Joan Robinson and Modern Economic Theory

In the 1930s there were three great waves in economics: the Keynesian revolution, the imperfect (monopolistic) competition revolution, and the ‘fruitful clarification of the analysis of economic reality resulting from the mathematical and econometric handling of the subject’ (Samuelson, 1977, p. 890). Joan Robinson was a creative participant (as a member of Keynes’s Circus) and a generalizer of the first wave and one of the two independent (and complementary) architects of the second. Her position in the third is ambivalent. While she was innocent of modern mathematical techniques and showed some hostility towards their use in economics, her own theoretical writings (especially her major pre-war (1933) and post-war (1956, 1966) books) are very formalistic and abstract. She casts her argument in what may be called the axiomatic method, even though she is tinged with the ‘Marshallian incubus’ in execution.

Throughout my lifetime as an economist Joan Robinson was always there at the frontier of science. We began together in 1932: I as a student at Chicago; she in her first publishing phase. Many times, in print and elsewhere, I expressed the considered opinion that the corpus of her work richly deserved a Nobel Prize.

The fundamental nature of the underlying economic problem, i.e. economics in the abstract, has never really been fathomed because in the 200 or more years since Adam Smith’s Wealth of Nations we have had a continuous outpouring of fresh textbooks on ‘Principles’ without ever finding the definitive statement. True enough, Alfred Marshall and Paul Samuelson both survived many editions, but there are always new statements forthcoming, attempting to displace the masterful renditions. Many of these new attempts succeed. There seems to be an ever ready market for a new packaging of the fundamentals of our subject.

The steady course on which mathematical economics has held for the past four decades sharply contrasts with its progress during the preceding century, which was marked by several major scientific accidents. One of them occurred in 1838, at the beginning of that period, with the publication of Augustin Cournot’s Recherches sur les Principes Mathématiques de la Théorie des Richesses. By its mathematical form and by its economic content, his book stands in splendid isolation in time; and in explaining its date historians of economic analysis in the first half of the nineteenth century must use a wide confidence interval.

When Roy Harrod (1948) said that, since investment raised demand and higher demand led to more investment, (a) he completed the Keynesian system; (b) he demonstrated the instability of capitalism; and (c) he contradicted a basic tenet of traditional economic analysis. When a firm or an individual makes an investment, any influence on its or his demand is negligible. This apparent contradiction between micro and macro has to be resolved and its resolution should be illuminating. The physicist faces a somewhat similar problem to that of the economist: he has a very large number of variables, the individual behavior of which he cannot hope to predict. The German physicist Haken (1983) has proposed a suggestive approach called ‘synergetics’. One or a few parameters may act in such a way as to produce uniform changes in the behavior of very large numbers of independent entities; he calls this ‘self-organization’.

In her first published paper, Joan Robinson stated the following methodological position: ‘a serious subject, in the academic sense, is neither more nor less than its own technique … [gradually developed in] an austere and disinterested search, not ‘for the Truth’, but for a single self-consistent system of ideas’ (1932, pp. 3–4).1 She saw the intellectual activity of providing unreal answers to unreal questions as part and parcel of the laborious task of finding a set of assumptions at once tractable and realistic. This view of economic theory ‘as a sequence of conceptual models that seek to express in simplified form different aspects of an always more complicated reality’ is defended by Tjalling Koopmans, who remarks that ‘most economists when pressed will agree to one formulation or another of such a view of the logical structure of economic knowledge’ (Koopmans, 1957, pp. 142–43). In defense of general equilibrium theory, Frank Hahn takes a similar position, arguing that ‘the Arrow-Debreu machinery gives us the best base camp for sallies into new territory’ (Hahn, 1984, p. 10).2

It is not my purpose in this brief note to alter anything in that restrospective treatment of Joan Robinson’s economics in which I had a share (Gram and Walsh, 1983), but rather to add some further thoughts on the specific topic of her objections to the notion of ‘getting into equilibrium’, in both short periods and long. In particular, I shall argue that her well-known objections to ‘getting into equilibrium’ have been vindicated by some of the most recent theoretical work on both short-and long-period models. What is both sad and ironic is that she was probably seldom in the minds of the (relatively) younger theorists whose herculean mathematical labors have led them at last to reach a ‘Pole’ (an image of hers) where her by now somewhat tattered flag has been flying for so many years.

The notion of equilibrium and the distinction between two kinds, long run and short, has been fundamental to almost all economic analysis of the past century. By contrast, the classics and Marx sought for the ‘laws of motion’ of capitalism. These latter do not define a long-run equilibrium, and it will be the claim here that a full-fledged, long-run equilibrium is not possible in a class society. This is not exactly Joan Robinson’s position, but it is close, and it builds on some of her favorite themes.

When I was asked to contribute to this volume I thought that the following work, done back in 1976 as part of a discussion with Joan Robinson, might be of some interest to a wider public. It is presented here in the spirit of the kind of commemoration she would have preferred — a discussion of her work in order to further the construction of a more satisfactory economic theory: the aim she single-mindedly pursued in her work, earning the respect and admiration of both those who did and those who did not share her critical stance.

It is not only during the last few decades that Sraffa has acquired an important place in discussions of economic theory. An article on the laws of variable returns published as early as 1926 (Sraffa, 1926) when he was 28, in which he was already pointing towards a criticism of the demand and supply theory of value, attracted much attention as soon as it appeared. At that time however the interest focused on some remarks on the theory of the firm, which were to be developed later by Joan Robinson in her theory of imperfect competition. The work by Sraffa which is at the heart of the discussion today belongs mainly to the period since 1951. Iam referring to the Works of Ricardo (1951–73) which Sraffa edited for the Royal Economic Society, and to his book, Production of Commodities by Means of Commodities, (1960). These works relate to two connected themes which are important (more important, perhaps, than is immediately apparent) for an understanding of the present situation of economic theory. The two themes are the criticism of the dominant demand-and-supply theory of distribution and relative prices, and the revival of a different approach to these problems: that of the British classical economists down to Ricardo. An attempt, however brief, at reporting Sraffa’s contribution therefore requires some preliminary words about the dominant and the classical approaches to distribution, and the criticism levelled against the former.

As Joan Robinson’s views matured, the study of expansion through the accumulation of capital moved more and more to central stage, and she increasingly sought to group other questions around it. Yet it proved a difficult nut to crack. Accumulation has been analyzed by economists in two very different ways, between which she moved uneasily The most common has been to see it as the expansion of the productive potential of an economy with a given technology, which may be improved in the process — this was her basic stance. But it has also been understood as the outright transformation of the technical and productive organization of the economy, an outlook of which she thoroughly approved, but which played almost no role in her analytics.

This presentation is a token of homage to Joan Robinson, from whom I learnt the method of analysis which goes with her name. Her life was a constant effort to rationalize the manner of thinking in economics and she insisted on rigor in the formulation of ideas.

The problem of economic growth was a major preoccupation of Joan Robinson. She was a major contributor to the post-Keynesian theory of economic growth that followed the publication of Harrod’s seminal dynamic model. She wrote extensively and critically on the foundations of neoclassical growth theory; her concern for logically sound argument lay behind her extensive writings questioning the validity of an aggregate capital concept and the corresponding notion of an aggregate production function. Her critique of neoclassical theory placed her at the forefront of the still raging ‘Cambridge controversy’ in capital theory. Joan Robinson’s work on capital and growth also showed a real concern that economic dynamics be studied as a process in real time. The dynamics of the stationary state might be a useful starting point, but the more serious concerns were the relationship between the short- and long-run periods in the process of accumulation and growth. She also stressed the importance of understanding the evolution of economies outside equilibrium positions.

A first approximation to the study of the intertemporal aspects of resource allocation (capital theory from now on), consists of concentrating on the steady states (the rest points) of associated dynamical systems. Provided one does not lose sight from the fact that this is not an end in itself, there is much useful information to be gleaned from steady-state analysis — indeed, one of the prime tools of economics. No doubt, the pervasiveness of the notion of the stationary state in classical economics and the fact that there is so much one can do without invoking powerful mathematics, have also added to its popularity.

Joan Robinson frequently argued that neoclassical general equilibrium theory could not determine the rate of interest in intertemporal models (see, for example, Robinson, 1973). There were two aspects to this critique: First, neoclassical marginal productivity theory depended on the notion of an aggregate capital stock. Because of aggregation problems, notably reswitching, this concept could not be defined without resorting to circular reasoning except in the most unrealistic of models. Second, for any rate of interest there is a different short-period equilibrium in a neoclassical model. There are not enough equilibrium conditions to determine what this rate of interest is.

An editorial note in the Economic Journal (May 1930) reported the death of Frank Ramsey, and his 1928 paper was described as ‘one of the most remarkable contributions to mathematical economics ever made’. In the same issue the editor organized a symposium on increasing returns and the representative firm. This symposium seems to be a natural follow-up of a number of papers published by the Journal during 1926–8, including the well-known article of Allyn Young (1928) that is still available, and duly remembered. The problems of equilibrium of a firm under increasing returns, or more generally, of designing price-guided resource allocation processes to cope with increasing returns, has since been a topic of continuing interest. Ramsey’s contribution was enshrined as a durable piece with a resurgence of interest in intertemporal economics in the fifties. But neither John Keynes, the editor of the Economic Journal who was most appreciative of Ramsey’s talents, neither the subsequent writers on ‘growth theory’ in Cambridge, England (nor, for that matter, those in Cambridge, Massachusetts), have made any precise suggestion towards incorporating increasing returns in a Ramsey-type exercise.

Many problems in capital theory — particularly ‘Austrian’ capital theory —take the following form: an asset has an intrinsic value X(t) at time t. If he takes a particular action at time T, then the asset’s owner gets X(T) at T. In anticipation of future usage we shall call the action taken at T stopping and refer to T as a stopping time. This set-up raises two natural, and related, questions. When should the intrinsic process be stopped? What is the present value of the asset? The standard examples are when to drink the wine whose quality at t is given by X(t) or when to cut down the tree which contains lumber with a value of X(t). If the discount rate is r then these questions may be simply answered. The optimal stopping time T* maximizes e-rTX(T) and the present value of the tree is its discounted value 1<math> <mrow> <mi>V</mi><mo stretchy='false'>(</mo><mi>t</mi><mo stretchy='false'>)</mo><mo>=</mo><msup> <mi>e</mi> <mrow> <msup> <mrow></mrow> <mrow> <mo>&#x2212;</mo><mi>r</mi><mrow><mo>(</mo> <mrow> <mi>T</mi><mo>*</mo><mo>&#x2212;</mo><mi>t</mi></mrow> <mo>)</mo></mrow></mrow> </msup> </mrow> </msup> <mi>X</mi><mrow><mo>(</mo> <mrow> <mi>T</mi><mo>*</mo></mrow> <mo>)</mo></mrow></mrow> </math>]]</EquationSource><EquationSource Format="TEX"><![CDATA[$$V(t) = {e^{^{ - r\left( {T* - t} \right)}}}X\left( {T*} \right)$$ To distinguish from intrinsic value, call this latter quantity the market value of the asset.

An enduring theme in Cambridge economics — and Joan Robinson’s work in particular — is that the overall macroeconomic situation is intimately linked with income distribution. Following in Robinson’s footsteps, Row-thorn (1982), Taylor (1983), and Dutt (1984) have recently presented models of how growth and distribution interact in a one-sector setting, under two alternative modes of macroeconomic adjustment. The first is based on changes in output or capacity utilization, along lines stressed by Keynes in the General Theory, and Kalecki (1971) and his followers. One striking Kaleckian result is that income redistribution favoring workers may lead to higher capacity utilization and faster growth. The second, ‘neo-Keynesian’ adjustment mechanism functions via income redistribution to create forced saving in response to higher investment demand. In such models (as proposed by Keynes in his Treatise on Money; Kaldor, 1960; and Robinson, 1962, for example) faster growth is necessarily associated with an endogenous distributional shift away from labor. The tenor of these results carries over when inflationary mechanisms and financial markets are brought into the models — see Marglin (1984) and Taylor (1985) for neo-Keynesian and Kaleckian variants, respectively.

The literature on productivity devotes considerable and deserved attention to a variety of measurement problems and to distinctions such as that between labor productivity and total factor productivity. However, some of the basic definitional issues that arise implicitly in many of the discussions do not seem to have been examined to the degree they merit. In this paper, we contrast three different basic concepts of productivity, discuss the differences in their interpretation and significance, and then demonstrate empirically that use of different notions of productivity can give rise to great differences in measurements of productivity growth.

Measures of the rate of technical advance, sometimes called sources of growth, total factor productivity, residuals, or measures of our ignorance are now commonplace. It is, however, not a commonplace to find recognition that there are basically two such measures: (i) the traditional measure, based on the early work of Hicks, and (ii) newer ones, based on the theoretical work of the late Sir Roy Harrod and Joan Robinson.

Joan Robinson subscribed to an elegant model of the real wage in which the wage is related to the average product of labor in the wage-goods sector and the proportion of the total workforce which is engaged in the production of wage-goods. Clearly, this model also suffices to account for the distribution of income (between wages and profits) within the wage-goods sector. Elsewhere (Dixon, 1981), I have set out a model of the wage-share in the economy as a whole, using Robinson’s model of the wage-goods sector as a starting point. In that paper it was shown that the proximate determinants of the economy-wide wage-share are: the proportion of the total workforce which is engaged in producing wage-goods, the price level of wage-goods relative to the general price level, and the level of the average product of labor in the wage-goods sector relative to the average product of labor in the economy taken as a whole. In this contribution I wish to demonstrate that these last two items (i.e. relative prices and relative labor productivities), taken together, are a measure of the rate of exchange of the labor time expended in the production of the various outputs produced in the economy. We are thus able to derive an expression for the economy-wide wage-share in terms of the sectoral composition of employment and the (implicit) rate of exchange of labor time.

In one of her many lucid critiques of mainstream economics, Joan Robinson wrote:

Micro questions — concerning the relative prices of commodities and the behaviour of individuals, firms and households — cannot be discussed in the air without any reference to the structure of the economy in which they exist and to the process of cyclical and secular change. Equally, macro theories of accumulation and effective demand are generalizations about micro behaviour: the relation of income to expenditure for consumption, of investment to the pursuit of profit, of the management of placements, in which financial wealth is held, to rates of interest and of wages to the level of prices results from the reactions of individuals and social groups to the situation in which they find themselves.

Even the

artificial conception of a stationary state has to be specified in terms of the behaviour of its inhabitants.

[Emphasis added.]

She went on to add that ‘if there is no micro theory, there cannot be any macro theory either’ (Robinson, 1977, p.1320). She approvingly quoted Hicks in the same article to the effect that models of steady growth (i.e. growing economy counterparts of the classical stationary state) are futile and the whole point of a steady state is to study disturbances (including policy shifts) that take the economy away from it.

Economic development was not a major focus for Joan Robinson’s research. Nevertheless, she considered the subject important and returned to it throughout her life.

In her well-known essay ‘Disguised Unemployment’ Joan Robinson coined this term for a situation widely observed in the Great Depression in which men, thrown out of regular employment, crowded into occupations like carrying bags, rendering small services or selling matches in the Strand. The reasoning underlying her argument can be brought out by a simple two-sector model: in one sector money wages are rigid downwards; in the other, where self-employment is common, incomes are flexible. In competitive full employment equilibrium, the marginal productivity of labor is the same in both sectors. If then a fall in aggregate demand below the full employment level occurs, men will be thrown out of work in the rigid wage sector, but, rather than become unemployed, will move into the flexible income sector. Money income per man in this sector will fall as more men are accommodated to spread a smaller work load. Productivity differentials (measured in terms of man-years, man-weeks, or man-days, but not in terms of man-hours, for productivity of hours not worked is not meaningful, though it is not clear how hours spent waiting for work, or in search of work, should be counted) will increase, but no visible unemployment will appear. The difference between a situation of general low labor productivity (say due to absence of skills) and a situation of disguised unemployment in this sense is that a rise in the level of effective demand will shift workers back into the high-productivity, rigid-wage sector and remove the disguised unemployment. The workers are adapted to the requirements in this sector and, if the time spent in the flexible sector has not been too long, so that they have not forgotten their skills, have remained well fed and healthy and have not been demoralized, a rise in effective demand is a sufficient remedy.

Two circumstances inspire this venture into ancient history, or prehistory if you will. The first circumstance is the current (1985–6) relapse of the international monetary system into an oligarchy of Central Bank and Finance Ministry discretion (and indiscretion) for however long the oligarchs can agree among themselves and leave the rest of us bedazzled by their perspicacity. The second circumstances is only the fortieth anniversary of my own initial attempt to teach International Economics to undergraduates (at Roosevelt University in Chicago, in the spring of 1947). Before my ageing memory fails completely, I shall try to reconstruct what was being taught to American college students of that day about the macroeconomic side of international economics in that epoch of intractable — some even said, of permanent — dollar shortage.

This essay analyzes a model that, for over a hundred years, has lurked behind discussions of the transfer mechanism and the theory of foreign exchanges, yet has never been made explicit. The model originates with John Stuart Mill (1848, vol. II, book III, ch. XXI) and was developed by Taussig (1917, 1927), Keynes (1930), and Robinson (1937a) among others. It underlies what Samuelson (1952) called the ‘orthodox presumption’ in the theory of the transfer problem; and it also lies at the basis of the simplest expositions of the so-called ‘elasticity approach’ to the balance of payments. Of course, one cannot say definitely that the model presented here faithfully represents the notions adhered to only subconsciously by all these writers; but it does reproduce (and make plausible) some of the most prominent propositions associated with these doctrines, as well as expose (and explain) some of the fallacies involved in their misapplication and misinterpretation.

It is a striking fact (and, in view of her strong social sympathies, a puzzling one) that during no phase of her long professional career did Joan Robinson seriously interest herself in the economics of labor unions. Thus in The Economics of Imperfect Competition there are chapters on the monopolistic and monopsonistic exploitation of labor but there is almost nothing about the monopolistic exploitation by labor. In her middle years, Keynesian preconceptions prevented her from attaching much significance to the practices of unions; see, for example, her 1945 essay ‘Obstacles to Full Employment’ (Collected Essays, vol. 1, pp.105–14). And in the final capital-theoretic phase of her career she worked almost exclusively with Ricardian models which impose blinkers of another kind.

Underlying much of Joan Robinson’s work, particularly in the later years of her life, was the aim of explaining the process of economic growth. She considered that the theory of allocation of resources with given endowments and technology was largely vacuous and that ‘a dynamic long-run analysis of how resources can be increased is now what we require’ (Robinson, 1962, p. 100). In formulating such a theory of growth both she and her opponents in the neoclassical camp agreed that the major sources of growth of output per head had to be accumulation and technical progress. Their disagreement was about the effects and the determinants of accumulation, and the causes and character of technical progress. Very little progress, however, has been made in explaining economic development by either the Cambridge or the neoclassical schools. The causes of the poverty of nations remain as obscure as they were when Joan Robinson called for a rediscovery of the problem of economic growth in the 1950s, despite the enormous intellectual energy devoted to the problem since then. In the 1980s most intellectual effort in economics is once again being devoted to static allocation problems of an even more obscure variety than Robinson decried.

In Economic Philosophy Joan Robinson argues that economic theorists have been perennially concerned with whether economic behavior and its outcomes can be morally justified and that this concern has had a major impact on economics.

Over forty years ago, Joan Robinson (1942, p. 22) wrote:

no point of substance in Marx’s argument depends upon the labor theory of value. Voltaire remarked that it is possible to kill a flock of sheep by witchcraft if you give them plenty of arsenic at the same time. The sheep, in this figure, may well stand for the complacent apologists of capitalism; Marx’s penetrating insight and bitter hatred of oppression supply the arsenic, while the labour theory of value supplies the incantations.

In particular, Robinson’s point applies to the theory of exploitation, which was the centerpiece of Marx’s analysis of capitalism. He developed the theory of exploitation in Volume I of

Capital

, with a model postulating the labor theory of value. (It was there assumed that prices of commodities are proportional to their embodied labor contents, their ‘values’) But exploitation of workers by capitalists, in the Marxian sense, remains true even in more intricate models where prices are of the usual market clearing sort, and are not proportional to embodied labor values. While Robinson understood this in 1942, the point has now been accepted by all but the most dogmatic writers, due to its elaboration in a series of mathematical models developed in the last fifteen years.

1

It is almost impossible to separate Joan Robinson the economist, Joan Robinson the person, and Joan Robinson the Cantabrigian, that is, a participant in the extraordinary group of economists who flourished at Cambridge University in the decades following World War I. I find it also difficult to separate my own reactions as an economist, as a person, and as a non-Cantabrigian. So this note perhaps must be discounted in academic terms for its strong personal flavor.

The quality that stands out in my memory of Joan Robinson is something very rare in England and the United States, so rare indeed that there is not even a word for it. The Austrians call it Zivilcourage. The translation of ‘moral courage’ is somewhat too pompous, and ‘civilian courage’ does not catch the flavor. Courage and fearlessness are certainly involved, and it is a form of courage that has little to do with the courage displayed by wartime heroes in battle. Men and women of great physical courage can turn out to be utter cowards when it comes to displaying Zivilcourage. A test for it is the following setting: you attend a meeting of important people, whose judgment and power you respect. A topic is discussed and views are voiced round the table. A consensus emerges. Finally, it is your turn to give your views. You utterly disagree with the emerging consensus. Do you have the guts to say what you think, at the risk of being regarded a fool or a knave by all the other distinguished people, or do you modify your views or suppress them? In my experience, people with no physical courage can display great Zivilcourage on such an occasion, and vice versa.

I came to King’s College, Cambridge, in November 1944 lured mainly by two names, Keynes and Mrs Robinson. My University of Bombay Prize consisted of three books which I had tried to read carefully soon after graduation: Keynes’s General Theory, Mrs Robinson’s Imperfect Competition, and James Meade’s Economic Analysis and Policy. The first aroused my curiosity about a number of things which were not clear enough to me then. The second was crystal clear and intellectually exciting. The third convinced me that Economics was a useful subject. Someone told me that if I liked these books, I should go to King’s because both Keynes and Mrs Robinson were there — little did we know of Cambridge then! I was, therefore, delighted when I learnt on arriving at Cambridge that Mrs Robinson at least was not drafted into the war effort — whether because of her radicalism or simply because of her sex, I had wondered.

I am no student of Joan Robinson’s: my interests lie elsewhere. But I was one of her supervisees when I got my first taste of economics at Cambridge in 1930, at age 19. I went to Cambridge only for a short time to learn English and international law; but, getting tired of law, switched to economics for my last two terms. I took Dennis Robertson’s and Maurice Dobb’s lectures and was assigned Joan Robinson as my supervisor, for whom I had to write fortnightly essays.

As an undergraduate at McGill University I was privileged to be guided by my professors B. S. Keirstead and Benjamin Higgins through a rich fare of classics such as Böhm-Bawerk’s Positive Theory of Capital, Hayek’s Pure Theory of Capital, Hicks’ Value and Capital, Lerner’s Economics of Control, Marshall’s Principles of Economics, Chamberlin’s Monopolistic Competition, Robinson’s Imperfect Competition and Introduction to the Theory of Employment, Keynes’s General Theory, Haberler’s Prosperity and Depression, and Hansen’s Fiscal Policy and Business Cycles. To me it was all a wonderful discovery that gave me a lifelong love of economic analysis. The two works by Joan Robinson, particularly the The Economics of Imperfect Competition, made a deep impression on me. I was fascinated by the tools of marginal analysis, and awed by the brilliance and depth of the work — most especially the ‘Digression on Rent’.

I made Joan Robinson’s personal acquaintance in September 1950, soon after my arrival in Cambridge, where I spent an academic year on an Italian research scholarship. At that time I was studying the business cycle both from a theoretical and from an empirical point of view. (During the 194950 academic year I had been in the United States, first in Chicago, then at Harvard, where my supervisor was Joseph Schumpeter, the great theorist of the cyclical development of capitalism.) In Cambridge — the old Cambridge — I asked for and obtained, as my supervisor, Dennis Robertson, author of one of the best books ever written on industrial fluctuations. From the United States — at that time under the strong theory and policy influence of Keynesian ideas (quantum mutatum ab illo!) — I had written a long letter to an economist friend of mine in Rome, where I was criticizing rather sharply certain theses worked out by Keynes in his General Theory; that letter was then published (Sylos Labini, 1949).

My first encounter with Joan Robinson was when I was a graduate student and read her Review of Economic Studies article on the concept of capital and the production function. This left me (and apparently many other people) bewildered; in my case not only because I had difficulty in following her argument, but also because due to defects in my education in economics, I was not very familiar with the concept of capital she was attacking. What I had studied of economic theory, general equilibrium theory, made me doubt that it was possible to summarize general production sets in a two-dimensional space, but I was in enough awe of those very successful theorists who asserted that it was possible, that I decided to avoid thinking about the problem too hard.

The effects of the capital theory controversy were first felt in the teaching of economics when I was a graduate student in the late 1960s. At the University of Wisconsin there were both defenders and critics of neoclassical theory. The critics did not appear to me to have any political axe to grind. They simply had an argument to present. I suffered a certain loss of faith in the logical coherence of the neoclassical theory I had learned, but the excitement generated by the debate was a great compensation.

I first met Joan in 1971, and in an important sense I met her at just the right time to be open to her influence. One of the odder characteristics of the United States in the late sixties and early seventies was the existence of younger economists (and graduate students) who, although they had become deeply distrustful of establishment policies, notably the war in Vietnam, still accepted more or less without question orthodox economic theory. Although not nearly as young as most of the people whom I have in mind, in other respects I was a case in point. I had just published what can certainly be described as a neoclassical book.

When I first met her in the fifties, I liked to think of Joan playing opposite Humphrey Bogart in a John Huston film about the Resistance. She’d have been perfect — black clothes and red stockings, a commanding presence, cool, tough, single-mindedly loyal, fiercely partisan. Strong progressive opinions, but no nonsense about accepting dogmas, Marxist or otherwise. Critical of bourgeois society and its hypocrisy. A fighter, fearless and tough as they come.

Joan was a great lady. This impression, once received, never dimmed with closer acquaintance. As a student in Oxford, when I first saw her she appeared wearing a bright green pajama suit; she was young, vigorous and beautiful; she addressed us in abstruse terms about the long-period theory of employment. I understood little but sat spellbound, and, in a sense, have remained so ever since. She had already published a great, influential book — a work which subsequently should have ensured her a Nobel prize, had she not had the two stigmas of being both a woman and a radical. The contrast between her and Edward Chamberlin is striking and highly significant. He spent the remainder of his life in self-hagiography, collecting all reviews, references and comments on his one work. Once she had published her first book, Joan more or less never mentioned it again: she charged into fresh fields, attempting to complete the Keynesian rectification of the shortcomings of Marshallian partial equilibrium.

The role of Joan Robinson in the writing of Keynes’s General Theory is amply displayed in the publication of his Collected Works, and most particularly in volumes XIII, XIV and XXIX. Even in 1931, she was a prominent figure in that small group — the Circus — which was so influential in the debates that moved Keynes from the Treatise on Money to take that revolutionary step — the development of macroeconomics as set out in the General Theory.