Graziani, A., (2003), The Monetary Theory of Production

This page intentionally left blank

The Monetary Theory of Production

In mainstream economic theory money functions as an instrument for thecirculation of commodities or for keeping a stock of liquid wealth. In neithercase is it considered fundamental to the production of goods or the distri-bution of income. Augusto Graziani challenges traditional theories of mon-etary production, arguing that a modern economy based on credit cannot beunderstood without a focus on the administration of credit ows. He arguesthat market asset conguration depends not upon consumer preferences andavailable technologies but on how money and credit are managed. A strongexponent of the circulation theory ofmonetary production, Graziani presentsan original and perhaps controversial argument which will stimulate debateon the topic.

Augusto Graziani is Professor of Economics in the University of Rome LaSapienza. He is the author of Teoria Economica (4th edition, 2002).

Federico Caffe` Lectures

This series of annual lectureswas initiated to honour thememory of FedericoCaffe`. They are jointly sponsored by the Department of Public Economics atthe University of Rome, where Caffe` held a chair from 1959 to 1987, and theBank of Italy, where he served for many years as an adviser. The publicationof the lectures will provide a vehicle for leading scholars in the economicsprofession, and for the interested general reader, to reect on the pressingeconomic and social issues of the times.

The Monetary Theoryof Production

Augusto Graziani

Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, So Paulo

Cambridge University PressThe Edinburgh Building, Cambridge , United Kingdom

First published in print format

isbn-13 978-0-521-81211-5 hardback

isbn-13 978-0-511-07304-5 eBook (EBL)

Augusto Graziani 2003

2003

Information on this title: www.cambridge.org/9780521812115

This book is in copyright. Subject to statutory exception and to the provision ofrelevant collective licensing agreements, no reproduction of any part may take placewithout the written permission of Cambridge University Press.

isbn-10 0-511-07304-6 eBook (EBL)

isbn-10 0-521-81211-9 hardback

Cambridge University Press has no responsibility for the persistence or accuracy ofs for external or third-party internet websites referred to in this book, and does notguarantee that any content on such websites is, or will remain, accurate or appropriate.

Published in the United States of America by Cambridge University Press, New York

www.cambridge.org

-

-

-

-

Contents

Acknowledgements page ix

1 Introduction 1

2 Neoclassical monetary theory 33

3 A monetary economy 58

4 The creation of bank money 82

5 The distribution of income 96

6 The role of nancial markets 114

7 Real and monetary interest 129

8 Implications for monetary policy 138

9 Concluding remarks 144

References 159Index 170

Acknowledgements

The present book reproduces, in an expanded and revisedversion, the content of two lectures given in December 1998in the Faculty of Economics of the University of Rome LaSapienza, in the framework of the series of yearly lecturesdedicated to the memory of Federico Caffe`. A rst syntheticpresentation of the same ideas was published in 1989 as TheTheory of the Monetary Circuit, in the Thames Papers inPolitical Economy series. A subsequent enlarged version waspublished in Italian as La teoria monetaria della produzione(Arezzo, Banca Popolare dellEtruria, 1994).Thanks are due to the Faculty of Economics at the Univer-

sity of Rome La Sapienza, and in particular to my colleaguesDomenico Tosato and Mario Tiberi, for inviting me to givethese lectures. Special thanks are due to Philip Arestis, atthe time editor of the Thames Papers in Political Economy,who in 1988 rst encouraged me to contribute a paper on thetheory of the monetary circuit.

Rome, May 2002

ix

1Introduction

1.1 The theory of the monetary circuit

Over the last twenty years, mostly owing to research car-ried out by French and Italian scholars, a new formulation ofmonetarymacroeconomics, the so-called Theory of themon-etary circuit, also denominated The circulation approach(Deleplace and Nell 1996), has been gaining ground. The ba-sic theoretical tenets of the theory can be synthesised in threemain propositions: rigorous distinction between banks andrms, endogenous determination of the money stock, and re-jection of the marginal theory of distribution.1

The circulation approach in the early Swedishand German literatures

Under a strictly chronological criterion, the rst descriptionof a monetary circuit is found in Knut Wicksells rightly cel-ebrated monograph on Interest and Prices.2

1 A general presentation of the circuit approach is contained in Lavoie 1987,Graziani 1989, Halevi and Taouil 1998. An implicit description of the cir-cuit mechanism can be found in Bossone 2001. An excellent review andcritical assessment of the post-Keynesian reading of the macroeconomicmodel is given byArestis 1997, chapter 3.Adetailed analysis of the conceptof endogenous money and of the debate between accommodationists (sup-porters of endogenous money) and structuralists (accepting endogenousmoney only under severe qualications) is contained in Fontana 2001.

2 Wicksell 1936 [1898], chapter 9, section B. In Wicksells wake, theSwedish school has analysed the monetary circulation along the same

1

2 The Monetary Theory of Production

Wicksells analysis strongly inuenced a number of authorsbelonging to the Austrian and German schools, both hav-ing a long tradition in the analysis of money and banking.3

The very term circuit, introduced in contemporary litera-ture by French authors, reproduces the German Kreislauf, aterm used by German writers to describe the circulation ofmoney and of real goods (Schumpeter 1934 [1911], chapter 1).Neisser devoted two works to the analysis of money circula-tion. The rst one (Neisser 1928) gives ample space to therelationships between banks and rms. The second one(Neisser 1931) is specically devoted to the analysis of cir-culation among rms and between rms and wage earners.N. Johannsen, the famous amateur economist recalled byKeynes in the Treatise on Money (1971 [1930], chapter 27),analyses in detail the monetary circuit in his book TheCircuit of Money published in 1903 under the pseudonymof J.J.O. Lahn (an analysis of Johannsens book is containedin Hagemann and Ruhl 1987). The German contributions tothe analysis of the circular ow between the 1930s and the1960s are analysed in detail by Schmitt and Greppi (1996).More recently, a revival of the circulation approach in

Germany has been carried out by the so-called School ofMonetary Keynesianism, headed by Hajo Riese in Berlin. TheBerlin school describes the market mechanism as a mone-tary circuit, rejects the marginal theory of distribution anddenes money as an institutional entity and not as a spon-taneous product of the market (Luken Klassen 1998; Riese1998).

lines (Lundberg 1937). The Introduction by L. Berti to the Italian edi-tion of Myrdal 1939 is an excellent guide to the Swedish monetary theoryconsidered in this perspective.

3 Schumpeter 1934 [1911]; vonMises 1934 [1912]; Hahn 1920; Neisser 1928,1931 and 1950 [1934]; Schneider 1962, chapter 2. A detailed analysis ofSchumpeters monetary thought is contained in Messori 1984. De Vecchi1993 is a most important piece of research centred on works written bySchumpeter before he moved to the United States.

Introduction 3

The circulation approach in France

In many aspects, the French school of the circuit had a pre-cursor in an isolated French scholar, Jacques Le Bourva. Tohim is due one of the rst andmore lucid presentations of themonetary circuit as well as of the process of money creationand destruction, both viewed as endogenous phenomena(Le Bourva 1962; reprinted with a Comment by Marc Lavoie1992).More recently, the revival and analytical development of

circuit theory in France has been due to three main groupsof authors. The so-called Dijon school is headed by BernardSchmitt, an author who has given a precise formulation ofthe principles of the theory, dened a particular terminol-ogy and constantly applied both of them in his works. Theresearch by Schmitt goes beyond mere theoretical analysisand is largely concerned with problems of both internationalpayments and developing countries, which he examinesfrom his very individual theoretical point of view (Schmitt1972).A second set of scholars gathers around Alain Parguez, for

many years the editor of the series Monnaie et Production,published under his editorship by ISMEA of Paris between1984 and 1996. The series contains contributions by schol-ars from various countries. So long as it was published, itwas the only really international connection established be-tween French followers of the circulation approach and theircounterparts in Anglo-Saxon countries. The group headedby Parguez is strictly connected to French-Canadian authors,among whom the best known are Marc Lavoie and MarioSeccareccia from the University of Ottawa (in fact, one ofthe rst reviews of circuit theory and of the contributions ofthe main authors belonging to it is due to Lavoie (1987)). TheParguez group is not as particular as Bernard Schmitt in ad-hering to the conceptual and terminological subtleties on


which he often insists, and is largely concernedwith present-day problems of economic policy in advanced countries.Among the French and French-Canadian representatives ofthe circuit theory, Parguez and Lavoie are the two who moveclosest to the post-Keynesian approach (Parguez 1975 and1984; Lavoie is himself the author of a handbook titled Foun-dations of Post-Keynesian Economic Analysis, Lavoie 1993).A third group, active mostly in the 1980s, was formed in

Bordeaux around Francois Poulon. Starting from the basicideas of circuit theory, Poulon has endeavoured to constructa complete macroeconomicmodel. Poulon is the only Frenchfollower of the theory to have written a complete handbookof macroeconomics (Poulon 1982).

The circulation approach in Italy

Among the Italian precursors, a special mention is due toProfessor Paolo Sylos Labini who, in contrast to the domi-nant Italian doctrine, has always maintained that the moneystock is endogenously determined thanks to the creation ofmoney by the banks in response to the demand for creditfrom rms (Sylos Labini 1948). In more recent years, the doc-trine of themonetary circuit has arousedwide interest amongItalian scholars. A detailed analysis of circuit theory is givenby Graziani 1989; a typical circuit analysis is performed byMessori 1985.

The circulation approach inAnglo-Saxon countries

Approaches very similar in content to the circuit approachare to be found in the so-called Anglo-Saxon high theory ofthe 1930s. An analysis of money circulation identical in sub-stance to the circulation approach is to be found in Keynessworks, in particular in theTreatise onMoney (1930) aswell asin the 193739 essays which followed the publication of the

Introduction 5

General Theory (this point is illustrated in detail in Graziani1991). A similar approach was followed by Joan Robinson inan often neglected chapter of The Accumulation of Capital(Robinson 1956: 25, The meaning of money), as well asby other contemporary Anglo-Saxon authors (Dillard 1980;Godley and Cripps 1981; Godley 1990;Wray 1993; and, alongthe same lines, Eboli 1991).

1.2 Theoretical vicissitudes

Any elementary presentation of monetary theory makes clearthat money, besides being a numeraire used for measuringprices, performs twomain functions: (a) money is an interme-diary of exchange, since, in present-day economies, paymentis nearly alwaysmade inmoney, barter having practically dis-appeared; (b) money is a form of wealth, since anybody canhold the whole or part of his or her own wealth in the formof liquid balances, while waiting to establish what seems tobe the most protable placement.Money as an intermediary of exchange is the older and

more intuitive notion of money. In fact, in the imagination ofthe person in the street, money is nomore than ameans of en-abling agents to buy commodities. If money, instead of beingspent in the market, is kept as an idle balance, this is com-monly understood as being a merely temporary destination,connected to the uncertainty of the moment and acceptedonly by agents waiting to make use of it in its natural func-tion: being exchanged for real goods.The conception ofmoney as an intermediary of exchange is

the rst to appear in the history of economic thought. AdamSmith explains how the adoption of money is a consequenceof the division of labour and a spontaneous reaction of themarket to the practical problems that direct barter would cre-ate. After telling the long story of primitive money, Smithconcludes: It is in this manner that money has become inall civilised nations the universal instrument of commerce,


by the intervention of which goods of all kinds are boughtand sold, or exchanged for one another (Smith 1993 [1776],book I, chapter 4: 34). Similarly, in StuartMillswords,moneyis the medium through which the incomes of the differentmembers of the community are distributed to them, and themeasure bywhich they estimate their possessions (Mill 1909[1848], book III, chapter 7, 3: 487).If money is a mere intermediary of exchange, and if, as

is postulated in general economic equilibrium theory, eachagent keeps a strictly balanced budget (equality between therespective values of goods and services bought and sold), thenal outcome is that all that an agent buys is paid for bymeans of real goods or services supplied (this is why sup-porters of this view insist on the fact that money, if properlyunderstood, while being an intermediary of exchange, is nomeans of payment in itself). The whole market mechanismappears to be in the nature of a general barter, made easier bythe intermediation of money, possibly obscured by the veilof money, but not altered in its substance.4

Carl Menger, a stauch supporter of the denition of moneyas an intermediary of exchange, used to consider money asbeing the spontaneous product of market choices. Accordingto his historical reconstruction of the origin of money, amongall goods traded in the market, one of them emerged becauseof its being scarce, durable and easy to carry.5 Gradually all

4 Patinkin and Steiger 1989 critically examine the character of the veil as-signed to money. Paradoxically, some circuit theorists, like Schmitt andCencini, come very close to the neoclassical approach in dening moneyas amere technical instrument allowing goods to be exchanged on themar-ket. In this view, payments made by an individual are actually completedonlywhen the budget is perfectly balanced so that the purchase of each sin-gle commodity has been paid for by means of other commodities. Moneyis a pure instrument of circulation. It is no wealth, nor is it endowed withpurchasing power. It is a mere numerical instrument having the functionof measuring and making exchange possible (Cencini and Schmitt 1992:115).

5 Menger 1892. Mengers teaching was followed by Hicks, who adds that,as soon as a specic precious metal became a recognised intermediary ofexchange, the state was ready to come in and take over the coinage ofmoney (Hicks 1989: 63ff.).

Introduction 7

agents came to demand that particular good exclusively aspayment for any other goods supplied, with the consequencethat that good nally became the general intermediary of allexchanges. In Mengers view, paper money is (and should be)no more than a representative of metal money, this being theonly real and sound money.While being adequate at the intuitive level, the concept of

money as a mere intermediary of exchange was abandonedbecause of two serious analytical problems associated withit, the rst being the correct denition of the utility of money,the second being the possibility of considering money itselfas an observable magnitude. Both aspects deserve detailedexamination.

The controversy concerning the correct denition of the util-ity of money, which took place at the end of the nineteenthand the beginning of the twentieth century, was one conse-quence of the dominance of the theory of value based on util-ity. At the time, according to the dominant theory, the valueof any good was determined by its marginal utility. Money,being used not for direct consumption but as an instrumentfor acquiring other goods, was not considered to be the sourceof any direct utility. The utility of money was therefore de-ned as an indirect utility, determined by the utility of thebundle of commodities that could be purchased by meansof a given money stock. This point, already put forward byvon Wieser and Bohm-Bawerk, was formulated with specialvigour by Maffeo Pantaleoni in his famous Pure Economics.When introducing his analysis of money, Pantaleoni writes:[Money] may be absolutely destitute of all direct utility . . .The more the particular thing we use as money is destitute ofdirect utility, the more essentially it is money . . . Money isonly endowedwith an indirect utility, consisting in its powerof obtaining for us, solely by means of exchange, some di-rect commodity (Pantaleoni 1898: 221). The same principlewas nally codied by Ludwig vonMises in his famous 1912Theory of Money: In the case of money subjective use-value


and subjective exchange-value coincide . . . The subjectivevalue of money always depends on the subjective value ofthe other economic goods that can be obtained in exchangefor it (von Mises 1934 [1912]: 97, 98).However, as Helfferich convincingly remarked, the volume

of goods that a unit of money can buy depends on the level ofmoney prices and therefore on the exchange value of money.Thus, in order to measure the utility of money and its value,one should already know its value. We are clearly arguingin a circle (Helfferich 1919; a detailed discussion of the sameproblem is contained in Schumpeter 1954, part IV, chapter 8:108691).In fact von Mises himself was fully conscious of the prob-

lem and, in a somewhat devious way, tried to nd a solutionto it. Von Mises tried to introduce a distinction, which sub-sequently entered into common usage, between individualexperiments and market experiments (Patinkin 1965, Math-ematical Appendix, n. 1). Individuals, when entering themarket, ignore the ruling prices. This does not prevent themfrom preparing a strategy of action (their demand or supplyschedule) or from determining the quantities that they areprepared to buy or sell as functions of all possible prices.What consumers decide upon when entering the market isnot the quantity that they will actually buy (a quantity thatwill only be determined once the prevailing price is known),but their demand function, in which prices appear as param-eters. In any possible price constellation, money will havea different purchasing power and therefore a different util-ity. The individual is ready to face any possible set of pricesand therefore any possible value of money. Individuals ig-nore the actual level of prices; but, by considering prices asparameters, they are ready to consider their own money bal-ance as being endowed with a marginal utility which willitself depend on the actually prevailing set of prices. On thebasis of plans previously drawn, individuals will start negoti-ations, thus contributing to the determination of equilibrium

Introduction 9

prices. Once the set of prices that makes demand and supplyequal in each market has been reached, negotiations cometo an end, equilibrium prices are known to all participantsand the marginal utility of money is also determined. It isa well-known principle of demand theory that the recipro-cal interdependence of prices and quantities exchanged doesnot make the problem indeterminate. In the sameway, the in-terdependence between prices and value of money does notlead to a circular argument.Unfortunately, von Misess presentation (1934 [1912]: 97

107) wasmade obscure by his attempt (this one surely wrong)to demonstrate that a single individual is able to know theutility of money even before the market has reached an equi-libriumposition. In order to show that an individual is able toplan hismarket strategy before knowing the equilibrium levelof prices, von Mises imagines that the individual, when en-tering the market, assumes present prices to be equal to thoseprevailing in the previous period. The same prices should de-termine the value of money, and therefore its utility. Any pe-riod thus relates to the previous one, back to an initial time inhistory when commodity money was used not as money butas a material good having a direct utility. The value of moneythus comes to depend on the value of gold as a commodity.Von Misess initial intuition was correct. But the develop-ment of his reasoning was unfortunate and he was himselfaccused of arguing in a circle (Patinkin 1965, appendix D).From this moment onwards, the theory of money took a

different route. Instead of reformulating von Misess reason-ing in a more correct way, it seemed simpler to modify thetheoretical approach at its very root. Thus the idea was in-troduced that the utility of money is not an indirect one (de-rived from the utility of goods that money can buy), but thedirect utility that an agent draws from having a money hold-ing. By this denition, money is considered to yield utilitynot when spent but when kept idle. An individual who de-mandsmoney in order to spend it is considered as demanding


goods, notmoney; a true demand formoney is only expressedby individuals wanting money in order to keep it as a liquidbalance.Clear traces of a similar idea can be found in Marshall

(1975 [1870]: 1667) and Wicksell (1936 [1898], chapter 6,section A). The rst to give a rigorous analytical formula-tion of this approachwas the almost forgotten economist KarlSchlesinger. In an essay published in 1914 on the Theory of aMoney and Credit Economy (Schlesinger 1914), he suggestedthat the need that money satises, rather than being a needfor real goods that money can buy, is the need of having a liq-uid balance as protection against uncertainty. In Schlesingersown words: Let us suppose that chance decits cannot becovered by credits. They can then be covered only by sellingthe rm . . . or else by cash reserves held against such con-tingencies . . . The individual loss in not earning an intereston these cash reserves can be regarded as a risk premium(Schlesinger 1914: 967). Schlesingers book went unnoticedand remained totally ignored for many years.Indications along similar lines are given by Irving Fisher,

who writes: . . . in a world of chance and sudden changes,quick saleability, or liquidity, is a great advantage . . . Themost saleable of all property is, of course, money: and asKarl [sic] Menger pointed out, it is precisely this saleabilitywhich makes it money. The convenience of surely being able,without any previous preparation, to dispose of it for anyexchange, in other words its liquidity, is itself a sufcientreturn upon the capital which a man seems to keep idle inmoney form (Fisher 1930: 21516; similar statements are inFisher 1963 [1911]: 8ff.). Finally J. R. Hickss famous articleof 1933 made it clear that money yields utility in the formof protection against uncertainty, and that consequently theutility of money comes not from spending but rather from notspending it. The demand for money is therefore present onlyin conditions of uncertainty and is a demand for a stock of

Introduction 11

wealth. This result nally overcame the problems connectedto the denition of the utility of money andmeant that moneycould be considered capable of yielding direct utility. Thanksto his rigorous presentation, Hicks was credited as havingbeen the rst to resolve the utility-of-money controversy.

The second kind of problem connected to the denition ofmoney as an intermediary of exchange emerges most clearlyin the analysis of money circulation in modern times.Nowadays,money is papermoney introduced into themar-

ket bymeans of bank credit. The banking system grants creditto single agents having to make a payment, for instance rmshaving to hire labour and pay wages. The moment wagesare paid, the rms become debtors and the wage earners be-come creditors of the bank. The result of the operation is theemergence of a stock of money equal in amount to the creditgranted to rms. The money stock stays in existence as longas the debt of the rms is not repaid. Once the debt is repaid,the money circuit is closed and the money initially createdis also destroyed.Let us now assume a world free from uncertainty and pop-

ulated by perfectly rational agents. In this world, any agentwill go into debt only at the very moment in which he has tomake a payment. Similarly, any agent who receives a moneypayment tries to spend it as soon as possible on goods or onsecurities. Both kinds of expenditure bring themoney back tothe rms, who immediately repay their debt to the bank. In ahypothetical world free from uncertainty and from frictions,the aforementioned steps would take place in an immediatesuccession with no time lag. This means that money is cre-ated, passed on from one agent to the next, and destroyedin the same instant. If this is the case, money is no longeran observable magnitude and the paradoxical result emergesof a monetary economy being dened as an economy inwhich money, in spite of its being by denition necessary for


exchanges to take place, escapes any observation and anypos-siblemeasurement. If all agents behaved as J.B. Say imagined,namely spending any amount of money as soon as received,the velocity of circulation would be innite, money wouldbe destroyed as soon as it was created and any attempt tomeasure the money stock in existence at any given momentof time would invariably produce a zero result. As a para-doxical consequence, the image would emerge of a monetaryeconomy (in the sense of an economy having ruled out barterand in which all payments are regulated in terms of money)in which money did not exist.6

A consequence is that it is almost impossible to reconcile asimilar denition with the Walrasian model of general equi-librium. If, as is typical of the Walrasian model, the negotia-tions for the denition of equilibrium prices precede actualexchanges, and if all exchanges take place at the same mo-ment at equilibriumprices, all agents simultaneously sell andbuy goods having an identical total value. Thus the wholeprocess of exchanges takes the form of a great barter, whichno longer requires the use of money. If the model is ex-tended to a number of periods, but the assumption of pre-determined prices is preserved by assuming the presence offuture markets, equilibrium prices are simultaneously deter-mined for the current as well as for all future periods. Oncemore, the model depicts an economy which can work with-out money. The theoretical approach of theWalrasian model,owing to the simultaneous determination of all present andfuture prices, ignores any possible uncertainty, thus ruling

6 Knut Wicksell saw this problem and envisaged a model structured so asto avoid it: in Wicksells model, wage earners buy nished products notfrom producers but from traders, who sell the product of the previous pro-duction cycle and pay the revenue into bank deposits earning the currentrate of interest. At the end of the current production cycle, traders buythe nished product and replenish their stock. In this case, an amount ofmoney equal to the initial liquidity requirements of producers is alwaysin existence.

Introduction 13

out any possible demand for liquid balances.7 General equi-librium theory, owing to the problem of reconciling it withthe theory of a monetary economy, was downgraded to thetheory of barter economy.For money to be an observable magnitude, it must be kept

by single agents for a nite period of time, no matter howshort, thus taking the form of a cash balance, be it notes orbank deposits. Since, as mentioned earlier, liquid balancesare kept as a protection against uncertainty, this means that,for money to be an observable magnitude, the market mustbe operating under uncertainty. If we move in a hypotheti-cal market free from uncertainty, liquid balances disappear,and with them the possibility of observing and measuringthe money stock in existence. As Benetti and Cartelier haveremarked, once one decides to abstract from uncertainty, thevery existence of money balances is ruled out, except whenthe economy is out of equilibrium (Benetti and Cartelier1990). In fact, when Keynes, in the General Theory, denedmoney as a cash balance having the function of protectingagents from uncertainty, he was choosing the only analyt-ically satisfactory solution and accepting the only possibleconception ofmoneywhich couldmake it an observablemag-nitude. It is no surprise that theKeynesian approach tomoneyhas been considered for over half a century the nal conclu-sion of a long controversy.

7 A similar result is well known. As long ago as 1930, Erik Lindhal, who wasworking in the framework of a general equilibriummodel, had noticed thatmoney creation on the part of the banking system is only possible out ofequilibrium (Lindhal 1930, part II, chapter 1). The same remark can befound in later authors (Debreu 1959; Arrow and Hahn 1972: 338; Hahn1982). An indirect proof of this point is that Clower, in order to give arole to money in a general equilibrium context, builds a model in which,in contrast to the typical structure of Walrasian models, exchanges arenot synchronised and can be started only if at least some of the agentsdispose of an initial money balance to nance their initial expenditures(since Clower does not consider bank credit, the nature of money in hismodel remains undened; Clower 1969: 20211. An excellent review ofthe problem is given by Villieu 1993).


The denition of money as a stock of wealth was con-sidered unobjectionable and became universally accepted.8

However, once universally adopted, the denition of moneyas a stock of liquid wealth went through gradual alterationsand through a parallel degeneration. Since money was nolonger considered in its role as a means of payment and wasconsidered only as a part of the stock of wealth, it was nolonger identied with the ow of payments performed overa period of time. A consequence was a tendency to considerthe stock of money as a given parameter.Nothing would have prevented, in principle, money be-

ing placed on the same footing as any other commodity, andthe production of money being analysed along with the pro-duction of other commodities. By proceeding in this direc-tion, it would have been possible to analyse the formation ofthe money stock as the result of negotiations between banksand rms in the money market. In fact the very denitionof equilibrium was an obstacle in this direction. A generalequilibrium is dened not only by the objective conditions(equality of supply and demand in all markets), but also bythe so-called subjective conditions, requiring that the bud-get constraint be satised for all agents so that all individualbudgets are rigorously balanced. The budget constraint be-ing interpreted in its most restrictive meaning (not only asan equality between assets and liabilities but as a strict bal-ance of current income and expenditure), equilibrium wasmade to coincide with a position in which all agents have

8 An aside is in order. As previouslymentioned,Menger emphasised the factthat a specic commodity emerges as money as the consequence of a spon-taneous choice made by market agents. In the Treatise on Money, Keynes,following Knapp, had dened money as a means of payment recognisedby the state: . . . it is a peculiar characteristic of money contracts that it isthe State . . . which decides what it is that must be delivered as a lawful orcustomary discharge of a contract . . ., Keynes 1971 [1930], vol. 1: 4, 6. IntheGeneral Theory, while trying to demonstrate that, owing to uncertainty,once money is present it may become the more convenient form of wealth,Keynes doesnt give any explicit denition of money. Knapps denitionhas been recently revived in a most convinced way by Wray 1998.

Introduction 15

extinguished any possible debt, including their debt to thebanking sector. If in equilibrium all debts have been extin-guished, the money stock has disappeared. Once more itseems impossible to reconcile a rigorously dened equilib-rium position with the presence of money, with the only dif-ference that, in this case, the disappearance of money doesnot depend on an inconsistency between a demand formoneybalances and rational behaviour in the absence of uncer-tainty, but on the simple fact that agents automatically elim-inate the presence of money by simply respecting their ownbudget constraints (on this point more will be added later,see 2.2).This time, the way out was found by enlarging the model

and including in it the government sector. In principle, thegovernment sector is not held to have a balanced budget.Consequently, the presence of a current decit in the gov-ernments budget is not incompatible with a general equilib-rium.More precisely, a position of full equilibrium is denedas one in which the amount of government debt not nancedby placement of securities (and therefore the amount of le-gal money outstanding) is equal to the amount of money de-manded in the market. In fact nowadays this kind of solution(money entirely created by the government decit and beingin the nature of an exogenous magnitude) is presented as anobvious truth in any introductory presentation.However, this solution is in itself weak. To begin with, it

modies the very nature of scal policy in that the level ofthe government decit is conceived as determined not by therequirements of the community in terms of government ser-vices, but by the money stock needed to ensure the smoothcirculation of goods in the market. The government is nolonger viewed as a supplier of social services but as a supplierof liquidity (Riese 1998: 56). If both roles of the governmentare to be satised at the same time, the nal level of govern-ment decit has to be such as to fund the provision of socialservices while at the same time supplying the required stock


of money two targets hard to reach with the same level ofgovernment decit (Sawyer 1985: 16; Tobin 1986: 11).In addition to that, if, as is customary in most of mod-

ern macroeconomics, monetary theory considers money asa given stock, this leaves unexplained how the available pur-chasing power is distributed among single agents or amongsocial groups present in the market. This is no great loss foranyonewho is a true follower of neoclassical theory. In fact, ina neoclassical theoretical perspective the purchasing powerat the disposal of each individual does not depend on themoney stock in his possession but on the amount of real goodsor services that he is willing to supply and able to sell. Theinitial distribution of the money stock among single agents,in itself, is not a relevant factor. The same is no great loss ei-ther to the followers of the post-Keynesian school, especiallyto the followers of Kaldor. To them, the banking system per-forms a totally passive role vis-a`-vis the demand for creditcoming from producers. The rms can consequently carryout their production plans free from any nancial constraint.The same loss becomes, however, a substantial limitation to

anyone who thinks that, when creating liquidity, the bankingsystem operates a selection process. In this case, the agentsendowedwith an autonomous and potentially unlimited pur-chasing power are not all possible agents present in the mar-ket but only thosewho are considered eligible for bank credit.These usually belong to the class of entrepreneurs, to the ex-clusion of wage earners. Circuit theorists subscribe to thistrain of thought (this point is dealt with later on in thischapter, 1.4). To them, the denition of money as a meansof payment remains an essential element in the analysis ofmacroeconomic equilibrium.

1.3 The circuit version

In opposition to the dominant Keynesian view of money asa stock of wealth, circuit theorists remark that the rst and

Introduction 17

foremost role of money is to make possible the circulation ofcommodities. Therefore money appears in its authentic ca-pacity only when a good is exchanged against money andmoney passes from the balances of one agent to the bal-ances of some other one. In this perspective, the more rig-orous among circuit theorists insist on the fact that whenmoney is kept idle, even if only to cover future payments, itis no longer an instrument of circulation but rather a stock ofwealth (Schmitt 1996: 132ff.).In opposition also to the neoclassical view of money as a

mere intermediary of exchange, circuit theorists emphasisethe fact that money should be viewed as an authentic meansof payment. Money, they remark, enters the market by way ofbank credit. When a rm is making use of a bank overdraft, itis in fact acquiring commodities or labour without giving anyreal good in exchange; which means that it is using money asa means of payment. In terms of substance, circuit theoristsremark that money exerts its primary inuence on macroeco-nomic equilibrium when it is used for buying commoditiesand not when it is kept as an idle balance. Under this aspect,circuit theorists clearly differ from the followers of Keynes,who insist on the fact that money makes its presence felt justbecause it can be kept as a liquid balance and become idlemoney.The decision of circuit theorists to shift their attention

away from the time that liquid balances are kept as such andto concentrate on the time that money is used in order tomake a payment displaces the focus of theoretical analysis.The dominant theory of money, when analysing the demandfor money, enquires about its motivations and possible uc-tuations; when analysing the money supply, the theory oftenconsiders the money stock as the result of independent de-cisions taken by the monetary authorities. Circuit theoristsinstead concentrate their analysis on the chain of payments,starting with the initial creation of liquid means, going onto the successive utilisations of money in the market, andending with the nal destruction of money. The very term


monetary circuit draws its origin from the fact that thetheory examines the complete life cycle of money, from itscreation by the banking system, through its circulation inthe market, to its being repaid to the banks and consequentdestruction.

1.4 Circuit theories and neoclassical analysis

The approach adopted by circuit theorists opens a deep the-oretical cleavage separating the circuit doctrine both fromneoclassical and from Keynesian theory.It is a well-known fact that neoclassical theory has been

the object of a number of analytical criticisms: starting withthe older objections from students of welfare theory (lackof perfect competition, external economies or diseconomies,increasing or constant returns, presence of indivisibilities),down to Sraffas critique concerning the possibility of mea-suring capital and therefore of applying the marginal theoryof distribution, and on to the modern theories of asymmetri-cal information and of interdependence between quality andprice.Most if not all of these criticisms do not reject the individ-

ualistic approach typical of neoclassical theory. On the con-trary, according to circuit theorists, so long as this approach ispreserved, the fundamental limits of neoclassical theory arenot overcome. The rst and most important of those limits,according to circuit theorists, is that any theory based on anindividualistic approach is necessarily conned tomicroeco-nomics and is unable to build a truemacroeconomic analysis.A proof of that is given by the fact that all theories based onan individualistic approach have in common the denition ofmacroeconomics as the result of an aggregation performed ona microeconomic model and not as an independent analysisbased on new and different assumptions.In the perspective of circuit theory, a simple aggrega-

tion of the individual behaviour functions doesnt turn a

Introduction 19

microeconomicmodel into a truemacroeconomic theory. Thestarting point for a construction of a macroeconomic modelcan only be the identication of the social groups present inthe community, followed by the denition of the conditionsnecessary for their reproduction and perpetuation over time.An example of a truemacroeconomic approach is given by theclassical economists,who started froman apriori subdivisionof society into a number of social classes, each of them havinga different initial wealth endowment (landowners own pro-ductive resources, entrepreneurs are able to organise produc-tive factors, wage earners can only supply their own labour).The same can be said of Marxian analysis, based on the dis-tinction between capitalists and proletarians, a distinctioncorresponding to the separation between labour and meansof production. The same is also true of Keynes, who madeuse of a sort of a priori distinction between consumers, whoevaluate consumption goods according to their immediateutility, and investors, who evaluate capital goods accordingto subjective and uncertain prot expectations.In a similar perspective, circuit theorists introduce a pre-

liminary distinction between producers and wage earners,producers having access to bank credit and wage earners be-ing excluded from it. The two groups enter the market hav-ing different initial endowments. Entrepreneurs, being ad-mitted to bank credit, can rely on a potentially unlimitedpurchasing power, while wage earners can only dispose ofas much money as they have previously earned. The two so-cial groups have to comply with totally different budget con-straints, which makes a basic difference in the denition oftheir own behaviours.The contrast with neoclassical theory appears even more

clearly if one thinks of the fact that in a perfectmarket, such asthe one assumed in the neoclassical model, the fact of havingmoney actually available does not create a constraint to thepurchasing power of the agent. As already mentioned, in theneoclassical model, while it is true that no agent can violate


his own budget constraint, the purchasing power of eachindividual is determined by magnitudes which are not mon-etary but real in nature, such as his working ability and theamount of his real fortune. In a perfect market, any real re-source can be converted intomoney at the ruling price, when-ever the opportunity arises to exchange it for a different good.In this setting any possible liquid balance is but one of thevarious kinds of wealth pertaining to the agent and derivingfrom real income previously produced. Even more: any agentexpecting to get in the future a higher ow of income, or tobecome the owner of new wealth, has access to bank creditand can get immediate liquid resources against the promiseto repay the debt when due. Therefore, the purchasing powerof an agent is not limited by his present wealth but is deter-mined by his ability to produce real goods in a much widertime horizon.Circuit theorists start with a totally opposite vision. In their

view, in a monetary economy, precise mechanisms prevailwhich bring purchasing power into the hands of some agentsrather than others. To begin with, since the market does notguarantee full employment, the purchasing power of an agentis never determined by the simple ability to perform produc-tive work but rather by the fact of being actually employedandof beingpaid in terms ofmoney. The same is true of credit,which is not granted to anyone presumably able to repayhis debt, but only to selected agents, usually being produc-tive rms. Only rms have actual access to bank credit andtherefore enjoy a purchasing power exceeding their presentwealth. As a rule, instead, wage earners can enter the marketonly after they have sold their own labour and received thecorresponding pay.9 A similar assumption clearly echoes the

9 Such an assumption is explicitly made by Benetti and Cartelier (1990)and by Cartelier (1996). This is by no means a new idea. In the eighteenthcentury it was commonly accepted that money and power should gohand in hand (a most persuasive analysis of this point is performedby Giacomin 1994). In some sense the same assumption is to be found

Introduction 21

Marxian distinction between a class of property owners anda class of propertyless workers, a distinction now turned intothe division between a class able to spend beyond its currentincome and another class that is subject to a budget constraintdetermined by already earned income. Once this assumptionis accepted, the theorem of the neutrality of money is clearlyrejected in point of principle, since any creation of money in-creases the spending ability of a well-dened group of agents,whichmeans that the effects it exerts on the price level cannotbe neutral.10

Similarly, circuit theory parts company with the neoclas-sical approach in the denition of the money stock. Neoclas-sical theory concentrates on the equilibrium position whichis reached when the stock of money equals the liquid bal-ances permanently required by the public. Circuit theory triesinstead to consider the whole life cycle of money, startingwith its creation by means of bank loans and ending withits destruction when those loans are repaid. In a traditionalsetting, based on agricultural production, money can be con-sidered as being created at the beginning of the productioncycle and gradually destroyed as bank loans are repaid. In thiscase the average money stock would equal half of the initialnance. In a modern setting, where synchronised industrial

in Marxian thought. Marx denes a capitalist as an agent having theproperty of means of production and being adequately endowed withmoney. Access to bank credit assures precisely the necessary endowmentof liquidity. Nowadays this is no longer a common assumption. Moreoften than not, contemporary literature insists on the fact that creditgranted to households equals or even exeeds credit granted to rms(a well-documented argument is contained in Howells 2001). It ishowever highly debatable whether credit granted to households is reallygiven to consumers or rather is in fact indirectly granted to rms, byallowing consumers to buy nished products.

10 Schumpeter, when discussing the traditional principle according towhich money is just a veil that can and must be removed whenever theanalysis concerns the fundamental features of the economic process, justas a veil must be drawn aside if we are to see the face behind it, warnsthe reader that the essential features of the capitalist process may de-pend upon the veil and that the face behind it is incomplete without it(Schumpeter 1954, chapter 6: 277, 278).


production is the rule, loans are continually granted and con-tinually repaid. In a stationary economy, the money stockwould be equal to the amount of initial nance created bythe banks.

1.5 Circuit theory and Keynesian analysis

The very fact of considering the unequal initial distributionof purchasing power creates a conict between circuit theo-rists and the Keynesian school. It has already been recalledthat the Keynesian school gives a special weight to moneyonly because money can be kept as an idle balance. In fact, inKeynesian models, sudden shifts in the speculative demandfor money are responsible for demand failures and for bothrecurrent andprolongedunemployment crises. To circuit the-orists, on the other hand, uctuations in liquidity preferenceand in aggregate demand, while being an indisputable ele-ment of historical experience, are not the most relevant as-pect of a market economy. The path of an economy is inu-enced in a much deeper way by money and credit ows, bythe investment decisions emerging from the negotiations be-tween banks and rms, by the proportions in which aggregateproduction is divided between consumption and investmentgoods, and by the consequent distribution of national incomebetween wages and prots. These are the basic mechanismsexplaining, among other things, uctuations in aggregate de-mand, the same uctuations that the Keynesian school tracesback to the simple action of liquidity preference.Todays macroeconomic theory, while originating from

Keyness General Theory, is usually presented in the formof the ISLM model, a version due to Hicks, Hansen andModigliani. The ISLM model is rejected by circuit theoristson the ground that, by considering themoney stock as a givenmagnitude, it omits the analysis of the creation of money,neglects the relationships between banks and rms (in theHicksHansen model, banks and rms are in fact aggregated

Introduction 23

into one single sector), and consequently ignores the interde-pendence between the IS and the LM lines. Not all of thesecriticisms are raised against Keyness original model as ex-pressed in the General Theory; yet, circuit theorists make nosecret of the fact that, between the twomainKeynesianworks,they assign a clear preference to the Treatise on Money overthe General Theory.

1.6 Circuit theory and post-Keynesian theory

On one single point another divergence emerges between cir-cuit theorists and the post-Keynesian school. It is a well-known fact that the post-Keynesians of the rst generation(Nicholas Kaldor, Joan Robinson, Richard Kahn) do not con-sider as relevant the problem of the money supply, a problemthat circuit theorists consider to be a fundamental one. Justi-cation by the post-Keynesians for neglecting themoney sup-ply is grounded on twomain arguments. Therst is that, sincethe monetary authorities are forced to satisfy the demand forliquidity coming from the market, the money supply tends tobe innitely elastic. The second argument is that, even if themonetary authorities could limit the amount of credit beinggranted by the banks, the market would nd other forms ofliquidity,mainly consisting in creditmutually granted by sin-gle agents. This would allow the agents to carry out their ownproduction plans whatever the credit policy run by the mon-etary authorities. The conclusion drawn by post-Keynesiansis that liquidity is never a serious constraint.11

A consequence of a similar approach is that post-Keynesians never analyse the relationships between banksand rms. Here a clear divergence emerges between the post-Keynesian and the circuit schools. The relationships between

11 Kaldor 1985; Leijonhufvud andHeymann 1991. Keynes himself expresseda totally different idea: . . . the investment market can become congestedthrough shortage of cash. It can never become congested through shortageof saving (Keynes 1973c [1937]: 669 [222]).


rms and banks are fundamental to circuit theoreticians,since such relationships determine the amount of liquidityavailable. Under this aspect, smaller divergences emerge be-tween the circuit doctrine and the second generation of post-Keynesians (as represented by S.Weintraub, P. Davidson, H.P.Minsky, J. Kregel, and B.J. Moore), who pay instead great at-tention to the problems concerning the money supply andthe nance available to rms.Apoint of convergence between circuit theory and the post-

Keynesian school can be found in the analysis of income dis-tribution. Here circuit theorists follow closely what was oncenamed the KeynesKalecki formulation. First sketched byD.H. Robertson (1926) andbyKeynes in hisTreatise onMoney(Keynes 1971 [1930]), this approach to income distributionwas developed by Kalecki (1991 [1942]), Kaldor (1956), andJoan Robinson (1956). According to this approach, rms candecide the activity levels and the nature of production (con-sumption or investment goods), while wage earners, what-ever the level of their money wage, can only buy real con-sumer goods in the amount made available by producers(Kregel 1973, chapter 10). Therefore the actual level of realwages is determined by producers and not by regular marketnegotiations. Atmost it can be the object of a political conict.This point, emphasising the widely different market powerof producers and wage earners, is sometimes expressed bysaying that, in the labour market, wage earners can only ne-gotiate money wages, never real wages. This doesnt dependonwage earners being irrational agents, or on their being sub-ject to money illusion. It is instead a typical feature of anymarket economy that producers can decide the proportionsof consumer goods and investment goods produced, whilewage earners can only decide how their money wages willbe spent. If wage earners could actually negotiate the level ofreal wages, producers would correspondingly lose the powerto determine the level and the nature of production. Assum-ing that producers can control the real side of production is

Introduction 25

tantamount to assuming that wage earners can only negotiatethe money level of wages. As Francois Simiand once said,the level of money wages is a fact; the level of real wages isan opinion (Simiand 1932, vol. 1: 160).It is worth noting that the same mechanism indicated by

Keynes and Kalecki for income distribution within the pri-vate sector is admitted by the neoclassical theory for thedistribution of income between the private and the govern-ment sectors. This is the mechanism of the so-called inationtax, by which the government sector, without applying anyexplicit tax, gets hold of a portion of national product (onthis point more will be added in chapter 2, 2.5). There ishowever a difference in that, according to neoclassical the-ory, the ability to get hold of real goods by using purelymonetary instruments is strictly limited to the governmentsector, and the mechanism making this possible is consid-ered a wholly negative deviation from the ideal workingof a market system. In the KeynesKalecki formulation, thesame mechanism is viewed instead as the normal workingof a market economy and is considered to work in favour ofprots.

1.7 The basic ideas of the theory of themonetary circuit

The ideas lying at the basis of the theory of the monetarycircuit can be summarised in the following propositions:

a) Money is in the nature of credit money and in moderntimes is represented by bank credit.

b) Credit money is created whenever an agent spendsmoney granted to him by a bank and is destroyed when-ever a bank credit is repaid.

c) Money being produced and introduced into the marketby means of negotiations between banks and rms is anendogenous variable.


d) The community is divided into two different groups ofagents. The rst, represented by producers, has accessto bank credit and as a consequence enjoys a purchasingpower which is not constrained by the level of real in-come or by ownership of real wealth. The second group,represented by wage earners, can only spend already-earned income.

e) Since agents are not granted credit on the same footing,the system of relative prices reects the way in whichpurchasing power has been granted to different agents.As a consequence, money is never neutral.

f) A complete theoretical analysis has to explain thewholeitinerary followed by money, starting with the momentcredit is granted, going through the circulation ofmoneyin the market, and reaching the nal repayment of theinitial bank loan. Money being created by the bank-ing sector and being extinguished when it goes backto the same sector, its existence and operation can bedescribed as a circuit.

g) Since macroeconomic analysis explains how the cre-ation of money determines both the division of produc-tion between consumption and investment goods andthe distribution of income between wages and prots, italso shows that the role of money goes far beyond thatof making exchanges easier or improving the technicalworking of the market. To the social group being ad-mitted to bank credit, money is, at the economic level,a source of prots and, at the social level, a source ofpower.

1.8 A synthetic description of the monetary circuit

The different phases of the monetary circuit can be easily de-scribed. In this rst synthetic representation, only four agentsare considered: the central bank, commercial banks, rms,

Introduction 27

and wage earners. The government sector will be introducedlater on.12

Step one: A decision is taken by the banks to grant credit torms, thus enabling them to start a process of production. Theamount of credit supplied by the banks at this stage can bedenominated initial nance.13 As already mentioned, circuittheorists usually assume that only rms are admitted to bankcredit.If we consider rms as one integrated and consolidated

sector, the only purchase rms have to make before startingproduction is to hire labour, and their only payment is thewage bill. All other exchanges can be neglected, being inter-nal to the rms sector. Therefore the demand for bank creditcoming from producers depends only on the wage rate andon the number of workers that rms intend to hire (Moore1983, 1984; Graziani 1984).It is clear that if, instead of considering rms as awhole, we

were considering a single isolated rm, the situation wouldbe different. A single rm has to cover not only the wagebill but other current costs, including possible purchases ofdurable goods, such as machinery and other forms of in-vestment. Still the simplied aggregate representation by nomeans alters the substance of the picture. Let us considerthe case of a single rm having been granted credit for the

12 A similar description can be found in Wicksell 1936 [1898], chapter 9,section B. Going back in the past, a remarkable description of the circu-lation of money is given by Galiani. His is not a commodity money butan immaterial at money, very similar to Wicksells pure credit (Galiani1780, book II, chapter 1). In more recent times, the same description canbe found in Parguez 1981; Lavoie 1987, 1993: 15169, and 1996; Graziani1989;Wray 1996. It is debatable whether the description of the circulationof money as a monetary circuit is implicitly present in Keyness thought.In favour of this interpretation: Barre`re 1979: 160, 1988a: 22; Poulon 1982,chapter 11: 300; Graziani 1987, 1989; against it Kregel 1986b.

13 This step was analysed by Keynes when introducing his nance mo-tive for holding money, an innovation which gave rise to a long debate(Graziani 1984). Cesaroni (2001) puts forward a number of remarks criticalof Keyness construction.


construction of a plant. As soon as the rm starts operating,the liquidity is passed on to the rms employed in the con-struction, who pass it on to their own employees. Thus, albeitin an indirectway, the credit initially granted is totally turnedinto wages. At the end of the process, the rm being grantedthe initial credit is in debt to the banking system, while wageearners are creditors of the banks: exactly as though the rmhad itself carried out the construction for which the credithad been initially granted. We can therefore conclude thatconsidering rms as one integrated sector simplies the rea-soning without altering its substance.Since the payment of the wage bill is followed by the pro-

duction of some kind of nished or semi-nished product,there is a correspondence between the wage bill paid andthe cost of produced commodities. The initial requirement ofbank credit can therefore be measured both by the amount ofthe wage bill and by the value of inventories in possessionof the rms. Some authors in fact prefer to say that the creditrequirements of the rms are measured by the money cost ofcommodities being produced (this is the denition initiallygiven by Hawtrey (1923), and later on adopted by Godley andCripps (1981)). To some extent the two denitions are equiva-lent. If we refer to the bank debt of the rms in a single instantof time, it is correct to set it equal to the money value of com-modities produced and not yet sold, namely semi-nishedproducts plus inventories. If we refer to the initial credit re-quirement of the rms, and therefore to their demand forbank credit, it seems more correct to make credit require-ments equal to the wage bill corresponding to the plannedlevel of production.The last denition has the advantage of bringing to light

the clear correspondence existing between the credit mar-ket and the labour market. Any increase in money wagesor in employment gives rise to a higher credit requirementand to a consequent renegotiation of the agreements betweenbanks and rms. This explains why rms, when negotiating

Introduction 29

with unions for the determination of money wages, alwaystry to predict the possible reactions of the banking system,since such reactionswill nally determinewhether or not theagreed wage rates can be paid. The wage policy of the rmsthus ultimately depends on the credit policy of the banks.14

Negotiations between banks and rms determine theamount of credit and the level of the interest rate.Step two: The second phase of the economic process con-

sists in the decisions concerning production and expendi-ture. A basic assumption of circuit theory is that rms enjoytotal independence when deciding upon the real aspects ofproduction, namely employment levels and the amount ofconsumer goods and investment goods produced.Wage earn-ers can only take decisions on how to distribute their moneyincomes between consumption expenditure, addition to cashbalances (bank deposits or notes) or purchase of securities(in the present simplied presentation, since we are neglect-ing the government sector, securities can only be issued byproducers).15

Step three: In the third phase, commodities produced areput on sale. Consumer goods are sold to wage earners, whileinvestment goods are exchanged inside the rms sector (rmshaving producedmeans of production sell them to rmsmak-ing use of them).The money that wage earners spend in the commodities

market, as well as money spent in the nancial market on thepurchase of securities, ows back to the rms, who can useit to repay their bank debt. To the extent that bank debts are

14 Some authors perform an analysis of the supply curve of the single bank(Messori 1988; Screpanti 1993: 134ff.). Stiglitz (1999) concentrates on theanalysis of the behaviour of the banks under asymmetrical information.This aspect will be omitted in the present analysis.

15 As already mentioned, the image used in circuit theory of an economicmechanism in which rms decide upon the real magnitudes while wageearners can only determine their ownmonetary expenditures is borrowedfromKeynessTreatise onMoney; Keynes 1971 [1930], chapters 10 (i): 136,and 20: 31517; also Keynes 1973c [1937]).


repaid, an equal amount of money is destroyed. To the extentinstead that wage earners use their money to increase theirown cash balances, an equal amount of money remains inexistence in the form of rms debt and wage earners credittowards the banks.Step four: Once the initial bank debt is repaid and the

money is destroyed, the monetary circuit is closed. Newmoney will be created when the banks grant new credit for anew production cycle. This can take place almost automati-cally if rms, instead of repaying their bank debt, use the re-ceipts from the sale of commodities and from the placementof securities to begin a new production cycle. In this case,as Keynes remarked, bank credit becomes a sort of revolvingfund of constant amount.16 However, this doesnt mean thatthe rms have become nancially independent: the very factthat they are allowed by the banks to make use of liquiditygranted for a previous production cycle implies a renewal ofcredit on the part of the banks. This can by nomeans be takenfor granted and in fact is tantamount to getting new nancefrom the banks.If wage earners spend their incomes entirely whether on

the commodity market or on the nancial market rms willget back the whole of their monetary advances and will beable to repay the whole principal of their bank debt. In thiscase, as somewould say, the circuit is closed without losses.If instead wage earners decide to keep a portion of their sav-ings in the form of liquid balances, rms are unable to repaytheir bank debt by the same amount. As a consequence, atthe end of the production cycle the money initially createdwill not be entirely destroyed. If banks are now intending tonance a new production cycle equal to the preceding one

16 Keynes 1973b [1937]: 209 [247]. The inverse is also true, namely that aconstant supply of credit is necessary to ensure a constant level of pro-duction. The last formulation points out the fact that, as remarked a longtime ago by F. Machlup, circulating capital is in fact only a different formof xed capital (Machlup 1932).

Introduction 31

by granting the same nance, the total money stock will beincreased: precisely, it will be equal to the wage bill plus thenew liquid balances set aside by wage earners at the end ofthe previous cycle.Stepve: So far, the abovedescriptionhas omitted the prob-

lem of the payment of interest to the banks. It is self-evidentthat since the only money existing in the market is the moneythat banks have lent to the rms, even in the most favourablecase, the rms can only repay in money the principal of theirdebt and are anyhow unable to pay interest. In order to getthe money needed to satisfy their interest payments, the onlything they can do is to sell part of their product to the banks,which is tantamount to saying that interest can only be paidin kind.Aparallel solution, notwidely different in substance,is reached if the banks buy equities issued by the rms.17 Thepresence of a government sector would not make things eas-ier. A government decit might bring to the rms the moneynecessary to pay interest to the banks; but a government debttowards the central bankwould remain pending. There seemsto be no way out: either a debt equal to the interest paymentsremains unsatised, or interest is paid in kind.A similar solution, in spite of its being more reasonable, is

rejected bymore than one author on the ground that it ignoresthe desire to accumulate wealth in monetary form, which isthe object of all accumulation in a monetary economy (Wray1996: 452). It is indisputable that the banks, like any otherentrepreneur, may want to earn prots in money form. Butthis is only a transient form, while the nal destination ofmoney is to be spent in order tomake a protable investment.

1.9 A nal remark

A conclusion to be drawn from the preceding analysis is thatin a closed economy there is only one circumstance which

17 A similar solution is suggested by Bossone 2001: 869, 873.


can produce losses to rms as a whole, and this is the de-cision of savers to hoard part of their savings in the form ofcash balances. It might seem that in an open economy thedecision of savers to place part of their savings on the foreignmarket might be an additional source of losses to nationalrms as a whole. In fact, in an integrated nancial market,national rms might issue securities not only on the nationalmarket but also on the nancial markets of other countries,thus capturing the liquidity that the decisions of savers mighttake away from them.Therefore, if we abstract from the case of an increase in

liquidity preference, rms as a whole run no risk of globalnancial losses. On the other hand, since it cannot be ruledout that single rmsmay incur losses, this means that, alwaysabstracting from increases in liquidity preference, any loss in-curred by a single rmmust be balanced by an identical protearned by some other rm. Mistakes in management made byone single rm, far from making the rms sector weaker, pro-duce higher prots for other rms. This conclusion leads toa careful revisitation of the idea that in a market economyprot can be taken to be an indicator of efciency. To beginwith, the fact that rms as a whole dont make losses is nosign of efciency, since rms as a whole will always have bal-anced budgets, whatever their ability to reduce costs or meetconsumer preferences. On the other hand, if we consider eachsingle rm separately, the fact that one single rm is makingprots is no sign of an especially efcient management, sinceprots earned by one rm may simply be the mirror image ofinefciencies and consequent losses incurred by other rms.

2Neoclassical monetary theory

2.1 Introduction

One of the main theoretical conclusions of the neoclassicaltheory of money is that, in a totally frictionless and perfectlycompetitive market, money performs a wholly neutral role.While allowing exchanges to take place in a more efcientand rational way, and while avoiding the inconveniences ofa barter economy, it alters in no way the nal equilibrium po-sition (Tobin 1992: 775). As we know, this would only be trueif all priceswithout exception couldmove at exactly the samespeed and in the same proportion something which wouldleave relative prices unchanged. Experience shows insteadthat the reaction of money prices is different in speed andmeasure, whichmeans that, at least for a while, any ination-arymovement brings about an alteration in relative prices. Asa consequence, market prices will no longer be equilibriumprices and distortions in the allocation of resources will ap-pear.The monetary theory of the neoclassical school tries in fact

to show that, if we compare a barter economy and a monetaryeconomy (available resources and individual preferences be-ing the same), quantities produced and relative prices are thesame. If this is true, money as a result becomes perfectly neu-tral, a mere veil, as already mentioned, hiding but not chang-ing the substance of the equilibrium position (1.2). The veil

33


consists in the fact that in a monetary economy prices are nolonger quoted in terms of a commodity chosen as numerairebut in terms of money. Since, provided that relative prices areunchanged, the level of money prices is wholly irrelevant,moving from a barter to a monetary economy only makesthe market more efcient without modifying the results ofnegotiations.The relevance of the neoclassical conclusion is strictly con-

nected to the fundamental theorem of welfare economics, ac-cording to which, in a fully competitive market, the equi-librium position coincides with an optimal allocation ofproductive resources. By optimal allocation neoclassical the-orists mean that the set of technologies applied and the setof commodities produced allows, within the constraints setby available resources, the highest satisfaction of consumerspreferences.While neutrality of money is the main theoretical thesis of

neoclassicalmonetary theory, price stability is itsmain policyrecommendation.

2.2 Money in the neoclassical school:truism and theory

The principles of the neoclassical school are usually pre-sented in a trivial form, unobjectionable in itself, but havinga merely denitional content.The principle of neutrality is expressed by saying that, if

the money stock is an exogenous variable (determined bythe monetary authorities), and the quantities of commoditiesavailable in the market are given, any variation in the quan-tity of money in circulation realised by means of an equipro-portional variation in the liquid balances held by each sin-gle agent produces a proportional variation in all moneyprices, relative prices and quantities produced remaining un-changed.A few words are in order to explain why the neoclassical

school assumes the money stock to be an exogenous variable.

Neoclassical monetary theory 35

Of course, any variable can be assumed to be given from out-side merely by way of simplication. But in the neoclassicalmodel the assumption of a given money stock may have adeeper analytical justication.In a position of full equilibrium, equality between demand

and supply prevails in all markets and the budget constraintsof each single agent are satised, which means that, at theend of the chosen period, all debts have been repaid. If, alongwith other debts, bank debts have also been fully repaid, eachloan initially granted by the banks has been extinguished andbankmoney has been totally destroyed.1 If an amount of bankdeposits is still in existence, this means that at least one agenthas a debt pending with at least one bank. In the neoclassicalmodel the government is the only agent allowed, also in anequilibrium position, to be indebted to one bank, the centralbank. Therefore, in equilibrium, the only money in existenceis money created through the government channel, namelyby way of a loan granted to the government by the centralbank. This is the meaning of the usual identity dening themoney stock typical of all neoclassical models:

M(t)=M(t1) + G T (BG(t) BG(t1)),where M is the money stock, G is government expenditure,T the tax yield and BG is the stock of government bonds.Since it is assumed that in equilibrium agents want to

keep a positive money balance, the creation of money inthe amount required by the market, far from being a kindof seigniorage, is no more than a fair exchange of real goods(acquired by the government sector) against a different kind ofcommodity yielding utility to single agents (real balances). Ifthe only stock of money existing in equilibrium is suppliedby government expenditure, it becomes legitimate to viewit as an exogenous magnitude determined by the monetaryauthorities. The role of the government decit thus becomes

1 As De Viti De Marco would say, the only role of money is to ensure perfectcompensation between the value of what is brought to the market and thevalue of what is taken out of it (De Viti De Marco 1990 [1934]: 23ff.).


one of providingmoney to themarket (as already remarked in1.2, this might give rise to some analytical problems: Tobin1986: 11; Riese 1998: 57ff.).2 If the quantities of commodi-ties, the existing money stock and the velocity of circulationare all considered as given magnitudes, the inevitable con-clusion is that the general price level, P, varies in the sameproportion as the quantity of money and that the stability ofmoney prices depends on the stability of the money stock oron the equiproportional increase inM andQ. Hence the tradi-tional prescription concerning the stability of money prices:for stable money prices to prevail, the money stock shouldbe stable or increase at the same rate as the amount of com-modities exchanged on themarket, a result synthesised in thefamous identity: PQ = MV rst put forward by Irving Fisher(1963 [1911]: 248).Similar results are often criticised as clear truisms. It is in

fact self-evident that if the increase in the quantity of moneytakes place byway of an equiproportional increase in the cashbalances of each single agent, the distribution of realwealth isunchanged. The consequent equiproportional increase in allprices leaves the real income of each agent unchanged. If thedistribution of both real wealth and real income is unaltered,relative prices will also be unchanged.However, a similar interpretation doesnt do full justice to

the neoclassical school, a school that has always made aneffort to keep its results far above the mere tautological level.A correct formulation of the two principles of neutrality andstability cannot bemade to descend from ad hoc assumptions

2 It should, however, be remembered that the most serious and rigoroussupporters of the quantity theory of money, such as Wicksell, von Mises,and Hayek, refused to consider the money stock as a given magnitude, theamount of which the monetary authorities can vary at will. They all con-sidered instead the existingmoney stock as originating from bank credit. Intheir models, the money stock is therefore an endogenous variable. Whatthey tried to do was to ascertain the conditions for price stability andneutrality of money to be valid in a model endowed with endogenousmechanisms for the creation of money.


specially introduced for the purpose of demonstrating theproperties of money. Both principles should depend on thesame assumptions on which the general theory of prices isbased.The basic assumptions of the neoclassical model can be

expressed by three main propositions:

a) A behavioural assumption: each individual is endowedwith a set of personal and fully independent prefer-ences.

b) A technological assumption: each individual knows thetechnologies allowing the transformation of productiveresources into nished products.

c) An institutional assumption: perfect competition pre-vails in each single market, where all agents rigorouslyobserve their budget constraint.

When dealing with a monetary economy, the denition ofthe budget constraint requires some clarication. If the bud-get constraint has to be satised at each single instant, thismeans that each agent simultaneously buys and sells com-modities having the same market value. In this case, the sub-stance of the economy is more in the nature of a barter even ifworking as a monetary economy. The same would be true ifa commodity money is used, for instance a gold money hav-ing equal intrinsic and market value. In this case, the budgetconstraint of each single agent would necessarily be satisedat each single instant. This is why many traditional thinkersbelieve that a commodity money is the only true money: tothem, as Keynes once said, the preference for a metallic cur-rency, or for a fully convertible papermoney, is a sort ofmoralprinciple, essential for exchanges to be really fair (Keynes1983 [1914]).If the budget constraint in its looser formulation is ac-

cepted, money, being in the form of cash balances, appearsas an observable variable within each single period and dis-appears only when the nal position of full equilibrium is


reached. It can be said that money, as Benetti and Cartelier(1990) have emphasised, performs the role of nancing tem-porary disequilibria in the budgets of single agents. On theother hand, provided that the budget constraint of each agentis nally satised, individual behaviour is not altered by thepresence of temporary debt and credit positions and conse-quently the nal equilibrium position is not modied either(Tobin 1980).

2.3 The equality between savings and investment

In the equilibrium position, demand should equal supply inall markets. Since in a macroeconomic model only two com-modities are assumed to be present (consumption goods andinvestment goods), if one of the twomarkets is in equilibriumthe othermust be in equilibrium too. The general equilibriumcondition for the commodities market is therefore just one,namely the equality between savings and investment. Theway in which such equality is described in terms of demandand supply deserves some attention.Walras knew all too well that while on the supply side a

number of different investment goods are produced, on thedemand side savers express a general demand for a commod-ity yielding the maximum possible return. Walras thereforeintroduced a general market, where an imaginary commod-ity denominated e (the initial of the French excedent, or theexcess of income over consumption) is sold by savers andbought by investors (Walras 1954 [1926], lesson 23, n. 240).As we shall see in what follows, followers of the circulationapproach adopt a different solution, at the same time clearerand more adherent to reality, consisting in the introductionof a market for securities. In this market, producers sell se-curities to savers wishing to place their savings and earn areturn.None of these solutions is followed by the neoclassi-

cal school. In the neoclassical model, the equality between


savings and investment is reached thanks to the action of thebanks. This is one of the most delicate and debatable aspectsof the neoclassical model.In order to fully understand the problems raised by the neo-

classical formulation of a monetary equilibrium, it should berecalled that a clear difference exists between the classicalschool (Smith, Ricardo, Malthus) on one side and the neo-classical authors on the other in an apparently remote eldof economic theory, namely the theory of wages.The classical school considers real wages as being rigor-

ously paid in advance. Even if wages are usually paid afterthe corresponding labour has been supplied (for instance, atthe end of the week, or month, or after the delivery of com-modities), this does not mean that they are paid after theproduction period has been completed. The assumption ofthe classical school was in fact that wages are paid beforethe nished product has been completed. A similar assump-tion is connected to the fact that classical authors had inmindagriculture as the typical sector of production, a sector wherewage earners would bewaiting for a whole year before gettingtheir means of subsistence, which is clearly impossible. Thesame assumption of real wages being paid in advance mightalso be connected to reasons of image. If wages are paid in ad-vance, they are paid from the output of the previous period,i.e. from the capitalists savings, which can be seen as thesource of the famous and debated wage fund. The level ofthe wage fund, clearly depending on the level of output in theprevious period and on the propensity to save of capitalists,determines the wage bill of the current period. Now a worldin which capitalists limit their own personal consumption inorder to give employment to workers yields the image of anequitable world in which prots appear as the well-deservedremuneration for forgone consumption.A similar image of the production process lies at the basis

of the denition of investment given by the classical school.In the classical model, where real wages are paid in advance,


investment is not dened as the amount of newly producedcapital goods (this was to be the denition adopted by theneoclassical school). Classical economists instead dene in-vestment as the amount of resources engaged in current pro-duction. If we consider capitalists as a class, total investmentequals the wage bill (if we consider also landowners, invest-ment will be the sum total of the wage bill and of the rentpaid to landowners, since what capitalists invest in produc-tion is given by their outlays in favour of any member of adifferent class). If we neglect landowners, and remember thatworkers are paid a subsistence wage, investment is equal tothe wage earners current consumption. In fact, in the clas-sical approach, the same goods comprise the consumptionof wage earners and the investment of capitalists. As AdamSmith said: What is annually saved is as regularly consumedas what is annually spent and nearly in the same time too;but it is consumed by a different set of people (Smith 1993[1776], book I, chapter 4: 34).While the classical school, motivated by realism as well

as by reasons of image, had assumed real wages to be paidin advance, analytical considerations push the neoclassicalschool to prefer the opposite assumption of real wages beingpaid at the end of the production process.The pivot of the neoclassical model is the marginal the-

ory of distribution. For an equilibrium position to prevail,producers should have found out the optimal set of prod-ucts (optimal quantities to be produced of each commod-ity), as well as an optimal resource allocation. Since, in anequilibrium position, each resource is paid according to itsmarginal product, this implies that labour, as well as all otherresources, can only be paid when the productive process hasbeen completed and the product has been sold on the mar-ket. Of course one could argue that wages and other rewardsare determined beforehand on the basis of past experience.But this can be argued of a stationary economy; if the econ-omy is changing, the level of the marginal products cannot be


deduced from past experience (this point is analysed ingreater detail in Graziani 1994).If rigorously dened, the marginal theory of distribution

can therefore only apply to an economy in which resourcesget their remuneration only after production has been com-pleted. This point was seen most clearly by Alfred Marshall,when he said that Capital in general and labour in generalco-operate in the production of the national dividend anddraw from it their earnings in the measure of their respective(marginal) efciencies.3

Once the principle of wages paid in advance is abandonedin favour of a theory of wages paid at the end of the produc-tive process, new and noteworthy analytical consequencesemerge. If wages are no longer paid in advance, productioncan be realised without any previous saving. The contrast be-tween a capitalist class taking care of saving and investmentand a working class condemned to live at the subsistencelevel disappears. What in the classical approach used to becalled investment, namely the subsistence advanced to wageearners, simply disappears from the model and the very terminvestment acquires a newmeaning, in that it now indicatesthe portion of the national product produced in the form ofcapital goods.In the neoclassical model, decisions to save and to invest

are no longer made by the same set of agents, as in the clas-sical model, being respectively taken by savers and investorsacting in a totally independent fashion. The problem then

3 Marshall 1961 [1920], book VI, chapter 2, 10: 544. It is interesting to notethat Marshall, in line with the marginal principle, rejects the concept ofa xed wage fund. However, when commenting on industrial production,he tends to consider wages as paid in advance. His argument is, though,rather weak: industrial production, he says, in contrast to agriculture, doesnot proceed through yearly crops but by way of a synchronised cycle.This means that at the beginning of each production cycle, the amountproduced in the previous cycle is already available and can be used forpaying wages to workers engaged in the following cycle (ibid., chapter 2,n. 6 and appendix J). A similar solution, while unobjectionable in itself,leaves unexplained the payment of wages at the beginning of the rst cycle.


arises of ascertaining whether a mechanism exists by whichthe market brings saving and investment to equality. No suchproblem exists in the classical model, where capitalists asa class decide independently the share of net output to besaved and used as a wage fund. In the new model some formof mechanism equilibrating the respective choices of saversand investors has to be found.Let us neglect the solution given by Walras in what funda-

mentally is the model of a barter economy and let us considerthe problem in the framework of a monetary economy. It wasWicksell who rst raised the problem and provided a fullsolution to it.It should be recalled that Wicksell, while following the

marginal theory of distribution, does not abandon the typ-ical assumption of the classical model, according to whichreal wages are paid in advance. Just as the classical authorsused to do, Wicksell imagines that, at the beginning of theproduction period, an amount of consumption goods is avail-able, having been produced in the previous period and saved(in Wicksells simplied model, savers are at the same timetraders, or shop-keepers, and are indicated by the genericterm capitalists).Wicksell imagines as a starting point that no money is

present in the economy. All means of payment ar

Date post:	11-Nov-2015
Category:	Documents
Upload:	carlos-vazquez
View:	18 times
Download:	0 times

Graziani, A., (2003), The Monetary Theory of Production

Documents