Harcourt (2006) Structure of Post-Keynesian Economics

8/17/2019 Harcourt (2006) Structure of Post-Keynesian Economics

1/216

http://www.cambridge.org/9780521833875


2/216

The Structure of Post-Keynesian Economics

This book is a major contribution to post-Keynesian thought. Withstudies of the key pioneers – Keynes himself, Kalecki, Kahn, Goodwin,Kaldor, Joan Robinson, Sraffa and Pasinetti – Geoff Harcourt empha-sises their positive contributions to theories of distribution, pricing,accumulation, endogenous money and growth. The propositions of earlier chapters are brought together in chapters 6 and 8 in an inte-grated narrative and interpretation of the major episodes in advancedcapitalist economies in the post-war period, leading to a discussion of the relevance of post-Keynesian ideas to both our understanding of economies and to policy-making. (Chapter 7 is concerned with theoriesof growth from Adam Smith to the present day.) The appendixesinclude biographical sketches of the pioneers and an analysis of theconceptual core of their discontent with orthodox theories. Drawing onthe author’s experience of teaching and researching over fifty years, thisbook will appeal to undergraduate and graduate students interestedin alternative approaches to theoretical, applied and policy issues ineconomics, as well as to teachers and researchers in economics.

G . C . H A R C O U R T is Emeritus Reader in the History of Economic

Theory, University of Cambridge, Emeritus Fellow of Jesus College,Cambridge and Professor Emeritus of the University of Adelaide.


3/216


4/216

The Structure of Post-Keynesian EconomicsThe Core Contributions of the Pioneers

G. C. Harcourt


5/216

CAMBRIDGE UNIVERSITY PRESS

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

Cambridge University PressThe Edinburgh Building, Cambridge CB2 8RU, UK

First published in print format

ISBN-13 978-0-521-83387-5

ISBN-13 978-0-511-25028-6

© G. C. Harcourt 2006

2006

Information on this title: www.cambridge.org/9780521833875

This publication is in copyright. Subject to statutory exception and to the provision ofrelevant collective licensing agreements, no reproduction of any part may take place

without the written permission of Cambridge University Press.

ISBN-10 0-511-25028-2

ISBN-10 0-521-83387-6

Cambridge University Press has no responsibility for the persistence or accuracy of urlsfor external or third-party internet websites referred to in this publication, 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

hardback

eBook (NetLibrary)eBook (NetLibrary)

hardback

http://www.cambridge.org/9780521833875http://www.cambridge.org/http://www.cambridge.org/9780521833875http://www.cambridge.org/


6/216

Contents

List of gure s page viiPrefac e and acknowledge ments ix

1 Introduction: why post-Keynesian economics and who wereits Cambridge pioneers? 1

2 Post-Keynesian macroeconomic theories of distribution 6Kaldor’s ‘Keynesian’ theory 6Kalecki’s ‘degree of monopoly’ theory 11Kalecki’s review of Keynes’ General Theory 21The eclecticism of Joan Robinson 25Hahn’s nest hour: the macroeconomic theory of employment and

distribution of his PhD dissertation 29

3 Post-Keynesian theories of the determination of the mark-up 32Wood’s ‘Golden Age’ model 32The choice of technique in the investment decision: orthodox and

post-Keynesian approaches 37Harcourt and Kenyon’s model in historical time 43Why is internal nance to be preferred?: Kalecki’s theory of increasing risk 50

4 Macroeconomic theories of accumulation 55Keynes’ theory: right ingredients, wrong recipe 55Lerner’s internal critique 57Kalecki’s, Joan Robinson’s and Asimakopulos’ Keynesian critique 60 Joan Robinson’s banana diagram 61

5 Money and nance: exogenous or endogenous? 66

6 The complete model: its role in an explanation of post-warinationary episodes 72

7 Theories of growth: from Adam Smith to ‘modern’endogenous growth theory 84

Introduction 84Smith and Ricardo 85Marx 95Harrod 102

v


7/216

Solow–Swan 109Kaldor, Mark 1 114 Joan Robinson (as told to Donald Harris) 119Goodwin’s eclecticism 121Pasinetti’s grand synthesis 123Kaldor, Mark 2 134Endogenous growth theory 139

8 Applications to policy 145The vital link between ‘vision’ and policy 145‘Package deals’: a solution to the Kaleckian dilemma? 147

Appe ndix 1: Biographical sketc hes of the pioneers: Keynes, Kalecki, Sraffa, Joan Robinson, Kahn, Kaldor 158

John Maynard Keynes, 1883–1946 158Michal Kalecki, 1899–1970 160Piero Sraffa, 1898–1983 164 Joan Robinson, 1903–1983 166Richard Kahn, 1905–1989 170Nicholas Kaldor, 1908–1986 172

Appe ndix 2: The conceptual core of the post-Keynes ian disc ontent with orthodox theorie s of value, distribution and grow th 177Bibl iography 185Inde x 199

vi Contents


8/216

Figures

2.1 Kaldor’s ‘Keynesian’ theory of distribution page92.2 Joan Robinson’s diagram of Kalecki’s model 122.3 Kalecki’s rm in the short period 232.4 Accumulation and internal nance: systemic effects 282.5 Hahn’s short-period model of income distribution and activity 313.1 Wood’s opportunity frontier 353.2 Wood’s nance frontier 353.3 The optimum p, g combination 363.4 The family of opportunity frontiers 36

3.5 The family of nance frontiers 373.6 Choice of technique in Wood’s model 383.7 The best-practice isoquant, with constant returns to scale 393.8 Choice of technique by the POPC , by Bob Rowthorn as told

to Geoff Harcourt 403.9 Choice of technique by three different rules 41

3.10 Expected marginal costs of production of existing vintages 453.11 Price, output and output shortfall 46

3.12 Price, quantity from new capacity 473.13 Determination of extra capacity needed 473.14 Determination of investment expenditure 483.15 Determination of price and investment expenditure with

choice of technique 494.1 Lerner’s determination of Keynes’ theory of investment 584.2 Keynes’ second argument as to why r is lower, the greater is

investment now 614.3 Joan Robinson’s banana diagram 646.1 The relationship between growth and protability 756.2 Distribution and growth determined in Marglin’s model 756.3 Neo-Keynesian model of growth and distribution 776.4 Overdetermination of g and r 786.5 The uneasy truce engineered by sustained ination 796.6 Episode 1: higher growth, higher price ination 81

vii


9/216

6.7 Dynamic stagation 816.8 Episode 3: lower growth, lower ination 827.1 The individual capitalist tenant farmer 927.2 Total ‘corn’ production in the agricultural sector 93

7.3 Harrod à la Sen 1067.4 Harrod’s model with an autonomous term in the saving

function 1097.5 Swan’s Way 1117.6 Kaldor’s technical progress function, rst model 1167.7 Kaldor’s representative rm 1177.8 Class war, accumulation and crisis 1207.9 Kaldor’s two-sector complementary model 135

7.10 Growth for ever: 1 1407.11 The steady state in Solow’s model 1417.12 Growth for ever: 2 144A2.1 Short-period utilisation possibilities doubling up for

long-period accumulation possibilities 179A2.2 Samuelson’s (1996) example of Wicksell effects in the

simple Austrian model 183

A2.3 Demand for capital (per unit of labour) in Samuelson’s(1966) model 183

viii List of gures


10/216

Preface and acknowledgements

While writing this book, I have had in mind two sets of readers: rst,undergraduate and graduate students who may be looking for alternativeapproaches to thinking about theoretical, applied and policy issues ineconomics. By presenting a structure of the thought (and its origins) thatI have found so helpful over my working life I hope to at least interestand possibly even enthuse this rst set. Secondly, I also hope that whatI have written may interest teachers and researchers in economics, notso much perhaps for the details of the analysis, with which many willbe familiar, but for the way in which one person at least sees the inter-

connections and interrelationships which have emerged as our disciplinehas evolved and developed.The ideas in the book themselves have evolved and developed for me

over the past fty years, in both lectures and research. My model is notexactly Dennis Robertson’s three volumes of lectures on EconomicPrinciples in Cambridge, Robertson (1957, 1958, 1959); but I supposeit has something in common with them, even with his admission that ‘if it is all wrong, it can’t be helped now’ (Robertson 1957, 7). I trust,

though, that I have not written in quite as querulous a tone as that intowhich Robertson sometimes lapsed, for I remain, as ever, a happy andenthusiastic, even optimistic, person who nevertheless is willing to admitthat he may be wrong.

I wrote the rst draft of this Preface in April 2005, in the ftieth yearsince I rst came to Cambridge in September 1955. Half my workinglife has since been spent here (the other half in Adelaide, most happyyears) and I count myself most fortunate to have studied and taughtin such a stimulating and satisfying, even if sometimes–no, often–socantankerous an environment.

Much more than this, though, this year Joan and I will celebrate ourGolden Wedding anniversary on 30th July. As ever, her love and supporthave made possible the writing of the book, much of which occurred inthe study she imaginatively prepared for me in our New Square homewhen, having had three years’ grace over and above the obligatory

ix


11/216

seventy years’ constraint, I no longer had a room at Jesus. I would like todedicate the volume to her with my love.

I am much indebted to many cohorts of pupils who have listenedto my lectures, to my graduate students and colleagues in Adelaide,

Cambridge and Toronto and to friends in many countries who havecontributed greatly in discussions and their own writings to my under-standing of economic issues. I hope I will not cause offence if I thankexplicitly the people who have most directly inuenced what I havewritten here: Mauro Baranzini, Stephanie Blankenburg, Wylie Bradford,Giuseppe Fontana, Prue Kerr, Tom Russell, Sean Turnell and theanonymous readers of the manuscript for Cambridge University Press.

Finally, may I thank Rhona Watson for her generous and efcientsearches and for answers to my obscure queries, and Susan Cross,Frances Thomson, Frances Flood, Debra Armstrong and Janet Nursewho cheerfully typed the manuscript and never once complained aboutmy atrocious handwriting.

Thanks are also due to Macmillan for the extract from Joan Robin-son’s Essays in the Theory of Economic Growth (1962) reproduced inchapter 3.

GCHCambridgeOctober 2005

x Preface and acknowledgements


12/216

1 Introduction: why post-Keynesian economicsand who were its Cambridge pioneers?

Maynard Keynes, Richard Kahn, Richard Goodwin, Nicholas Kaldor,Luigi Pasinetti, Joan Robinson and Piero Sraffa all started initially,at least in some degree, within the mainstream of their time. They allmoved well and truly outside it, attempting to create either a revolution-ary alternative or to rehabilitate the classical–Marxian tradition, in mostcases in the light of the Keynesian revolution. The one exception isMichal Kalecki, whose personal history and independent mind com-bined to place him virtually always outside the mainstream. Thisvolume, though, is not principally concerned with why and how the

discontents that led them to change their minds arose. Rather, its prin-cipal object is to set out the structures of their alternative approaches inorder to suggest mo d es of thinking about theoretical and policy issuesin political economy. 1

The structures presented here are based on over forty years of teachingand researching under the rubric of what is now called post-Keynesianeconomics. I certainly was not aware that it was so called when I startedon this track in the 1950s. In fact, I have much sympathy with the stance

of my old friend, the late Athanasios (Tom) Asimakopulos, who declinedan invitation to be included in the rst edition of Philip Arestis andMalcolm Sawyer’s admirable A Biographical Dictionary of Dissenting Economists (1992 ), because he regarded his views and contributions asbelonging fully within the tradition of economics proper, not in a dis-senting stream. It was only in order to provide a suitable tribute to hisinuential contributions and splendid personal example as a teacher andhuman being that his widow, Marika, allowed the entry on Tom to beincluded in the second edition of Arestis and Sawyer (see Harcourt2000 ). However, it must be admitted that at the time of writing (August2004), though something of a backlash/comeback may be discerned (see

1 Some of the reasons for their discontent are given in the appendixes to the volume: thesecontain short intellectual biographies of the main contributors (appendix 1, pp. 158–76)and a sketch of some of their principal arguments (appendix 2, pp. 177–84).

1

http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-


13/216

Harcourt 2001a for reasons why), the views and approaches taken in thisvolume still continue to be regarded by the bulk of the profession asthose of dissenters.

The most succinct denition of post-Keynesian economics comes

from Joan Robinson ( 1978 ; CEP , vol. V, 1979b, 210)2

:

To me, the expression post-Keynesian has a denite meaning; it applies to aneconomic theory or method of analysis which takes account of the differencebetween the future and the past. (emphasis in the original).

I obviously have no quarrel with this; but, as I try to be ever-mindful of historical developments, I also wish to stress that the approaches to politicaleconomy which reect post-Keynesian thought are there partly for histor-

ical reasons and partly because of logical associations. Post-Keynesianismis an extremely broad church. The overlaps at each end of a long spectrumof views are marginal ( sic), often reecting little more than a shared hostilitytowards mainstream neoclassical economics and methodology, IS/LM Keynesianism and the ‘x-price’ Keynesianism of the ‘New Keynesians’and certain French economists. Some post-Ke ynesians are working ac-tively towards a synthesis of the principal strands. 3 Others regard the searchfor a synthesis, for a general all-embracing structure, as a profound mistake:to quote Joan Robinson ( 1974 ; CEP , vol. V, 1979b, 119), a foundingmother, a misguided attempt to replace ‘one box of tricks’ by another.Post-Keynesianism should be a situation-and-issue-specic method of doing political economy, a ‘horses for courses’ approach, itself an all-embracing structure at the methodological level (see Harcourt 2001a ,Essay 19).

The principal object of analysis is the advanced capitalist economies of

the twentieth and twenty-rst centuries. The central aim is to provide aframework within which to understand and explain their macroeco-nomic and/or microeconomic processes over time. It must be admittedthat the tradition within which they are presented objects vigorously tothe microeconomic/macroeconomic dichotomy of mainstream econom-ics (see Joan Robinson 1977b ; CEP , vol. V, 1979b, 4–5 for a typically

2 The Convention in this book is to separate by a semicolon the date of the cited work fromthe date of the collected work(s) where it is reprinted. 1978 here is therefore thepublication date of Joan Robinson’s ‘Keynes and Ricardo’, which is reprinted invol. V of her Collected Economic Papers (CEP ) in 1979 ( CW is the siglum for Keynes’Collected Writings ).’

3 The deepest and most profound example of the attempts to provide a coherent synthesisis the splendid monograph by Heinrich Bortis, Institutions, Behaviour and EconomicTheory: A Contribution to Classical–Keynesian Political Economy (1997 ). Reading succes-sive drafts of Henry’s book taught me so much. If I were ever to be persuaded that asynthesis were possible, it would be because of his arguments.

2 The Structure of Post-Keynesian Economics

http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-


14/216

forceful argument why). Basically, neither individual nor group/classbehaviour may be understood without making explicit the economy-wide structures and relationships that provide the backdrop to theirbehaviour. Similarly, economy-wide structures and relationships not

only inuence but also are inuenced by individual and group/classmotivations and behaviour. Thus the microeconomic foundations of macroeconomics must always be complemented with – indeed, itcould be argued, dominated by – the macroeconomic foundations of microeconomics, see Crotty ( 1980 ).

The particular subsets of the mainstream literature that this happyband became increasingly dissatised with were the theory of distribu-tion, especially the marginal productivity theory in its aggregative form(but also the supply and demand approach in general, see Bharadwaj1978 ); the theory of pricing at the level of the rm and the industry,principally as it came down from Marshall and Pigou; the theory of investment behaviour and expenditure that is implied in Marshall andPigou and, and more explicitly, in the writings of Irving Fisher; and thetheory of growth, to which is allied the theory of the trade cycle (thebusiness cycle to our North American cousins), as it has been developed

in the post-war period by leading neoclassical economists (some of whom, such as James Meade, Robert Solow, and Trevor Swan were/ are also leading Keynesians). In doing so, they were inspired and stimu-lated – even irritated – by Roy Harrod’s and Evsey Domar’s seminalcontributions in the late pre-war and early post-war years. The nalobjective of the volume is to show how the alternative theories of thepost-Keynesians under each of these heads may be combined into anoverarching general framework that may then be applied in explanations

of post-war happenings in the advanced capitalist world. This sameframework, together with its constituent parts, may be used to rationalisevarious policy proposals which tackled, or should have been used totackle, some of the major malfunctions of these economies in the sameperiod.

An equally important aim of the volume is to rescue the pioneeringcontributions of this rst generation from the benign neglect and misun-derstandings that are starting to occur as the time from their respectivedeaths lengthens. It is important to have recorded for posterity the back-ground and the nuances to the making of the theories by people whoknew these pioneers personally and who were present for at least partof the time when the ideas were developed, not only to restore them totheir correct place in the narrative but also to correct the misconceptionsand often neglect they suffer or experience as the third and even fourthgeneration of post-Keynesians increasingly come to constitute the

Introduction 3

http://-/?-http://-/?-http://-/?-http://-/?-


15/216

post-Keynesian literature and canon. I do not mean to denigratethe contributions of the latter groups; but I would like to restore to theirrightful plac e the fundamental pioneering contributions of the rstcontributors. 4

The structure of the volume is as follows: In chapter 2 I discuss post-Keynesian macroeconomic theories of distribution. I start with Kaldor’s1955 –6 paper, as it is the best known. I use it and its characteristics asthe backdrop to discussions of Kalecki’s earlier contributions, includinghis review of Keynes’ General Theory , Joan Robinson’s eclectic approachand Frank Hahn’s macro theory of employment and distribution whichwas initially developed in his PhD dissertation at the LSE in the later1940s and early 1950s.

Post-Keynesian theories of the determination of the size of the mark-up are discussed in chapter 3. Adrian Wood’s ‘Golden Age’ model istaken as the benchmark against which are assessed the ‘historical time’model developed by Peter Kenyon and myself and the choice of tech-nique in the investment decision in both the orthodox and the post-Keynesian approach. The chapter closes with a discussion of whyinternal nance is usually preferred to other forms of nance of invest-

ment expenditure. Kalecki’s principle of increasing risk is taken as themost insightful explanation.Chapter 4 is concerned with macroeconomic theories of accumula-

tion. It starts with a critique of the details of Keynes’ theory in TheGeneral Theory and after. The critique stems from the writings of Kalecki, Joan Robinson and Asimakopulos. All the ingredients involvedin it come together in Joan Robinson’s well-known banana diagram, anexposition of which ends the chapter.

Chapter 5 contains a brief discussion of money and nance – whetherthey are exogenous or endogenous in theory and real life. In chapter 6 allthe previous developments are brought together in an explanation of post-war inationary episodes, drawing on the conict ination modelsof Steve Marglin ( 1984a , 1984b ) and Bob Rowthorn ( 1977 ).

Theories of growth from Adam Smith to ‘modern’ endogenousgrowth theory are discussed in chapter 7. We start with Smith andRicardo’s theories, move on to Marx and then to Harrod’s theory. Thereaction to Harrod’s ndings and problems by Solow and Swan, on

4 Paul Davidson ( 2003 –4) has written a most idiosyncratic review article of John King’shistory of post-Keynesian economics since 1936 (King 2002 ). It was entitled ‘Setting therecord straight. . . ’. I was tempted to write a reply with Luigi Pasinetti entitled ‘ Reallysetting the record straight’ but desisted after I read the courteous but powerful replies toDavidson by Marc Lavoie and King himself.


http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-


16/216

the one hand, and Kaldor and Joan Robinson, on the other, are thendiscussed together with Richard Goodwin’s eclectic theories and Pasi-netti’s grand synthesis. The chapter closes with discussions of Kaldor’slater views in which he scraps many of his earlier ideas, and of endogen-

ous growth theory, emphasising how it relates to previous discussionsfrom Smith on.

The concluding chapter 8 uses the approaches developed in earlierchapters to examine their application to policy issues. It discusses how‘vision’, approach and method interrelate with policy recommendations.It closes with a proposed ‘package deal’ solution to a crucial dilemmaraised by Kalecki in his classic 1943 paper on the political aspects of full employment, especially how it may be permanently sustained asopposed to attained from a deep slump.

The volume ends with two appendixes: biographical sketches of thepioneers and an account of the conceptual core of the post-Keynesiandiscontent with the orthodox theories of value, distribution and growth.

Introduction 5



17/216


18/216

the Cambridge economists’ (Keynes 1933 ; CW , vol. X, (1972 , 71),that is to say, the rst to think like Keynes.)

By the time we get to the mid-1950s when the ‘Golden Age of Capitalism’ was already in full swing, the advanced capitalist economies

were experiencing full employment and growth, real wages were farabove subsistence and so it was possible in Kaldor’s view to make an180 turn and allow the prot-receiving capitalist class to have rst biteof the cherry, as it were, leaving wage-earners to receive the residual afterprots, accumulation (and rentier consumption) had been accountedfor. (Arthur Lewis 1980 , 257 told a not dissimilar story but used aneoclassical approach to analyse the distribution of income betweenprots and wages in the phase of development when there were nolonger unlimited supplies of labour.)

Despite arguing that only a fully employed economy could continueto grow over the long term, Kaldor nevertheless called his theory‘Keynesian’, for at least three reasons. First, he located the origins of his theory in Keynes’ analogy of the widow’s cruse in A Treatise on Money(1930 ; CW, vol. V, 1971, 125), whereby the more prot-receivers spent,the more prots they received:

If entrepreneurs choose to spend a portion of their prots on consumption . . .,the effect is to increase the prot on the sale of liquid consumption goodsby an amount exactly equal to the amount of prots which have been thusexpended . . . Thus, however much of their prots entrepreneurs expend onconsumption, the increment of wealth belonging to entrepreneurs remains thesame as before. Thus prots, as a source of capital increment for entrepreneurs,are a widow’s cruse which remains undepleted however much of them may bedevoted to riotous living. When . . . entrepreneurs are making losses . . . bysaving more, the cruse becomes a Danaid jar which can never be lled up.

(emphasis in original)

Secondly, Kaldor took the Keynesian view that (planned) investmentled and saving, determined by income and its distribution, responded.Thirdly, he argued that the Keynesian multiplier was a short-periodconcept in The General Theory model, with changes in income neededto bring planned investment and planned saving into equality, be-cause money-wages and prices were sticky in the short period. In thelong period, however, the multiplier applied to the distribution of long-period, full-employment income, principally because, in the long period,prices were relatively more exible than money-wages, and the marginalpropensity to save out of prots was greater than the marginal propensityto save out of wages.

We should also see Kaldor’s contribution within a context of thedevelopment of the peculiarly Cambridge (England) contributions to

Macroeconomic theories of distribution 7

http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-


19/216

growth theory, stimulated by Harrod’s 1939 article and 1948 book andby the awakening of interest in the post-war period in developmentitself. This meant that Kaldor, in tackling the problem of Harrod in-stability (see chapter 7, pp. 102–9 for a discussion of Harrod’s model

and problem) and the processes by which the warranted rate of growth, g w , and the natural rate of growth, g n , were equalised, assumed for histheory of distribution that planned investment, if realised, was such as togive the economy over the long term the necessary capacity to allow it togrow at g n . This required forces at work which took g w to equality with g n and allowed planned investment to become actual investment.

The value of the share of investment in long-period, full-employmentnational income was therefore predetermined in Kaldor’s model. In latermodels he attempted, not ever successfully, to show why the economyshould be at full employment and the share of investment in nationalexpenditure should be endogenously determined at the share consistentwith producing g n over time.

Kaldor assumed that the long-period equilibrium position of a grow-ing capitalist economy is a full-employment one, Y f . He assumed simpleproportional saving functions with sp > sw 0, where sp is the marginal

propensity to save ( mps ) of prot-receivers and sw is the mps of wage-earners. Let P be total prots, W be total wages. 1 Then Keynes’ saving– investment equilibrium condition determines the distribution of Y f ratherthan the level of actiivty and income.

Thus:

S ¼ sp P þ swW ¼ I (2.1)

where I ̄ is given, autonomous:

sp P þ sw Y f P ¼ I i.e.

P ¼I

sp sw

swsp sw

Y f

and

II Y f

¼1

sp sw I Y f swsp sw (2.2a)1 Subsequently, the distinction between saving from different classes of income – prots,

wages – and by different classes of persons – prot-receivers, wage-earners who save – was analysed explicitly by Pasinetti ( 1962 ) and followed up by Kaldor (in, for example,Kaldor 1966a ). See also Harcourt ( 1972 , chapter 5).


http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-


20/216

As

sw ! 0 ;P

Y f !

1sp

I Y f

(2.2b)

clearly the Keynesian multiplier relationship.We now use a simple diagram (see gure 2.1 ), to illustrate Kaldor’s

arguments. 2 (In lectures, I have always tried to use words for thepoets, algebra for the mathematically inclined and geometry forthe in-betweens.)

On the vertical axis is measured the share of (given or autonomous)investment, I ̄ , in long-period, full-employment national income, Y f, andthe share of (planned) saving, S , in Y f . On the horizontal axis is meas-ured the share of prots ( P ) in Y f . Because sp > sw and prices are moreexible than money-wages in the long term, if the economy is notinitially at the distribution of Y f where planned saving and plannedinvestment are equalised

2 The diagram was suggested to me by the late Hugh Hudson, who edited the rst twovolumes of Kaldor’s collected papers.

Figure 2.1. Kaldor’s ‘Keynesian’ theory of distribution.


http://-/?-http://-/?-


21/216

P

Y f ethere are zones of excess demand to the left of the intersection of S Y f andI Y f , and of excess supply to the right. These impact on prices relatively tomoney-wages and redistribute the given level of Y f so as to raise (to theleft) or reduce (to the right) the share of prots, thus changing the shareof saving appropriately.

P

Y f eis therefore a stable equilibrium position, for if the economy is notinitially there, appropriate signals and processes will take it there. Atthat point the value of

S Y f

ð¼ sÞ

is such as to make

g w ¼ sq (where q is the desired incremental capital–output ratio) equal to g n . Theeconomy thus has the desired amount of investment expenditure andcapacity creation to allow it to grow at g n , realising its full-employmentpotential by employing all its labour force and the expanding capacity of its stock of capital goods over time.

Kaldor provides two provisos: the share of prots must not be so lowas to make the prot-receiving capitalists feel that accumulation andprot-making are not worth the candle (this is shown in gure 2.1 asthe prot-receivers’ capitalists’ jaundiced range. It corresponds to Ricar-do’s argument that there must be at least some minimum rate of protreceived to keep capitalism going.) Correspondingly, the share of protsmust not be so high as to entail a share of wages and a level of real wagesthat are unacceptable to the wage-earners, who are assumed passively to

accept whatever residual of national income is left for them after theprot-receivers have received their share. In this situation, the wage-earners no longer passively accept the residual but respond by causing awage–price and wage–wage inationary spiral and the distribution of income will no longer be determined by the Kaldor process. This rangeis designated as the wage-earners’ revolting range in gure 2.1 .



22/216

That in outline is Kaldor’s ‘Keynesian’ macro theory of distribution.The other theories we now examine share much of what he argues – thatinvestment leads and saving responds, that sp > sw – but they do notconne themselves to the long period nor to a situation of long-period

full employment. By contrast, they allow distribution and employmentto be determined together.

Kalecki’s ‘degree of monopoly’ theory

We illustrate Kalecki’s theory rst with his theory of the simultaneousdetermination of employment, output and the distribution of incomein the short period, a contribution that dates from the 1930s: seeKalecki ( 1938 ; CW, vol. II, 1991). Indeed, a version of it was alreadypresent in Kalecki’s review of The General Theory in 1936 . Unfortu-nately, the review was published in Polish and a full English translationdid not become available until December 1982 (see Targetti and Kinda-Hass 1982 ). Kalecki’s review is discussed on pp. 21–5; here we use JoanRobinson’s exposition of his theory. It exploits a very simple, cleverand illuminating diagram (see gure 2.2 ), on which generations of

Cambridge undergraduates were brought up by Joan Robinson herself and then, later, by me. It appeared in print in Joan Robinson’s MichalKalecki Memorial Lecture in the Oxford Bulletin memorial issue inhonour of Kalecki in 1977 ( Joan Robinson 1977a ; CEP , vol. V, 1979b,191).

In Kalecki’s writings on capitalism two principal forms of price for-mation are recognised: Marshallian market-determined prices for rawmaterials, and mark-up pricing for industrial products. The size of themark-up is determined by what Kalecki called ‘the degree of monopoly’,a portmanteau term to take in the factors determining the degree of discretion in price-setting that imperfectly competitive rms in differentindustries have: patents, barriers to entry, dominance as a price leaderand so on. 3 In Joan Robinson’s version it is the second form of pricingthat is used, combined with Kalecki’s view that in the short term the costcurves of individual rms are reverse L-shaped – that is to say, marginalcosts are equal to short-term average variable costs up to capacityoutput.

3 Peter Kriesler ( 1987 ) has written the denitive account of the development of Kalecki’sviews on these and related matters.


http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-


23/216

Joan Robinson uses this assumption, a not unrealistic one, of course,to introduce a very neat measuring trick. All quantities in the shortperiod concerned in her analysis are measured in terms of the labourrequired directly and indirectly to produce one unit of the consumptiongood. (We have a two-sector model, the consumption goods sector andthe investment goods sector.) This implies that the wage cost per unit of production and the wage cost per unit of labour employed coincide, as

do the price per unit of production, employment and the amount of value added per unit to be shared between prots and wages. We dealwith a closed economy without a government so that all intermediatepurchases net out and only values added of the private sectors areconsidered.

On the vertical axis of gure 2.2 we measure prices, costs, prots andwages per unit. On the horizontal axis we measure employment in termsof the units dened above. The vertical line indicates capacity output/ maximum employment offered in the consumption goods trades.

Though it may not be found as such in any of his writings, it is wellknown that Kalecki’s macroeconomic theory of distribution may beneatly expressed in the aphorism that wage-earners spend what theyearn and prot-receivers receive what they spend. To see this, considerthe simplest case where wage-earners do not save and prot-receivers donot consume. Write the national accounts:

Figure 2.2. Joan Robinson’s diagram of Kalecki’s model.



24/216

E C þ I Y C þ S W þ PC W by assumptionSo S Pand I P

Kalecki argues that prot-receivers as a class (as well as individually) canalways decide what to spend on investment (up to a nancial constraint)but not how much as a class (or individually) they will receive. Thereforecausation must run I ! P ! S . In equilibrium, planned I is not onlyrealised but it is also equal to planned S and an equivalent level of Pcreated by the level of I and its consequent effect on employment andincome (and its distribution). Y (and its distribution) must be such asto give a level of P (¼ S ) to match the injection of planned and nowrealised I .

Moreover, it may be shown that the prots of the consumption goodstrades are equal to the wages bill of the investment goods trades. Thewages bill of the consumption goods trades, even when fully spent,covers only the wage costs of the consumption goods trades, that is tosay, itself. In order to receive prots there must be an injection of

spending from outside the sector, that is to say, the spending from thewages received by the wage-earners in the investment goods trades.Prot per unit in the consumption goods sector will be determined bythe size of the mark-up in that sector.

As it is a short-period analysis, we may provisionally take the level of planned investment as given. (Of course, Kalecki spent the whole of hisprofessional life improving his theories of investment and accumulationbecause he always stressed that accumulation was the most important

process at work and to be explained in an analysis of the motion of capitalism over time.) The given level of planned investment impliesthat we know the amount of employment needed to produce the corres-ponding amount of investment goods. For simplicity, and also becauseit is not a bad rst approximation, we assume that all investment goodsare made to order and that agreed prices, and so prot margins, themark-ups over wage costs in the investment goods sector, have beenagreed upon between buyers and sellers.

The amount of employment needed to make the investment goods isshown as BC in gure 2.2 ; B is the point where capacity working (0 B )occurs in the consumption goods sector. The money-wage is also as-sumed to be given and, again for simplicity, is assumed to be the same inboth sectors. Wage costs per unit and the wage itself are shown as 0 w.The rectangle BCDE therefore measures the wages bill in the investmentgoods sector. We draw a rectangular hyperbola which subtends the



25/216

same area as BCDE under each point on the left-hand side of the verticalline with a vertical axis starting at w and a horizontal axis emanatingfrom w as shown. The rectangle 0 ABEw is, therefore, the maximumwages bill payable by the consumption goods trades if all available

capacity there were to be operated. The area under the rectangularhyperbola may be thought of as the ‘marzipan icing’ on the consumptiongoods sector’s wages bill Christmas cake, the thickness or thinness of itsspread depending on the size of the mark-up set, itself determined by theintensity of the degree of monopoly ruling in the short period concerned.

If free competition ruled in the consumption goods sector, excesssupply pressures would mean downward pressures on the now market-determined margin until full-capacity working prevailed and the com-petitive price of 0 pc with a prot margin of wp c were established at theintersection of the rectangular hyperbola with the vertical line. Noticethat there is no suggestion that just because prices of consumption goodshave been assumed to be exible and market-determined that full em-ployment of labour and capital automatically results. The combinedwage bills of both sectors would have to provide enough demand toabsorb the available labour force. But if the given level of investment

was ‘low’ – due, say, to sluggish ‘animal spirits’ of the investors – thenthere still could be involuntary unemployment occurring overall eve nthough there is full capacity working in the consumption goods sector. 4

Suppose, though, that the existing ‘degree of monopoly’ implies amark-up of wp m . Then employment and production in the consumptiongoods sector will be 0 A and, again, demands and supplies will match inboth sectors with W ¼ C and I ̄ ¼ P ¼ S . Moreover, we have nowdetermined simultaneously output, employment and the distribution of

income between wages and prots in the short period without the needto assume full employment for the mechanism to ‘work’. So Kalecki’sadaptation of Marx’s schemes of production and reproduction hasallowed him to present both the production and expenditure aspectsand interrelationships of capitalism, and to make explicit the role of price-setting behaviour in macroeconomic processes. This is a majorreason why Joan Robinson came increasingly to prefer Kalecki’s ap-proach to Keynes’ more Marshallian framework, which she thoughtrequired indirect and rather convoluted arguments to make points whichare crystal clear in Kalecki’s exposition. (An example of what she hadin mind may be obtained by comparing the structure of Kalecki’smodel with Keynes’ passages about the widow’s cruse which we quotedon p. 7.)

4 We discuss all this more fully in chapter 4 .


http://-/?-http://-/?-http://-/?-


26/216

Kalecki’s framework also allows another fundamental aspect of theworkings of indeed any economy and social system to be brought out.For not only are employment and output determined by aggregateexpenditure on investment and consumption, they are also directly

related to the productivity of labour in the two sectors and to the realwage in the following sense. Given the productivity of wage-earners inthe consumption goods sector and the value of the real wage (which isdetermined once the prices of consumption goods have been deter-mined, the money-wage being taken as given for the current shortperiod), 5 the surplus of consumption goods per unit of labour over andabove the real wage is known. Then, given employment in the invest-ment goods trades, enough units of labour must be employed in theconsumption sector to ‘feed’, as it were, themselves and to provideenough consumption goods from the surpluses simultaneously createdto ‘feed’ the investment goods wage-earners as well.

The framework may be used to analyse both developing economiesand democratic socialist economies as well as capitalist economies. Forexample, Kalecki argued that in a democratic socialist society, the plan-ners should decide the level of consumption per head and overall for

a fully employed labour force, possibly through deciding money-wagelevels and the relative wage structure, and by a judicious use of salestaxes added to labour and material costs to set prices. Those wage-earners not required in the consumption goods sector would be availablefor the investment goods sector. There, the composition of the totalproduction of investment goods would be decided by socialist managersfollowing rational rules for the choice of technique, given the allocationof available resources between the combination of various capital goods

to make both consumption goods and capital goods themselves.Kalecki himself always advocated a modest rise from period to period

in the level of consumption goods per head. Citizens then would havemore jam today rather than being fobbed off with promises of morejam in tomorrows that never came, as tended to happen too often inStalinist regimes, not least in Kalecki’s own country, Poland. There,when he returned to Poland in the late 1950s, he was soon sidelined asfar as policy was concerned, just because of his progressive democraticsocialist views (see Steindl 1981 , 592).

5 In a short-period by short-period analysis such as Kalecki developed in his model of cyclical growth over time, the money-wage may also be made an endogenous variable.Thus, wage-earners react to realised real wages in any period by bargaining for money-wages in the following periods that are designed to obtain the real wages they wish tohave, the pressure of real wage resistance, for example. These and related themes aretaken up again in chapter 6 .


http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-


27/216


28/216

can be written as

1l m

w l c þ l m m c

l m 6

This can be converted to consumption good units per worker bymultiplying by

l ml c þ l m m c

the ratio of labour time per machine to labour time per consumptiongood, to give

1l c þ l m mc

w

the original result.The two approaches serve to emphasise two important factors stressed

by Joan Robinson: rst, the need to specify the units in which economicquantities are measured, and, secondly, that accumulation has both asaving and an investment aspect. The result is quite general and relates to

the surplus of consumption goods per worker over the real wage rate,whether or not the latter is the subsistence one. It also illustrates thedependence of the rate of growth of capital goods on the level of the realwage rate that we noted above. The lower is the real wage rate, the

6 This result is derived as follows: assume that wage-earners spend all their wages onconsumption goods and that prot-receivers spend all their prots on machines; then

L ¼Cp cw

¼ pcw

L cl c

¼ 1w

L cl c

where C is the quantity of consumption goods produced in a period, pc is the price of consumption goods, w is the money-wage rate and L c is labour employed in consumptiongoods production. Lm ¼ L – L c ¼ (1 – w̄ I c )L , where Lm is labour employed in machineproduction.

Now

M ¼ 1l m

L m ¼1 wl c

l mL

and R ¼ m c C ¼ w̄ mc L . (Machines are assumed to last for one period only and the gestationperiods of machines and consumption goods are assumed to be equal.)

I ¼ M R ¼1 wðl c þ l m mcÞ

l mL

where I is net investment

M RL

¼ 1l m

w l c þ l m mc

l m :



29/216

greater is the surplus of consumption goods available per worker foremploying workers in the investment goods sector on net investment,and the greater is the command of any given surplus over labour timein that sector.

Johnson’s model may be used to illustrate the technical limitationson the growth of capital goods, supposing there is to be available anunlimited supply of labour, an d that production each period is plannedby (say) Worswick’s dictator. 7 Assume that the dictator is planningthe production of the current period, t, and that he wishes to maximisethe rate of growth of the stock of machines over time. He has availablefrom period t – l, a stock of machines ( M t – 1 ) which is the gross invest-ment of the previous period. M t – l machines allow the employment of L c wage-earners in the consumption goods sector and the production of C ¼ L cl c consumption goods. If wage-earners receive a given money-wagew, all of which they spend, the dictator will choose a price of consump-tion goods, pc , such that the resulting real wage rate, w ¼ w pc will be thesubsistence one.

The technical surplus available for gross investment ( S ) – that is, theconsumption goods available with which to employ wage-earners in the

investment goods sector – isI cI c

L cw

i.e.

S ¼ L c1l c

w : 1l c w is consumption goods production per person employed in the consump-tion goods sector less the real wage per person. The surplus buys

L cw

1l c

w units of labour time in the investment goods sector, and makes possiblegross investment ( M t ) of

L cl m w

1l c

w 7 Allowance being made for the assumption that machines last for one period only, the

results of the following three paragraphs are identical with those of Worswick ( 1959 ,128–30).


http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-


30/216

machines. Net investment ( I ) is

L cl m w

1l c

w M t 1machines. (Machines are assumed to last for one period only, and thegestation periods of machines and consumption goods are equal.) It isclear that the s m aller is the real wage rate, the greater are both gross andnet investment. 8 Establishing the subsistence real wage rate maximisestotal employment ( L), employment in the inves tment goods sector(L–L c ), and the production of machines in period t .

9 (It is assumed thatmachines are made by labour alone.)

The rate of growth of capital goods ( g ) may be derived as follows:

g ¼ I

M t 1

This can be written as

L cl m w

1l c

wh i M t 1

1

Now

M t 1 ¼ mcC ¼ mcL cl c

8 Both@ M t @ w

and@ I @ w

¼ Lcl cl m

1

ðwÞ2 < 0

The use of the calculus may be questioned, since two different situations, rather thanchanges over time, or at a point in time, in the one situation are being compared.

9 From n. 5

L ¼ 1w

L cl c

As@ L@ w

¼ L cl c

1

ðwÞ2 < 0

L will be greatest when w̄ is as small, as is consistent with wage-earners either staying aliveor being willing to work in the industrial sectors of the economy.


http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-


31/216

Therefore

g ¼ 1mcl m

1 l cww 1

That is, g is inversely related to the real w age rate and to the indirectlabour cost per unit of consumption goods. 10 The rate of increase of thestock of machines depends upon both the real wage rate and the realamortization element of consumption goods. With a given real amort-ization element of consumption goods, the rate of increase is greatestwhen the real wage rate is the subsistence one.

In conclusion, two points can be noted. First, as the current period’sproduction of machines and consumption goods become available onlyat the end of the period, it must be assumed that wage-earners carry,on average, a stock of consumption goods equal to half the currentperiod’s production. 11 If the stock of machines is growing, this cannotbe done; in these circumstances, M t– 2 will be less than M t– 1 and,therefore, C t– 1 will be less than C t . This difculty can be overcome bysupposing that the dictator offers for sale on the rst day of the currentperiod the consumption goods produced in the previous period. Simul-

taneously, he offers employment and pays money-wages in the twosectors. Employment in the consumption goods sector is still determinedby the stock of machines there ( M t– 1 ) and, provided that the dictatorchooses a price of consumption goods such that the subsistence realwage rate is established, the surplus of consumption goods, total em-ployment in the investment goods sector, gross and net investment, andthe rate of growth of the stock of machines will still be maximised.However, the surplus of consumption goods must now be regarded asthe difference between the stock of consumption goods accruing fromthe previous period and the real wages bill in the consumption goodssector established at the beginning of the current period.

The second point is that the above results depend upon the dictatordirecting production. There is no guarantee that the decisions of privatebusinesspeople would bring about the same total employment, produc-tion, and rate of growth of machines each period. For this to occur,businesspeople in the investment goods sector would have to plan toproduce the quantity of machines implied in the equation for M t above;businesspeople in the consumption goods sector would have to buy

10

@ r @ w

¼ 1

mcl m wð Þ2 < 0

11 Johnson ( 1962 , 28.)


http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-


32/216

this quantity, offer sufcient employment to keep M t – 1 machines run-ning and, supposing w to be given, set the price of consumption goodswhich establishes the subsistence real wage rate. Failure to do any of thesewould mean frustration of plans and lower levels of employment and

production.

Kalecki’s review of Keynes’ General Theory

Before Kalecki published his macro theory of distribution in Econome-trica in 1938, he wrote in Polish a quite extraordinary review article of Keynes’ General Theory (Kalecki 1936 ). Unfortunately, as we have seen,the review was not available in full i n a n English translation until 1982(see Targetti and Kinda-Hass 1982 ). 12

The review shows conclusively that Kalecki had derived independ-ently the principal propositions of Keynes’ book, albeit by a differentroute: for it seems beyond the bounds of credibility that anyone couldhave written such a masterful account of the new theory and display sucha complete command of the issues unless either he had derived themindependently or, as in the case of, for example, David Champernowne

(1936 ) and Brian Reddaway ( 1936 ), who also wrote remarkable revie ws ,had been Keynes’ pupils when The General Theory was being written. 13

Kalecki’s title in English is ‘Some remarks on Keynes’ theory’ – Kalecki was never one to waste words. He thought the book was ‘withoutany doubt, a turning point in the history of economics’ (245, all pagereferences are to the 1982 Targetti and Kinda-Hass translation, unlessstated otherwise). Kalecki identies two fundamental parts to the theory.The rst is the determination of short-period equilibrium with what he

calls ‘a given production apparatus’ (245), once the level of investment(per unit of time) is given. The second is the determination of thevolume of investment itself (this was the principal preoccupation all of his life in Kalecki’s writings on the capitalist economy). Kalecki judgedthat Keynes had solved the rst point ‘very satisfactorily’, thoughKalecki in the review decided to give his own interpretation and follow

12 Ferdinando Targetti was a research student of mine in 1972–3 in Cambridge andBogulslawa Kinda-Hass is a Polish economist who married Targetti. I asked them totranslate the review and write a commentary on it for Australian Economic Papers (AEP) ,of which I was then a joint editor. I regard Kalecki’s article as the most important articlewe ever published in AEP .

13 Austin Robinson also wrote a most perceptive review for the Economist (Robinson1936 ). It is the only signed review ever to appear in the Economist and, even then, itwas only his initials, EAGR. Austin complained to Keynes that the journal had changedthe title, making it too narrow (‘Mr. Keynes on Money’) and had altered some of thetext. Keynes said it served him right for publishing in the yellow press.




33/216

a different path to arrive at Keynes’ basic conclusions. Kalecki detec-ted serious deciencies in the construction as well as the exposition of Keynes’ arguments under the second heading, leaving problems thatremain still at least partly unsolved. We discuss the second part of

Kalecki’s essay in chapter 4 on accumulation, in which we set outKeynes’ theory of investment and the deciencies in it as revealed byKalecki’s and others’ criticisms. Here we concentrate on the rst fun-damental part of Keynes’ contribution, assuming, as we did above, thatthe level of planned investment is given.

Kalecki starts by clearing out of the way some preliminary issues:rst, the concept of a given production apparatus. Because actual accu-mulation will change the inherited production apparatus – the stockand capacity of existing capital goods – over any period of time (unlessit is assumed that we are in a stationary state), we must choose a period of analysis that is short enough to allow us to ignore the effects of a change inthe apparatus on output, employment and so on. As output and incomeare measured per unit of time, they are independent, Kalecki argues, of the length of the period we choose. Kalecki says that Keynes considers aclosed system, an economy without either an overseas sector or a govern-

ment sector. Kalecki follows suit and further assumes that wage-earnersdo not save so as not to ‘obscure some typical char ac teristics of thefunctioning of the capitalist economy in general’ (246). 14

Kalecki refers to Keynes’ use of the wage unit as his numé raire .Though he adopts it for expositional purposes in the review, he isnot happy about using it because, in his opinion, it eliminates fromthe analysis one of the most important factors in the working of theeconomy – the general movement of prices. Probably he had in mind

the proportional movement in prices when money-wages change in theshort period.

Kalecki also follows Keynes’, in many ways unfortunate, procedurein the light of subsequent interpretations of The General Theory , of assuming a con sta nt money-wage. He also assumes that there is a reservearmy of labour. 15

Now the analysis proper begins. The level of production, Kaleckiargues, depends upon employment and its allocation to particular

14 This was always a characteristic of Kalecki’s procedures – make simplifying assumptionsthat allow us to bring out starkly and clearly the processes at work without leaving outessential determinants of the economy’s behaviour which are relevant for the particularissue being analysed – surely the mark of an inspired theorist and an example of howimportant is the role of judgement in the making of a really great economist.

15 On the face of it, this is a strange assumption to make for though, of course, it – unemployment – was very much in evidence then – and now – yet one of the aims of the analysis of both Keynes and Kalecki was to explain why it was there.


http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-


34/216

sectors of the p ro duction apparatus. We then go to the basic unit of ana-lysis, the rm. 16 In every rm the level of production (and so theemployment offered in each short period) is determined by the inter-section of the curve of marginal revenue which, Kalecki points out,is horizontal in a freely competitive market – that is to say, the rm isa price-taker. It is signicant that Kalecki, as with Keynes and, later on,Paul Davidson, Jan Kregel and Nina Shapiro (see, for example, Shapiro1997 ), felt then that the new theory of effective demand was independ-ent of the market structures of the economy. Kalecki conrms this bysaying that he will ‘deal with a more general case, which includes alsoimperfect competition’ (247). Raw material costs and depreciation(which is associated with use) are deducted in order to construct thecurves of marginal value added ( MVA ) and marginal labour costs( MLC ), see gure 2.3 . Production is determined by the intersection of the two curves, with all quantities measured in wage units.

Kalecki denes short-period equilibrium as the state where none of the curves of all the rms moves. The position of the MLC curve is

16 It is signicant that Lorie Tarshis, who was Kalecki’s contemporary and Keynes’ pupil,also always argued that the rm should be the basic unit of analysis. Tarshis’ Ph Ddissertation, written in the second half of the 1930s for Cambridge (Tarshis 1939 ), wasan independent discovery of Kalecki’s theory of distribution. It drew on Joan Robinson’sThe Economics of Imperfect Competition (1933a ) and Kahn’s lectures on the economics of the short period, embedding their ndings in the system of The General Theory . It is atragedy that it was never published; for the reason why, see Harcourt ( 1982 , 1995a ;2001a ).

Figure 2.3. Kalecki’s rm in the short period.


http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-


35/216

settled because of the assumptions of a given money-wage and produc-tion apparatus. So the MVA curves when they no longer shift are res-ponsible for equilibrium. 0 ABC is the total value added provided by therm when it is producing 0 C . The shaded area formed by the two curves

is the income of the capitalists-entrepreneurs and rentiers, those wecalled prot-receivers above. S 0 ABC , the sum of the values added of all rms in the economy, is the national income measured in thewage units; the sum of all the shaded areas is the capitalists’ income,prots; and the sum of all the unshaded areas is the global income of thewage-earners, the wage bill.

But social income, Kalecki argues, is the value of consumption ( C )plus investment ( I ) and because wage-earners are assumed neither tosave nor to borrow, the sum of the non-shaded areas coincides with thevalue of the wage-earners’ consumption. The sum of the shaded areascoincides with the prot-receivers’ expenditure on consumption andinvestment – they receive what they spend.

A spontaneous change in the wage-earners’ spending cannot occurbecause they spend what they earn, neither saving nor able to dissave.A spontaneous change in spending by capitalists is possible because

they can spend their reserves, or contract new debt, or issue new shares.Suppose they raise their expenditure per unit of time. This will shiftthe MVA curves up to the point where the sum of the shaded areasmatches the higher value of their expenditure and ‘will force’ a higherincome of the same amount for them. (The same process could beillustrated in Joan Robinson’s diagram above, see p. 12, by supposingthat there is a higher level of investment planned so that the wages bill of the investment goods sector is greater and the rectangular hyperbola

corresponding to it moves out to subtend the greater area associatedwith this greater wage bill.)

In the new resulting short-period equilibrium, employment and in-come of the wage-earners and the value of consumption will be higher,activity having expanded in all branches of industry. It is though, Kaleckistresses, the level of capitalists’ spending that is the crucial factor inthe determination of short-period equilibrium. The new level of invest-ment also ‘forces’ saving, the value of which is equal to the value of thisinvestment (even though the capitalists who do the investment are notnecessarily those who save – the latter’s saving arises from, is created by,the investment of the former). Kalecki adds a proviso: to determineshort-period equilibrium in all its details the composition of the magni-tudes of consumption and investment needs to be known, but this is aminor consideration relative to the determination of the overall levelsof employment and income.



36/216

Kalecki writes, therefore, Y ¼ f(I) and dY dI ¼ f 0(I Þ, where f 0(I ) is the

Kahn–Meade–Keynes multiplier. Investment is therefore the factorwhich decides both short-term equilibrium and the amounts of employ-ment and social income at a given moment . Hence why we have high

or low levels of employment and production depends on the analysis of the factors governing the amount of investment.

Kalecki stresses that saving does not determine investment but thatinvestment creates saving so that the ‘equilibrium between demand for“capital” and supply of “capital” always exists’, whatever determines therate of interest (250).

Kalecki also discusses the effect of a change in the wage unit on short-period equilibrium. He supposes that businesspeople do not infer im-mediately the consequences for the expected protability of plannedinvestment of, say, a fall in the value of the wage unit (the relevant caseto consider in a situation when the conventional wisdom saw cuts inmoney-wages as the cure for unemployment). They do nothing immedi-ately, prices therefore fall in the same proportion as the wage unit andthe ‘improvement’ in protability turns out to be ‘illusionary’ (251). 17 Itwould be an interesting thought experiment to do Kalecki’s case in

reverse by considering the impact of a rise in the value of the wage unit.Summing up, Kalecki’s review article is a remarkable contributionboth then and now. It sets out succinctly, clearly and persuasively, atheory of the determination of the levels of employment, income and the distribution of income in the short period.

The eclecticism of Joan Robinson

Though Joan Robinson was to end up adopting Kalecki’s framework,especially in short-period analysis, nevertheless over the years she took amost eclectic stance on the issues of distribution, especially when con-sidering the long period. In the 1930s, while already an enthusiasticKeynesian, when she came to extend the system of The General Theoryto the long period (Robinson 1937 ), she was still content to use aneoclassical framework for the distribution aspects of her long-periodmodel. (There is, of course, nothing that surprising about this – bothKeynes and Joan Robinson were brought up on Marshallian–Pigovianconcepts and approaches.) Joan Robinson exploited the then highlyfashionable concept of the elasticity of substitution between capital and

17 If it is asked: ‘what about the Pigou or real balances effect?’ it should be recalled thatKalecki devastatingly criticised it in 1944, see Kalecki ( 1944 ; 1990 –7).


http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-


37/216

labour to determine the distribution of the product between protsand wages and so, because of the different values of the marginal pro-pensities to save from the two income sources (and correspondinglyclasses), the overall rate of saving.

Even in the 1950s, though she was by then criticising both neoclas-sical methodology as she saw it and marginal productivity theory, ina wide-ranging article published initially in French and reprinted inEnglish in her second volume of Collected Economic Papers (1960), sheused a straightforward neoclassical argument, together with argumentsrelated to Harrod’s theory of growth and Kalecki’s and Kaldor’s macro-economic theories of distribution, to illuminate some possible long-period and longer-term theoretical possibilities and happenings. Shedid conclude with an enigmatic statement – ‘It is at the points wherethe theory breaks down that it begins to become interesting’ (Robinson1960, 158). Moreover, she also used Kalecki-type arguments for short-period analysis as well as considering the role of bargaining in the labourmarket between capital and labour as a major determinant of distri-bution in specic situations. In addition, she examined the link betweenthe determinants of the rate of exploitation in Marxian analysis and the

distribution of income between prots and wages in both a Keynesianand a Srafan framework. Her suggestions here have been most simplyand persuasively exposited by one of her younger colleagues, DonaldHarris, in his article in the American Econom ic Review (Harris 1975 ) and,more fully, in his later book (Harris 1978 ). 18 Her nal views are to befound in a joint article with Amit Bhaduri (Bhaduri and Robinson 1980 )where, signicantly, it is Marx, Kalecki and Sraffa who are named asthe key inspirations.

Here we concentrate on the short period, in particular on an especiallypithy paragraph in an article Joan Robinson wrote in the New Left Reviewon Piero Sraffa’s 1960 book:

In any given situation, with given productive capacity in existence, a higher rateof investment brings about both a higher level of total gross income (through ahigher level of employment and utilization of plant) and a higher share of grossprot in gross income (by pushing up prices relatively to money-wage rates).Thus, within reason, investment generates the saving that it requires.

Joan Robinson ( 1965b ; CEP , vol. III, 1965, 177)

18 The details of this synthesis are presented in Chapter 7 (pp. 119–21), where all thestrands of the discussions above, not only on the distribution of income but also ontheories of price-setting and accumulation, are brought together.


http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-


38/216


39/216

It is immediately obvious ( ex post , all things in economics are obviouswhen someone else has explained them) that a higher level of accumula-tion need not necessarily be associated with a higher level of income suchas we have come to expect from the operation of the simple Kahn– Meade–Keynes multiplier analysis (and has occurred in gure 2.4 ).Whether income will be higher, the same, or lower obviously dependson the respective sizes of the movements upwards in the nominalplanned investment and corresponding saving schedules. The equilib-rium condition is as ever saving equals investment, but now saving maybe changed by both the higher level of income and the higher share of prots at any given level of income. Indeed, the result is the privatesector near-equiv alent of the balanced budget multiplier theorem inthe public sector. 19 By itself, a higher level of planned investment hasan expansionary effect; by itself a higher level of planned saving out of

a given level of income has a contractionary effect, an example of theold-fashioned paradox of thrift.

19 My former graduate student Jorge Araujo even went so far as to dignify this seeminglynegative multiplier result with the title ‘the Harcourt effect ’ (see Araujo 1999 , 338,emphasis in original!).

Figure 2.4. Accumulation and internal nance: systematic effects.


http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-


40/216

Hahn’s nest hour: the macroeconomic theory of employment and distribution of his PhD dissertation

In the lat e 1940s, early 1950s Frank Hahn wrote a PhD dissertation at

the LSE.20

He was supervised for a short while by Kaldor and then byLionel Robbins. His contribution is in some ways the most satisfyingversion of th e Keynesian macro theory of distribution that we have (seeHahn 1972 ). 21

The inspirations for Hahn are not only classical – the subtitle is AnEnquiry into the Theory of Distribution – but also the Keynesian revolution,Samuelson’s Foundations and the then current theories of entrepreneur-ial behaviour in situations of uncertainty. The method is Popperian – theory must produce inferences that may be empirically falsiable, atleast in principle. It is an early example of the attempts to providemicroeconomic foundations for macroeconomics. It emphasises theKeynesian forces of effective demand and the analysis is set emphaticallyin the short period, for Hahn will have no truck with long-period equi-librium as a guide to an explanation of the level and changes in the levelof the share of wages in the national income. With this volume he stakes

a just claim to being an originator of the Keynesian macro theory of distribution. He recognises that Kalecki was there rst, but dismisseshis theory once we are not analysing the distribution of income in adepression in which there is an unlimited amount of excess capacity, as‘not very helpful when net investment is positive, for then there is noreason why [Kalecki’s] basic assumptions [that there is continuousexcess capacity, that marginal “overhead” costs are small enough to beignored and that the concept of an average degree of monopoly is

meaningful] should hold’ (Hahn 1972 , 44–5). Neither does he likeKalecki’s ‘degree of monopoly’ determination of the size of the mark-up because it does not predict the observed changes in the share of prots over the cycle.

The basic tools are the IS/LM version of the Keynesian model, usedby Hahn to build up the aggregate demand side of the story. Theaggregate supply side is built up from the behaviour of the principaldecision-makers in the economy, those who decide on the levels of out-put and employment and on rates of accumulation in the short period,

20 It was not published until 1972 in the Weidenfeld & Nicholson series of famousdissertations unpublished on their immediate completion.

21 I often teased Frank that it is the best thing he ever did, that it has been downhill eversince.


http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-


41/216

operating in an environment of uncertainty and having suitably speci-ed utility functions to reect this. Simply put, the decision-makersmust expect to receive more prots, both absolutely and as a share of higher levels of national income, in order to offset the increasing disu-

tility associated with organising higher levels of output and employmentand of accumulation to provide the capacity to make these higher levelspossible. This leads, through complicated aggregation procedures, to asimple relationship between the share of wages in short-period nationalincome and levels of activity – income – themselves.

On the aggregate demand side, the IS/LM construction is adapted totake into account the distribution of income and accompanying differ-ences in the marginal propensities to save from wages and prots, assum-ing prices to be more exible than money-wages in the short period (forsimplicity, money-wages are assumed to be constant). Expected levels of income play two roles. First, they induce certain rates of accumulationonce the level of output is such that existing capacity cannot produce it,making allowance for some extra capacity to cope with unexpectedevents. There is then a relationship between the share of wages in eachlevel of income and income itself, such that both the levels of income and

their distribution are such as to provide overall levels of saving that arejust offset by the injections associated with the rates of accumulationthat have in turn been induced by each level of income. Unrealistically,though Hahn is in good company with Harrod at this point, currentaccumulation also provides the capacity needed to produce the currentincome that induced it. (The factors behind the LM curve do not reallyget a look in, at least explicitly.)

The two relationships are shown in gure 2.5 . On the vertical axis

we measure the share of wages in national income, on the horizontalaxis, income itself. The SS line shows the supply side of the story, theIS line the demand side. Where they intersect, we have a stable short-period equilibrium position, with the share of wages (and thereforeprots) being such as to justify both the production of the associatedlevel of income (aggregate supply) and to provide sufcient aggregatedemand for it. It is a stable equilibrium position, because to the rightof the intersection the forces of excess supply resulting producesignals that take the economy back to the intersection; similarly, tothe left of the intersection excess demand situations bring about thesame result.

To sum up, all the approaches examined in this chapter share manyof the same ingredients - though each, of course, takes its own unique



42/216

slant on the issues involved. What is clear is that all these macroeco-nomic theories of distribution are consistent with quite different postu-lates about individual rms’ behaviour and motivation. What is missingfrom them, at least explicitly and with the exception of Kalecki’s contri-bution, is a theory of the determination of the size of the mark-up, thesubject of chapter 3.

Figure 2.5. Hahn’s short-period model of income distribution andactivity.



43/216

3 Post-Keynesian theories of the determinationof the mark-up

Wood’s ‘Golden Age’ model

There are a number of papers on the determination of the size of themark-up in the post-Keynesian literature – Ball ( 1964 ); Eichner ( 1973 ,1976 ); Harcourt and Kenyon ( 1976 ; Harcourt 1982 ); Wood ( 1975 ), forexample. I concentrate here on the contributions of Adrian Wood andPeter Kenyon and myself because each in their own way reveals thestrengths and the limitations of the analysis. James Ball’s contributionmust be accorded a pioneering role and his contribution has been shame-fully neglected in subsequent discussions. The other contributions arerelatively well known, with Alfred Eichner’s articles and books prob-ably the most widely known. However, I shall not discuss his particularversion of the theory, while acknowledging its originality and inuence,because it is dependent upon Keynes’ theory of investment expenditure,the mec and all that as set out in The General Theory . As we argue inchapter 4 , there are serious aws – or at least unnecessary limitations – inthe details of Keynes’ theory. Subsequently, these have been removed bythe criticisms and contributions of Abba Lerner, Kalecki, Joan Robinsonand Tom Asimakopulos.

We use Wood’s model and then Harcourt and Kenyon’s model becauseWood’s analysis is explicitly ‘Golden Age’, logical time analysis, whileHarcourt and Kenyon attempt to make an analysis set in historical time,as advocated by Kalecki (implicitly) and Joan Robinson (explicitly).

Both procedures have their place in the development of theories asKahn ( 1959 ) and Joan Robinson ( 1962a ) have argued – the need to ex

intellectual muscles by doing the necessary preliminary work of ‘GoldenAge’ analysis before tackling the more complex task of process analysis,that is to say, an analysis in historical time. Kahn puts the issuesextremely clearly:

when one speaks of a Golden Age being preferred [to another one], it means thatit would be preferable to be in it . . . to be in it involves having been in it for along time past, and enjoying the legacy of the past in terms of the accumulated

32



44/216

stock of capital and the degree of mechanisation. The desirability of a movementfrom one . . . to the other, and the manner in which it might be smoothlynegotiated is . . . [an] important and difcult [problem]. What I have said inthis paper . . . is no more than prolegomena to the solution of real problems.

(Kahn 1959 , 206–7, emphasis in original)

‘Golden Age’ analysis allows denitions of concepts and relationshipsto be set out precisely and exactly, valuable mind-clearing exercises.Historical time analysis, by its nature, is often more fuzzy and vaguebut has a closer hold on the processes that are actually observed in reallife economies.

Joan Robinson ( 1962a , 23–6) speaks of:

two kinds of economic arguments, each of which is useful in analysis providedthat it is not stultied by being confused with the other.[The rst] proceeds by specifying a sufcient number of equations to determineits unknowns, and so nding values for them that are compatible with each other. . . The other . . . species a particular set of values obtaining at a moment intime, which are not . . . in equilibrium with each other, and shows how theirinteractions may be expected to play themselves out. . .There is much to be learned from a priori comparisons of equilibrium positions,but they must be kept in their logical place . . . cannot be applied to actualsituations . . . a mortal certainty that . . . particular actual [situations are] not inequilibrium. . .A model applicable to actual history has to be capable of getting out of equilib-rium . . . normally not . . . in it. To construct such a model . . . specify thetechnical conditions obtaining . . . and the behaviour reactions . . . and then, soto say, dump it down in a particular situation in historic time and work out whatwill happen next. . .At any moment in logical time, the past is determined just as much as the future

[, there is no causation] . . . In an historical model, causal relations have to bespecied. Today is a break in time between an unknown future and an irrevoc-able past. What happens next will result from the interactions of . . . behaviour. . . Movement can only be forward.

Both Wood’s and Harcourt and Kenyon’s contributions are written ina Marxian vein, involving a vision of capitalist society and especially of the role of the rm in capitalism as an institution in which accumulation,growth and prot-making are ends in themselves for the decision-makersin charge. In order not only to survive but also to dominate, price-settingand attempted sales are necessary means to these ends. Both sets of authors postulate a direct connection between price-setting on the onehand and the supply of nance through internal funds, given the avail-ability of external funds, needed for planned accumulation on the other.

So what determines the size of the mark-up over costs in such aregime – or, in Wood’s case, the rate of prot in sales revenue or target

Theories of the determination of the mark-up 33



45/216

prot margin? Wood’s analysis is explicitly long-period in the sense of Joan Robinson’s discussion and refers to a price-leader rm in an oli-gopolistic industry in which the r m ’s expectations of the future for allrespects and purposes are fullled. 1 The rm’s aim is to maximise the

growth of its sales revenue subject to certain constraints – the growth inthe demand for its product, growth in its capacity and the availability of nance.

Wood takes growth in aggregate demand as exogenous so that thegrowth in demand for the product of the rm relates to the effects onsuch growth of its selling policies vis-à -vis those of its competitors. Woodformulates these constraints as two frontiers – the opportunity frontier(OF ) and the nance frontier ( FF ).

First, OF : the opportunities of the rm for growth may be related tosets of alternative pricing, investment and sales policies. Each set isassociated with a given average prot margin, a particular rate of growthin sales revenue and a particular level of planned investment expend-iture. At some point the rm must encounter a trade-off between ahigher prot margin, on the one hand, and a higher rate of growth of sales, on the other. This trade-off is the OF of the rm; its position and

shape depends upon the efciency of the rm in controlling its costs.Investment expenditure is the clue to growth by providing additionalcapacity and lowering costs (by embodying the latest best-practice tech-niques through investment in the capital stock). This means that for agiven rate of growth of sales revenue, a higher prot margin may beachieved; for a given prot margin, a higher rate of growth of salesrevenue may be achieved. For every possible investment coefcient orincremental capital–output ratio, we may dene a unique opportunity

frontier, OF , see gure 3.1 , where p is the prot margin and g is the rateof growth of sales revenue. Once the frontier has been reached, the rmhas either to cu

Harcourt (2006) Structure of Post-Keynesian Economics

Documents

Transcript of Harcourt (2006) Structure of Post-Keynesian Economics