ever accommodating central bank faithfully playing its role of lender of last resort. This is 'fine tuning' brimming with a sparkling and effervescent optimism. One thing that comes out of this Post Keynedian position is a belief that the central bank will have no other choice but to undertake a fully accommodating monetary policy, should the financial fragility of the system tip over into a serious economic recession. It is another way of arriving at the basic tenet of the American Post Keynesian school that the 'money' supply is endogenous. At any rate, the industrial sphere has been pushed into the background and the financial tail now seems to be wagging the economic dog. The fact is that recent developments in the now dominant financial sphere bear with them the continuing potential for a disaster that conventional economic policy tools may not be able to prevent. In a situation where the financial sphere is crowding out and 'stealing' resources from the industrial sphere, what are the duties of a central bank? In the Treatise, the central bank is faced with a dilemma. Either it increases the quantity of money in the financial sphere, thus further encouraging the speculative binge, or it reduces the money supply thus raising the rate of interest and having a negative effect on real investment in the industrial sphere. There is a way out, however, and that is to 'discriminate in the te~ms of lending (either in the rate charged or by rationing the amount lent) between financial and industrial borrowers'.8 This implies the use of selective controls to influence the flow of credit Stephen Rousseas 679 (a much broader term than the 'money' supply that includes, in addition to banks, other financial institutions such as insurance companies and pension funds) between, as well as within, the two spheres of industrial and financial circulation. This is not something that Post Keynesian have taken up and that flies in the face of their full-endogeneity argument. They are content to fall back on the central bank's role of lender of last resort and the compensatory, stabilizing policies of the federal government. v In all his major writings, Keynes was deeply troubled by the underlying tendency in capitalism, exacerbated by the separation of ownership and control, for the speculative function to overwhelm and replace the enterprise function or, in other words, for the Financial Sphere to be caught in the throes of a speculative bubble that pushes the Industrial Sphere to the side, with output and employment the resulting casualties. This concern was most directly and eloquently expressed in the General Theory: Speculators may do no harm as bubbles on a steady stream of enterprise. But the position is serious when enterprise becomes the bubble on a whirlpool of speculation. When the capital development of a country becomes a by-product of the activities of a casino, the job is likely to be iII~done. The measure of success attained by Wall Street, regarded as an institution to direct new investment into the most profitable channels in terms of future yield, cannot be claimed as one of the outstanding triumphs of laissez-faire capitalism.9 Matters have broadened since the time of Keynes. It is not just Wall Street; it is the whole financial structure, with banks at the core of things, that has become involved in the highly destabilizing speculative activities of the 1980s - especially with the orgy of deregulation under the Reagan-Bush administrations and the apotheosis of 'greed' that was its end result. It was the time of Ivan Boesky who, before his fall in the insider trading scandals of 1986, gave the commencement address at the University of California's Graduate School of Business Administration. Without the slightest trace of embarrassment or whimsy, he told the young men chafing at the bit that 'Greed was good'. And from what happened subsequently in the investment banks and brokerage houses of New York and Chicago, it seems they took him at his word. It was also the time when the students at Harvard University'S Graduate School of Business Administration waved dollar bills over their heads on commencement day in anticipation of the millions to be made in a society where an unbridled free-wheeling mentalite was based on the swinging Cole Porter style of 'Anything Goes'. As under Louis Phillipe's Guizot, the message was clear: enrichissez vous. 680 Spheres of Industrial and Financial Circulation Nor did the 'bottom line' preoccupation with short-run profits, that is so rampant today, escape Keynes's attention in 1936. '[H]uman nature', he wrote 'desires quick results, there is a peculiar zest in making money quickly, and remoter gains are discounted by the average man at a very high rate.' to Unfortunately, in these highly unstable and troubled times, the only policy tool that American Post Keynesian economics came out with was TIP, or the wellknown tax-based incomes policy of Sydney Weintraub and Henry Wallich,u With both big business and big labor opposing it in a system of PAC-financing of politicians, it was predestined to failure. As it is, with the union-busting activities of the Reagan and Thatcher administrations, money wages recently have not been an inflationary factor. Yet without an incomes policy in place, at least on a standby basis, conventional Keynesian monetary and fiscal policies cannot work. But even if, somehow, an effective incomes policy could be put in place to resolve the long-standing problem of price-level stability, this in no way precludes a resource- stealing speculative binge in the Financial Sphere of circulation. The monetary side must be brought into the analysis and somehow tamed. Perhaps we should go even further back into Keynes, into his Tract on Monetary Reform (1923), where his major concern was with offsetting and controlling changes in the velocity of money. The problem of controlling the flow of credit along predetermined paths is something Post Keynesian monetary theorists have largely ignored. Incomes policy and selective credit controls combined are a prerequisite for the effective operation of Keynesian monetary and fiscal policies, if full employment is to be realized in the short run and the economy maintained, over time, on its long-term growth path. Post Keynesian economics must, in other words, move in the direction of economic planning within a capitalist system that entails a social contract for the distribution of income - to be worked out jointly by capital, labor and government. It must also require, as Keynes urged, the 'socialization of investment'. Keynesianism 'without tears' does not exist. It is the product of a bowdlerized neo-Classical version of Keynes. The need for the willful coordination of economic policy on all levels entails the realization that the allocation of resources through an impersonal market system (with minimum compensatory government action) is nothing more than the fiction of textbooks. Yet as much as one may sympathize with American Post Keynesian theory, it seems to be going nowhere. It violates the basic methodology of Keynes, himself. Keynes, confronted with a problem, always looked for those policy variables that were most amenable to political control. This is what American Post Keynesian economics has not been able to do. It either lapses into a blind faith in the ability of government to maintain the system on an even keel through the use of conventional policy tools (in which case it is not much different from the 'bastard' Keynesianism of Paul Samuelson and James Tobin), or it comes up with policy tools (TIP or selective credit controls) that are dead in the water. In the latter case its policy variables are not subject to political control in the capitalism that we Stephen Rousseas 681 know. Nor does it seem likely that they ever will be. From a policy point of view, American Post Keynesian economics does not have much to say that the real power-holders are prepared to take seriously. To return to the opening gambit of this chapter, the purpose of theory is not to predict but to explain. That is where the true strength of American Post Keynesian economics lies. But in explaining the system of American capitalism, it demonstrates, at the same time, that the system is perhaps beyond reform. That is the basic conundrum. This can be seen in a situation in which the economy is headed into a serious recession, the government raises taxes (because of a debt crisis) and the central bank, obsessed with a fear of inflation, responds by raising interest rates, or not lowering them far enough. And over the twelve years of the Reagan-Bush administrations (and one might add in the Thatcher-Major governments of Britain) the major policy objective has been to undo fifty years of social progress by skewing the distribution of income in favor of capital in the supply-side belief that it is the rich who are the cutting edge of a dynamic capitalism, a program that has been extended with a vengence by the 'Contract with America' of Newt Gingrich. As Charles de Gaulle once said, 'Reforme oui, Chienlit non': American Post Keynesian economics clearly demonstrates that we cannot get the reform. If that is so, then we will have to settle for the 'chien lit' . VI It remains to add that the arguments of this chapter have a special relevance to the future of the European Community under the Maastricht Treaty. In January of 1994 the second phase of the Treaty on European Unity was promulgated. The transitory European Monetary Institute (EMI) of Phase 2 will give way to the European System of Central Banks (ESCB) and the European Central Bank (ECB). The business of the ECB will be to establish 'an open market economy with free competition, favoring an efficient allocation of resources.' And this is to be done with a minimum of interference from outside political forces. The primary objective of the ECB will be 'to maintain price stability' and no national central bank or any community government shall seek in any way to impose its will on the ECB. Complete political independence of the community central bank is to be guaranteed and respected under the Maastricht Treaty. This move towards a community-wide monetarism repudiates the notion that in a democratic society, when responsibility for something increases, responsibility to it must also increase. The target of the ECB is to be the community-wide price level or, as others have suggested, the community-wide level of nominal gross domestic product. In either case what is being followed is a form of Volckerian 'pragmatic' monetarism. Fiscal policy is not to be coordinated on a community-wide basis; that 682 Spheres of Industrial and Financial Circulation will be left to the discretion of the member states. Still, national fiscal policies will be constrained by two considerations: (i) the ratio of the planned or actual government deficit to gross domestic product which is not to exceed 3 per cent, and (ii) the ratio of government debt to gross domestic product which is not to be greater than 60 per cent. There is no mention of an incomes policy to control nominal wages or a credit policy to guide the flow of credit in the overall economy of the common market. The private market mechanism is to prevail unencumbered by government interference. It is significant that the socialist government of France announced in 1993 that the Bank of France was henceforth to be insulated from all political interference and influence, as in the case of the Deutsche Bundesbank. Presumably, the two major central banks of Europe are to dominate the ECB despite its broader configuration under the Maastricht Treaty. There is to be, in short, no demand management in real terms; that is, no planned rate of real economic growth or any unemployment-rate target. There is, furthermore, an implicit assumption that the community-wide velocity function will be stable and well behaved. If, as a result of community-wide deregulation and liberalization, a wave of financial speculation and innovation takes place (as was the case in the United States during the 1980s), the community will be faced with marked shifts of the velocity function that will counteract and make a mockery of ECB monetary policy and its target of price-level stability. The Financial Sphere could well engage in 'stealing' resources from the Industrial Sphere. There is a very large probability that the lessons and warnings of the Treatise and the General Theory are about to be ignored, once again. There is no easy fix, no gimmick that can avoid the need for a community-wide coordination of all facets of economic policy - an incomes policy and a credit policy adequately supported by the traditional Keynesian tools of monetary and fiscal policy. Short of this the errors of the past are bound to be repeated. Notes I. 2. 3. 4. 5. 6. 8. Nicholas Kaldor (1970), 'The New Monetarism', Lloyds Bank Review, July, p. 5. See the exchange with Paul Davidson and Stephen Rousseas (1989), 'On the Endogeneity of Money Once More' Journal of Post Keynesian Economics, Spring, 11(3),474-490 John Maynard Keynes (1930), A Treatise on Money, vol. I (London: Macmillan p.243. Ibid. p. 254. Ibid. For two early articles on this problem, see Hyman Minsky (1957), 'Central Banking and Money Market Changes', Quarterly Journal of Economics, May, and Stephen W. Rousseas (1960), 'Velocity Changes and the Effectiveness of Monetary Policy, 1951-57', Review of Economics and Statistics, February. John Maynard Keynes, A Treatise on Money. p. 255. Stephen Rousseas 9. 10. 11. 683 John Maynard Keynes (1936), The General Theory of Employment Interest and Money (New York: Harcourt-Brace), p. 159. Ibid. p. 157. 'A Tax-Based Incomes Policy', in Sidney Weintraub (1978), Keynes, Keynesians, and Monetarists (Philadelphia: University of Pennsylvania Press, ch. 10, pp.259-80. 27 Profit Squeeze, Rentier Squeeze, and Macroeconomic Policy under Fixed and Flexible Exchange Rates * Gerald Epstein 1 INTRODUCTION It has become increasingly apparent that international economic arrangements, and particularly international financial integration, are critical determinants of macroeconomic outcomes. In this chapter I develop a model and present empirical results which identify several ways in which two major institutional features of so-called 'Open Economy Macroeconomics' - the nature of the exchange rate regime and the degree of international capital mobility influence the choice of macroeconomic policy and the effects of that policy on corporate profits. In particular, I argue that under flexible exchange rates expansionary policy is less likely to generate an industrial profit squeeze than under fixed exchange rates. The reason for this difference is that, with flexible exchange rates, firms are able to pass along increased costs in higher prices without being hurt by foreign competition as depreciating exchange rates maintain industry's competitiveness. Thus industrial capitalists are more likely to support expansionary macroeconomic policy under flexible than under fixed exchange rates. Financial capitalists, however, are likely to be wary of expansionary policy under both fixed and flexible exchange rates. No matter what the exchange rate system, after some point expansionary policy is likely to increase inflation which, in tum, may harm financial profits. As a result, under fixed exchange rates industry and finance will be relatively united in their support of contractionary policy, whereas under flexible exchange rates industry will support more expansionary policy than will finance. I 684 Gerald Epstein 685 By placing them in an 'open-economy' context, this chapter helps to resolve an apparent contradiction between two classic analyses of the political economy of macroeconomic policy: Kalecki's (1971) seminal article on 'Political Aspects of Full Employment' and Boddy and Crotty's (1975) well-known analysis of 'Class Conflict and Macro Policy': The Political Business Cycle. 2 Kalecki asks whether a capitalist state and economy would be able to tolerate sustained full employment. He does not believe it can: The rise in wage rates resulting from the stronger bargaining power of the workers is less likely to reduce profits than to increase prices. and thus affects adversely only the rentier interests. But 'discipline in the factories' and 'political stability' are more appreciated by the business leaders than profits. (Kalecki, 1971: 141; emphasis added) Hence rentiers, concerned about inflation and political stability, will join industrialists whose profits are not squeezed but who also fear the political power of workers. Boddy and Crotty (1975), on the other hand, argue that as expansion proceeds industrial profits will in fact be squeezed. Thus, industrial capital has a 'narrow' economic interest in supporting contractionary policy.3 The profit squeeze rather than the 'rentier squeeze' will be the dominant economic basis for contractionary policy. This chapter argues that the Kalecki and Boddy-Crotty analyses are each applicable depending on the nature of the exchange rate regime. Under fixed exchange rates, as existed under the Bretton Woods System of the 1950s and 1960s, the Boddy-Crotty analysis is more likely to apply as firms' profits get squeezed between rising wage costs and increased international competition which cannot be eased by depreciating exchange rates. Under flexible exchange rates, on the other hand, industrial profits are less likely to be squeezed as industrial firms will be able to raise their prices under the cover of depreciating exchange rates, without running as great a risk of being undercut by international competition. 4 Under these conditions, however, inflation will harm rentiers as Kalecki suggests. This chapter, none the less, departs from Kalecki by arguing that rentiers might have sufficient political power to convince the government, and particularly the central bank, to impose restrictive macroeconomic policy, even before industrialists become concerned about the increasing political power of labor. Thus a 'rentier squeeze' can help to explain contractionary policy under flexible exchange rates, despite the absence of squeezed industrial profit margins, or even threatened political instability. Note, however, that such contractionary policy might reduce industrial profits. 686 Profit Squeeze, Rentier Squeeze In short, under flexible exchange rates capitalists preferred policy will differ. As a result, under flexible exchange rates, economies may experience wide swings in macroeconomic policy as one group or the other gains the upper hand in macroeconomic policy-making. These conflicts are likely to be exacerbated by highly mobile and destabilizing short-term international capital flows, which produce cumulative and selffulfilling swings in exchange rates. Under these conditions, expansionary policy helps industrial capital more and hurts rentiers more while contractionary policy has the opposite effects, at least in the short run. This analysis helps to explain the recent move to more fixed exchange rates and less expansionary policy among the industrialized nations. In the face of intra-capitalist conflicts over macroeconomic policy under flexible exchange rates, both industrial and financial capital are likely to opt for relatively fixed exchange rates and thus pressure their governments to adopt them. Since under fixed exchange rates both industry and finance prefer macroeconomic policy which avoids sustained full employment, they also pressure the government to avoid 'excessively expansionary' policy. The chapter is organized as follows. Section 2 presents a simple model which illustrates the role of the exchange rate regime in determining the intra-class effects of macroeconomic policy. In section 3 I indicate how destabilizing speculation exacerbates the conflicts between industry and finance. Section 4 provides some empirical evidence which supports the model for the case of the United States. In section 5 I suggest why industrial and financial capital might opt for more fixed exchange rates. The last section presents conclusions. 2 EXCHANGE RATE REGIMES AND THE CONNECTIONS BETWEEN FINANCE AND INDUSTRY The Model I begin by developing a simple open economy model which has Keynesian, Kaleckian and neo-Marxian features. 5 I start with aggregate demand. y = A(y) + T(y, T) product market locus6 (1) where A is the domestic absorption = C + I + G; T is the trade balance; y is real income; T is terms of trade; e is the exchange rate (domestic/foreign currency); p* is the foreign price level, which is assumed to be fixed; P is the domestic price level. Gerald Epstein 687 Equilibrium output is equal to domestic demand (absorption), which depends on real income,? plus the trade balance, which depends on income and the terms of trade. T = (eP*/P) terms of trade (competitiveness) (2) Note that an increase in T is a deterioration of the terms of trade or an improvement in competitiveness. 8 The trade balance is given by exports minus imports, both measured in the domestic currency. (3) T= X(r) -M(r, y) The yy line is drawn in r-y space as an upward sloping product market locus (see Figure 27.1). The locus shows combinations of T and y where aggregate demand for output equals income. Assuming that the so-called Marshall-LernerRobinson condition holds, an increase in r improves the trade balance measured in domestic currency and increases aggregate demand; this increases the level of output required to clear the market. 9 Above the yy line, improvements in the trade balance lead to excess demand. Below the yy line, there is excess supply. Similarly one can draw a trade balance locus where exports equal imports. T= 0 = X(r) -M(y, r) (3)' Equation (3)' yields an upward sloping curve in T-y space. Increases in T increase competitiveness and the trade balance T.IO Income, therefore, has to increase to increase imports and bring the trade balance back down to O. Points above the line are points of balance of trade surplus as improvements in competitiveness increase net exports. Points below the line are balance of trade deficits (Figure 27.1). To keep the model as simple as possible, monetary factors are left in the background. I merely assume that monetary or fiscal policy affects aggregate demand. Under the flexible exchange rate regime macroeconomic policy can also affect the exchange rate. Contractionary macroeconomic policy, for example, thus moves the yy curve up and to the left. II The next step is to develop the domestic pricing equation. I assume a Kaleckian mark-up over unit labor and raw materials costs. P = (aW + bPm)(1 + x) Price Equation (4) where P is output price; a is labor per unit of output; b is raw materials per unit of output; W is nominal wage per cost of labor; 12 P", is raw materials price; x is mark-up. The degree to which workers bargain over real (as opposed to nominal) wages is, of course, a hotly contested theoretical and empirical issue. I use a wage equation which allows for either nominal or real wage bargaining or some combination of the two. Let Q represent the consumer price index, which reflects both 688 Profit Squeeze, Rentier Squeeze y' y T=O y y Figure 27.1 domestic (P) and import (eP*) prices. depending on the shares of domestic (f3) and imported (1- (3) goods in consumers' consumption baskets. Q = pf3 (ep*)I-/3 (5) The wage equation is then given as: W = 8(nQ)aW (I - a) (6) 8 reflects the reserve army effect which I will describe momentarily. W represents the nominal wage-workers bargain for ex is a parameter between 0 and 1 which reflects the degree to which workers bargain over real or nominal wages. If (l is O. workers bargain over dominal wages. In this case equation (6) reduces to: (6)' W=8W If. on the other hand. becomes: W=80Q (l is equal to 1. workers bargain over real wages and (6) (6)" and in this case the real wages (WIQ) equals 8.0. ot might vary from country to country. and its size will be an important determinant of the effects of macro economic policy on the rate of profit. Gerald Epstein 689 Following Marx, I assume that workers' wages are affected by the 'reserve army' (Marx, 1967, ch.25).13 I represent the reserve army affect as operating through the parameter 8. 8 = a(ylf) = 8(y) 8' > 0 Reserve Army effect on wages (7) where 8 is the reserve army parameter 8 > 0; y is full employment level of output = I. Thus as output goes up and reserve army gets smaller, wages increase. 14 Raw materials may be domestically produced or imported. Here I make the assumption that all raw materials are domestically produced or priced in the domestic currency. This assumption, is relevant to the case of the US. 15 In note 22 below, I analyze the model regarding imported raw materials, showing that in the case of nominal wage bargaining imported raw materials reduce the competitiveness maintaining effects of exchange rate changes, only vitiating them in the cases of extreme dependence on raw materials priced in foreign currency. 16 The next step is to take into account the effects of competition on profit margins. 11 A common way to model import competition is to use the import penetration ratio (Bowles, Gordon, Weisskopf, 1989; Goldstein, 1986a; Pugel, 1978; Shepherd, 1982). The import penetration ratio measures imports as a share of domestic demand. 4> = M/y + M - E = M( T, y)/[A(y)] import penetration ratio (8) The higher is ~, the greater the competition from imports. The import penetration ratio falls when 7' increases, that is, when the terms of trade deteriorate. An increase in income will increase, have no effect, or decrease the import penetration ratio as the elasticity of imports with respect to demand is greater than equal to or less than one. IS Hence: 4> = 4>(7', y) 4>. ~ 0 (9) Goldstein has shown that the mark-up in the United States is reduced by increases in the import penetration ratio (Goldstein, 1986a).19 I thus follow previous empirical work in taking the mark-up as a negative function of import competition. 2o Hence, we can write the mark-up (x) as a negative function of international competition (import penetration ratio) (4)). x = x(4)('I', y» (10) We can write (10) in the following way: X=X(T,y) x.>O,Xy~O (11) Equation (II) says that as the terms of trade deteriorate, holding output constant, competitiveness improves, holding the terms of trade constant, the mark-up may rise, fall or stay the same, depending on the effect of output increases on the import penetration ratio. Thus, international competition serves as a constraint on mark-up behavior, which can be off set by changes in the terms of trade. 690 Profit Squeeze, Rentier Squeeze The pricing behavior of firms combined with the wage-setting behavior of labor have implications for the terms of trade and competitiveness. Substituting (5), (6) and (7) into (4) gives: r- = [(as (y)o,a (W / ep*)I-a + (bPm )(ep*)-I(l-a »(1 + x(r ,y)fC-!JI ) (12) Equation (12) represents the terms of trade consistent with mark-up pricing and reserve army wage setting. The consistent terms of trade varies with the level of output. For example, an increase in the level of output y leads to an increase in w which leads to an increase in P and, therefore a reduction in T, or a deterioration of competitiveness. 21 This function can be drawn in T - Y space. It is downward sloping. As output increase, wages rise and domestic prices go up through the mark-up. With exchange rates and foreign prices fixed, domestic prices increase relative to foreign prices. Thus, the terms of trade improve, while domestic competitiveness and the trade balance deteriorates (Figure 27.2). The effects of exchange depreciation on the consistent terms of trade depends on the wage-setting process. If, for example, wage setting is entirely in nominal terms, then 0: = 0 and equation (12) reduces to: y t = He) t=t(e) y Figure 27.2 Gerald Epstein 1'= ep*/[(aB(y)W + bPm )(1 + x)] 691 (12)' Notice that in this case as the currency depreciates (i.e. the exchange rate goes up), the consistent terms of trade line shifts up. This occurs because the depreci ation of the currency increases the cost of imports and improves competitiveness. The curve shifts up because workers cannot protect their real wages. On the other hand, if workers bargain over real wages (a = I), then (12) reduces to: T- = [(al) (y)O + (bPm )(1 + x)f(.-~ ) (12 )" If workers completely protect real wages, the curve does not shift up when the exchange rate depreciates because workers' wage demands would go up with increases in the costs of imported goods. Such wages increases would lead to increases in domestic prices through the mark-up. In this case, exchange rate depreciations are ineffective in improving competitiveness. In general, the wages will not be set either in completely nominal or real terms. Workers' wages will partly increase to compensate for exchange rate depreci ations. Hence the 1'1' line will shift out to some extent in response to depreciations of the exchange rate, but not as much as in the case of complete nominal wage bargaining.22 The final relationship is the profit rate for industrial firms. 7T/PK = r = [y(P - aW - bPm)/PK] Profit Rate (13) let K = I, assuming that the capital stock is fixed in the short run. The profit rate can be written as: r = y[(x/l + x)] Profit Rate (14) The profit rate equals real output times the profit share. One can derive iso-profit curves from (14) in 'T, y space. substituting in (11) into (14) gives: r = r(T, y) (15) The rr line is downward sloping and convex to the origin under reasonable conditions 23 (see Figure 27.3). Figure 27.4 presents the whole model. Point A represents maximal industrial profits given the constraints of the economy. A state attempting to maximize industrial profits would try to set policy to make the yy curve go through A. Note that in general the economy will be at a point of less than full employment. Now assume that there is an investment boom. An increase in output (shift in yy to y'y') results in a deterioration in the profit rate (point B) - a high employment profit squeeze. A state attempting to increase industrial profits will not accommodate this increase in wages or deterioration in competitiveness. Industry would want it to contract monetary or fiscal policy to move the economy back to A. 692 Profit Squeeze, Rentier Squeeze r' xt -+ r' r y Figure 27.3 't y' Figure 27.4 Gerald Epstein 693 Finance would similarly support contractionary policy. They oppose such inflationary expansion and exchange rate depreciation because they harm bank profits. 24 Inflation tends to harm bank profits because banks are net monetary creditors (Petersen, 1986; Santoni, 1986; Wallich, H. C. 1975). If inflation were fully anticipated, banks might be able to protect their earnings in inflationary times. Yet evidence suggests that in the United States at least interest rates have not fully compensated for inflation (Summers, 1983). and indeed, that. inflation has harmed real bank earnings (Epstein and Schor, 1990 (a); Petersen, 1986).25 In short, finance and industry are unified over macroeconomic policy under fixed exchange rates: they both oppose a highly expansionary policy. Flexible Exchange Rates This unity between industrial and financial capital evaporates under flexible exchange rates. Assume that exchange rates are set in forward-looking asset markets. As a benchmark case, assume that the exchange rate changes to offset expected relative price differences. I relax this assumption below. This assumption yields a purchasing power parity exchange rate theory of equation (16). 1\ 1\* e=p-p A (\6) where 1\ denotes proportional rate of exchange (e.g. dele). Equivalently, from (2). the assumption yields a constant terms of trade. 'T = eP*IP = constant (17) An increase in investment leads to a depreciation of the exchange rate as assetholders expecting higher inflation sell domestic currency. This depreciation causes competitiveness to improve, shifting the consistent terms of trade upwards as long as there is not complete real wage resistance (see Figure 27.5). This depreciation, in turn. protects domestic producers from foreign competition. 26 Industrial capitalists are thus able to achieve a higher rate of profit. 27 Even though industrial profits are not squeezed as output increases, Kalecki's modern day rentiers, the banks, are hurt. Thus flexible exchange rates bring about a conflict of interests between industry and finance with respect to macro economic policy. 3 INTERNATIONAL CAPITAL MOBILITY, SPECULATIVE BUBBLES AND INTRA-CLASS CONFLICT The simple model outlined above suggests that with flexible exchange rates conflicts will arise between industrial and financial capital over macroeconomic 694 Profit Squeeze. Rentier Squeeze y' 't* r' t = t (e') t =t (e) y' y Figure 27.5 policy. The model assumed that exchange rates adjusted to maintain purchasing power parity. However. as critics of financial capital mobility and floating exchange rates have long argued. short-term financial speculation can move exchange rates in self-reinforcing bubbles. which drive exchange rates far away from purchasing power parities. 28 Hence, exchange rates are more likely to be determined by a process where some forces (represented by 0) move the exchange rate back to the exchange rate given by purchasing power parity (e*), while other~, fepresented by E, keep the exchange rate moving in whatever direction it happens to be moving at the time. 29 (18) In the 1970s and 1980s, during significant periods of time, the bandwagon or bubble effects (E) have been more important than the competitiveness effect (a). Figure 27.6 illustrates the effects of an economic expansion accommodated by expansionary macroeconomic policy where exc~ange rates continue to depreciate in a cumulative fashion. The exchange rate depreciates by more than is necessary to maintain competitiveness. The inflationary effects of the expansion are wors- Gerald Epstein 695 Figure 27.6 ened, however, as is the decline of the currency, both of which harm financial interests to a greater extent than before. Contractionary policy has more serious detrimental effects on industrial capital with destabilizing speculation. Where exchange rates simply change to maintain competitiveness, as in the previous section, contractionary policy has no detrimental effects on profit margins. 3o With destabilizing speculation, however, (:ontractionary policy results in an increase in international competition which squeezes profit margins further, at any level of output (see Figure 27.7). Starting at point A, contractionary policy under fixed rates would improve profit rates (B). With purchasing power parity, profit rates would fall as a result of the decline in output (C). With cumulative movements in the exchange rate leading to a deterioration in competitiveness, profit rates would fall even more as competitiveness and output further decline (D). Hence policy taken to protect the interests of financial or industrial capitalists can be magnified through cumulative speculative capital flows which generates further conflicts between finance and industry. 696 Profit Squeeze, Rentier Squeeze 't T=O y r t = 'i (e) y 'i = 'i (e') y Figure 27.7 4 EMPIRICAL EVIDENCE: THE CASE OF THE US IN THE POSTWAR PERIOD The model described above suggests that where workers are unable to protect their real wages completely in the face of price hikes. increases in unit labor costs are more likely to reduce profit margins and rates under fixed than under flexible rates. Furthermore. it claims that banks and other rentiers are harmed by inflation associated with expansionary policy. Finally. the model suggests that with speculative capital mobility. contractionary policy may be particularly harmful to industrial capital under flexible exchange rates. and expansionary policy will be particularly harmful for financial capital. In this section I illustrate these points with data from the manufacturing and banking sectors for the postwar United States. The statistical results of this section are broadly consistent with the implications of the model described above.3' Figure 27.8 presents data for producer prices and unit labor costs in manufacturing. During the fixed exchange rate period. increases in unit labor costs were Gerald Epstein 697 6.0 5.25 J - _ .... 5.5 / 5.00 I I I 4.75 LOGP 4.50 4.25 5.0 I I LOGULC ... --..,.---~ 4.00 1960 1965 4.5 1975 1970 LOGULC : Log of unit labor costs in manufacturing LOGP : Log of producer prices in manufactruring - 1980 1985 LOGULC - - - LOGP I Figure 27.8 Unit labor costs and prices in US manufacturing (1960.1-1985.1) reflected in prices to much less a degree than under the flexible rate period of dollar depreciation. The econometric results presented in Table 27.1 support the general results presented in the graph. Table 27.1 reports regression equations where manufacturing producer prices are estimated as a function of unit labor costs in manufacturing, raw materials prices, and manufacturing capacity utilization, where the latter is included to test for demand effects on prices. 32 The results indicate that for manufacturing firms, producer prices increased by smaller amounts with respect to unit labor costs during the fixed rate period than during the flexible rate period as the model above suggests. Indeed, during the fixed rate period changes in unit labor costs appear to have had no effects on pricing. 33 Increases in raw materials' prices, on the other hand, were passed on in the form of higher prices under fixed rates, possibly because such increases tended to be worldwide (SylosLabini, 1979).34 Demand, as measured by capacity utilization, appeared to have the same effects over the entire postwar period, holding unit labor and raw materials costs constant. To investigate the hypothesized relationship between relative unit labor costs, exchange rates and prices under the different exchange rate regimes, I estimated a vector auto-regression (VAR) of monthly data for the US exchange rate, and unit labor costs and prices in manufacturing. 3' The model suggests that increases in unit labor costs ought to have less effects on prices and exchange rates under fixed exchange rates than under flexible exchange rates. 698 Profit Squeeze, Rentier Squeeze Table 27.1 Unit labor costs and producer prices in US manufacturing Dependent variable! time period Producer prices 1960.01-1985.03 1960.01-1971.09 1971.1 0-1985.03 1971.10-1980.03 1980.04-1985.03 (1960.01-1985.12) C Independent variables ULC(-I) RMP(-l) CAP(-l) -.04 (-.04) -.07 (-.06) -.03 (-.04) -1.49 (-3.41) 1.84 (1.43) .27 (2.35) -1.99 (-5.50) -2.02 (-4.88) -2.04 (-5.07) -2.13 (-4.49) 2.72 (2.31) .60 (5.21) .20 -.to (-.92) .50 (2.81) 1.63 (4.03) -0.7 (-.46) .77 (8.74) 1.11 (4.65) .57 (3.59) -.14 (-.49) .74 (5.46) R2 DW .99 2.2 .99 1.7 .99 2.3 .99 2.4 .99 2.0 .99 2.0 .99 2.0 .99 2.0 .99 2.1 .99 1.9 II 1960.01-1985.03 1960.01-1971.09 1971.10-1985.03 1971.1 0-1980.03 1980.04-1985.03 (1.5) .74 (4.78) 1.37 (3.54) -.12 (-.46) .61 (7.64) 1.09 (8.72) .49 (4.38) .03 (.12) .73 (6.06) .29 (4.91) .17 (2.62) .31 (4.26) .20 (1.93) -.14 (-1.11) Note: t-statistics in parentheses. Monthly data. The dependent variable is the log of the producer price index for all manufacturing. The independent variables are, ULC, the log of unit labor costs in manufacturing; RMP, the log of price index for raw materials and components for manufacturing; CAP, the log of the capacity utilization rate in manufacturing. The regressions were estimated with third order poly nominal distributed lags, 12 lags, no end point constraints. the regressions under section I of the table were corrected for the first order serial correlation; the regressions under part II were corrected for first and second order serial correlation. Source: see appendix for data sources. Figures 27.9 and 27.10 show the effects, one to twelve months later, of a one standard deviation increase in manufacturing unit labor costs on unit labor costs, prices and exchange rates during the period of fixed exchange rates and flexible exchange rates. The graphs indicate that, as the model suggests, increases in unit 699 Gerald Epstein 0.7 " , "" 0.6 0.5 1 1 1 1 0.4 " ... ... ... ... EXRUK \ \ \ PLM \ 0.3 , ,, 0.2 0.1 0.0 -0.1 I \ --- -- --- --PWM \ \, " "" , " '" \ \ \ \ \ \ --- ----- 3 5 4 PLM : Unit labor cost, manufacturing EXRUK : Dollar prices exchange rate ($/Sterling) PWM : Producer prices, manufacturing 8 7 6 I- ... , --- -0.2 2 ... ... 9 10 12 11 I PLM - - - EXRUK - - PWM Figure 27.9" Response to one standard deviation shock ofPLM (1960.01-1971.09) 1.25 ,-----~---------------., ,r - -"", I 1.00 \ EXRUK " \ I / )/ 0.75 PWM " -----",/ 0.50 \ \ I " / / \ \ \ 1 I I I " 1 / \ 1/ , r I I I , I I 1 0.25 0.00 -- / - - - / I +-----.r---,--..--r---y--,--,-----.,....---,--..---i 2 3 4 PLM Unit labor cost, manufacturing EXRUK Dollar-sterling exchange rate (Slsterling) PWM Producer prices, manufacturing 5 6 1-" 7 8 9 JO II 12 PLM - - - - EXRUK - - - PWM 1 Figure 27.10 Response to one standard deviation shock ofPLM (1973.01-1980.09) " Profit Squeeze, Rentier Squeeze 700 labor costs are associated with exchange rate depreciations and price increases under flexible exchange rates. They also tend to generate wild swings in exchange rates as we would expect from a regime of destabilizing speculation (see section 3 above).36 Under fixed exchange rates, increases in unit labor costs had relatively small predicted effects on prices. The predicted effects on exchange rates are perverse, with exchange rates appreciating initially rather than depreciating as one would expect under the flexible rate period. These results support the general thrust of this chapter which suggests that in the United States during the postwar period, under fixed exchange rates increases in unit labor costs were more likely to squeeze profits than under flexible exchange rates. Morever, they suggest that under flexible exchange rates changes in unit labor costs and prices induce wild swings in exchange rates, which, par ticularly in contractionary times, can harm industrial profitability. It remains to be shown that in the United States workers' wages are not highly sensitive to price or exchange rate changes. Some evidence can be gathered along those lines from Figures 27.11 and 27.12 for the VAR presented above. These figures show the effects on unit labor costs and exchange rates of a one standard deviation increase in prices. The results suggest that, indeed, unit labor costs are not highly sensitive to changes in prices, as one would expect if there is little real wage resistance. 0.6 0.5 ,, 0.4 0.1 I I --- 0.0 I PWM --------------~~-----PLM \ \ I \ -0.2 \ EXRUK,' \. .... _.... --- , , I , -0.3 2 3 4 5 PLM : Unit labor cost, manufacturing EXRUK : Dollar-sterling exchange rate ($/sterling) PWM : Producer prices, manufacturing Figure 27.11 " \ -0.1 -0.4 --- I 0.3 0.2 , 6 , I I I 7 1-- PLM 8 9 10 1\ 12 EXRUK - - PWM 1 1.-_ _ _ _ _ _ _ _ _ _- - - - ' Response to one standard deviation shock of PWM (1960.01-1971.09). Gerald Epstein 701 4.--------------------------------------, EXRUK 3 / /~ / - -, , / ... ;.::.-=..-- 2 ".......... -~;::./ - PWM :-- / / PLM I I ,, -I I I , ,/ 2 --------- ------------ / I I 0 ,, -- / 3 4 5 I- PLM Unit labor cost, manufacturing EXRUK Dollar-sterling exchange rate ($/sterling) PWM producer prices, manufacturing . 6 7 8 9 10 II 12 PLM - - - EXRUK - - - PWM Figure 27.12 Response to one standard deviation shock ofPWM (1973.01-1980.09) Similarly, Figures 27.13 and 27.14 show the effects on prices and unit labor costs of a one-standard deviation change in exchange rates. Unit labor costs are not highly sensitive to changes in the exchange rate, though prices are. These results support the hypothesis that, in the United States, exchange rates can be an effective means of altering product wages. These purely suggestive results are consistent with a great deal of empirical work which indicates that wages in the US are quite sticky. (For example, see Gordon, 1983, 1989, and the vast numbers of references cited there.) In Table 27.2, I estimate wage equations for US manufacturing during the post-war period which are consistent with previous research. 37 These equations suggest that wages lag behind changes in producer prices, which, in tum, suggests that manufacturing firms in the United States could have profited from expansion and exchange rate depreciation during the flexible exchange rate period. 38 Exchange Rates and Rentiers While the effects of expansion on industrial profits depends on the wage setting structure and the exchange rate regime, the effects on rentiers depends on the inflationary effects of expansion. The results of this section indicate that, US banks' profit rates fall in inflationary periods, holding other factors constant. Figure 27.15 shows the evolution of US inflation and the rate of return for US commercial banks over the post-war period. 39 Casual observation suggests that 702 Profit Squeeze, Rentier Squeeze Table 27.2 Wages and prices in US Manufacturing (1960.1-1985.1) Dependent variable! regression equation Hourly compensation of workers Independent variables C P 1. Levels 39.40 (.41) .48 (6.10) .99 2.1 2. Log/log 2.60 (2.33) .61 (7.62) .99 2.0 3. Rate of change .01 (10.67) .50 (8.76) .61 1.9 Note: T-statictics in parentheses. Quarterly data. Dependent variable: hourly compensation of workers in manufacturing. Indenpendent variable (P) is the manufacturing producer price index, a third order polynomial distributed lag, with four lags and no endpoint constraints. In tha first equation, all variables are measured as levels; in the second equation, all variables are measured as natruallogs; in the third equation, all variable are measured as rates of change. Source: see data appendix. 3.0 - r - - - - - - - - - - - - - - - - - - - - - - , I 2.5 EXRUK ' I 2.0 1.5 1.0 0.5 0.0 --=...:_-_---..;:::-'-=-'=-='-='=-=~~PW:...:M:..::.-=-::=-=.=..::~.=....:::..:._=_:=-=I­ PLM 2 3 4 5 PLM : Unit labor cost. manufacturing EXRUK : Dollar-sterling exchange rate ($/sterling) PWM : Producer prices. manufacturing Figure 27.13 1971.09) 6 7 8 9 10 11 1- PLM - - -EXRUK - - 12 PWM I Response to olle standard deviation shock of EXRUK (1960.01- Gerald Epstein 703 6,--------------------------------------, 5 r _ _ .... I I EXRUK 4 --_/ _-_........ ..... 3 2 PWM _- ______ - - - --------- O~~~----------~~------------­ PLM -I+---._--r--,--~--_.--_r--_.--,_--._--._~ 2 3 5 4 6 7 I- PLM Unit labor cost, manufacturing EXRUK Dollar-sterling exchange rate ($/sterling) PWM Producer prices, manufacturing 10 9 8 II PLM - - - EXRUK - - 12 PWM I Figure 27.14 Response to one standard deviation shock of EXRUK (1973.011980.09) 0.15 1\ 1 I I ' I ,., / 0.100 / / , I 1 \ \ \ \ \,. 0.10 \ I \ \ ,. \ "" \ 0.05 ,, ,. 0.00 0.Q75 0.050 0.025 62 64 66 68 70 72 74 76 - 78 80 82 84 RRC - - - RINFL 86 I Figure 27.15 Rate of return on bank capital (RRC) and rate of inflation (RINFL) 1962-86 Profit Squeeze, Rentier Squeeze 704 Table 27.3 The effects of inflation on bank profit rates (1962-86) C RINFL UXINF RGNPG R2 DW .09 (l0.3) -.18 (-1.95) -.31 (-1.94) -.003 (-2.33) .38 2.0 Note: T-statistics in parentheses. Annual data. Dependent variable: rate of return on comercial bank capital, adjusted for effects of inflation on net income; all insured bank. Independent variables: RINFL: rate of change of the consumer price index, lagged one period. UXINF: expected inflation, lagged oun period, measured as residuals from a ARMA forecasting equation; RGNPG: real GNP growth, lagged one period. Source: Epstein and Schor (1990a). commercial banks' rate of return is inversely associated with inflation. As the figure indicates, the inflationary period of rapid dollar depreciation in the late 1970s was particularly damaging to bank profits, as argued in sections 2 and 3 above. Casual observation is strengthened by the following econometric results. Table 27.3 presents a regression equation which indicates the relationship between bank profits, measured as the rate of return on capital, and inflation. The results indicate that the rate of return on bank capital are harmed by inflation (RINFL) and unexpected inflation (UXINF). Increases in real GNP also harm bank profits. 4o 5 CAPITAL'S PREFERENCE FOR MORE FIXED EXCHANGE RATES WHEN INTERNATIONAL CAPITAL FLOWS ARE DESTABILIZING The previous sections argue that flexible exchange rates and financial speculative flows can cause wild gyrations in exchange rates and intensify intra-class conflict. Under these conditions, both financial and industrial capital might find it in their interests to adopt a fixed exchange rate system, which, in turn, will unify their interests around less expansionary policy. This analysis can help to explain both the recent move toward fixed exchange rates and, perhaps, the more contractionary policy implemented in many advanced capitalist countries in the 1980s. This section illustrates this argument with a simple example. 41 Assume that finance and industry vie to dominate macroeconomic policymaking. As a benchmark case, assume first that financial and industrial interests are equally likely to control macroeconomic policy. I alter this assumption below. Table 27.4 shows the value of industry and finance's profit rates depending on the stance of macroeconomic policy and the exchange rate regime. Gerald Epstein Table 27.4 705 Profit rates in industry and finance by macroeconomic policy and exchange rate regime Sector Policy Regime Industry Expansionary policy (profit rate) Fixed 10 10 Flexible 20 5 Contractionary policy (profit rate) Fixed 15 20 5 20 Flexible Finance Notes: Hypothetical rates of profit. See text. Table 27.4 represents a situation where industry prefers flexible exchange rates if policy is expansionary, but prefers fixed exchange rates if policy is contractionary. Finance, on the other hand, prefers contractionary policy under both regimes. If we assume that finance and industry will control macroeconomic policy with equal probability (p = 112), then the expected profit rates of industry and finance under fixed and flexible exchange rates can be easily calculated and are given in Table 27.5. 42 Under these assumptions, industry will clearly support fixed exchange rates, to protect itself against possible finance dominance of macroeconomic policy. Note that with the numbers used in the example above, the choice of regime is a close call for industry.43 However, if the country has a relatively independent central bank, and if that central bank tends to be dominated by finance, as in the United States, industry's preference for fixed exchange rates becomes much stronger (Epstein, 1981). Indeed, with finance able to veto expansionary policy through its control of the central bank, the likelihood that finance will dominate policy approaches one. In this case, industry's preference for fixed exchange rates becomes overwhelming. (In this example, under flexible rates, industry's expected profits is 5 whereas under fixed rates it is 15.)44 This example helps to explain the recent move toward more fixed exchange rates and policy coordination among the advanced capitalist countries. One schooled in rational expectations might claim that the foregoing argument should also have precluded the earlier move to flexible exchange rates, since both industry and finance would have predicted that their profits would be higher under the fixed exchange rate regime. As Post Keynesians correctly stress, however, learning takes time in an uncertain world. US industry had to experience the 1980s to believe it. 45 Profit Squeeze, Rentier Squeeze 706 Table 27.5 Expected profits under fixed and flexible exchange rates' Regime Industry Sector Finance Fixed exchange rates 15 20 Flexible exchange rates 12.5 12.5 Note: Numbers are derived from Table 27.4. Expeted profits are calculated under the assumption that finance and industry are equally likely to control macroeconomic policy. For an explanation of the calculations, see endnote 30. 6 CONCLUSION Mainstream theory assesses the move from one exchange rate regime to another on the basis of which yields higher 'social welfare'. One implication of this chapter is that the mainstream approach will miss powerful political-economic dynamics - rooted in the effects of these regimes on macroeconomic policy and sectoral profitability - which are likely to affect strongly the choice of exchange rate regime. For example, in his classic brief for flexible exchange rates, Milton Friedman argued that flexible rates reduce international conflict by allowing each country to pursue its desired macroeconomic policy without adversely affecting other countries. However dubious this proposition, it suffers from another defect: by ignoring class and intra-class conflict and their strategic interaction with macroeconomic policy, Friedman fails to see that, even if flexible exchange rates reduce international conflict, flexible exchange rates and high capital mobility intensify domestic intra-capitalist conflict.46 With the capitalist class divided over macroeconomic policy, under flexible exchange rates, nations will probably experience periods of macroeconomic policy that will appear to be overly expansionary followed by periods that will seem excessively contractionary. Ultimately, a class-wide policy that favors relatively fixed exchange rates and less expansionary policy is likely to coalesce. I would suggest that this is roughly what the United States has witnessed over the last fifteen years. 47 This argument implies an interpretation of the increasingly common view that high levels of international financial integration has rendered macroeconomic policy less effective (see, for example, Crotty, 1989; Wachtel, 1987.) An important sense in which macroeconomic policy is less effective under flexible Gerald Epstein 707 exchange rates and high capital mobility is that macroeconomic policy which benefits one section of the capitalist class harms another; thus Keynesian macroeconomic policy has become less able to forge and maintain a political coalition that can sustain Keynesian policy. That is not to say that in an era of high capital mobility, fixed exchange rates can salvage Keynesian policy. On the contrary, high capital mobility which can generate massive speculation against fixed rates, creates an environment in which capitalists do reach a consensus, but it is a consensus to abandon Keynesian policy not to restore it. Data Appendix CAP: Capacity Utilization in Manufacturing; per cent of capacity, seasonally adjusted. Board of Governors of The Federal Reserve System. Statistical Release G.3 (402). EXRUX: Dollar/Sterling Exchange Rate. International Financial Statistics. International Monetary Fund. Hourly compensation of workers: Wages and salaries of employees plus employers' contributions for social insurance and private benefit plans. US Department of Labor; Bureau of Labor Statistics. Producer prices: Producer Price index; total manufactures; 1967 Labor; Bureau of Labor Statistics. =100. US Department of RGNPG: Rate of growth of real GNP; Department of Commerce. Bureau of Economic Analysis. RINFL: Rate of inflation. Rate of change of the consumer price index. US Department of Labor; Bureau of Labor Statistics. RMP: Raw materials prices. Materials and components for manufacturing. Monthly 1967 = 100. US Department of Labor; Bureau of Labor Statistics. ULC: Unit labor costs in US manufacturing; monthly data. 1967 = 100. Department of Commerce. Bureau of Economic Analysis. Business Conditions Digest. UXINF: Unexpected inflation, calculated as the residual from an ARMA forecasting equation. Notes I. I would like to thank Samuel Bowles and two anonymous referees for extremely helpful comments on previous versions of this material. I also appreciate the helpful comments of Thomas Weisskopf, Juliet Schor, Diane Flaherty, Thomas Michl, John Goldstein, and members of the Center for Democratic Alternatives Macro-Working Group. Thanks also to Emily Kawano and Elizabeth Kruse for research assistance and the Department of Economics of the University of Massachusetts for financial assistance. Remaining errors are mine. The terms 'industry' and 'finance', while having a long and distinguished lineage are none the less vague. I use industry here to refer primarily to businesses which hire 708 2. 3. 4. 5. 6. 7. 8. 9. 10. II. Profit Squeeze. Rentier Squeeze workers to produce and sell non-financial products. By finance I refer to banks and other net creditor businesses which make the bulk of their profits through their credit relations. Admittedly, this is a highly simplified distinction which, however, is useful at the level of abstraction at which this chapter is operating. In more concrete applications. more refined work on intra-capitalist divisions is necessary. See, for example, Epstein (1981), Epstein and Schor (1986, 1988), Ferguson and Rogers (1986). See also the work by Glyn and Sutcliffe (1972) who develop the profit squeeze argument in a secular context. See Weisskopf (1979) for a general discussion and empirical tests of the profit squeeze theory in the context of Marxian crisis theory. More generally, the chapter helps to resolve analogous disagreements between the Post Keynesian and neo-Marxian understandings of the relationship between macroeconomic policy and income distribution. Weisskopf makes a different argument, suggesting that the capitalist class as a whole will oppose inflation. 'Sooner or later a capitalist government must confront the choice between encouraging accelerating inflation or allowing a cyclical economic contraction to take place '" it is generally preferable to the capitalist class (and to a pro-capitalist government) over the alternative of accelerating inflation with unpredictable consequences' (Weisskopf, 1978). It is not clear, however, why full employment must lead to accelerating inflation. For an early argument along these lines, see Epstein (1982). Goldstein (1985, a; fn. 24) notes the offsetting effects of flexible exchange rates but does not develop the point, which, by contrast, is the focus of the discussion here. This model is related to Dornbusch (1980), ch. 4. and Bowles (1987). Layard and Nickell (1986), develop an open economy NAIRU model similar to Dornbusch's. Like Dornbusch's analysis, however, the Layard-Nickell model differs from mine in ignoring the effects of international competition on the mark-Up. Their paper also assumes away short-run flexibility of the exchange rate, which plays a central role in my analysis. I assume throughout that profits have no short-run effects on investment. This assumption could be dropped without fundamentally altering the results of the chapter, as long as it does not lead the model to be unstable. See Bowles and Boyer (1988), Goldstein (1985, a) and Marglin and Bhaduri (forthcoming) for related models where investment does depend on profits. By ignoring the effects of the terms of trade on absorption, I am assuming away the so called Laursen-Metzler effects (Laursen and Metzler, 1950), as well as possible macroeconomic effects of distributional changes associated with changes in exchange rates (e.g. Krugman and Taylor, 1978). I assume that foreign prices are fixed. The results would go through under the minimal assumption that foreign firms' prices do not completely offset domestic price or (in the later part of the chapter) exchange rate changes. For a discussion of this issue of so-called pass through, see, for example, Mann (1986). The so-called Marshall-Lerner-Robinson condition says that price elasticities of demand for imports and exports are sufficiently high that an increase in T will improve the trade balance. There is a good deal of econometric evidence that these conditions hold for most advanced capitalist countries, though the lags might be long and variable (Goldstein and Kahn, 1986). This result also assumes the Marshall-Lerner-Robinson condition holds. I do not distinguish in most of this discussion between fiscal and monetary policy. Some readers might find this problematic for several reasons. First, according to the socalled Mundel-F1eming model, fiscal and monetary policy have very different effects depending on the exchange rate regime: monetary policy is ineffective in determining output under fixed exchange rates while fiscal policy is ineffective under flexible Gerald Epstein 12. 13. 14. 15. 16. 709 exchange rates. However, for a large country like the US, which is the main object of study here, fiscal and monetary policy will both affect output under both regimes in the standard model (see, for example, Dornbusch (1980) for a discussion of these points). Secondly, one might argue that, under flexible exchange rates, expansionary fiscal policy will increase the value of the currency while expansionary monetary policy will lower the value of the currency, and thus I must distinguish between them in the discussion that follows. However, there still seems to be great controversy about the actual effects of fiscal policy on the exchange rate. So this concern is not obviously warranted. However, any reader who prefers can simply substitute 'monetary policy' for 'macroeconomic policy' throughout. I also leave the definition of monetary policy open. What monetary policy can and will effective will depend on the institutional context. In some contexts, interest rate policy like that described by Kaldor (1982) will suffice; in others, policies which directly affect the availability of credit will be necessary. At the level of abstraction here, it is not necessary to specify the precise mechanisms, though I should emphasize no monetarist mechanism is assumed .. Here I am abstracting from the distinction between labor and labor power in the accounting. Some modern versions emphasize the 'cost of job loss' as an important determinant of workers' bargaining power (see Bowles and Boyer, 1988; Bowles, Gordon and Weisskopf, 1989; Schor and Bowles, 1987). I could have labor productivity decreasing with respect to the depletion of the reserve army and nothing essential would change. See Rebitzer (1989) for evidence that unit labor costs are negatively associated with unemployment in the US. Much of the raw materials imported by the United States is probably denominated in dollars. While I have not been able to find direct evidence for this much indirect evidence exists. It is well known that oil prices have been denominated in dollars or most of the period discussed here. Scharrer (1980) estimates that 95 per cent of German imports of oil where priced in dollars. Presumably the percentage US oil imports priced in dollars is at least as high. He also estimates that 75 per cent of imports to Germany from Latin America are also priced in dollars, and 51 per cent from Asia. Again, one would expect US percentages to be at least as high. Kenen (1983) estimates that even industrialized countries price much of their exports to the US in dollars. For example, in 1976, in the middle of the sample period for my empirical results below, Japan priced 89 per cent of its exports in dollars and the UK priced 52 per cent in dollars. Dollar pricing will not completely insulate import prices from exchange rate changes, unless exporters have complete money illusion. However, it should reduce the price effects at least for moderate changes and at least for the short run. In addition to dollar pricing the share of imported raw mater ials in US production is relatively small. Total imports of crude materials and fuels as a percentage of total manufacturing output was 9 per cent on average for the period 1970-86 (US President, 1988, Table B-105; and B-IO). This, moreover, is probably an overestimate for manufacturing since probably not all of these materials were used in that sector. Michl (1989) estimates that raw materials, including imports plus domestically produced materials, fluctuated between 8 and 13 per cent of manufacturing output in the US between 1958 and 1982. Thus, as a simplification, it does not do serious damage to assume all materials produced domestically or priced in domestic currency. Moreover, note 22 shows that one could relax this assumption significantly and still not fundamentally affect the results of this chapter. Only in the case of extreme import dependence will the results change. There I show that the larger the share of imported raw materials in unit costs, in less exchange rate depreciations improve competitiveness. In the limit, exchange depreciations are useless in restoring competitiveness following an increase in unit labor costs. 710 17. 18. 19. . 20. 21. Profit Squeeze. Rentier Squeeze See Boddy and Crotty (1975). Crotty and Rapping (1975), G1yn and Sutcliffe (1972), Sylos-Labini (1979). Goldstein (1985a, 1986a, b) and Kotz (1982) have developed the microfoundations for the full-employment profit squeeze. Both domestic and international competition playa role in the squeeze in Goldstein's work, while Kotz focuses on domestic competition. T = M/A, where M = imports, measured in domestic currency; A is domestic demand. T < 0 as long as the elasticity of imports with respect to the terms of trade is greater than one. This is a stricter condition than the Marshall-Lerner-Robinson conditions, but is likely to hold for the US. The effects of a small change in income on the import penetration ratio is given by: y = [AMy - MAy)/A2, where My is the marginal propensity to import and Ay is the domestic marginal propensity to spend. As the expression for 4>), makes clear: CPy ~ 0 as AM/MAy ~ 1. Also see Pugel (1979). Bowles, Gordon and Weisskopf (1989) have shown that the import penetration ratio is negatively associated with profit rates in the US which provides further empirical support for my use of it here. See, however, Funke (1986) for the case of the United Kingdom. We could also take the mark-up to be independently affected by output. However, there is little agreement about the effects of output on the mark-up independently of international competition. For studies which have found a positive association between demand and the mark-up, see for example, Eckstein and Fromm (1986), Eckstein and Wyss (1972), Goldstein (1896a. b), Gordon (1975) and Bowles, Gordon and Weisskopf (1989). For studies which have not found a relation, see, for example, Coutts, Godley and Nordhaus (1978), Godley and Nordhaus (1972), and Sawyer (1983). I thus focus here on international competition, though, one could include the effects of domestic output in equation (II) below. Taking the case of nominal wage bargaining, (ex 22. =0), dTldy =-[T(ao (y)W (1 + x»)/[P(1 + k lL. ») < 0, T = where k lL . T is the elasticity of the mark-up factor (/L I + x) with respect to the terms of trade (T). It can be easily shown, though with messier algebra. that TT line is also downward sloping in the case where real wage resistance is present. I can now describe the effects of imported raw materials. Assume that all raw materials are imported. Moreover, to isolate the effects of raw materials, assume that workers engage in nominal wage bargaining. I relax this assumption below. P=(aoW +bePm*)(I +X(T,y» ( 1a) Where Pm* is imported raw material price and all other variables are as defined in the text. In this case, the consistent terms of trade, T, is: 'F = ep*/[(aBW + bePm*)(1 + X(T, y))) (2a) To determine the effects of a change in the exchange rate on the consistent terms of trade for every income level, differentiate (2a) with respect to e: (3a) where k r. u is elasticity of the consistent terms of trade with respect to the exchange rate; k lL • u = unit labor costs; C = unit costs = unit labor costs + unit raw materials cost. elasticity of markup factor (u I + x) with respect to the unchange rate, ULC unit labor costs. = = Gerald Epstein 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 711 As equation (3a) makes clear, the lower the share of labor costs, i.e. the higher the share of raw materials costs in total costs, the smaller the change in consistent terms of trade with a change in the exchange rate. in the extreme case where imported raw materials costs constitute I ()() per cent of the cost, there will be no change in the consistent terms of trade when the exchange rate changes. Of course, as discussed in the text, the United States is quite far from this extreme. Where bargaining is completely over real wages, and production depends on imported raw materials, exchange rate depreciations will lead to a shifting in of the consistent terms of trade which worsen competitiveness and harms profit margins. Where bargaining is only to some extent based on real wages, the consistent terms of trade will shift less than in either of these two cases. dTldy (r constant) -{XyY + x + x 2)/(xTy). A sufficient (but not necessary condition for a negative slope is Xy ~ O. A sufficient condition for the second derivative to be positive, is that xyy' x TT < O. In the United States, large multinational banks also opposed excessive expansion because they feared that the international role of the dollar would be undermined by US inflation and exchange rate depreciation. (See, for example, Epstein, 1981.) Also see the regression results presented below. As discussed above, the greater the degree of real wage bargaining, the less protection is provided by depreciation. One might think that there is an inconsistency between the 'micro' price-setting process and the 'macro' exchange rate model developed here. How do firms know that if they raise their prices, they will be protected by exchange rate depreciation? There are many routes by which increases in unit labor costs can lead to exchange depreciations. The most plausible route is through forward-looking speculative exchange markets. Forward looking exchange markets, predicting that there will be future prices rises, will depreciate exchange rates immediately as unit labor costs rise. Another mechanism is through firm price setting which operates on a rule of thumb - trial and error basis. When unit labor costs rise, firms raise prices through a rule of thumb. If their market share erodes too much, they might adjust in the next period. If their market share does not erode, because, for example, exchange rates depreciate, then they will keep using the rule in the future. A large theoretical literature has developed which supports the idea that exchange rate speculation can be destabilizing. See Frankel and Meese (\ 987) and Krugman (1989) for recent surveys. For a similar formulation, see Frankel and Meese (1987). Note that I am assuming that reductions in output, per se, do not lower the profit margin. Note that because of cross-country variations in wage setting, raw materials use, and other structural factors, the political dynamics described here might not apply to all countries. For a preliminary attempt to analyze these structural differences and some of their implications for comparative macroeconomic policy, see Epstein (1989), and Epstein and Schor (1986,1988, forthcoming (b». The equation assumes the raw material price input coefficient is stable over the period involved. Sylos-Labini (1979), for example, makes the same assumption in his estimates. Chow tests confirm the change in regime. One can most clearly reject the hypothesis that there was no structural break between 1960.01-1971.09 and 1971.10-1980.03; for regressions I, F(l0,223) 3.82; for regressions II, F(l5,213) 3.52 with the criticallevel for a I per cent significance level being less than 2.5 in both cases. For the entire fixed and flexible exchange rate period, 1960.01-1971.09 and 1971.1 0- = = = = 712 Profit Squeeze, Rentier Squeeze 1985.01, one can almost reject the stability hypothesis for regressions I; F(lO,283) = 1.84, with the critical value for 95 per cent significance level of approximately 1.86; for the regressions in part II, F(15,273) 1.84, with a critical level of approximately 1.65 for the 95 per cent significance level. Moreover, one can reject the hypothesis of stability between the two subperiods of the flexible rate period, 1971.10-1980.09 and 1980.04-1985-03; for regressions I, F(10.142) 2.02, with critical level of 1.89 at the 95 per cent significance level; for regressions II, F(15,132) 2.81, with the critical level of approximately 2.23 at the I per cent significance level. The results for raw materials during the 1970-80 period appears anomalous. The insignificant or low coefficients might be due to that fact that the raw materials coefficients have not been included in the regressions and, perhaps, were changing during this period. A vector autoregression is simply a set of regressions of an endogenous variable on lags of itself and other endogenous and exogenous variables. For a description of the uses of vector autoregressions, see, for example Sims (1980). I use these techniques simply as descriptive statistics, rather than to test the model formally or to assess 'causality' in the model. Using VARs for the latter purposes is highly controversial. The exchange rate is the dollar/sterling exchange rate. I used this, rather than a trade weighted exchange rate because that series was not available for the entire period. Estimates with the dollar/Deutsche-Mark rate gave similar results for the entire period, and the IMF trade-weighted exchange rate gave similar results for the period after 1973. I estimated equations for the fixed and flexible exchange rate periods separately, as well as for different subperiods within the flexible rate period. Using Chow tests and standard significance levels, I could not reject the hypothesis of coefficient stability over the whole period. I found similar results when I used consumer prices instead of producer prices. These data are adjusted for the effects of inflation on the real value of assets of banks. For a description of the data, see Epstein and Schor (forthcoming(a», the Federal Reserve Bank of New York (1986) and the references therein. This result may seem anomalous. It may result from the truncated lag structure of the equation. Increase in the rate of growth of real GNP may take a longer time than one year to improve bank profits. The numbers used in this example are, of course, purely hypothetical. But they are roughly consistent with the empirical results presented in the previous section. The following example illustrates how the expected profit rates are calculated: = = 34. 35. 36. 37. 38. 39. 40. 41. 42. = Fixed Exchange Rates Looking at Table 27.4, it is readily seen that under fixed rates industry and finance prefer contractionary policy. Using that information and the assumption that finance and industry are equally likely to control policy, we can calculate expected profits under fixed exchange rates. For example: Expected Industrial Profits =(probability that industry controls policy x industry's profits given its preferred policy under fixed rates) + (probability that finances will control policy x industry's profits under finance's preferred policy given fixed rates. = 112 x (industry's profit rate under contraction) + 112 (industry's profit under contraction) = 1/2 x (15) + 112 x (15) 15. (EXPECTED INDUSTRIAL PRofiTS UNDER FIXED EXCHANGE RATES) = The other entries in Table 27.4 can be calculated similarly. Gerald Epstein 43. 44. 45. 46. 47. 713 If industry is risk averse, it is more likely to prefer fixed rates because the variance of their profits is greater under flexible rates. Similarly, if workers are able to protect real wages in the face of depreciation, then industrial capital will have even less interest in preserving flexible exchange rates. This might apply to some of the European countries which have chosen to join the European Monetary System. In addition, because of vast structural changes during the postwar period, it is highly unlikely that the old Bretton Woods system could have survived even if both industry and finance had foreseen the consequences of the move to flexible rates. Fixed exchange rates might, on the other hand, intensify conflict between capital and labor. If labor.is relatively weak politically, as in the US, however, conflict within the capitalist class is likely to be more disruptive of macroeconomic policy. For historical and institutional evidence in support of this interpretation see Epstein (1981, 1985), and Epstein and Schor (forthcoming (b». References Blecker, R.A. (1989). 'International Competition, Income Distribution and Economic Growth', Cambridge Journal of Economics. Boddy, R. and J. Crotty (1975), 'Class Conflict and Macro Policy: The Political Business Cycle' , Review of Radical Political Economics, 7 (I), 1-19. Bowles, S. (1987), 'Price Wars: Contested Global Exchange and Contradictory Organizing Rules', mimeo, University of Massachusetts. Bowles, S. and R. Boyer (1988), 'Labor Discipline and Aggregate Demand: A Macroeconomic Model', American Economic Review, 78, 145-50. Bowles, S. and R. Boyer (1990), 'A Wage-led Employment Regime: Income Distribution, Labor Discipline, and Aggregate Demand in Welfare Capitalism', in S.Marglin and J.B. Schor (eds), The Golden Age of Capitalism (Oxford: Oxford University Press). Bowles, S., D.M. Gordon and T.E. Weisskopf (1989), 'Business Ascendancy and Economic Impasse: A Structural Retrospective on Conservative Economics, 1979-87', Journal of Economic Perspectives, 3 (I), 107-34. Coutts, K., W Godley and W. Nordhaus (1978), Industrial Pricing in the British Manufacturing Industry (Cambridge: Cambridge University Press). Crotty, G.R. and L.A. Rapping (1975) 'The 1975 Report of the President's Council of Economic Advisors: A Radical Critique' American Economic Review, 65 (5) December, 791-811 Crotty, J. (1989), 'The Limits of Keynesian Macroeconomic Policy in the Age of the Global Marketplace', in A. MacEwan and W.Tabb (eds), Instability and Change in the International Economy (New York: Monthly Review Press). Dornbusch, R. (1980), Open Economy Macroeconomics (New York: Basic Books). Eckstein, O. and G. Fromm (1968), 'The Price Equation', American Economic Review, December, pp. 1159-83. Eckstein, O. and D. Wyss (1972), 'Industry Price Equations', in O. Eckstein (ed.), The Econometrics of Price Determination (Washington, DC: Publication Services). Board of Federal Reserve System. Epstein, G. (1981), 'Domestic Stagflation: The Federal Reserve in the Hidden Election', in T. Ferguson and J. Rogers (eds), The Hidden Election (New York: Pantheon), 141-95. Epstein, G. (1982), 'Money and the Profit Squeeze in the Post-War U.S.', Paper presented to the Williams College Conference on Applied Macroeconomics. Epstein, G. (1985), 'The Triple Debt Crisis', in World Policy Journal, II (4), 625-57. 714 Profit Squeeze, Rentier Squeeze Epstein, G. (1989), 'A Political Economy Model of Comparative Central Banking', mimeo, University of Massachusetts at Amherst. Epstein, G. and J.B. Schor (1986), 'The Political Economy of Central Banking', Harvard Institute for Economic Research, Discussion Paper No. 1281. Epstein, G. and J.B. Schor (1988), 'The Determinants of Central Bank Policy in Open Economies', in B. Jossa and C.Parico (eds), Monetary Theory and Central Banking (Naples: Liguori Press), 225-64. Epstein, G. and J.B. Schor (1990a), 'Corporate Profitability as a Determinant of Restrictive Monetary Policy: Estimates for the Post War U.S.', in T. Mayer (ed.), The Political Economy of American Monetary Policy (New York: Cambridge University Press). Epstein, G. and J.B. Schor (l990b), 'Macropolicy in the Rise and Fall of the Golden Age', in S. Marglin and J.B. Schor (eds) The Golden Age of Capitalism: Reinterpreting the postwar experience (Oxford: Oxford University Press). Federal Reserve Bank of New York (1986), Recent Trends in Commercial Bank Profitability (New York: Federal Reserve Bank of New York). Ferguson, T. and J. Rogers (1986), Right Turn: The Decline of the Democrats and the Future of American Politics (New York: Hill & Wang). Frankel, J.A. and R. Meese (1987), 'Are Exchange Rates Excessively Variable l' National Bureau of Economic Research, Working Paper No. 2249. Friedman, M. (1953), 'The Case for Flexible Exchange Rates', in M. Friedman, Essays in Positive Economics (Chicago: University of Chicago Press), pp. 157-203. Funke, M. (1986), 'Influences on the Profitability of the Manufacturing Sector in the UK An Empirical Study', Oxford Bulletin of Economics and Statistic, 48 (2), 165-87. Glyn, A. and B. Sutcliffe (1972), British Capitalism, Workers and the Profit Squeeze (Harmondsworth: Penguin). Godley, W. and W. Nordhaus (1972), 'Pricing in the Trade Cycle', Economic journal, September, 853-82. Goldstein, J.P. (1985a), 'The Cyclical Profit Squeeze: A Marxian Microfoundation', Review of Radical Political Economics, 17 (1/2), 103-28. Goldstein, J.P. (1985b), 'Pricing, Accumulation, and Crisis in Post Keynesian theory', journal of Post Keynesian Economics, 8 (1),121-34. Goldstein, J.P. (l986a), 'Mark-Up PriCing over the Business Cycle: The Microfoundations of the Variable Mark-Up', Southern Economic journal, 53 (I), 233--46. Goldstein, J.P. (I 986b), 'Markup Variability and Flexibility: Theory and Empirical Evidence', Journal of Business, 59(4), 599-621. Goldstein, M. and M. Kahn (1986), 'Income and Price Effects in Foreign Trade', in R. Jones and P. Kenen (eds), Handbook in International Economics, Vol. II (New York: North Holland, pp. 1041-99. Gordon, RJ. (1975), 'The Impact of Aggregate Demand on Prices', Brookings Papers on Economic Activity, 613-{)2. Gordon, RJ. (1983), 'A Century of Evidence on Wage and Price Stickiness in the United States, the United Kingdom, and Japan', in J. Tobin (ed.), Macroeconomics, Prices and Quantities (Washington, DC: Brookings Institution), 85-133. Gordon, RJ. (1989), 'US Inflation, Labor's Share, and the Natural Rate of Unemployment', National Bureau of Economic Research, Working Paper No. 2585. Kaldor, N. (1982), The Scourge of Monetarism (New York: Oxford University Press). Kalecki, M. (1971), 'Political Aspects of Full Employment' , repr. in M. Kalecki, Selected Essays on the Dynamics of the Capitalist Economy, 1933-1970 (Cambridge: Cambridge University Press). Kenen, P.B. (1983), The Role of the Dollar as an International Currency, Occasional Paper No. 13 (New York: Group of Thirty). Kotz, D.M. (1982), 'Monopoly, Inflation, and Economic Crisis', Review of Radical Political Economics, 14 (4),1-17. Gerald Epstein 715 Krugman, P. (1989), Exchange-Rate Instability (Cambridge: MIT Press). Krugman, P. and L. Taylor (1978), 'Contractionary Effects of Devaluation', Journal of International Economics, 8, 445-56. Laursen, S. and L. Metzler (1950), 'Flexible Exchange Rates and the Theory of Employment', Review of Economics and Statistics, 32, 289-99. Layard, P.R.G. and S.J. Nickell (1986), 'Unemployment in Britain', Economica, Supplement, 53 (210, s), SI21-70. Mann, C.L. (1986), 'Prices, Profit Margins, and Exchange Rates', Federal Reserve Bulletin, June, 366-79. Marglin, S.A. and A. Bhaduri (FI990), 'Profit Squeeze and Keynesian Theory', in S.A. Marglin and 1.B. Schor (eds), The Golden Age of Capitalism: Reinterpreting the Postwar Experience (Oxford: Oxford University Press). Marx, K. (1967), Capital, vol. I (Moscow: Progress Publishers). Michl, T.R. (1989), 'Why is the Rate of Profit Still So Low?', mimeo, Colgate University. Page, S.A.B. (1981), 'The Choice of Invoicing Currency in Merchandise Trade', National Institute Economic Review, November, 60-1. Petersen, W.M. (1986), 'The Effects of Inflation on Bank Profitability', in Federal Reserve Bank of New York, Recent Trends in Commercial Bank Profitability (New York, Federal Reserve Bank of New York). Pugel, T.A. (1978),lnternational Market Linkages and US Manufacturing: Prices, Profits, and Patterns (Cambridge, MA: Bal\inger). Pugel, T. (1979) 'US Promotions of Manufactured Goods Expert', Working Paper, New York: NY University Graduate School of Business Administration. Rebitzer, J.B. (1989), 'Unemployment, Long-term Employment Relations and the Determination of Unit Labor Cost Growth in US Manufacturing Industries', International Review ofApplied Economics, 3, pp, 125-47. Santoni, G. (1986), 'The Effects of Inflation on Commercial Banks', Federal Reserve Bank ofSt Louis Review, March, 15-26. Sawyer, M. (1983), Business Pricing and Inflation (London: Macmillan). Scharrer, H.E. (1980), 'Currencies and Currency Hedging in German Foreign Trade', Studies on Economic and Monetary Problems and on Banking History, 18, Deutsche Bank. Schor, J.B. and S. Bowles (1987), 'Employment Rents and the Incidence of Strikes', Review of Economics and Statistics, 69 (4), 584-92. Shepherd, W.G. (1982), 'Causes of Increased Competition in the US Economy, 1939-1980', Review of Economics and Statistics, 64, 613-26. Sims, C. (1980), 'Macroeconomics and Reality', Econometrica, 48, 1-48. Summers, L.H. (1983), 'The Nonadjustment of Nominal Interest Rates: A study of the Fisher Effect', in J. Tobin (ed.), Macroeconomic Prices and Quantities (Washington, DC: Brookings Institution), 201-41. Sylos-Labini, P. (1979), 'Prices and Income Distribution in Manufacturing Industry', Journal of Post-Keynesian Economics, 2 (1), 3-25. United States President (1988), Economic Report of the President (Washington, DC: Government Printing Office). Wachtel, H. (1987), The Money Mandarins (New York: Pantheon). Wallace, H. (1977), 'Inflation is Destroying Bank Earnings and Capital Adequacy', Bankers Magazine, Autumn, 12-16. Wallich, H.C. (1975) Statement to Congress, in Federal Research Bulletin 6 (7) July, 411-413,6 (8) August, 480-486, 6 (12) December, 848-851. Weisskopf, T.E. (1978), 'Marxist Perspectives on Cyclical Crisis', in Union for Radical Political Economics (ed.) US Capitalism in Crisis, Economics Education Project (New York: URPE), 241-260. Weisskopf, T.E. (1979), 'Marxian Crisis Theory and the Rate of Profit in the Postwar US Economy', Cambridge Journal of Economics, I (3),341-78. 28 The European Plan for the Creation of a Single Currency Suzanne de Brunhoff* It will be recalled that the older economists of the nineteenth century were internationalist and generally favoured a world currency. (R. Mundell)· It has been an inherent characteristic of the automatic international metallic currency ... to throw the burden on the countries least able to support it, making the poor poorer. (Keynes)2 Since 1970, the instability of flexible exchange rates, the development of international finance unrestricted by national borders, and the difficulties of national monetary policies have inspired numerous reform projects. With the dollar effectively our global currency for want of any other, the question of a world currency, as a unit of reference and of payment, frequently arises. 'Get ready for a world currency,' runs a headline in The Economist (911/89). This chapter is a commentary on two EEC projects, carried out in 1989 and 1990 respectively, that examine the prospect of a single currency for Member States. European Monetary Union seems to take precedence over matters relating to the budgets of Member States, or joint social measures. We will discuss not only the problems posed by the creation of a single currency, but also those resulting form the very fact that priority is given to the subject. Any events that could thwart the projects of 1989 and 1990 are disregarded here. 1 A SINGLE EUROPEAN CURRENCY The European Monetary System, in operation since 1979, appears to have lost effect. It does not involve all the Member States, and above all seems to be a sort of 'mark area'. Of the documents studied here - 'Report on Economic and Monetary Union in the European Community' (Brussels, 1989),3 and 'Economic and Monetary Union: The Economic Rationale and Design of the System' (Brussels, 1990)4 only the final stage will be presented: that of a single currency issued by a supranational central bank, Eurofed. The objective, in the context of a single market-oriented economy, is that of a common monetary policy. 716 Suzanne de Brunhoff 717 The Choice of a Common Monetary Policy In the second report (pp. 4-5), three possible perspectives are successively ruled out: l. 2. 3. 'Monetary Union with Competing Monetary Policies': This is one of the features of the proposal put forward in November 1989 by the UK Treasury'. Yet with each country trying to exercise its own monetary sovereignty, the system would have no irrevocable fixing of exchange rates, and could fall apart should disagreement arise concerning the fundamental objective of price stability. 'Monetary Union without Economic Union': 'This is to a large extent the approach followed by the Advisory Council to the Federal Ministry of Economic Affairs in Germany.' The following criticism is made in the report: 'The lack of an economic policy complement to monetary union could lead to an excessive burdening of monetary policy.' Furthermore, 'the Community monetary institution would have an over-dominant role'. This point will be dealth with later on. 'Monetary Union with a Centralized Economic Union': The report considers that this would presUppose a form of supranational centralization which 'does not reflect the political preferences of the Community' ... 'For example, there is no need to transfer to the Community level large expenditure functions of public budgets, such as defence and social security.' In fact, even if the second perspective is discussed, and economic union does involve some Community budget expenditure, monetary union is none the less favored as a means of using the market to strengthen integration within the Community. European Currency and Eurofed Which Single Currency? The future European currency would be the ECU (European Currency Unit) - at present a sort of basket of EMS currencies - which would be transformed at the opportune moment. Two questions arise: that of the monetary sovereignty of Member States, and that of the form the single currency would take. The Brussels Commission passes over any propositions similar to the Keynes Plan (1940-4), no matter how these may be interpreted by various authors. Only external transactions are carried out in a common currency, which determines the exchange rates of the national currencies of Member States; the latter are maintained for internal transactions. B. Schmitt (1988) perceives the Member States as regions of production and exchange, and the international arena, where only 718 Creation of a Single European Currency exchange relationships take place. This is not the EEC's viewpoint, although the two reports studied here do not define a European production area. This implies that the single market, strengthened by a common currency, lies at the heart of economic integration. TheNewECU The single currency or the new ECU, representing the final stage in monetary union, would replace the national currencies of Member States. However, the actual notion of a currency is not clearly defined in either of the texts, and can only be inferred from passages dealing with the advantages that the Brussels Commission attributes to a single currency. The first advantage would be the elimination of transaction costs and exchange rate risks within the Community; the second is that a boost would be given to European financial market integration. Also, on an international level, an 'Ecu area' would give Europe a more important role to play: at present, it is the dollarmark-yen relationship which is taken into account. As a single currency, the Ecu would also be the means of payment and unit of account of the new monetary zone. First and foremost, the Ecu should ensure price stability within the Community - in other words, it should be a stable unit of account. However, the current texts do not reveal the form this unit of account would takeS, nor is there any analysis of the origins of inflation. One therefore has to resort to a quantitativist conception of monetary stability, dependent on the management of the money supply by the Ecu issuing bank. The question then shifts from the currency to the monetary institution,6 the model for this being the Bundesbank. Eurofed: Independence and Creditworthiness Before going on to Eurofed itself, it should be said that both reports underline the need for monetary policy to impose a certain discipline on the financing of national budgets. There should be no monetary financing of public deficits. Neither should public borrowing be compulsorily subscribed by certain institutions; it should depend on the financial markets. One none the less wonders whether this could not lead to heavy competition between borrowing States to the detriment of their weaker neighbours, according to the difference of 'creditworthiness and financial power' (Padoan, 1986) that would subsist. (b) The central European bank, Eurofed, would to some degree be federal in nature. It would comprise two bodies: a Council, consisting of the twelve governors of the national central banks of the EEC, and certain members of a second body, the Board. The Council's task would be to define 'the stance of monetary policy'. The Board would be made up of specialists whose appointment would be long term and would not coincide with (and thus be (a) Suzanne de Brunhoff 719 influenced by) general elections. It would be responsible for the day-to-day running of EU monetary policy. Everything depends on the national central banks and Eurofed, despite the fact that 'the institutional organization of financial intermediaries varies dramatically across countries'. (Grilli, 1989) It is not clear what the role of the national central banks would be if a single currency came into force. In the second text, their role is seen as one of banking policy, involving relations with national financial institutions, 'business interests', supervision and surveillance of markets and so on. They would be responsible for 'the smooth functioning of the national system of payments'. However, there appears to be no question - neither for the national central banks nor for Eurofed-of the central bank acting as a lender of last resort in the event of a slump like the October 1987 crash. Another point which is not entirely clear is that of the fixing of the various central banks' intervention rates with a single currency in place. It is difficult to see how these rates could remain decentralized. This question will be discussed when we look at the fixing of an exchange rate for the European currency. In any case, the second text states that Eurofed could have all the necessary tools of monetary policy intervention, notably those of the open market. The report suggests that a secondary objective be defined as a reference point. We come across here (albeit indirectly) the question of the 'nominal anchor' of the unit of account. 7• At any rate, Eurofed would have exclusive control over the issue of Ecus and ultimate responsibility for the European payment system, which would imply a centralization that would modify the role of the national central banks. On an international level, it would be the responsibility of Eurofed to manage the exchange rate of the Ecu in relation to currencies of non-EEe countries. In doing so, it would respect 'the final objective of monetary union. i.e. price stability'. The inherent difficulties of the fixing of a balanced exchange rate (the notion of an exchange rate being no more clearly defined in the text than that of a unit of account) are all too familiar. The problems encountered on a national level end up applying to the whole region, and are further complicated by the differences between Member States. One of these difficulties is linked to the creation of a European trade balance, in the absence of conformity among the various national economic policies. Furthermore, the question is not just one of monetary union in Europe; it is also important that the Ecu should have comparable status to that of the dollar and the yen. Finally, the danger of creating monetary 'target zones', as suggested by Williamson (1986), is two fold: first, the confrontation of monetary 'blocs' and second, that of a hierarchic monetary system (Padoan, 1986), divided between 'core nations' belonging to the monetary zones, and 'peripheric countries' .8 In the two reports, the objective of the creation of a single European currency, which would strengthen the integration of Members States, is accompanied by a 720 Creation of a Single European Currency number of suggestions for economic cooperation. None the less, the single currency project is given pride of place and is disconnected from a common fiscal policy. The principal objective is that of a stable currency; though the definitions of unit of account and exchange rate remain unclear. Everything hangs on the creditworthiness of Eurofed. We will now turn to the limits of this perspecti ve. 2 THE LIMITS OF EUROPEAN MONETARY UNION Democratic Deficiency Eurofed is designed as an independent central bank for the governments and authorities of the EU, and is thus deemed creditworthy. We have looked at the privileged role it would have, in particular with regards to budgetary policy. The criticism is sometimes made that Eurofed enjoys too much power without democratic control (by this we mean representative democracy); hence the notion of a 'democratic deficiency'. In the first of the two reports, a form of post-monitoring is suggested: reporting would be in the form of an annual report of Eurofed to the European Parliament and the European Council. Its administration could be supervised by an independent committee; this latter suggestion, however, is not to be found in the second report, which proposes instead that the Presidents of the European Commission and the European Council be present at meetings of the bank's decision-making group. Yet this has more to do with the liaison between monetary and European policy than with democracy. One of the problems put forward relating to Eurofed concerns the role of the European Parliament, the limits of whose power are generally acknowledged. However, on a national level, in a country like France, the institutions of the Fifth Republic and the way in which political practices have evolved over the years mean that Parliament has restricted power. Monetary policy is independent of it. Even if there were important changes at European Parliamentary level, it would still not necessarily solve the problem of Eurofed's democratic deficiency. It was mentioned earlier that Eurofed is designed in such a way that the appointment of its directors would not be affected by general elections. One of the arguments supporting this is that the credibility of monetary policy depends on competence and on the acceptance of rules such as the overriding need for a stable currency, and not on democracy. In any case, the possible discontent of economic agents towards a monetary policy considered too lax, or else too restrictive, is not necessarily reflected at electoral level - as factors like capital flight, wage claims and the currency black market show only too well. The idea of Eurofed's 'democratic deficiency' seems more likely to come from criticism of a supranational monetary discipline that is too remote from choices concerning budgetary policy, and from growth requirements. Suzanne de Brunhoff 721 Budgetary Policy and Welfare Payments In the second report, the Bundesbank and especially the American Fed are put forward as blueprints for the more or less federal character Eurofed could have. But the comparison is only limited, as the activity of the German and US central banks is accompanied by national budgetary policy, whatever the budgets of their separate regions or states may be. Budgetary policy doubtless declined in importance during the 1980s, monetary policy, despite its limits, being shifted to the centre of economic policy. In either case, the situation is not the same. In the USA, budgetary policy has become relatively uncontrollable - as that nation's large debt indicates. On the other hand, German budgetary policy - after the 1979-82 episode - has been subjected to the discipline of monetary policy, which for several years created deflationary pressure, despite a high unemployment rate. But the existence of a national budgetary policy enables movement between regions. It should also be pointed out that in Germany, as in most of Western Europe, social policy is important. One question that arises at EU level is that of how to maintain social policy, bearing in mind the prospect of monetary union without budgetary union. The two reports dismiss the creation of a Community budget of any importance, which would be responsible for welfare payments - while the pressure on wages threatens to rise. Examining the case of Europe on the basis of the American example, M.J. Peck (1989: 293) raises the question of 'the impact on wages of increased competition in the product market .... Deregulation and increased import competition ... pushed wages down ... in diverse industries'. The EU could face the same risk. Common monetary discipline would further be called for to reduce welfare payment in national state budgets. Social inequality, having increased considerably throughout the 1980s, is in danger of rising even more over the next decade. Another problem is that of the closing of EU borders to all non-EU workers, with certain exceptions. To justify this measure, attention is drawn to the fact that the level of unemployment is still high. However, neither of the texts anticipates a particularly generous common policy towards poor, debt-burdened developing countries. Furthermore, there is much talk about penalizing employers exploiting an illegal workforce that has no social cover; yet current measures do not prevent the existence of a black market dominated by the employers. The 1990 report raises the issue of the adaptation of the labor market. It suggests the encouragement of the best practices for labour management. This concerns especially mobility, permanent education, training systems and the introduction of new technologies and processes of production .... The Community will also seek to give substance to the Social Charter, while enhancing the capacity of national labour markets to adjust to competitive pressures in an efficient way. 722 Creation of a Single European Currency It can be seen that the question raised by MJ. Peck, and by those who are concerned about the future of welfare payments, relates to market efficiency, which of course affects the situation in the Member States. Meanwhile, the Social Charter remains without substance. Limits of the convergence of economic policies The planned monetary centralization does not take into account the national diversity of EEC Member States whose size, resources and relationship with the outside world vary enormously, to say nothing of political and institutional difference. This is no doubt why K. Pohl, President of the Bundesbank and of the board of governors of the central banks of the Community, suggested (Le Monde, 13/6/90) undertaking the monetary union of the five countries that already share similar objectives: Germany, France, Belgium, Luxembourg and the Netherlands. In the 1990 report, the question comes up of the differences between national budgets, and of three models for the relationship between them: 'impulsion', 'cooperation' and 'cohesion'. The Community and its budget are implied here to a greater extent - though with prudence, the text being the result of a compromise between widely differing tendencies (p. 10). The rules applying to the financing of budgetary deficits have already been seen, as well as factors relating to impulsion in the case of labour markets. Cooperation, which is above all defined by a common respect for methods of financing, could include expenditure on infrastructure, for instance. Such expenditures could be financed by Community borrowing, within prudent limits. Cohesion would go even further. 'It is essential to ensure that the beneficial effects of economic and monetary union are felt in all parts of the Community .... In the final stage of EMU, there might be a need to further strengthen Structural Policies. Instruments and resources would be adapted to the needs of union' and would be 'in favour of widening eligibility criteria for support to the least favoured Member States'. Finally, a 'shock absorber mechanism' would be set up in the event of economic crisis. Cooperation and cohesion would therefore affect the Community1s budget, which would be increased selectively while respecting 'the stability-orientated monetary policy'. Meanwhile pressure would be exerted on the size of national budgets. 'The Community should indicate a presumption in favour of budgetary limitations to ensure that deficits would be sustainable without undue pressures on either national or Community interest rates' (1990 report: to). As social expenditure - as well as many other forms of expenditure - would remain the responsibility of national budgets, and as there is no mention of a common taxation system, it can be seen that despite the budgetary projects discussed in the reports, a common monetary policy, Eurofed and the single currency none the less take precedence. Thus, even though they are unable to restrict any inequality - they could end up sanctioning it. The disparity of economic power between Member States would lead to a disparity of political power. Any patterns seen early on (the difference between Germany and Portugal, for instance) would be repeated - or exacerbated - at the final stage. Suzanne de Brunhoff 723 Unless there is a creative boost to the market and to the single currency, together with secondary economic measures, there will be a danger of political tension and of national (or even nationalistic) reactions, especially in the event of economic crisis. The discrepancy between the schemes for monetary and for economic union, as presented in the two EEC texts, reminds one of the limits of the institution of a single currency. Yet the second report develops to a considerable extent the notion of cohesion, and that of a 'policy mix'. Also, new compromises will no doubt be negotiated. However, the only new Community structure recommended in 1990 is that of Eurofed, which would exercise greater monetary control than does the EMS8- which, in France's case, has made the franc dependent on the mark, the latter being managed by the Bundesbank in a more stable and creditworthy manner. Whence the economic and political limits of the EEC projects examined here. It is difficult to know when or how a single European currency issued by Eurofed will come into force, in view of the various upheavals of 1989 and 1990. If the scheme is carried through and the Ecu is transformed to replace the national currencies of Member States, a European monetary area would replace the mark area of the EMS. However, the relationship between Member States would remain unbalanced, with Germany in the lead and the other countries following behind. This disparity could be accompanied by an increase in social inequality throughout the Community, which would particularly affect wage earners as far as wage levels or social protection are concerned - even if wage disparity remains at its present level; but especially if it increases, of course. In this respect, a single currency, which in the EEC texts is favored for its stability resulting from its institutional nature, is also a 'social relationship', to use Marx's expression. It would impose a form of macroeconomic cohesion (in a 'quantitativist' sense) that would be neither politically nor socially neutral; and new conflicts could arise. (Translated by Nicola Jones) Notes I. 2. 3. 4. 5. 6. This paper was completed in 1991, before crises broke up the European Monetary System, and before the Maastricht Treatise was ratified. In spite of this, I think its argument is still valid. In 'A Theory of Optimal Currency Areas' American Economic Review, September, 1961 In 'Postwar Currency Policy', 1941, reprinted in the Collected Writings of 1.M. Keynes, Vol. XXV (London: Macmillan, 1980). Called 'The Delors Report'. Called 'The Christophersen Report'. The theoretical problem (in terms of an international monetary standard) is often confused with the technical debate, as follows: 'The monetary reform is a simple redefinition of units' (Giovannini, 1990). The role of European institutions is currently overestimated. 'One Money, One Market' , a study by the European Commission (1990), attaches too much relevance to the role of supranational bodies. 724 7. 8. Creation of a Single European Currency The choice of nominal anchors, in the neo-Classical framework, concerns strategies of the central bank, which can be defined in terms of an exchange rate rule or mon etary targets, or some combination. That question, however, leads to the non neoClassical problem of changes in the value of national currencies, whether they are pegged or not pegged to an external asset (for instance an international money). Kindleberger (1981) assumes that there can be only one international money (the dollar after 1945 and before 1971), produced by 'a darwinian uneven social process' and providing 'a stable monetary system in which one nation serves as leader'. International monetary management 'requires for its stability, the acceptance of the hegemony of the (n-minus 1) country and its policy', that is monopolistic hegemony (Llewellyn, 1990). 'At the present time, the world economic system is plagued with uncertainty .... The dollar is finished as international money, but there is no clear successor' (Kindleberger, 1981). Kindleberger thinks that a tripartite leadership would not succeed in providing stability. Other economists, however, assume that 'cooperation under oligopoly' is a possible route to a stable international system (Guerrieri and Padoan, 1988). At the present time (1991), the EMS is 'a regional subsector of the global monetary system', or a 'regional bloc'. Besides, it includes an internal asymmetry. 'Although Germany is a member of the EMS, it behaves like an independent floater in that the other EMS currencies are pegged to it while the exchange rate is effectively determined by Germany' (Lewellyn, 1990). It is doubtful whether a single currency (ECU) would put an end to that hierarchy. References Aglietta, M., A. Brender, and V. Coudert (1990), Globalisation financiere: I'aventure obligee (Paris: Economica). Cohen, B.1. (1977), Reorganizing the World's Money (New York: Basic Books). Commission of the European Communities (1990), One Market, One Money (Luxembourg: Office of Official Publications of the European Community). Eichengreen, B. (1990), 'One Money for Europe? Lessons from the US Currency Union?', Economic Policy 10. Giovannini, A. (1990), 'The De10rs Report and the Problems of Transition towards Monetary Union' De Pecunia, October. Grilli V. (1989), 'Financial Markets and 1992' , Brookings Papers on Economic Activity, N.2. Guerrieri, P. and P. Padoan (eds) (1988), The Political Economy of International Cooperation (London: Croom Helm). Kindelberger, C.P. (1981), International Money: A Collection of Essays (London: Macmillan). Llewellyn, D.T. and C. Milner (eds) (1990), Current Issues in International Monetary Economics (London: Macmillan). Padoan, P.C. (1986), The Political Economy of International Stability (London: Croom Helm). Peck, M.1. (1989), 'Industrial Organization and the Gain from Europe 1992' Brookings Papers on Economic Activity, No.2. Schmitt, B. (1988), L'Ecu et la Souverainete nationale en Europe (Paris: Dunod). Williamson, J. (1986) 'Target Zones and the Management of the Dollar', Brookings Papers on Economic Activity, No. 1. Wyplosz C. (1990), 'Les implications budgetaires de l'union monetaire', Revue de /' OFCE,33. Afterword: Why and How to Replace the Microeconomic Theory of Money Ghislain Deleplace and Edward J. Nelli Our Introduction to this volume ended with a section entitled: 'Convergence Towards a Common Research Program'. The absence of a question mark revealed the generally optimistic inclination of the editors, but, after 28 diverse chapters, the reader may be unsure of the exact message and may feel dubious of our positive outlook. However, diverse as the chapters are, two minimum convictions are shared by all the authors who contributed to the volume: the present state of mainstream monetary theory is not satisfactory, for reasons rooted in its approach to money; and the critique of that approach provides material for another view. The most advanced strand of mainstream monetary theory today is a microeconomic theory of money. It will be worthwhile to summarize the reasons why it seems to be trapped in a dead end (section 1). These reasons prove not to be contingent, but stem from the framework offered to monetary analysis by General Equilibrium Theory (hereafter GET). Therefore it is useful to look at other frameworks of analysis, still in the mainstream tradition, which, in one way or another, keep GET at a distance. These will be scrutinized on those matters which in our Introduction to this volume we considered as central to the Post Keynesian and Circulation approaches (section 2). Finally, the latter approaches will be contrasted to the various mainstream strands of thought on three fundamental issues concerning money (section 3). 1 THE MICROECONOMIC THEORY OF MONEY: A DEAD END This theory may conveniently be traced back to John Hicks' famous 'Suggestion for Simplifying the Theory of Money' (1935). Hicks called for a 'marginal revolution' which would consist in extending the theory of value to the theory of money, hence in focusing on the choice made by economic agents to use money: We now realise that the marginal utility analysis is nothing else than a general theory of choice, which is applicable whenever the choice is between altern a- 725 726 Afterword: Microeconomic Theory of Money tives that are capable of quantitative expressions. Now money is obviously capable of quantitative expression, and therefore the objection that money has no marginal utility must be wrong. People do choose to have money rather than other things, and therefore, in the relevant sense, money must have a marginal utility. (Hicks, 1935: 2-3) This suggestion implies a precept which is not innocent: the theory of money is a theory of the demand for money. Of course, this was already the case in previous monetary theories (e.g., the Cambridge Quantity Theory of money), and this would also be the case in other theories rejecting a micro economic approach to money (e.g., standard Keynesian monetary theory). But analyzing the demand for money in the framework of a theory of choice means that money is not only preferred to 'other things', but that it is preferred to them to perform the same function. To renounce using money to pay for goods or to store wealth cannot mean renouncing the activities of purchasing goods or owning wealth; otherwise, the choice would not concern the use of money but other economic activities. This means that if 'people do choose to have money rather than other things' , these 'other things' must be considered as whatever may be used to perform the functions traditionally ascribed to money: goods, bonds, contracts, etc. So the main task of monetary theory is to explain why and to what extent economic agents prefer to count, exchange and store wealth in money rather than in any other way that could be analyzed by the theory of value. A typical microeconomic theory of money then approaches the subject through the comparison between a non-monetary economy and a monetary economy. The research program is as follows: 1. 2. 3. consider the above functions in an economy without money, as it is described by the 'real' theory of value; show that they are performed in a non-satisfactory way; derive the existence of a demand for money. This is the so-called integration of money in the theory of value, which means that money is introduced in a theory which is logically built without it (thus creating a dichotomy). This may look rather strange but is consistent with a long tradition of considering money as an answer to 'the inconveniences of barter' or a 'veil'. After 60 years, the results achieved by this approach are found disappointing by its proponents themselves (see for example Hahn, 1988, or Starr, 1989). A closer examination suggests that it is in fact a dead end, because it is left with two alternatives: 1. Either: the model of a non-monetary economy is assumed to possess all the resources needed to perform the functions which should be performed by money. In this case there is no room for money in the model. Hence there will be no demand for money, unless arbitrary assumptions are introduced to Ghislain Deleplace and Edward J. Nell 2. 727 depreciate the model and create some inefficiency in the non-monetary system that can be rectified by money. But these assumptions cannot be explained by individual preferences. Or: the model of a non-monetary economy does not incorporate some features which are considered essential to a market economy, and which are accounted for by money. In this case money definitely has a role, but such a model cannot provide the appropriate framework for a microeconomic theory of money: for, in a market economy so modelled, people could not choose not to use money. In modern mainstream economics, the most widely accepted model of a nonmonetary economy is the theory of value contained in the Arrow-Debreu version of General Equilibrium Theory (GET). It should be noted that alternatives I and 2 exhaust the ways to consider money in this model. Either GET is a correct description of a market economy (in which case 1 holds) or it is an inappropriate one (so 2 holds); there is no third option. In particular, one must discard the idea that, although the model works, it can be made to work better with money. One cannot improve on perfection. The eqUilibrium in GET is optimal, without the use of money. It is incoherent to suggest that an optimal outcome could be improved by introducing money. If money improves the outcome, then alternative 1 has been rejected, and some sort of imperfection has been assumed, which can be corrected by money. Trying to escape this fatal choice by drawing on the distinction between existence and stability of general eqUilibrium is not more helpful. The idea is that an eqUilibrium might exist in exchanges defined without money, but that the use of money would make reaching this eqUilibrium easier. But why? There would have to be shortcomings in the model, costs or inefficiencies in regard to transactions, which the use of money would remedy - in effect implying alternative 2 above. But such shortcomings have nothing to do with stability per se, but rather concern the implementation of transactions (see below). Thus the examination of alternatives I and 2 provides an exhaustive appreciation of the integration of money in GET. It happens that they have been illustrated by an attempt to introduce respectively the store-of-value function and the medium-of-exchange function of money in this model. The difficulties faced by both alternatives mean that neither textbook function of money has so far been able to provide the foundations for a theory of the demand for money in GET, i.e. for a theory of money as defined by Hicks. Besides, in a theory of value conceived as a theory of relative prices, the function of measuring prices is so obviously fulfilled by a commodity numeraire that the third textbook function of money is not even considered. On the contrary, the attempts to look for a third microeconomic route and to overcome these difficulties through a radicalization of the integration of money in a market theory (monetary laissez-faire) have exacerbated this non-monetary character of the measure of prices in GET. 728 Afterword: Microeconomic Theory of Money Let us quickly describe the reasons why no one should be happy with any of these three principal directions of research in mainstream monetary theory. Money as a Store of Value If Hicks' 1935 article is a standard reference for the microeconomic theory of money, Patinkin's 1956 book has been most influential for providing an analysis of money in the general-equilibrium framework. Among other insights, it brought into the light two points which would be decisive for later work in monetary theory. • • First, the integration of money in the theory of value cannot be achieved by specifying a commodity money with the transactions or cash-balance quantity equation and simply adding it to the system of eqUilibrium equations in the commodities markets; this attempt leads to a contradiction. Secondly, the emphasis put on the real-balance effect is only consistent with a demand for money as a stock, which presupposes that money in GET is a store of value. Hence the problem of money in GET has been identified during the following decades with the integration of a fiat money through its store-of-value function. But, as Frank Hahn pointed out: The most serious challenge that the existence of money poses to the theorist is this: the best developed model of the economy cannot find room for it. The best developed model is, of course, the Arrow-Debreu version of a Walrasian general eqUilibrium. A world in which all conceivable contingent future contracts are possible neither needs nor wants intrinsically worthless money. (Hahn, 1981: 1) Dealing with a monetary economy then requires to 'find room' for money in the GET framework. But this not enough. Suppose that an actor has established his reputation by playing young romantic heroes. He has been away from the stage for some time, and he wants once again to attract the applause of the audience. Therefore he will have to choose carefully the play for his comeback. First the plot and circumstances must provide a suitable venue for him: a play located in a retirement home will not do. Then he must be fit for the character: if the play takes place in Vienna at the time of the AustroHungarian monarchy, he'd better know how to dance the waltz. Finally he should check that no other character in the play (for example a seducer going grey at the temples) will steal the limelight. This is the same for money as a store of value in GET. Three conditions have to be fulfilled to make it a success. First, the analytical framework must be adapted in order to 'make room' for it. Secondly, one must be sure to avoid a Ghislain Deleplace and Edward 1. Nell 729 situation in which the price of money would be nil, thus demonetizing the model. Thirdly, money should not be outcompeted by another store of value. Let us examine these three conditions in a row. As suggested by the Hahn quotation, the reason why there is no room for money in the Arrow-Debreu version of GET is to be found in the existence of a complete set of contingent futures markets. By definition, these markets are held once and for all at the initial period in order to implement the intertemporal allocation of consumption and production; they never reopen. Hence there is no reason to carry over purchasing power into the future to make new transactions, because there will be none. There cannot be a demand for money as a store of value. This perfect model has to be made less perfect if one wishes to give a role to money. The way to do that is to assume either that the futures markets are missing for some commodities (defined by their physical properties, their location, their date of delivery and the state of nature when delivered), or that the operation of markets is costly, transaction costs being higher on futures than on spot markets. Then agents may be induced to use spot markets at each date, and such a 'sequence' economy may call for money. The existence of transaction costs on futures contracts is not by itself enough to 'make room' for money. These contracts have a great advantage: they allow agents to fulfil their budget constraintat each future date. However, because they are costly, this result is achieved at the expense of real resource costs, which represents an efficiency loss. Such costs will be worthwhile if the gain from fulfilling the budget constraint outweighs them. But money will be welcome if carrying it can provide the same result at a lower cost; a fiduciary money might be a good candidate - its stock diverts less resources than a commodity money. An appropriate structure of transaction costs on futures markets without money and on spot markets with fiat money is therefore the first condition to the integration of money as a store of value in GET. But of course to be held as a store of value money must have a value. If its value were nil, nobody would demand it and the room so carefully prepared for it would stay empty. Such a frightening possibility cannot be ruled out by simply putting money in the utility function of agents. In contrast with goods, the utility of money lies precisely in its value; but, according to the scarcity principle, one necessary condition for something to have a value is for it to have utility. With a commodity money, the price of money in equilibrium is tied to the price of the backing commodity, and it cannot fall to zero. But with an unbacked (fiat) money, one may not exclude a situation in which a zero-price for money would cause a demand for money equal to zero, which in turn would maintain a nil price ofmoney.2 This problem of the positivity of the price of money, launched by Hahn's 1965 article, is particularly acute in a model with a finite horizon. Since money is demanded as a store of value it must be worthless at the last period. As a result, 730 Afterword: Microeconomic Theory of Money however, it will not be demanded at the next-to-Iast period. But this will drive its price to zero. Its price being zero in the next-to-Iast period, it will not be demanded in the period before that ... and so on: the price of money is driven to nil in each period. Three solutions have been presented in the literature to overcome this difficulty. One is to assume that the money must be turned in at the end of the finite horizon, for example to pay taxes. This is obviously artificial; it prevents money from vanishing thanks to a function completely alien to the ones traditionally expected in a market economy. A second solution relies on the 'real-balance effect' in a temporary equilibrium model. In fact, this had already been proposed by Patinkin precisely to make room for money in GET. But Grandmont (1974) has observed that the real balance effect is not a sufficient condition to establish a temporary equilibrium with money, if agents have static expectations. It must be complemented by an intertemporal substitution effect, which requires that, if the price of money tends towards zero, the agents must expect that it will be positive (and even increase) in the succeeding period. However, no justification of such behavior is provided. The third solution has been widely exploited and consists in giving up the finite-horizon assumption, thanks to overlapping generations models. But this limits the use of money to intergenerational relations. Whichever of these solutions is chosen, this second condition for the integration of money as a store of value in GET (the positivity of its price) obviously requires imposing limitations on the generality of the models in which money can be introduced. Even if one accepts these limitations, however, they tum out not, in fact, to be sufficient to guarantee that there will be demand for a fiat money at a positive price! Hence the existence of such money in GET is not assured. The reason for this is simple. We have seen that if the price of a fiat money is nil, its demand is nil. But it does not follow that if its price is positive, its demand will be positive. As Starr puts it: Positivity is a necessary condition for usefulness. However, the property that money would be useful if its price were positive does nothing the ensure positi vity. (Starr, 1989: 295) If another asset outcompetes money in its store-of-value function, the demand for money will be nil, and its price will be driven to zero. Therefore a third condition is required: that there must exist a positive demand for money at its positive price, thus sustaining this price at a positive level. As Hicks had already observed in 1935, an alternative exists to money as a store of value: interest-bearing bonds. In a sequence economy, agents may purchase bonds spot and liquidate them later when they face a budget constraint; in fact they will prefer to do so, because the bonds bear interest, while government fiat money does not (the question of private banknotes will be studied below). This is the main critique of overlapping Ghislain Deleplace and Edward J. Nell 731 generations models: they establish the need for a store of value, but provide no reason why this store of value should be fiat money. One might imagine two ways to prevent money from being outcompeted by bonds. The first is again to call for transaction costs. This route is followed by Hahn (1988) in order to complement the Grandmont condition of positivity of price. One should note that mobilizing transaction costs to lead agents to discard the use of bonds in favor of money is less plausible than when such costs were called on to cause them to discard futures markets. As Hicks had already pointed out, discarding bonds in favor of money requires that the rate of return on bonds be lower than the rate of return on money. Primafacie, this means that the return on bonds must be negative! But there might be some sort of implicit return from money in convenience, or utility from liquidity. However, Hahn derives quite an awkward conclusion in respect to this issue: this condition - return on bonds less than the return on money - might be fulfilled for one agent and not for another, because 'there seems no satisfactory way of classifying assets by their liquidity which is independent of an agent's optimal plan and thus of his expectations' (Hahn, 1988: 964). Then how is it possible to justify that, because of its store-ofvalue property, all agents will always accept money in exchange for goods, as must be the case if money is to be universally acceptable? The other way to rule out bonds has been suggested by Wallace (1988Ho complement the overlapping generations models and 'oversimplify the theory of money'. It is indeed simple and radical: bonds will not be chosen instead of money because it is forbidden to do so. Governments impose 'legal restrictions' against private currency issue, because they want to be the sole beneficiaries of seigniorage. But remember the question was: why do agents prefer money to bonds as a store of value? Being twice off the mark is no guarantee for a right answer. In Wallace's theory, agents are admittedly not in a position to choose to use money or not, and governments forbid them to use bonds as a medium of exchange, not as a store of value. Summing up these difficulties found in the GET literature leads to a conclusion: GET provides no plausible foundation for a demand for money as a store of value. Moreover, the two last remarks suggest that attention must be shifted to another function of money: that of medium of exchange. Money as a Medium of Exchange Dealing with money as a medium of exchange amounts to asking the question: why do economic agents demand money in order to implement transactions? Neither the question itself nor the answer have a straightforward interpretation in GET. The reason is to be found in the treatment of market equilibrium in this theory. The Arrow-Debreu model determines the existence of the non-monetary equilibrium by setting the aggregate excess demand in each market equal to zero. It stops there. If one wishes to represent what is happening out of eqUilibrium (in 732 Afterword: Microeconomic Theory of MO(ley disequilibrium), one is naturally led to study a situation in which, for at least two markets (according to Walras' law), the aggregate excess demand in each is different from zero, and to analyze the conditions under which a process will make it evolve towards equilibrium (this is the problem of stability). The main condition is the implementation in each market of a rule of price adjustment according to the sign of the aggregate excess demand; the intervention of the auctioneer is supposed to ensure the following of that rule. This means that the 'practical solution' (in Walras' terms) to the problem of the determination of prices is completely identified with the solution to the stability problem, in contrast with the 'theoretical solution', which concerns the existence problem. The problem of the realization of transactions does not show up at all, in equilibrium or out of equilibrium. That is, the question of how diverse bundles of goods are actually tranferred from their various and many original owners to their equally varied final consumers is simply eliminated, thanks to the assumption that suppliers will deliver goods to, and demanders will collect goods from, a centralized location, termed by Walras, a 'chamber of settlement'. (In Debreu, 1959, this becomes a 'centralized system of accounts' .)3 This wholly ad hoc procedure is used to ensure that transfers proceed smoothly, either in equilibrium (tatonnement model) or in disequilibrium (non-tatonnement), and it guarantees the efficiency of the market (absence of non-zero individual excess demand on both sides or on the short side of each market, respectively). Of the two centralized but logically distinct features of GET, the auctioneer and the chamber of settlement, more attention has been devoted to the former than to the latter. This has led to the false impression that the analysis of the functioning of the market could be reduced to the stability problem as if, in equilibrium, no difficulty existed for the implementation of transactions. But, as long as transactions take place through a 'chamber of settlement', whether in eqUilibrium or in disequilibrium, there is no room for a medium of exchange between agents, because there is no such exchange: agents deliver goods to and obtain them from this institution. The chamber of settlement - not the auctioneer - appears thus to be the main obstacle to the introduction of money as a medium of exchange in GET. Its removal is justified if one feels unhappy with a representation of the market which does not contain one essential aspect: bilateral trade. But introducing bilateral trade is not an easy task, because it raises the spectre of a contradiction between market equilibrium (the very method of GET) and individual equilibrium. The reason for this is a consequence of the 'inconveniences of barter' in Smith's apologue of the butcher, the brewer and the baker; or it may be seen as following from the need for a 'double coincidence of wants' in Jevons' parable of the hungry tailor and the ill-clad baker. Whether it is inconvenience or the problem of double coincidence, bilateral trade causes difficulties; as a result, goods will be demanded for two different reasons: as use-values for consumption, and as means of payment for trade. By contrast, with a chamber of settlement, Ghislain Deleplace and Edward J. Nell 733 goods are only demanded for consumption, and the fulfillment of each individual budget constraint only requires that the value of the goods offered by an agent be equal to the value of the goods demanded by him. All individual budget constraints may be aggregated and the system of prices which clears the markets fulfills them automatically. But this is no longer true with bilateral trade: the budget constraint must be applied to each pair of agents, and it now concerns goods demanded for both consumption and trade. It has been shown by the work of Ostroy and Starr (see for example Ostroy and Starr, 1974) that these individual equilibrium conditions create an overdetermination which makes them inconsistent with the prices determined by the (aggregate) market equilibrium conditions. Various solutions may be imagined to circumvent this overdeterminacy, but they require some sort of centralization. A single medium of exchange (money) is the way to ensure the decentralization of trade, in the following sense: 'two traders, in deciding what trade to undertake, need only consult their own excess supplies and demands' (Starr, 1989: 44). This analysis has far-reaching implications. It shows that the presence of money is necessary to allow decentralization in bilateral trade; as such, it is a contribution to the theory of exchange in GET, taking a step towards doing away with the fiction of the chamber of settlement. But is it a contribution to a microeconomic theory of money? It could be so if it were proved that agents prefer to use money rather than other organizations of exchange because it is cheaper; but this is simply a petitio principi. As Starr puts it: Hence the use of monetary trade allows economizing on time and organization. Implicit in the treatment is the view that the real resource costs of the complex organization associated with centralized trade are prohibitive and that monetary trade is preferable. (Starr, 1989: 45; our emphasis) A role has been found for money in the market economy, but the existence of money is not deduced from a demand for money in Hicks' sense: people do not choose to use money in transactions. Money is a constituent part of the economy: although some room may be left for the choice by agents of which money to use, it will still be money, and the possibility to prefer exchanging through bonds or contracts is ruled out. If bilateral trade is decentralized, agents have no other alternative if they wish to trade. The choice theoretic approach to value cannot be extended to the medium of exchange. This conclusion is not only a negative one. Focusing on decentralized bilateral trading shifts the attention to monetary circulation: how can money accomplish all successive trades? Suppliers sell their goods or services for money, then pass that money along in producing or purchasing replacements, in order to supply again. The ground is then laid for a non-mainstream analysis of 'money in motion'. But before discarding the microeconomic approach to the theory of money, one should give it a last chance to capitalize on this notion of decentral- 734 Afterword: Microeconomic Theory of Money ization. The question is then refonnulated in the following way: if money allows a market economy to be decentralized, what sort of money may decentralize it best? This is the object of the laissez-faire approach to money. Monetary Laissez-Faire The failure of GET to integrate money in the equilibrium of a market economy is here interpreted as the consequence of the absence, not so much of money in the definition of a market economy, but of the market in the definition of money. The mistake was to try to introduce in market theory a traditional conception of money based on a kingly privilege which empowers the state to control the nature and the quantity of exchange media. On the contrary, it would be more stimulating to introduce in monetary theory (and, if the results are successful, in monetary practice) the traditional conception of the market. This is what Greenfield and Yeager (1983) called' A Laissez Faire Approach to Monetary Stability' . According to Selgin and White (1994), there are three ways in modem literature to answer the question: 'How Would the Invisible Hand Handle Money?'. The first one is 'free banking with a distinct base money', which in modem times originated in the work of Milton Friedman. As in conventional monetary regimes, a base money defines the unit of account and is the means of settlement between banks; it need not however be used by the public. The two main differences with current practice are the free note issue by banks and the absence of statutory requirements concerning the reserves in base money owned by note-issuing banks. Supporters contend that such a regime would provide an optimum quantity of money (Friedman, 1969). It is clear that this theory is not intended to solve the problems raised by the integration of money in GET, because it merely presupposes the existence of a base money. So doing, it is exposed to some of the above-mentioned difficulties. For example, in the absence of reserve requirements in base money, nothing would prevent banks from preferring to settle their reciprocal debts in an interestbearing asset. The demand for base money could then fall to zero. But, if this base money is a fiat money, a demand equal to zero will drive its price to zero, whereupon the unit of account vanishes. This result can only be avoided if the unit of account is not linked to a base money, but to a good deprived of any monetary function. To make sense of modem proposals for free banking, in view of the questions raised by the proposed integration of money into GET, it appears to be necessary to eliminate the unit-of-account function of money. This is precisely what is proposed by another strand of monetary laissez-faire, the so-called New Monetary Economics, which, according to Selgin and White, advocates 'competitive payments systems without base money'. The main source is Greenfield and Yeager (1983), who label their system 'BFH', in reference to previous works by Black, Fama and Hall. In this system, transactions between non-financial agents are made in exchange media (transaction accounts, notes) Ghislain Deleplace and Edward J. Nell 735 issued by competitive 'banks' (any type of financial firm). These exchange media are redeemable in any asset(s) agreed upon by banks, which also use it (them) to settle their reciprocal balances. All prices are measured in a unit of account which may be any chosen bundle of commodities, provided it is not used as medium of exchange, redemption or settlement: the unit of account is 'separated' from the payments system. All prices are freely adjusted by supply and demand, including the price of the redemption medium. The main advantage claimed for this system is that it would wholly eliminate the fluctuations of real activity produced by changes in supply and demand of base money, since there would no longer be such base money. The separation between the unit of account and exchange media can hardly be considered revolutionary because, as we noticed earlier, this is a typical feature of GET. It should not be expected that this assumption can, in itself, solve the problems faced by the integration of money in that theory. Two lines of critique can be distinguished. The first one focuses on imperfect or asymmetric information in the 'New Keynesian' way. If the market fails to reveal completely the respective qualities of the privately-issued exchange media, Gresham's law will operate: agents will try to force the bad ones into circulation, and keep the good ones. The working of the payments system will be affected; bankruptcies could lead to a shortage of circulating medium. If, on the contrary, information is perfect, as assumed by Greenfield and Yeager, a second critique holds: competitive payments systems are exposed to the same difficulties already mentioned about bilateral trade without money. On one hand, as long as nobody is forced to accept a medium of exchange issued by a particular bank, the need for a 'double coincidence of wants' is no weaker for 'competitive monies' than for goods. In fact, it is exactly the same need, because financial assets can be considered commodities like any others. But if such a 'double coincidence' is required, agents will be acquiring some exchange media to trade them for others, so their market equilibrium prices will in general not be consistent with what would be called for by the individual equilibrium budget constraints of the non-financial agents. On the other hand, even in the case of a single redemption medium, the market adjustment of its unit-of-account price which equalizes its aggregate demand (Le. the value of all the exchange media redeemed) and its aggregate supply (Le. the value of the reserves owned by all the banks) does not guarantee that each bank can face the demand for redemption addressed to it with its own reserves at this price. In the absence of a lender of last resort, bank failures are likely. It is now accepted in the literature that, because (as we noticed earlier) market equilibrium and individual equilibrium do not automatically overlap in GET, price flexibility of goods is not a sufficient condition to the viability of competiti ve markets without money. Similarly, price flexibility of exchange media issued by banks and of the redemption and settlement medium agreed upon by banks cannot be a sufficient condition for the viability of competitive payments systems 736 Afterword: Microeconomic Theory of Money without base money. Then the temptation is strong to drop any reference to GET altogether, and to root monetary laissez-faire in another vision of the market. Knowing the antipathy of Friedrich von Hayek towards GET, one would expect that his proposals for the 'denationalization of money' might go in that direction (Hayek, 1976). Focusing on the role of the relation between banking and the base money, Selgin and White list such as approach (with others) in a third category: 'competing non-commodity base monies'. Not being redeemable in anything, these competitive monies contrast with the above ones and may thus be considered as base monies. They also resemble fiat monies, although no state power enforces their use. Absence of reference to GET makes difficult an appreciation of the reasons why they would be demanded. According to Hayek, an appropriate definition of property rights protecting their 'brand names' would induce their private producers to pledge to hold their purchasing power constant in terms of a specified bundle of goods. Then they would offer a better protection against depreciation than commodity-based monies or government fiat monies, and would be preferred to them. But, in the absence of legal enforcement of the pledge and of redemption in the price-index basket, what would protect their users against further issuance without limit? A 'time inconsistency' problem appears here: it may be more profitable for the producers of monies to hyperinfiate once and for all than to honour their pledge and stay in business. As a consequence, this sort of money would have a zero-price in equilibrium. Although the route is different, a typical difficulty of the treatment of fiat money in GET shows up again. This overview of the present state of the microeconomic theory of money is instructive about the direction in which another approach to money might be found. Two main ideas have emerged. The first one is that the existence of money in a market economy is to be linked with the multiplicity of individual budget constraints, whether at a moment of time (because of decentralization) or through time (because of sequentiality). This revives the 'monetary constraint' put forward by Robert Clower in his 1967 article, as distinct from the single budget constraint. Unfortunately, the market disequilibrium models which have been derived from it have merely reformulated the budget constraint to incorporate spillover effects, and forgotten the main message: being a pre-condition of the implementation of exchange, money cannot be deduced from a theory of exchange without money, which at most may justify it. This begins to be recognized even by advocates of the microeconomic approach: 'The "Clower constraint" is not easily deducible from this kind of micro-theory. That does not mean that no story can be found or that it should be banished. But it would be nice to see it justified' (Hahn, 1988: 964). The radical conclusion which this observation calls for is to give up the method of comparing the working of a monetary economy with that of a non-monetary economy. But a positive develop- Ghislain Deleplace and Edward 1. Nell 737 ment along this line requires going beyond Clower's assumption that the monetary constraint applies equally to all agents, and discriminating between firms' and households' access to money. Because they sharply differentiate between firms and households, and subdivide households into classes, as well as stressing the endogeneity of money and/or finance, the Post Keynesian and Circulation approaches come into the picture at this point. The second idea which has emerged is the specificity of money: its role cannot be played by another asset or another organization of exchanges. But the explanation by lower transaction costs falls short, and this is not surprising in a theory focusing on the demand for money and neglecting the conditions of its production. The so-called 'New Keynesian' approach has developed micro-founded monetary macro-models based on capital markets with imperfect or asymmetric information, and it contributes to the analysis of the supply of credit and finance, but not so much of money, thanks to Stiglitz's precept that 'it is not money that makes the world go round, but credit' (Stiglitz, 1988: 320).4 Nevertheless, the question of money circulation cannot be left aside: if trading is bilateral and decentnilized, one has to understand how money may accomplish all successive trades. The microeconomic approach can only display the fact that an agent has received money from another agent, who had himself received it from a third one, and so on. But money creation remains outside the scope of theory. Here again, the Post Keynesian and Circulation approaches have something to say. Thus it happens that both ideas, which cannot be deepened in a microeconomic framework, are linked with constituent parts of the Post Keynesian and Circulation approaches. Instead of being treated grudgingly, putting money in a model which does not need or care for it, these two develop a conception of money as a basic feature of their models from the outset. Hence the bet that they may make an important contribution to the understanding of money. 2 COMPETING FRAMEWORKS FOR MONETARY ANALYSIS Professional economists have long dreamt that they might awaken one day to find themselves circulating in a stable, optimal and unique, garden of general equilibrium, in which they could move as they preferred through the use of money. But in the past two decades the fraternity of dreamers has been rudely awakened to find itself facing a much rockier landscape, covered with the debris of brokendown theories. For instance: • The Queen of all theories, General Equilibrium theory itself, appears almost to have self-destructed; it cannot accommodate money at all easily, it cannot be shown to be stable, except on strong and implausible assumptions, it tends to have multiple solutions, with no easy economic interpretation, and it is widely considered to be too abstract. 738 Afterword: Microeconomic Theory of Money A return to partial equilibrium thinking has taken place among many researchers. in particular those dealing with policy issues. • Among policy analysts. even those who accept the mainstream vision. there have developed attacks on aspects of the idea of equilibrium itself. as many have expressed dissatisfaction with the assumption that markets clear and that the market pressures which move economic variables arise from the process of market-clearing. This has led to focusing on 'internal' markets (e.g. the firm itself). or to introducing particular institutional features which affect market adjustment. e.g. credit restrictions. menu costs. or rigidities of various kinds. In either case. building a general theory appears more and more out of reach. and the retreat from equilibrium has ended up in bubbles and instability. leading to models of chaos. • Finally. all across the landscape of contemporary economics. there is a general dissatisfaction with the treatment of money. which apparently cannot be integrated into the mainstream theory of a market economy and. when considered apart (in relation to finance). finds expression in clashing schools of thought. The present volume has shown that contributions from the Post Keynesian (PK) and the Circulation (CA) approaches could build a research program which. so we think. could help to improve this situation. At the very least the program offers the prospect of providing a theory that integrates money. production and employment. Let us try to compare it. point by point. to the rest of the rubblestrewn research landscape. We shall consider only those schools of thought which have a distinctive theory of the articulation between the real sector and the monetary and financial sector of the economy: General Equilibrium theorists (GE). Disequilibrium Keynesians (DK). New Classicals (NC). New Keynesians (NK). Monetarists (M), Neo-Austrians (NA). and Neo-Ricardians (NR) Those other authors. e.g. Marxists. Regulationists, and the business cycle theorists most often associated with Goodwin who. on questions of monetary theory. are closely related to either the Post Keynesian or the Circulation approach. are not considered separately here. GE theorists seek to explain prices and outputs. for all agents. in all sectors of the economy on the basis of given endowments. given preferences and the established technological possibilities. Agents engage in maximizing behavior. reacting to parametrically given prices. Prices and outputs are determined simultaneously (in sharp contrast to the Neo-Ricardians.) Like NR. however. this approach develops its fundamental models without money. which is integrated at a later stage. But we have seen above the difficulties faced by this integration. While accepting the basic theoretical stance of GE, Disequilibrium Keynesians (DK) attempt to overcome its practical defects by introducing money and with it. constraints on buying and selling - the 'short-side' rule. They take some price or other - often money wages - as fixed. and then show that the resulting pattern of Ghislain Deleplace and Edward J. Nell 739 market adjustments will give rise to various kinds of systematic and persistent disequilibria, including 'Classical' and 'Keynesian' unemployment. Unfortunately, the 'cash-in-advance' constraint, also used in GE models, runs counter to economic rationality: if it existed in a system, someone could make money by arranging to lift it. One might invent banks, for example; that is what they do. New Classicals (NC), on the other hand, simply brush aside the defects of GE, and, using the device of representative agents, concentrate on partial analysis. They solve many of the most difficult problems, such as whether markets clear or are stable, in the most direct and effective manner - they assume the difficulties away. They then proceed to show, with devastating clarity, that if markets clear and equilibria are optimal, then factors will be fully employed and fluctuations can be nothing other than the best responses possible in the face of exogenous shocks. New Keynesians (NK) don't believe this, but since they also partially accept the GE framework, and otherwise often rely on representative agents, to obtain Keynesian results they are obliged to show that small but realistic deviations from the assumptions will generate large problems. As a consequence they are chiefly concerned with the micro foundations of macroeconomics, that is, with why there should be unemployment or various kinds of inflation - or other market disorders - when markets reflect the rational choices of optimizing agents, and tend to bring about the optimal allocation of scarce resources. If this is what markets do, if this is their social function, then how can they be responsible for unemployment? The answer is that markets systematically fail to achieve their objectives, owing to a variety of imperfections. These can be grouped under three headings: nominal rigidities include 'menu' costs, 'sticky' prices or money wages, fixed contracts, and adjustment costs; instrument uncertainties cover credit rationing, asymmetries in attitudes to risk, efficiency wages and various other kinds of uncertainties; suboptimalities concern the way small deviations from optimality at the micro level can translate into large macro effects. Exploring the relations between these is the province of New Keynesian theory. Monetarists (M) also pay lip service to the framework of GE, but don't actually use it. In practice they deploy one or another variant of the Keynesian IS-LM system, modified to allow for a version of the Quantity Theory. In the long-run variations in money affect nominal values; real variables are determined by real forces. In the short-run, however, changes in the quantity of money may affect real variables, and may do so directly, not just through the rate of interest. Neo-Austrians, (NA), like the Monetarists and the New Classicals, tend to be opposed to intervention and policy activism, but like the Neo-Ricardians, are critical of the GE framework and suspicious of claims that markets tend to equilibrium. Economic behavior is seen as the working out in practice, in the material world, of the process of rational choice. Such processes take place· over time; behavior develops and unfolds - later choices are conditioned by earlier ones. 740 Afterword: Microeconomic Theory of Money Markets are imperfect institutions which enable this to take place. Because behavior takes place over time and is path-dependent, and because markets are imperfect equilibrium is unlikely. Yet markets are forums designed to permit maximum individual free choice; hence intervention is liable to make things worse by restricting the free choices of some parties. Neo-Ricardian (NR) authors consider the basic distribution theory of neoClassical analysis to be flawed; hence they hold its account of costs and supply to be incorrect. This is the message of the Capital Critique. NR therefore proposes an alternative set of equations, relating prices to the rate of profits, given the production coefficients (the choice of which is also explained) and either the real wage, or the rate of profits. There are no 'endowments' in the NR system; means of production are produced, except for land, differences in the quality of which allow different methods of production to be, used simultaneously, giving rise to rents. Demand is not examined; the proportions of output are taken as 'given', that is, as parameters determined by processes explored in other branches of theory. Money is introduced by arguing that the rate of profits is determined by the rate of interest on money. A higher rate of interest drives up prices relative to money wages, thus lowering the real wage and raising the rate of profits; vice versa for a lower rate of interest. There are three orientations towards the basic neo-Classical vision here, with a number of cross-currents: • • • M, NC, and GE all represent variations on the basic Neo-Classical themes of market-clearing optimal equilibria. M sometimes adopts aspects of a Keynesian macro model, however, which both the New Classicals and GE reject. NK and DK both seek to support versions of the basic Keynesian results, in which markets tend to generate unemployment, but DK adopts a modified GE model of the economy, while NK usually adopts a partial equilibrium micro framework, and seldom ventures into macroanalysis at all. NR and NA both dissent from the Neo-Classical approach because they regard it as fundamentally misrepresenting the way the economy works; their positive analyses, however, are not at all close. The chapters in this book have developed a number of important areas in which common questions or similar analytical approaches tended to bring the Post Keynesians and the Circulation writers close together. Let us now look at the most important of these, to see how they are treated by the various mainstream schools of thought. We will begin with two issues concerning method, sequential analysis and uncertainty: Sequential Analysis We found that both PK and CA drew on sequential analysis, PK in connection with the multiplier and business cycle studies, CA in regard to the circulation of Ghislain Deleplace and Edward J. Nell 741 money. For both schools, a sequential analysis provided a convenient way to model out-of-equilibrium movements. For CA and for some PK it represented a step towards a new kind of analysis, not bound to equilibrium but not needing the detailed and particular assumptions that are usually necessary to permit fullfledged dynamics to reach reliable conclusions. The other schools, however, have shown little interest in sequential analysis. It appears from time to time but only in special circumstances, usually as a matter of dynamics or stability analysis. • • • • • GE: it has developed its own form of sequential analysis, reasoning primarily in terms of intertemporal equilibrium. High transaction costs are assumed to prevent the adequate formation of futures markets, requiring sequences of temporary equilibria. Such 'sequence economies', which eventually converge to long-term full equilibrium, are inefficient. Money enables trades to take place more smoothly in sequence by allowing the transfer of purchasing power over time, at minimal cost, thus satisfying the global budget constraint. No circuit is involved, however. NK and OK: Nominal rigidities may create what appear to be sequential patterns in the movements of prices and quantities, but in general no processes are involved. The objective is only to determine the equilibrium levels. The sequential movements are largely illustrative. In Clower's approach, the cashin-advance constraint generates an order of events, but the implicit sequence is largely undeveloped. The assumption is that it will proceed to equilibrium. M: It considers money multiplier sequences, and also examines sequences of responses in the case of the short-run Phillips Curve. But such cases are considered fluctuations around long-run equilibria. NR: Reliance on the 'long-period method' commits NR to the assertion that in the real world there must be tendencies to move in the direction of the long-period position. Hence convergence to such positions can be modelled. This is a dynamic process, and a form of sequential analysis, but it is designed to show convergence to an equilibrium (though not a market-clearing) position. It is an auxiliary analysis, not central to the method. All others: The rest do not find even a limited place for sequence analysis. Uncertainty Both PK and CA consider uncertainty to be the basic condition under which economic decisions are made. Moreover, both consider that uncertainty cannot be 'handled' by assigning probabilities to various possible outcomes. Both discuss two kinds of basic uncertainty - generalized uncertainty about the future involving technology, politics, etc. and uncertainty specifically about future market developments. By contrast most of the other approaches reject non-probabilistic uncertainty as a feature of the world which must enter into economic modelling. 742 • • • • • • Afterword: Microecollomic Theory of MOlley GE and NR: Both reject basic uncertainty. Both see the economy converging towards a definite position - equilibrium for GE and a long-term position for NR (which need not be market-clearing); such convergence requires that agents be able to obtain the appropriate information to make their decisions. Although NR does not use the expression, its analysis of the persistent forces leading to gravitation towards long-period positions is very similar to rational expectations. Moreover, the fact that such positions exist towards which the economy is assumed to be moving, requires that markets must have a certain degree of stability. NK: Here the concern is with imperfect information. The relevant information could be known, but institutions stand in the way. There is nothing wrong in principle with the idea that a general equilibrium exists towards which the economy will move, but in practice there. are serious difficulties. Basic uncertainty, however, is not one of them. NC: By contrast, the information can be known, and can be handled in a probabilistic way; this is regularly done in models of rational expectations. M: There is asymmetrical information in the short run, but market pressures lead to corrections in the long run. Uncertainty is normal, but probabilistic. DK: Disequilibrium results from price rigidities; these are unexplained. Uncertainty could be part of the explanation, but the effects of rigidity rather than the causes are the central concern. NA: Agrees that the world is uncertain in the non-probabilistic sense, and further that the crucial questions concern 'market uncertainty'. Learning takes place continuously, as does technological change: knowledge will reduce uncertainty, but it is costly and takes time to acquire knowledge; moreover, the state of knowledge changes all the time, so that information is continually becoming obsolescent. These various positions on information may be compared with PK and CA, who claim that there are no insurmountable obstacles to gathering information; there are no general reasons for information to be imperfect, other than 'that it is expensive to gather. The different positions of capital and labor may make asymmetries in information inesacapable, but such asymmetries are less important than the fact of uncertainty, which is the real problem - information about the future is simply not there to be found; there is nothing to know. Next consider the central idea of Keynesian analysis, that economies respond to, and only to, demand backed by money. Effective Demand Both PK and CA consider the economy to b~ demand-driven, that is, supply responds to the pressure of demand, specifically to expected demand backed by money. On the other hand supply does not automatically create a corresponding Ghislain Deleplace and Edward J. Nell 743 demand - although supply certainly creates corresponding purchasing power. (Neither PK nor CA examine value and productivity in detail.) But while supply creates purchasing power, such power need not be exercised. Demands can be divided into autonomous and dependent; the level of autonomous demand will determine the dependent demands, and together they will set the level to which output and employment will adapt. Nothing guarantees that this will emplov all the available resources or labor. With this basic idea only NR agrees. It is able to argue this since it bases value on reproduction and distribution. All the other approaches accept the basic neo-Classical doctrine that value rests on scarcity. Hence they hold that in the long run, in equilibrium, all available resources will be fully employed. (Otherwise, such resources would not be scarce; and if they were not scarce, they could not have a value. For basic optimizing theory tells us that if a resource is not fully employed in equilibrium, it must be assigned a zero shadow price.) However, in the short run, or in disequilibrium, the approaches taking Keynesian questions seriously all agree that effective demand matters and unemployment is possible. • • • • • NR: In the long run, the quantities of the system are determined by effective demand. There is no guarantee of full employment. In the short run there are too many extraneous influences to permit theorizing of a general kind, but it can be presumed that the system is demand-driven at that level also. (However, the system is sufficiently orderly that markets will eventually tend to gravitate towards the long-period position.) DK: Effective demand has a role in one of the unemployment regimes, the Keynesian, but not in the other, the Classical. NK: In the short run, there are rigidities which prevent effective demand from reaching the full employment level. In the long run, if the imperfections could be removed, the system would reach full employment. Before this hypothetical equilibrium is reached, however, the parameters of the system will shiftso the short run is the only revelant period. M: In the short run Keynesian demand relationships may hold, although demand is assumed to be chiefly determined by the money supply. In the long run, however, quantities are determined along with relative prices by the equations of supply and demand. NC, GE, and NA: Effective demand plays no role. Technology, preferences and endowments provide the basic data; markets adjust in response to prices. Finally, these is the question of the general picture of the economy implied by the approach. Mainstream pictures tend to emphasize harmony and optimality. Rational choices in conditions of scarcity are coordinated by the market to bring about an outcome in which everyone will be as well off as possible. Failures and imperfections no doubt abound, at least under some assumptions, but overall, the system is hard to improve on. 744 Afterword: Microeconomic Theory of Money By contrast, PK and CA, in the Keynesian tradition, see the system as inherently flawed, requiring control through policy. Instability Both PK and CA regard instability as one of the inherent problems of the modern economy. For some PK writers it develops as an aspect of the financial system, and grows over time; for others it is inherent in the processes of the business cycle, and has always been. CA authors tend to see it as a possibility inherent in the monetary system. There is simply no guarantee that the advances which enable production to be set in motion will be repaid. For both PK and CA instability tends to lead to a breakdown, which will require re-structuring. NK tends to agree, but the other schools tend to minimize instability or to regard it as a short-run problem only. • • • • • • GE: Equilibria can be shown to exist, but even on normal assumptions, there may be multiple equilibria, and there can be no presumption of stability. But the economic meaning of the instability cannot be clearly explained, nor its significance for policy. NC and M: Short-run instability may exist, due to external shocks or asymmetrical information. This may be compounded by the effects of increasing returns. But the normal responses of the market will pull the system towards its long-run equilibrium. For NC, the short-run responses should be considered optimal. For M, inflation is not a case of market instability but the result of improper policy or unwise behavior on the part of the monetary authorities. NK: Instability exists due to imperfections, nominal rigidities and peculiar market configurations. These may be compounded by the effects of increasing returns. Such imperfections, etc. may be substantial; even if small, they may have large effects. It cannot be presumed that any plausible dynamic processes would overcome them and move the system towards eqUilibrium. Interventionist policy is therefore necessary. DK: Instability is examined only in connection with the properties of nonWalrasian equilibria. The approach classifies a number of different possible configurations of markets when some prices are rigid; it analyses the existence, uniqueness and stability of the resulting equilibria. Instabilities are possible. NA: The system generates dynamic processes which do not necessarily converge. However these processes do not necessarily lead to reduced welfare or to positions that agents would regard as contrary to their desires. NR: Gravitation could imply instability, if the adjustment process is long enough to leave room for structural changes, but in general gravitation present a picture of the orderly behavior of persistent market forces. Full employment, however, is not implied, since prices do not reflect relative scarcities. Differences in frameworks of analysis lead to contrasts in the treatment of money, on which PK and CA may be compared with the other approaches. Ghislain Deleplace and Edward 1. Nell 3 745 CONTRASTING VIEWS ON MONEY In this volume, three issues concerning money have been examined, in which PK and CA are close to one another, but differ from the mainstream: the distinction between the industrial and the financial circulation, the respective roles of money and finance, and the endogeneity of the money supply. We have seen above the difficulties faced by GET in integrating money into its analysis. On the three issues, these difficulties affect the theories which incorporate the GE framework in full or in significant part. Circulation is understood merely as market exchange and, because of the dichotomy, the markets for goods and assets, on one hand, and for money, on the other hand, fall into separate categories. In particular, the interest rate is determined by time-preference and the technology (given the endowments), along with all other prices and quantities, and has no relation to money. The difference between money and finance is hard to conceptualize in this framework. Put most simply, there are no roles for money which finance could not play just as well. Finally, the emphasis on the demand for money does not mean that, when demanded, it will be accommodated by a corresponding supply; on the contrary, money supply remains exogenous, i.e. outside the theory. It seems, therefore, that a useful comparison on the three issues mentioned can only be made with the analytical frameworks which keep GE at a distance and/or do not refer explicitly to it. However, the case of NA is particularly complex: as noted above, the Hayekian view of competitive monies is only part of the landscape of monetary laissez-faire, which also borrows from M (Friedman's conception of modem free banking) and from GE (New Monetary Economics). It is unclear if what these views have in common -laissez-faire applied to money creation - is more important than their divergence about the general framework. In addition to the appreciation provided in section 1 above, it will be useful to note NA's position in regard to the three issues just mentioned. First the monetary system is seen as destabilizing an otherwise stable real system. Secondly, money is completely absorbed by finance. Thirdly, while there is 'endogeneity' (the supply of money created by banking institutions, in response to market incentives, adapts to the demand at the values of the underlying securities), it is not treated differently from the response of supply to demand in any other economic activity. We can now examine the various approaches with regard to these three issues: Industrial Versus Monetary and Financial Circulation Both PK and CA reject the idea that the 'real' sector can be adequately analyzed in barter terms. The 'monetary' and the 'real' aspects of the economy are intertwined and cannot be understood separately. But both schools nevertheless distinguish sharply between the activities of industry and labor, where the circulation 746 Afterword: Microeconomic Theory of Money of funds underwrites the production of goods and services, and the operations of the money and financial markets. CA in addition draws a sharp line between money and banking, on the one hand, concerned with the creation and circulation of money, and the financial markets, on the other, which deal with ownership, corporate control and rentier income. Most of the other approaches separate the 'real' and the 'monetary' quite sharply, treating them as distinct sectors, though some approaches also argue that the two interact to establish the equilibrium of the system as a whole. • • • • M: In the short run monetary changes affect the real system, tending to destabilize it, but in the long run, monetary variables only affect nominal values. The Classical dichotomy holds. DK: If represented at all, the monetary system is modelled according to Keynesian liquidity theory, allowing the monetary and the real sectors to be integrated. In some models disequilibrium is precipitated by the 'cash-inadvance' constraint. NK: 'Industrial circulation' is captured through the relationships between the IS locus and the labor market and the production function; the monetary system and financial relations are compressed in the usual way into the LM function. NR: The Central Bank fixes the long-term interest rate, which is a cost for firms. The effects are as follows: suppose interest rates rise - then money prices will increase, whatever the debt/equity structure of the industries; as a consequence, the real wage rate declines, and the rate of profit increases, catching up with the rise in the interest rate. But the money prices change differently according to the capital/labor ratio, hence there is a change in relative money prices, which must be the same as the price changes (as expressed in a real model) brought about by the decline of the real wage rate. The same adjustment holds in reverse for reductions of the interest rate. Thus distribution and prices are ultimately determined by banking policy, subject to various limitations, e.g. the power of unions. The approach ~ay be criticized both for implausible assumptions about the behavior of equity, and for an undue ascription of power to central banks. Money and Finance Both PK and CA regard money as essential to understanding the capitalist economy. 'Real' relationships are fictitious; all actual economic transactions in a modem system are carried out by means of money and/or credit. There is no such thing as an 'abstract' system of pure exchange. Production requires money, and money comes from banks, in modem capitalism. Money is essential in the short term. Finance, on the other hand, concerns the pattern of ownership, and of longterm indebtedness. Money and financial markets determine the rate of interest, Ghislain Deleplace and Edward J. Nell 747 which, in tum, affects real outcomes. Money also contributes in various ways to instability, and so is central to the explanation of unemployment, and an important factor in, but not the sole cause of, inflation. Inflation, in tum, may have real effects - in particular on the relation between returns to equity and to the holding of debt. For PK, money is closely related to finance, for CA there is a sharp line between them. The relation between money and finance is not however a central issue for the other approaches. • • • • M: The exogenously given supply of money and the transactions demand for it determine the general price level; the interest rate is determined in the real sector by the forces of productivity and thrift. Money may affect real outcomes in the short run, but not in the long. DK: The relationship between money and finance is not examined. NK: No specific theory of the currency is provided; the whole attention is concentrated on credit, which is analyzed in the same way as capital markets. Imperfect or asymmetrical information generates credit rationing which calls for policy corrections. NR: The financial market is dominated by the Central Bank, which determines the long-term interest rate. But it is unclear if this is supposed to take place tlirough the Bank's control over the quantity of money or through its management of the short-term interest rate (as in Tobin's view). Either the rate of interest or the growth of the money supply determines the rate of profit. But it seems that money dominates finance, and together they causally determine the long-period position of the real system. However, real relationships can be understood and modelled without any consideration of money. Endogeneity of Money Both PK and CA examine money as a medium of circulation; hence both understand how it responds to pressures from the market. None of the other schools pay attention to circulation; hence none have much to say about endogeneity. • • M: 'Endogeneity of money' is interpreted to mean that the Central Bank chooses to respond as a matter of policy to interest or income. Hence the money supply curve has a positive slope instead of being vertical. This is considered a special case of minor theoretical importance (although possibly significant in practice), analyzed to show that the general conclusions are not much affected. DK: The expenditures of firms are considered to be constrained by their prior or expected current sales. This is hard to understand if in normal times credit were always to be available to those with capital, as must be the case if the money supply responds to the needs of trade. 748 • • Afterword: Microeconomic Theory of Money NK: In line with the emphasis it puts on credit, rather than money, the question of endogenous responses is raised with regard to the behavior of commercial banks, not the Central bank. Hence no theory of endogenous money creation is provided. In that way NK comes closer to PK than to CA, and the modelling of credit rationing may contribute to an improved analysis of the banking and financial system. NR: The prices it determines are real prices. Money is treated separately, in so far as it is considered at all. There is presumably no reason to rule out an endogenous money supply; the question has simply not been considered. It is clear that PK and CA diverge significantly from the other schools, which likewise differ widely among themselves. NR is closest in some matters of theory - it rejects equilibrium based on rational choice and substitution; it rejects the scarcity theory of value, and it adopts a surplus approach. But in almost every other respect it tends to diverge. NK accepts all the neo-Classical paraphernalia, but by stressing the influence of imperfections, it arrives at conclusions that often seem to be the closest to PK and CA. But in fact the overlap is not very great. PK and CA treat money and finance as central to effective demand which, in tum, they regard as essential to understanding the instability of modem economies. None of the other approaches weaves together money, effective demand and instability in this way. In the resulting picture the economy is portrayed realistically, and money plays a coherent role. These are significant achievements, which appear to be beyond the reach of most of the schools of thought in contemporary economics. Notes 1. 2. 3. 4. Sections 1 and 2 were initially drafted, respectively, by Deleplace and by Nell, then cross-examined and revised jointly. Section 3 was drafted jointly and given final form by Nell, who also did the final editing and polishing. Special thanks to Carlo Benetti for his comments on section 1 and to Marc Lavoie and Alain Parguez for theirs on sections 2 and 3. The nature of the equilibrium associated with such a situation is another question; see Benetti's chapter in the present volume. Like the auctioneer, the administrators of the chamber of settlements must be assumed to work without economic motivation. Otherwise they would charge whatever fees the traffic would bear, and some traders might find it advantageous to contract privately, in which case the problem of bilateral exchange arises again (see below). Further, a single chamber would be a monopoly, and could discriminate. But if there were many competing chambers, the problem arises whether accounts would balance in each, calling for a 'super-chamber' to settle the accounts among the competing chambers. But this 'super-chamber' would then be a monopoly. For an attempt to overcome this limitation with a sequential model in a 'New Keynesian' framework, see Messori and Tamborini (1994). We thank Marcello Messori for having attracted our attention to the quotation by Stiglitz. Ghislain Deleplace and Edward J. Nell 749 References Benetti, C. (1992), 'The Ambiguity of the Notion of General Equilibrium with a ZeroPrice for Money', Chapter 12 in this volume. Clower, R.W. (1967), 'A Reconsideration of the Microfoundations of Monetary Theory', Western Economic Journal, 6, December. Debreu, G. (1959), Theory o/Value (New York: Wiley). Friedman, M. (1969), 'The Optimum Quantity of Money', in M. Friedman, The Optimum Quantity 0/ Money and Other Essays (Chicago: Aldine). Grandmont, J.M. (1974), 'On the Short Run Equilibrium in a Monetary Economy', in J.H. Dr~ze (ed.), Allocation Under Certainty: Equilibrium and Optimality (New York: Wiley). Greenfield, R.L. and L.B. Yeager (1983), 'A Laissez Faire Approach to Monetary Stability', Journal 0/ Money, Credit, and Banking, 15 (3), August. Hahn, F. (1965), 'On Some Problems of Proving the Existence of an Equilibrium in a Monetary Economy', in F. Hahn and F.P.R. Brechling (cds), The Theory 0/ Interest Rates (London: Macmillan). Hahn, F. (1981), Money and Inflation (London: Basil Blackwell). Hahn, F. (1988), 'On Monetary Theory', The Economic Journal, 98, December. Hayek, F.A. (1976), Denationalization 0/ Money (London: Institute for Economic Affairs). Hicks, J.R. (1935), 'A Suggestion for Simplifying the Theory of Money', Economica, 2 (5). Messori, M. and R. Tamborini (1994), 'Money, Credit, and Finance in a Sequence EconomY"(Rome: Universita 'La Sapienza'), mimeo. Os troy, J.M. and R.M. Starr (1974), 'Money and the Decentralization of Exchange', Econometrica, 42 (6). Patinkin, D. (1956), Money, Interest, and Prices (Evanston: Row, Peterson) (2nd edn, 1965). Selgin, G.A. and L.H. White (1994), 'How Would the Invisible Hand Handle Money?', Journal 0/ Economic Literature, 32, December. Starr, R.M. (1989), General Equilibrium Models 0/ Monetary Economies (Boston: Academic Press). Stiglitz, J.E. (1988), 'Money, Credit, and Business Fluctuations', Economic Record, 64, December. Wallace, N. (1988), 'A Suggestion for Oversimplifying the Theory of Money', The Economic Journal, 98, Supplement. Index of Names Abel, A. 630, 637 Aftalion, A. 10 Aglietta, M. 235, 339, 393, 394-5, 396,397 Airikkala, R. 486 Akashi, S. 627 Akerlof, G. 75,87 Amadeo, E.J. 539 Anderson, W.H.L. 628,629 Arnon,A. 543 Arrow, K. 55, 86, 203, 375 Artus, P. 486, 487 Asada, T. 627, 629 Asimakopulos, A. 152, 418,430, 432 Autume, A. d' 200, 534 Bagehot, W. 334,336,463, 552-3, 667 Bar-IIan, A. 527 Baranzini, M. 130 Barrere, A. 10, 38, 487 Barro, R.J. 512 Baumol, W. 525 Bendixen, F. 144, 152 Benetti, C. 11,39,234,235,236,328 Bernanke,B. 378,627,628,629, 630,631 Bernholz, P. 354 Bhaduri, A. 38, 197,708 Binder, P. 359 Bingham, T.R.G. 486 Black, F. 302, 339, 734 Blanchard, O. 39,300,610,627,628, 630,637 Blatt, G.M. 627 Bleaney, M. 405 Blimovich, A. 349 Boddy, R. 636,685,709 Bodin, J. 9,38 Boesky,l. 679 Boisguilbert, P. de 10 Boissieu, C. de 477,487 Bordes, C. 487 Bordo, M. 512-13 Bournay, 1 486 Bowles, S. 636,637,689, 708, 709, 710 Boyer, R. 708,709 Boyer-Xambeu, M-T. 235,327,328 Brainard, W.C. 85 Braudel, F. 302 Bruneel, D. 486 Brunner, K. 482,487 Buchan, A. 667 Cairnes, lE. 132 Campbell, 1 Y. 631 Cannan,E. 150,153 Canova, F. 627 Cant ilion, R. 10 Cartelier, 1 II, 35, 39, 234, 235, 236 Carter, P.J. 603 Caskey, J. 80,87 Cass, D. 234 Catephores, G. 300 Cencini, A. 407,428 Cesarano, F. 461 Chazelas, M. 487 Chick, V. 7, 152, 197,460,533 Clark, P.K. 628 Clarke, R.L. 603 Clower, R.W. 202,396,736 Condorcet, M. de 328 Copeland, M.A. 341 Coquelin, C. 454 Cornwall, J. 402,542 Coutiere, A. 487 Coutts, K. 710 Crick, F.W. 144 Cripps, F. 30 Crotty, 1 636,685,706, 709 Danziger, S. 64, 67 Dauvisis, J.F. 487 Davidson, G. 66 Davidson, P. 99,152,387,417,534, 682 aspects of money 457 contracts 62, 63, 460 751 Index of Names 752 Davidson, P. continued interest rates 537 investment 427 long-term funding 411 Marshallian prices 121 Post Keynesianism 6-7, 7, 57, 400 relative prices 54 safety net 66 uncertainty 378 unemployment 459 Day, R.H. 610,629 De Long, J.B. 80,87 De Vroey, M. 407 Debreu, G. 86, 375, 732 Deleplace, G. 235,327,328 Delli Gatti, D. 86 Denizet, J. 10,38,470,486,487 Diamond, D.W. 562 Doherty, M.R. 339 Dornbusch, R. 708 Douglas, P. 657 Dow, J.C.R. 99,543 Dow, S. 460,535,543 Dupont, F. 407 Dutt, A.K. 539, 543 Dybvig, B.A. 562 Dymski, G. 378, 395, 396 Earl,P. 542 Earley, J.S. 339,513 Eckstein, O. 602, 710 Edgeworth, F.Y. 307 Eichner, A. 7, 26, 38, 536, 537 Eisner, R. 628 Enfron, B. 487 Epstein, G. 693,704,705,707,708, 711,712,713 European Commission 723 Fagg Foster, J. 405 Fair, R.C. 608,614,615,629,630 Fama, E. 302, 382, 734 Fazzari, S. 80,87,628,629,631, 637 Feiwell, G. 420 Fels, E. 349 Ferguson, T. 707 Ferri, P. 88 Fetter, F.W. 330 Fischer, S. 300,610,627,628 Fisher, F. 200, 202 Fisher, I. 315,463,651,657 FUhl, C. 344, 352-3, 354, 354-5, 355-6 Foley, D. 301,339,396,610,619, 627,628,629 accelerator effects 635 cyclical variation 395 money contracts 444 WGE models 380 Forstmann, A. 354 Franke,R. 301,627,628,635 Frankel, J.A. 711 Friboulet, J.J. 428 Friedman, B. 493,619,629 Friedman, M. 74,513,673,734 fiat money 326 flexible exchange rates 706 general eqUilibrium 86 monetary aggregates 606 money and credit 662 money supply 491 wages 74 Fromm, G. 710 Fuerstenberg, G.M. von 630 Fullarton,J. 330,331 Funke, M. 710 Gale, D.D. 373-4, 375 Gallegati, M. 86 Gandolfo, G. 236 Gertler, M. 627, 628, 629 Ghilarducci, T. 603 Gillard, L. 235,327,328 Giovannini, A. 723 Glossman, D.B. 603 Glyn, A. 394, 708, 709 Godley, W. 30,710 Goldfeld, S.M. 512 Goldstein, J.P. 689,708,709-10, 710 Goodfriend, M. 99,100 Goodwin, R.M. 607, 610, 611, 628 Gordon, D.M. 636,689,709,710 Gordon, R.J. 602,640,701,710 Grandmont, J.M. 63-4, 730 Granger, C.W.J. 621 Index of Names Graziani, A. 38, 151, 152, 189, 197, 410,462,532,668 circuit theory 191, 429, 430, 431 ; money creation 659; money and credit 661 distribution theory II final finance 411 initial finance 400 interest payments/costs 184, 662 prices 167 reflux problem 406 Greenfield, R.L. 302, 328, 563, 734 Greenwald, B.C. 627,628,629,637 Greider, W. 99,100 Griffoul, L. 487 Grilli, V. 718-19 Guckenheimer,1. 641 Guerrieri, P. 724 Gurley, J. 74,86,382,470,486,628 Guttentag, ]. 558 Hahn,F. 152,203,374,375,726 axiom of absence of money illusion 73-4,86 challenge of existence of money 728 Clower constraint 736 money and bonds 731 money contracts 55 money and value 366, 368; zeroprice for money 367,371, 371-2,372-3 neutrality of money 67 savings and non-reproducibles 53-4 Hahn, LA 145 Hahnel, R. 636 Hall, RE. 302,514,629,734 Hallett, G. 109, III, 130, 137 Hanley, T.H. 603 Harris,D.G. 481 Harrod, RF. 49-50, 66 Hart, A.G. 87 Haslinger, F. 349 Hawtrey, RG. 316,328,336 Hayek, FA von 152,155-6,196-7, 669, 736 Heinsohn, G. 152,197,460,461,462 monetary contracts 441, 442 753 Herring, R 558 Hicks, J.R 71,85,297,339,402, 486 changes in value of money 300 decision-making 59, 60 hoarding 249 marginal utility analysis 725-6 non-neutrality 79-80 overdraft economy 10, 466, 532 statisticaUstochastic methods 68 Hixson, W.F. 660 Holmes, A. 495 Holmes, P. 641 Homer, S. 460,462 Hood, W.C. 402 Hotson,1.H. 197,543,660 Hubbard, R 640 Hume, D. 299,327,651,655 Ingrau, B. 87 Israel, G. 87 Jaffee,D.M. 481,487 Jarsulic, M. 99, 643 Jay, S, 88 Jevons, W.S. 307,315,328 Joffre, C. 487 Johnston, J. 514 ](jhr, W.A. 344 Jonung,L. 512-13 Kahn,M. 708 Kahn, R 302-3, 540, 543 Kaldor, N. 26,99, 151,236,339, 513,635.636.682.709 accelerator 607, 610 asset-portfolio theory 673 credit control 540 endogenous money 100,140.235 interest rates 539 investment function 628 pricing 538 Kalecki. M. 10, 236, 628, 636 accelerator 607, 610 full employment 537.685 investment 424,429,613.637; and Keynes 418-23 profits 197,218,669 Kashyap, A.K. 631 754 Index of Names Kaufman, H. 514 Keen, S. 86 Kenen, P.B. 709 Keynes, J.M. 10,66,85,86, 152, 153,236,302,328,432,462,629, 672,682,683,716 aggregate demand 667 asset price stability 665 eqUilibrium 72-3 euthanasia of rentier class 536 identities 131 idle balances 143 industrial and financial circulation 674-5 industrial and financial money demands 675-6 interest rate 94-5; real 67 introduction to the General Theory 70, 72, 73 investment: Kalecki and 418-23; saving and 151 Keynesian revolution 48-9 liquidity 63, 64 liquidity preference 457-8,459 monetary standard 307, 326; Ricardo's Ingot Plan and gold standard 308, 309-10 money 142, 144-7,461; circulation 145-7; government money as commodity money 447; purchasing power 147-8; unit of account 441 non-neutrality 55, 76 non-stationarity 68 object of production 451 relative price changes 52-4 speculation 679,680 taxonomic attack on Say's law 50-2 uncertainty 58, 62, 442 velocity of endogenous money supply 673-4 Kindleberger, C.P. 724 King, R.G. 629 Klamer, A. 85 Klein, L.R. 629 Knapp, G.P. 446,461 Kockesen, L. 627 Kohn, M. 396 Koopmans, T.C. 249 Kopcke, R.W. 628,630,637 Kotz, D.M. 543,709-10 Kregel, J.A. 11,422,427,429,653, 666,667,669 Krelle, W. 349 Kreps, D.M. 64 Kr61l, M. 344, 349-50 Krugman, P. 708,711 Kuh,E. 628 Kuttner, K.N. 629 La Geniere, R. de 534-5 Largentaye, J. de 198 Laroque, G. 63-4 Laudy, J. 487 Laursen, S. 708 Lavoie, M. 5,7,38,99,198,533,542 circuit theory 427,428 interest rates 537, 538, 539 loans make deposits 410 reflux problem 406 Lawson, T. 637 Layard, P.R.G. 708 Le Bourva, J. 540 Leijonhufvud, A. 85 Leonard, J. 409 Leonard, J.D. 603 Leontief, W. 341 Levine, D. 443,461 Levy, D. 88 Levy, M.A. 487 Levy-Garboua, L. 475,486 Levy-Garboua, V. 465,471,475, 486,487 Lexis, W. 362 Lilien, D.M. 514 Lipietz, A. 393, 394, 396 Liu, W. 643 Llau, P. 487 Llewellyn, D.T. 486-7,724 Lombard, N. 487 Lowe, A. 406 Lowe,J. 328 Lucas, R. 65, 74, 86 Maarek, G. 465,486,487 Malestroit, J. de 9,38 Index of Names Malinvaud, E. 417 Mangold, G. 349, 352 Mankiw, G. 637 Mann, C.L. 708 Marcuzzo, M.C. 245,308-9,314, 327,328 Marglin, S. 38, 197, 708 Marjolin, R. 10, 38 Marx, K. 10,328,341,454,463,689 capitalists' advancement of money 251,300 Mason, W.E. 308,326 Mayer, C. 640 McIntosh, M.K. 446 Meacci, F. 420 Meese, R. 711 Melton, W. 99, 100 Meltzer, A.H. 482,487 Mengle, D. 99 Messori, M. 430, 748 Metzler, L. 708 Meyer,l.R. 628 Michl, T.R. 709 Miles, M. 99 Mill,l.S. 301 Minsky, H.P. 7,36,85,86,88,151, 152,339,400,463,513,635,666, 682 credit controls 541 discount window 665 financial fragility 607 financial instability hypothesis 378,384,678 investment decisions 425-7 money as stock of wealth 140 money supply endogeneity 498, 499 securitization 527 Mints, L.W. 339,657,657-8,660-1, 667 Mjses, L. von 145 Mishkin, F.S. 538, 629 Mizrach, B. 628 Modigliani, F. 481,487 Moore, B.J. 39,92,99, 100,236, 297,406,414,457,501,513,517, 519,520,542,630,636 asset and liability management 498 755 convenience lending 94 credit controls 541 credit money 496 demand inflation 539 endogenous money 152,235,400, 533 fundism 402 interest rates 501-2,504-5,535-6 loans make deposits 410 'public' creation of money 659 Morgenstern, O. 60 Morishima, M. 134, 300 Mott, T. 86, 629, 637 Mundell, R. 716 Myrdal, G. 678 Neftci, S.N. 628 Neisser, H. 344 Nell, E.J. 4,5,7,21,26,35,38,198, 301,401,407,408,542,627 circuit of money in production economy 247,250,251,263, 264,275,287,293,300 interdependence and production 250,251 multiplier 34 private self-liquidating credit 407 Quantity Equation and aggregate income equation 39 Neufeld, E.P. 402 Neumann, J. von 60 Newbold, P. 621 Newton, I. 305,314 Nickell, S.J. 708 Niggle, C.J. 460,533 Nogaro, B. 10 Nordhaus, W. 710 O'Connell, S.A. 627,628,635 Orlean, A. 235 Os troy, 1.M. 200,375,733 Pacaud, A. 486 Padoan,P.C. 719,724 Palley, T.I. 496,501,519,530 Parguez, A. 152,197,198,278,339, 407,409,428,532,543,666, 668-9 . banks 663 Index of Names 756 Parguez, A. continued investment decisions 430 investment and saving 12 'probleme du profit' 668 Parke, W. 608 Parker Foster, G. 198.409. 542 Pasinetti. L.L. 537 Patat. J.P. 486 Patinkin, D. 151.396.420.730 demand for money 140 'invalid dichotomy' 299 model 371-2.374-5 money as store of value 728 Pecha. J. 487 Peck. M.I. 721 Pesaran. M.H. 381.384 Peter, H. 341.342-4.344-6.352. 356-7.361 Petersen. W.M. 693 Peyroux, C. 486 Phillips. R.I. 666.669 Pigou, A.C. 79. 87 Pindyck. R.S. 514.630 Pivetti. M. 536. 538 P1osser. C.1. 629 Polanyi. K. 461 Pollino R. 495.513.530.533 Potter, S.M. 627 Pou1on. F. 252 Preiser. E. 353-4 Pugel. T.A. 689.710 Piltz, T. 344 Quesnay. F. 10.341 Radcliffe Committee 494. 540 Ramsey. J. 628 Rapping. L.A. 709 Rappoport, P. 609 Rasmusen. E. 396 Rebitzer. lB. 709 Reichardt. H. 341.346-9.358 Reichlin. L. 609 Renversez. F. 487 Reus. E. 407 Ricardo. D. 245.308-10.314-15. 326.327 Rist. C. 10.512 Rizvi. S.A.T. 249 Robbins, L. 667 Robertson. D.H. 147.152. 153.326 gold standard 307. 327 money acquired to use 140, 151 Robinson. J. 3.378.410.411 Rodgers, C. 496 Rogers. J. 707 Romer. C.D. 619.629 Romer. D.H. 619.629 Roncaglia. A. 536. 669 Rosselli. A. 245.308-9.314.327. 328 Roubine. S. 487 Rou~seas. S. 99.495.513.517.540. 666.682 bank mark-up 410.411 credit controls 541 principle 181-3 Rubinfeld. D.L. 514.630 Samuels. W. 667 Samuelson. P.A. 131.132.300 multiplier 124.125.126 theory of money expenditures 105, 128-30 Santoni, G. 693 Sargent. TJ. 65 Savage. L. 60-1 Saville, I.D. 99.543 Sawyer. M. 710 Say, J.B. 143, 152 Sayers. C.L. 627 Scharrer, H.E. 709 Schmitt. B. 10,39,132,234,532, 542 cash balances as wealth 139, 151 European Community 717-18 virtual and realized magnitudes 428 wages II Schneider. E. 150,153,346,349 Schopenhauer, A. 113 Schor, J.B. 693,704,707,709,711, 712,713 Schumpeter, J. 10, 383, 409 Schwartz. AJ. 491.606 Scrope. P. 328 Seccareccia, M. 189,198,402,411, 537 757 Index of Names Selgin,O.A. 734 Semmler, W. 4,301,619,627,628, Tonveronachi, M. 666 Tooke,T. 10,218,235,330,403, Shackle, O.L.S. 378 Shafer, W. 610,629 Shaw, E. 382,470,486,628 Shell, K. 234 Shepherd, W.O. 689 Sherman, H. 602,636 Shiller, R.J. 513-14 Shleifer, A. 629 Sichel, D.E. 512 Sieveking, M. 619 Silber, W. 100 Simons, H. 652, 657, 667 Simpson, T.D. 512 Sims,C. 712 Sinai, A. 602 Skott, P. 635 Smith, A. 341,651,652,655-6,667 Solomon, A. 512 Solow, RM. 59 Sraffa, P. 217, 326, 328 Starr, RM. 202, 375, 726, 730, 733 Steiger, O. 152,197,460,461,462 monetary contracts 441, 442 Steindl,1. 287,419,424-5,432 Sterdyniak, H. 486 Stiglitz, J. 87,384,627,628,629,637 'New Keynesian' approach 737 Stock, J.H. 629 Strauss-Kahn, M.O. 487 StUtzel, W. 355-6,361 Sugden,R. 60,61 Summers, H. 629 Summers, L.H. 80,87,512,693 Sutcliffe, B. 708,709 Sylla, R 460, 462 Sylos-Labini, P. 697,709,711 credit controls 543 credit theory of money 493-4 role of banks 402 Trevithick, 1. 151 Tsiang, S.c. 151,152,396 629,635 Tamborini, R 748 Targetti, F. 38 Taylor, L. 38,610,627,628,635,708 Terzi, A. 197,542 Tessier, R. 486 Thomas, J.O. 153 Tobin, 1. 85,403, 470, 486, 525, 629, 673 neutrality of money 68 463 Vandevelde, F. 428 Veitch, J. 640 Vickers, D. 151, 387 Villa, P. 486 Viner, J. 85, 336, 339 Wachtel, H. 706 Wagner, A. 350 Wallace, N. 374,731 Wallich, H.C. 680,693 Walras, L. 117-18,344,356-7, 732 Watson, M.W. 629 Weintraub, E.R 105, 110, III, 118, 128 Weintraub, S. 7,152,400,513,683 TIP 680 wage-cost-mark-up equation 401, 536 Weiss, A. 87,629,637 Weisskopf, T.E. 636,689,708,709, 710 Weldon, J. 402 Weymuller, B. 486 Whalen, C. 669 White, L.H. 734 Wicksell, K. 10, 145,245,299,336, 375,512 circulation of money 152, 30 I interest rate and monetary growth 294-5 value of money 268 Wiggins, S. 641,643 Williamson,1. 719 Wittgenstein, L. 106, III, 130 Wojnilower, A.M. 339,500,541, 543 Wolfson, M. 641 Wood,A. 500,601,602,603,638, 26,38,198,543 758 Index of Names Wood.E. 339 Woodford. M. 628 Wray. L.R. 198.412.460.461.462. 463.535.542 development of banking 446 saving and investment 171 Wulwick. N. 543 Wyss. D. 710 Yeager.L.B. Zarnowitz. V. 302.328.563.734 630.638 Index of Subjects accelerator principle 610-11 accommodationism 517-19 comparison with structuralism 495-512 accounting constraint 347-51 accumulation circular flow 447-9,451,454 dynamics of 221,223-4,225-6, 637-8 aggregate demand 166-7 function 50-2 aggregate profits 171-3, 192-4 aggregate supply function 50-2 aggregation vs proportionality 28 Anti-Hayekian Theorem 194-5 arbitrage 271-2,295-6,335 asset management 520-4, 572 see also liability management asset-portfolio model 672-3 asset prices 72, 78-9 stability 665 asymmetric information 75,384-5, 388-9 auctioneer 207-8, 732 auto-economy 466-7 balance of payments 280-4 balances 203-6 non-zero 206-8 settlement and payment systems 208-16 Bank of Amsterdam 664 bank credit see credit bank debts, repaying 146-7, 162 bank deposits 289-90 cash and financial 675 runs on 561-4,568 Bank of England 313,327,575, 653-4,656 Bank for International Settlements (BIS) 566-7 capital adequacy standard 574-7 bank loans and demand deposits 518, 519 losses 596, 597 survey on lending practices 597-600 see also credit bank money. excess issue of 337-8 bank principle system 478 bankers 76-7 banking school view 330-40 bankruptcy 206 banknotes 334 bankslbanking system 22-3 comparison of PK and CA approaches 24-5,33 control of liquidity 482-3 credit network 160-1, 180-1 endogenous money and free inarket capitalism 651-71 finance. savings and 148-51 financial fragility 584, 591-7 institutions 336-8; structure 663 overdraft economy 472-5,481; balance sheet structure 468. 472.473 paper money 281-3 payment system 205 profit rates 701-4 role in circuit 191,252-3,254. 451-4; banks' special status 454-6 structuralism 520-4 supply of credit 585. 597-600 systemic risk, financial innovation and the safety net 552-81; rationale of bank risk 555-64 thriftiness: banks' own 180-7, 191; firms' 178-80 Banque de France 483, 682 barter 56 rebuttal 201-3 batch production 288 BFH system 563-4, 734-5 bifurcation analysis 640-1 bilateral trade 732-4 bilI-broking 334 759 Index of Subjects 760 bills of accommodation 268 bills of exchange 261, 268, 331 see also banking school view bimetallic regime 320-1 bonds 295-6, 730-1 borrowing, foreign 652-3 borrowing/GNP ratio 491,492 budgetary policy 721-2 'bullion controversy' 308-9 Bundesbank 682,721 business cycle financial crises and 582-603 financial-real interaction and investment 606-34 forces leading to 583-5 capacity utilization finance and investment 610, 611-12,614,617-21,625-6 thriftiness 175 capital constant and variable 259 entrepreneurial 418-19 financial and industrial 684-715 fixed 285-7,287-8 international mobility 693-6, 704-5 lending capital vs lending credit 333-4 mintage and credit 213-16 money, income and 133-6 money, sinking fund and 287-8 capital adequacy 599,603 BIS standard 574-7 capital expenditures 591,595 capital goods 122 capital markets, imperfect 613 capitalism endogenous money and free market 651-71 production circuit 273-80 secular 394-5 cash deposits 675 causal dynamics 16-17 causality 500-1,504-5,508-11 central bank( s) banking school view 331, 335-6, 338 industrial and financial circulation 678-9 interest rate: and banks' thriftiness 180-1,182-3,187,191; modem economy 290-1 monetary policy and financial crises 584-5,591, 600 national and Eurofed 719 overdraft economy 468; refinancing constraint 471-5 rise of 454-6 see also lender of last resort chamber of settlement 732-3 Chicago school 656-8, 663-5 CHIPS (Clearing House Interbank Payments System) 569-71 circuit theory 13 and current bank crisis 651-2, 659-63,664-5 investment decisions 427-32 and the Post Keynesians 532-5 scarcity and thriftiness 155-99 see also circulation approach circular flows 341-64 formal 342-6 in monetary theory 352-5 money in 440-64 real 346-56 repeated creations and destructions vs 132-7 circulating credit 332-3 circulation continuous 288-9 creation, destruction and 142-4; Keynes 145-7; monetary standard 317-18 demand and: effective 33-4; instability of monetary economy 34-7 financial and industrial 31,459, 672-83, 745-6 fixed capital 285-7 industrial vs monetary and financial 745-6 as institution 245-6 modeling 14-15 monetary see monetary circulation money as means of 24,276-7 Index of Subjects pattern of 291-2 payment system 203-6,209-11, 216-34 Post Keynesian fundism and 400--16 principles of 20-1 production economy 245-304 surplus value 265-6 velocity of 301 circulation approach (CA) 5,9-16, 46,241-2 common aspects 11-12 comparison with other approaches 737-48 passim debates 13-16 essentiality of money 157-67 monetary standard 318, 325-6 and Post Keynesian approach see Post Keynesian approach roots 9-10 scarcity and thriftiness 155-99 significance 297-8 classification 49-50 Clower constraint 371,379-85,394, 396,736-7 cohesion 722-3 collateral 384, 396 commodity money 307,326,444, 445-7,455 commodity standard 305,314-15 competition bank risk 574-5 Keynesian models of credit 388-91 competitiveness 687-93 consumer goods 122 consumption circular flow 351-2 Kalecki principle 168-9 multiplier theory 122, 124-5 consumption standard 307 Continental Illinois Bank 572, 582 continuous production 288-9 contracts see forward contracts; money contracts convenience lending 99 convertibility 334-5 Cooke Committee 575 761 cooperation 722 cooperative economy 56 coordination, lack of 27 corporate sector demand for credit 584, 591 financial fragility 584, 587-91 cost inflation 536-9 credit 11-12,23 capital mintage and 213-16 currency of 332-3 demand for 584, 591 endogenous finance 516-31 Keynesian models of 377-98 lending credit vs lending capital 333-4 metallic payment system with 212-13 vs money 332-3 money, banking institutional structure and 660-3 overdraft economy 467-8,469 private 407 securitization of 527-8,571-4,578 supply of 585, 597-600 see also bank loans credit contract 161 credit controls 540-1,542,543, 678-9,680 credit crunch 561 credit divisor 475-9 credit economy 403,661 credit flow 615,617 credit markets 558-61 credit money 18, 156,445-7,454-5 in the circuit 449-54 curves 495-500 fundism and monetary circulation 400--16 interest, accumulation and 447-9 credit network 160-1, 180-1 credit rationing 481-2,597-600 credit risk 567-71 credit theory of money 493-4 crises 15-16, 234, 460, 491 causes of 27-8 effective demand and 33-7 financial and the business cycle 582-603 762 Index of Subjects crises continued policy implications of current 651-71 Regulationism 393-5 triggered by speculative bubble bursts 556-8 crowding out 677 currency of credit 332-3 currency school view 331-2 cyclical variation 395 debt financing 80-1, 82-3 repaying bank debts 146-7, 162 standard 321-5 debt-capital ratio 638-9 debt crisis 565-7 debt economy 465-6 debt-equity ratio 587,589,676 decentralization 564, 732-4 decisions 57-65 classifying decision-making environments 58-65 investment decisions 417-33 demand see aggregate demand; effective demand; money demand demand deposits 518, 519 demand inflation 539-41 democratic deficiency 720 deposit banking 446 deposit insurance 338, 529 depreciation of money 310-14 devalorization of money 310-14, 322-4 'dirty' interest rate targeting 91,97 disaster myopia 558-61,565-7 discount rate 324, 325 disequilibrium 204-6, 232-4, and in Afterword individual 201,204-5 market 20 I, 204 micro/macro 391-2 Disequilibrium Keynesians (OK) 738-48 passim distribution goods market equilibrium and 636 profits and 171-3 economic system money contracts and 55-7 nature 17-19; money counts 17-18; structural di vision 18-19 effective demand 19-20 comparison of approaches 742-4 and crises 33-7 elasticity of productivity 52-3 elasticity of substitution 52-4 endogenous finance 516-31 endogenous money supply 31-2, 490-515,533, and in Afte/word accommodation ism and structuralism 495-512, 517-24 comparison of approaches 747-8 compatibility with free market capi talism 651-71 historical analysis 89,90-1,96-7, 98-9 and interest rate see interest rate overdraft economy 446-71,475, 477 . endogenous uncertainty 384, 385 entrepreneurial capital 418-19 entrepreneurial economy 56-7 equilibrium/equilibria 72-3, and in Afterword elementary dynamics model of monetary economy 219-34 goods market 636 identities and 130-2, 349-50, 357 monetary eqUilibrium 367 non-monetary equilibrium 367 overdraft economy 471-9 rejection of equilibrium method 16-17 see also general equilibrium equilibrium fallacy 131-2 equipment utilization rate see capacity utilization ergodicity 59-60 essentiality of money 157-67 European Central Bank (ECB)lEurofed 681-2,716,718-20,720 European Community 484,681-2 single currency plan 716-24 Index of Subjects European Currency Unit (ECU) 717-18 European Monetary System (EMS) 716,724 European Parliament 720 exchange banking school and currency school views 330-2 circular flows as 342-4 production as 251-5 exchange circuit 246-50 exchange rate BeU 719 regime and macroeconomic policy 684-715 exogenous money supply 89,90, 92-6,98,673 exogenous uncertainty 381,385, 386-8 expansion 234 characteristics of current 585-7 expectations profit' see profits rational 58, 58-60 and risk aversion 386-8 expected utility theory 60-1 expenditures capital 591,595 money 121, 128-30 export sector 277-80 Federal Deposit Insurance Corporation (FDIC) 599,600,603 Federal Reserve 337,377,503,562, 586-7,676,721 capital adequacy 599,603 credit restriction 600 interest rate causality 504, 508-11 Fedwire 569-70 fiat money 276-7,307,326,andin Afterward development of 444-5,445-7, 461 general eqUilibrium with zero-price for money 367-74 regime and standard debt 321-5 rise of 454-6 fiat standard 305,306-7 final finance 12 763 finance 148-51 endogenous 516-31 initial and final 11-12 internal see internal finance and investment 31-3; predictions from theories 612-17; regression results 617-25 money and 22,31-3,746-7 payment systems in monetary economy 206-8, 216-34 for production 146 revolving fund 147,451-2 self-financing 466-7 finance constraints 371,379-85,394, 396, 736-7 finance motive 22 financial capital 684-715 financial circulation 31,459,672-83, 745-6 financial crises see crises financial deposits 675 financial fragility 383-4 banking sector 584,591-7 corporate sector 584,587-91 systemic risk 558-61, 563 financial innovation 502, 529 systemic risk and the safety net 552-81 financial instability 378, 384, 529 aggregate detenninants 635-45 see also instability financial intennediaries 385,391-2 attitudes toward risk and exogenous uncertainty 386-8 competition and risk 389-90, 390-1 external financing roles 388-90 franchise 389-90 see also bankslbanking system financial markets 23, 146 bursts of speculative bubbles and crises 556-8 Keynes and Kalecki 418-19 source of liquidity 140-1 financing gap 591,594 finns' thriftiness 167-73 impact 173-80 fixed capital 285-7,287-8 fixed exchange rates 684-715 Index of Subjects 764 flexible exchange rates 684-715 flow of funds 224-5 foreign borrowing 652-3 forward contracts 442-3, 444, 445 fragility, financial see financial fragility free markets, endogenous money and 651-71 frown costs 91 fundism 400-16 future uncertainty 29 Geldwirlschafl 359 general equilibrium 380 and zero-price for money 366-76 general equilibrium theory (GET) 104-5,119-20,356-7,727 comparison with other approaches 738-48 passim demise of 106-19 Gibson's Paradox 280 globalization 554-5, 564-79 gold flows 280-1 metallic system 210-12 gold standard 308-14 model of regime 310-14 Golden Rule 275-6,291-7 Graziani Principle 184 Gresham's Law 268 gross substitution 52-4 growth, profits and 273-80 hedge finance 83 Hicksian non-neutrality 79-80 hoarding 249-50,271-2 Hopf bifurcation theorem 641 households saving 189-90,411-13 thriftiness 187-95 lOOper cent reserves 656, 657, 664, 665 identities 353-4 and equilibria 130-2, 349-50, 357 idle balances 143 imperfect capital markets 613 import penetration ratio 689 impulsion 722 income instantaneous incomes 132-3 modeling income payments 21 money, capital and 133-6 multiplier theory 122,123-4 national 348 spending and 458-9 income generating finance process 90-1,96-7 incomes policy 401-2,537,680 industrial capital 684-715 industrial circulation 31, 459, 672-83, 745-6 inflation 535, 586 bank profits and 701-4 cost 536-9 demand 539-41 incomes policy and 401-2 profit inflation 284 thriftiness and 175 information asymmetric 75, 384-5, 388-9 decisions, uncertainty and 57-65 Keynesian models of credit 379-85 initial finance 12 innovation, financial see financial innovation insolvency, risk of 572-3 instability comparison of approaches 744 financial see financial instability hoarding and 249-50, 271-2 of a monetary economy 34-7 paper money 283-4 regime instability 271-2 instantaneous incomes 132-3 interdependence 250-5 interest coverage ratio 587,588 interest rate 490, 535, 542 banks' thriftiness and 180-7 central bank see central bank(s) credit control 336 'dirty' interest rate targeting 91, 97 endogenous money and 25,498-9, 501-2; causality 500-1, 504-5, 508-11; structuralism 520-4 Index of Subjects distributive variable 536-7 exogenous variable 537-9 financial crises and the business cycle 591, 592, 593 investment and 613,614; interest rate spread 614,615,619-21, 622-3,626-7 investment decisions 421-3, 429-30 money, accumulation and 447-9 money and credit 662 money supply process 94-5, 96, 97 overdraft economy 478,482-3 payment systems 207-8,209, 230-1 relationship between short- and long-term rates 36-7 internal finance 80-1, 162, 403 constraint 192-4 international capital mobility 693-6, 704-5 see also globalization International Monetary Fund (IMF) 566-7 international monetary relations, standard of 305 international payments system 567-71 intra-class conflict 693-6 investment business cycle 606-34 cash flow constraint 637-8 circular flow approach 353-5 decisions 417-33 finance and see finance fluctuations 402-3,404 involuntary 641 money and 31-3 Post Keynesian economics 78-9, 80-2 saving and 170-1,184-5,189-90, 353-5,664-5 scarcity and thriftiness 160,170-1, 184-5, 189-90, 192-4; fall in investment 175-8 spending and non-neutrality of money 54-5 involuntary investment 641 765 Kalecki Principle 20,21 generalized 188-9 monetary foundation for 216-19 thriftiness and scarcity 168-71, 184-5 Keynesian models 71 of credit and choices 317-98 Keynesian non-neutrality 75-7 Keynesian uncertainty 61-5,387-8, 442 Keynesians, orthodox 493 Kreislaufaxiom 345-6, 347 laissez-faire, monetary 734-7 legal tender 321,447 lender of last resort (LLR) 208, 209, 456,665 globalization 565-7,568-9 historical development 552-5 overdraft economy 474-5 lending capital 333-4 lending credit 333-4 liability management 96,571-2 accommodationism 498, 502-3 structuralism 498, 502-3, 520-4 see also asset management liability structures gO-I liquidity loop of 615,616 Post Keynesian theory 54-5,63-4 liquidity preference 52, 53, 457-8 liquidity risk 567-71,573-4 liquidity trap 79, 95 loan-reserve proportionality 5~I, 503-4,507,511 loan supply curves 495-500 see also credit money loans make deposits 90-1 Maastricht Treaty 681-2 macroeconomic magnitudes 361-2 managed money 307, 326 market instability 502 market structure 385-92 market uncertainty 29 see also uncertainty Marxian principle 20-1 mass production 288-9 means of settlement 516, 524-8 Index of Subjects 766 medium of exchange 563-4 development of money 443-4 endogenous finance 516,524-8 microeconomic theory of money as dead end 731-4 mercantilist economy 263-4 merchant earnings 266-7 metallic payment system 210-12 with credit 212-13 micro/macro, treatment of 26-7 micro/macro disequilibrium 391-2 mintage payment systems and balance settlement 208-16 role of the mint 272-3 monetarism 74, 491, 738-48 passim monetary circuit as an institution 245-6 in a production economy 245-304 scarcity, thriftiness and 155-99 see also circuit theory monetary circulation 11 industrial vs financial and 745-6 Keynesian theory 142-4,145-7 monetary standard and 317-18 see also circulation monetary constraints 371,379-85, 394,396,736-7 monetary dependence 217 monetary economy 17-18 characteristic features 319 elementary dynamic model 219-34 instability 34-7 payment systems and dynamics 200-37 monetary equilibrium 367 monetary growth, stability in 294-6 monetary institutions 208, 209 see also central bank(s) monetary laissez-faire 734-7 monetary policy common European 717 economy with endogenous credit money 532-45 financial crises and business cycle 584-5,586-7,591,600 overdraft economy 482 Post Keynesian 678-81 monetary regime 319-21 monetary standard 305-29 monetary system cost of 268-71 development of monetary systems 442-5 equilibrium and overdraft economy 471-9 implications of Golden Rule 296-7 understanding 23-5 working of 22-3 monetary unit see unit of account money capital, sinking fund and 287-8 in the circuit 449-54 commodity money 307, 326, 444, 445-7,455 comparison of CA and PK approaches 23-4 contrasting views 745-8 creation 13-14; endogenous see endogenous money supply; Keynes 142-4, 145-6; monetary standard 317-18; public vs private 658-65 credit money see credit money credit theory of 493-4 credit vs 332-3 depreciation 310-14 destruction 15-16; Keynes 142-4, 146-7 devalorization 310-14,322-4 essentiality 157-67 fiat see fiat money and finance 22,31-3,746-7 and the Golden Rule 291-7 integration in theory of value 726-31 interest, accumulation and 447-9 managed 307,326 medium of circulation 24, 276-7 medium of exchange see medium of exchange medium of settlement 516, 524-8 microeconomic theory of, as dead end 725-37 nature of 441-5 neutrality 73-81, 156 non-neutrality see non-neutrality Index of Subjects as purchasing power 24, 139-54 Quantity Theory of 72,331-2 Regulationism 393-5 specificity 737 standard of 305; see also monetary standard as stock of wealth 139-54 as store of value 381-2, 728-31 uncertainty, money contracts and 61-5 value of 292-4; tabular standard 315-17 velocity of 491,492,614,615, 616,673-4 zero-price for 370-4 money contracts 55-7 development of 442-5 private property and 441-2 uncertainty, money and 61-5 money demand endogenous finance 525-8,528-9 industrial and financial 675-7 interdependence with supply 94, 98-9 option demand for money 63-4 transactions demands 267 money expenditures 121, 128-30 money illusion 74 money. prices, determination of 104-38 money supply endogenous see endogenous money supply exogenous 89,90,92-6,98,673 process 89-101; income generating finance 90-1, 96-7; 1930-60 92-6; 1960-80 96-7; portfolio change 90,92-6 for production circuit 267-73 moral hazard 559 multiplier 33-4 central error 125 crass contradiction 125-8 determination of money prices 121, 122-8 Naturwirtschaft 359 Neo-Austrians (NA) 738-48 passim 767 neo-Classical synthesis 4-5,71 Neo-Ricardians (NR) 738-48 passim neutrality of money 73-81, 156 New Classical approach (NC) 3-4, 74-5, 738-48 passim New Keynesian approach (NK) 3-4, 74-5,377-8,737,738-48 passim New Monetary Economics 302, 563, 734-5 'no-surplus' production economy 255-61 nominal anchor 719, 724 non-ergodicity 60-1 non-financial corporate sector see corporate sector non-monetary economy 726-7 non-monetary eqUilibrium 367 non-neutrality of money 54-5 Hicksian 79-80 Keynesian 75-7 non-zero balances 206-8 note issue 654, 655-6 numeraire 104-5,110-14 objective probability 58, 58-60 offshore clearance/settlement systems 569 option demand for money 63-4 output-capital ratio 610,611-12, 614,617-21,625-6 output prices 72, 78-9 outputs, physical 130 overdraft economy 465-88, 532-3 endogeneity of money 466-71, 475,477 equilibrium of monetary system 471-9 monetary regulation 479-85 overlapping generations models 371, 730-1 overlending 558-61, 563 papermoney 270-1,281-3 instability 283-4 payment systems 200-37 and balance settlement 208-16 concept 203-6, 325 monetary economy with wage relationship 216-34 Index of Subjects 768 payment systems continued non-zero balances and finance 206-8 payments 358-60 Peel's Banking Act 1844 656 Penn Central Corporation 582 Phillips curve 71 Ponzifinance 83.409-11 portfolio change process 90. 92-6 positivity of the price of money 729-30 Post Keynesian approach (PK) 5. 6-9.45.378 and circulation approach 16-37. 243; common ground 16-23; contrasts and differences 23-8; convergence towards a common research program 28-37; endogenous money 528-9; investment decisions 427-32; money in the circular flow 456-60 comparison with other approaches 738-48 passim endogenous money supply 89. 90-1.96-7,98.494-5,528-9 essential characteristics 70-88 fundism 400-16 industrial and financial circulation 677-81 monetary policy 532-5 monetary standard 318. 325-6 price-level flexibility 80 prices 164 circular flows 342-4. 360 determination of money prices 104-38 interest rate and 538. 662 linking industrial and financial circulation 31 neutrality of money 78-9, 80 producer and wages 696-701. 702-3 relative changes and gross substitution 52-4 see also relative prices primary indebtedness 465.469-71 private credit 407 private property 441-2 private property economies 446 probabilistic risk 57-61. 62 production continuous 288-9 as exchange 251-5 financing 146 interdependence and 250-5 real in circular flow 351-2 production economy 245-304 capitalist production circuit 273-80 economy with surplus 261-7 money supply for production circuit 267-73 'no-surplus' 255-61 productivity. elasticity of 52-3 profit consistency 164-6. 192-4 profit inflation 284 profit squeeze 684-715 profits aggregate 171-3.192-4 anti-Hayekian theorem 194-5 banks' 701-4 capitalist production circuit 273-80 circuit of 277-9 corporate sector 587. 590 essentiality of money 159-60. 162-4 expectations: finance and investment 613.614.617-25; investment decisions 420-1.424-5 fundism 406-8 industry and finance 704-6 Kalecki principle 168. 169-73. 218-19; profits and distribution 171-3 macroeconomic 355-6. 362 property markets 556-8 proportionality aggregation vs 28 loan-reserve 500-1,503-4,507. 511 purchased funds 585.597.598.602 purchasing power 24. 139-54 Quantity Theory of money rates of return banks' 701--4 effective 163 72. 331-2 Index of Subjects investment in the business cycle 613,614,621-5 required 183-7,194-5 rational expectations 58, 58-60 real balance effect 658, 730 real bases 466-7 Real Bills Doctrine 332, 335-6, 402 recession 234 recovery 234 refinancing constraint 471-5 reflux Law of Reflux 282, 334-5, 337-8 problem of monetary 401-14 regime instability 271-2 regulation free market and endogenous money 651-71 international 574-7,578 monetary in overdraft economy 479-85 Regulationism 393-5 relative prices circular flows and 342-4 GET 105, 107-8 Golden Rule and 292-3 render economy 179-80 rentier squeeze 684-715 rentiers 536-7 reserve army effect 688, 689 reserves 181 100 per cent 656, 657, 664, 665 required 483 reserves create deposits 95 structuralism 520-4 substitutability between borrowed and non-borrowed 500-1, 503-4,506-7,511 residual balance sheet 664 revolving fund 147,451-2 risk credit 567-71 finance and investment 613-14, 617-25 financial intermediaries' attitude 386-8,390-1 investment decisions 419-21, 424-5 liquidity 567-71,573-4 769 principle of increasing risk 424-5, 426 probabilistic 57-61, 62 systemic and the safety net 552-81 risk aversion 386-8 risk bearing 390-1 Rousseas Principle 181-3 generalized 191-2 safety net 552-81 international framework 577-9 saving 155 households' 189-90,411-13 and investment 170-1,184-5, 189-90,353-5,664-5 money and 148-51 profits and 170-1 see also thriftiness Say's Law Keynes's taxonomic attack on 50-2 validity of 357-8,360-1 scarcity 155-7 see also thriftiness Schumpeterian macro environment 383 seculat capitalism 394-5 securitization of credit 527-8,571-4, 578 seigniorage 269,311-14,327 self-finance 466-7 sequential analysis 29-31,528, 740-1 settlement finality 570-1 shares 295-6 simple credit 332-3 single European currency 716-24 sinking fund 287~8 social contract 537 social policy 721-2 specie-flow 280-1 specificity of money 737 speculation 83, 295-6, 395, 679-80 speculative bubbles bursts as triggers of crises 556-8 international capital mobility and intra-class conflict 693-6 stability, local 639-40 standard commodity 305,314-15 770 Index of Subjects standard debt 321-5 standard of international monetary relations 305 standard of money 305 see also monetary standard standard money 319-21 standard of value 305,314-15 stock of wealth 139-54 stocks 295-6 store of value 381-2, 728-31 structuralism 519-24 comparison with accommodationism 495-512 subjective probability 58,60-1 substitution, elasticity of 52-4 substitutability 500-1,503-4,506-7, 511 sun spot theory of bank runs 562-3 surplus 261-7 circulation of surplus value 265-6 emergence of 261-5 systemic risk 552-81 tabular standard 315-17 108 terms of trade 687-93 thriftiness 156-7 banks' 180-7, 191 firms' 167-73; impact 173-80 households' 187-95 time 26 Samuelson's theory of money expenditures 128-9 time-dimension of circular flows 361 time preference 52, 53· TIP (tax-based incomes policy) 680 trade, bilateral 732-4 transactions, order of 201-3 transactions demands 267 Treasury circuit 469,478 tt3tonnement uncertainty bank risk 555-6 cause of crises 27 choices in Keynesian models of credit 379-85, 387-8 comparison of approaches 741-2 endogenous 384;385 exogenous 381,385,386-8 future and market 29 information, decisions and 57-65 Keynesian 61-5,387-8,442 unemployment 80 unit of account 319, 443 BFH system 563-4 payment system 202, 203-4, 208, 209,213,216 unit labour costs 696-701,702-3 universal banking 663 value of money 292-4; tabular standard 315-17 standard of 305,314-15 theory: integration of money into 726-31; and monetary theory 366,368 velocity of circulation 301 velocity of money 491, 492, 614, 615,616,673-4 voluntary exchange 203 wage bill postulate 169-70 wage-cost-mark-up theory 401 wage relationship, economy with 216-34 wages 677 profit squeeze and rentier squeeze 687-9,696-701,702-3 reflux problem 401-5 'waiting' 64-5 Walrasian general equilibrium (WGE) 380 Walras's law 115-18 wealth 214-16 Weintraubian theory 401-3 welfare payments 721-2 Wicksellian 'pure credit' economy 403 widow's cruse 411 Zeno's paradox 124 zero-price for money 370-4