The Online Library of Liberty

A project of Liberty Fund, Inc.

• Acknowledgements
• Part One: Introduction
• Walter E. Grinder, In Pursuit of the Subjective Paradigm: Walter E. Grinder
• Austrian Economics In the Present Crisis of Economic Thought
• Part Two: Setting the Stage
• The Significance of the Austrian School of Economics In the History of Ideas
• The Role of Expectations In Economics As a Social Science
• Professor Shackle On the Economic Significance of Time
• The Science of Human Action
• Model Constructions and the Market Economy
• Part Three: the Market Process
• Some Notes On Economic Thought, 1933–1953
• Methodological Individualism and the Market Economy
• Economics As a Social Science1
• Ludwing Von Mises and the Market Process
• Part Four: Problems In Macroeconomic and Capital Theory
• Complementarity and Substitution In the Theory of Capital
• Mrs. Robinson On the Accumulation of Capital
• Sir John Hicks On Capital and Growth: (review Article)
• Sir John Hicks As a Neo-austrian: (review Article)
• A Reconsideration of the Austrian Theory of Industrial Fluctuations
• Part Five: On Economic Policy
• Causes and Consequences of the Inflation of Our Time
• The Market Economy and the Distribution of Wealth
• Cultivated Growth and the Market Economy
• Appendix: Economic Writings of Ludwig M. Lachmann

Sir John Hicks on Capital and Growth

(Review Article)

1.

For thirty years or so the appearance of a new book by Sir John Hicks has been an event eagerly looked forward to by the cognoscenti. The title Capital and Growth1 combines two subjects of peculiar interest today. The theory of Capital, after several decades of neglect, in which only investment, but not changes in the stock itself, had interested economists, has of late come into its own again. But this renaissance of the theory of capital is also closely connected with the other subject: economic growth can hardly be described, and certainly not explained, without reference to the composition of the capital stock as a whole.

In recent years the literature on Growth has grown to such an enormous size that a survey, or at least a guide for the baffled readers of economic journals, has become an urgent need. Professor Hicks is not the only one who has endeavoured to supply it.2 But there are certain reasons why his proved talents seem particularly suited to this task. For many years his success in setting economic ideas in historical perspective, and in blending his own analysis with the writing of the history of ideas, has impressed his readers. Never content merely to present his own thought, he has shown an ability (alas, only too rare among contemporary economists) to set it out in a perspective in which various, apparently disconnected, aspects of known ideas, suddenly acquire new meaning and become related to one another in unsuspected ways.

Reprinted from South African Journal of Economics 34 (June 1966).

It is noteworthy that our author must be the only prominent British economist who grew up in the 1920s, but was not brought up in the Marshallian household. A Paretian and Wicksellian, rather than a Marshallian,3 he displayed a certain aloofness towards the “Keynesian Revolution,”4 a fact which permitted him to see in Keynesian theory a variant of, rather than a contradiction to, the neo-classical tenets. Starting from a fundamentally neo-classical point of view, he has been able to absorb successive waves of thought, first Keynes, then Harrod-Domar, later on Linear Theory, and to master them all with remarkable success and alacrity.

In this way he has become a prominent mediator between different strands of thought, a broker of ideas whose influence has been far greater than is often realised today. In this role he has been much helped by another Hicksian characteristic, viz. a sturdy sense of realism, an aversion to those “heroic” assumptions which may simplify analysis but will not pass muster before a critical eye. To be sure, he builds his models, and does so with skill and evident relish, but he usually manages to keep them down to earth. In this book his insistence on the heterogeneity of capital is a good case in point.

2.

The book consists of four parts of which the first two are concerned with the theory of growth as such, while the third is devoted to “Optimum Growth,” the Welfare Economics of the subject, and the fourth to the implications of growth theory for other parts of economic analysis. It ends with five Mathematical Appendices.

The first part, “Methods of Dynamic Economics,” at a superficial glance looks like a historical background to the Hicksian Growth Equilibrium model set out in Part II, a brief summary of growth theory from Adam Smith to Sir Roy Harrod. But, as is so often the case with Hicksian prose, the first impression turns out to be deceptive. The eleven chapters of the first part, which comprise almost half of the text without appendices, are a veritable seed-bed of ideas. With almost incredible terseness the author sets forth his main points on dynamic theory, drawing frequently on the history of economic thought for illustrations, but also often interspersing the historical chapters with analytical matter which is then brought to bear on features of contemporary reality.

At the end of the first chapter, after paying tribute to Cassel as the much neglected originator of the idea of steady growth, the author warns us “Growth Theory ... is no more than a particular method of Dynamic Economics. It is not claimed (it ought not to be claimed) that it is the method—that there do not remain many dynamic problems to which some other approach would be more relevant. It may indeed be questioned whether it is ‘dynamic’ enough.” (14) On the other hand, “In statics, equilibrium is fundamental; in dynamics, as we shall find, we cannot do without it; but even in statics it is treacherous, and in dynamics, unless we are very careful, it will trip us up completely.” (15)

Chapter II, “The Concept of Equilibrium,” contains the important distinction between equilibrium at a point of time and equilibrium over a period of time. (24) The former is thus defined: “The system is in equilibrium in this sense, if ‘individuals’ are reaching a preferred position, with respect to their expectations, as they are at that point.” The latter equilibrium presupposes the existence of the former equilibrium at every point of time within the period. “But for period equilibrium there is the additional condition that these expectations must be consistent with one another and with what actually happens within the period.” But consistent expectations are not the only requirement of Growth Equilibrium.

There is another requirement the need for which we realise as soon as we abandon the assumption of the homogeneity of our capital stock. “An equilibrium path ... is a path that will (and can) be followed if expectations are appropriate to it, and if the initial capital stock is appropriate to it; both conditions are necessary.” (116)

In a sense, this sentence contains the basis of all subsequent Growth Equilibrium analysis in the book. Our author insists that, whatever may be legitimate in statics, in dynamic theory we must give up the notion of a homogeneous capital stock. “It is the big thing that was wrong with classical theory. If there is just one homogeneous ‘capital’, there is nothing to do with out savings but to invest them in this ‘capital’; there can be no problem of malinvestment—or of savings going to waste.” (35) The problem of the appropriate composition of the capital stock is thus shown to be one of the fundamental problems of all dynamic theory, whether of the growth equilibrium variety or otherwise. But the reader who had hoped that the causes and consequences of malinvestment in a world of uncertainty and divergent expectations would now be explored, is sadly disappointed. Except for the last chapter, in which Professor Hicks shows that technical progress will cause capital losses on specific resources, this remains the only time that malinvestment is mentioned in the book!

Two other matters of great significance are dealt with in the first part of the book. As others have done before him, Professor Hicks finds it necessary to stress, in his chapter on Marshall's method, that our world differs from that which Marshall took for granted in that we live in a world of prices “administered” by manufacturers, “but in those days even manufactured goods usually passed along a chain of wholesalers and retailers, each of whom was likely to have some independent price-making opportunity.” (55) Again, like others before him, our author attributes the cause of this change to the virtual disappearance of the wholesale merchant and his price-setting function after 1900.5 Formerly “the initiative would come from the wholesaler or shopkeeper, who would offer higher prices in order to get the goods which, even at the higher price, he could re-sell at a profit. Similarly, when demand fell, it would be the wholesaler who would offer a lower price. The manufacturer would have to accept that price if he could get no better.” (56) Hence, while Marshall's was a world of flexible prices, even though not of “perfect competition,” ours is a “fixprice world” with prices set on a “cost plus” basis and wage rates as ultimate price determinants.

The analytical significance of this historical change lies, on the one hand, in the fact that the “Temporary Equilibrium Method” which Hicks himself, following Lindahl, used in Value and Capital in 1939, has lost much of its validity. “The fundamental weakness of the Temporary Equilibrium method is the assumption, which it is obliged to make, that the market is in equilibrium—actual demand equals desired demand, actual supply equals desired supply—even in the very short period.” (76) Hence we have to look for another method of dynamic analysis. To find it we must move nearer to Keynes and his successors who are here given credit for having understood, earlier than others, that a fixprice world requires a fixprice method of analysis. Here the reader cannot help wondering why, if we are to choose our method by the criterion of realism, we should have any reason to prefer Growth Equilibrium to Sir John's erstwhile favourite of 1939. We shall return to this point at the end of our penultimate section.

In Chapters IX and X we find another significant change of a Hicksian tenet: Sir John explicitly revokes not merely the Acceleration principle, but any “Stock Adjustment principle” for which universal validity is claimed. The revocation is announced, it is true, in almost an undertone. “It is hardly a discovery to find that we are unable to “simulate” the behaviour of intelligent business management by any simple rules.” (102) But he adds significantly: “If we find—as we do find—that mechanical principles of adjustment do not offer a good representation, we shall have gained something in the way of scepticism about the use of such principles in more ambitious undertakings. And this ... will be quite useful to us later on.” It is possible to feel, however, that in putting this scepticism to work on some of the more esoteric growth models Sir John is practising the same excessive modesty as when announcing his recantation sotto voce.

3.

In Part II our author presents his own growth model. It is of the familiar 2-sector variety and, since constant returns to scale are assumed, relative prices are determined by cost of production. The wage level is “given” and the rate of profit thus determined as a residual. The rate of growth depends on the savings rate and the supply of labour. What happens if, in the Harrodian manner, “natural” and “warranted” growth rates diverge? It is shown that the ability of the system to adjust itself to such changes via price changes depends on the existence of a difference between the capital-labour ratios in the two sectors. In the following three chapters the author shows that these results are not seriously affected if we allow for a multiplicity of known techniques and of capital goods. But we are warned that technical progress is incompatible with a given growth path. “I insist that any particular growth equilibrium path is an equilibrium with respect to a given technology; changes in technology ... must imply a shift from one equilibrium growth path to another.” (171)

Chapter XVI, “Traverse,” is perhaps the most interesting in the book, as here the notion of Growth Equilibrium is put to its crucial test. We shall return to some of the fundamental problems raised in it in our next section. It opens on a cheerful note which soon proves deceptive. “Now at last we begin to emerge from Growth Equilibrium.... It has been fertile in the generation of class-room exercises; but so far as we can yet see, they are exercises, not real problems.... They are not even hypothetical real problems.... They are shadows of real problems, dressed up in such a way that by pure logic we can find solutions for them.” (183) Nevertheless there follow several pages of formal analysis in which the conditions of a successful Traverse are examined. Suddenly we are told, “Our analysis of the Traverse, in the one-capital-case, is no more than a bogy ... it is quite misleading. An actual economy—any actual economy—does not, indeed cannot, work just like that.” (190)

It appears that price flexibility is a major condition of a successful Traverse. “An economy which insists upon making its transitions on a Fixprice basis is doing so with ‘one hand tied behind its back’.” (196) But then there arises the question how, in the transition from old to new equilibrium path, the right new price system is to be found. This “cannot be an easy matter,” yet on it the success of the whole Traverse depends, since choice of technique and appropriate composition of the capital stock depend on relative prices. At the end of the chapter we find our author throwing up his hands in despair: “In an actual economic situation, all these problems arise at once, while (because of the advance in technology) the equilibrium at which the economy is aiming is continually shifting. No wonder that there is a problem of business management!” (197)

In Part III, “Optimum Growth,” Sir John turns with an audible sigh of relief from Positive to Welfare Economics, from the market place to the Turnpike. “The central problem of dynamic Optimum theory is the planning problem. Given an initial endowment of capital, embodied in particular capital goods ... what is the plan of production, in present and future, which will enable some given aim to be reached in the most efficient manner?” (203) But to maximize the rate of growth over a period may mean either of two things: we may either try to maximize the flow of consumable outputs during it, or maximize the size of the terminal capital stock. The Turnpike Theory, which is considered first, “is concerned with an optimization problem of the second type.” In Chapter XVIII the famous Neumann model which seeks to establish the conditions of continuous optimum growth (without consumption!) is set out in lucid language. In the next our author turns to the Turnpike Theorem itself. The problem here is: What is the optimum path to be followed by an economy which starts with a capital stock which is not appropriate to the balanced growth Neumann path; in what circumstances would it be better to discard the surplus parts of the stock for the sake of obtaining a balanced composition? Professor Hicks shows that it is largely a matter of time. Only over long periods would the advantages of balanced growth necessarily outweigh the capital losses from discarding surplus capital.

In the next three chapters he turns to the alternative type of Optimum theory which is concerned with a stream of consumption outputs. The argument here follows the line familiar from the second part of Value and Capital. With a given rate of interest, constant over time, there is an “intertemporal production frontier” in the sense that there is a limit to the substitution of future for present outputs. This frontier determines the optimum shape of the consumption output stream. At the end of Part III the author criticises Mr. Kaldor's Technical Progress function on the grounds that, to a large extent, technical progress stems from non-economic causes, such as scientific discovery. He therefore rejects a model “which would bring too much of the phenomenon into the strait-jacket of its ‘equilibrium.’” (276)

The last part contains what apparently are Hicksian afterthroughts on matters of contemporary interest. In Chapter XXIII money is introduced and liquidity preference comes up for review. An extension of the concept of stock equilibrium to assets in general enables us to “generalize the conception of demand for money, and assert its equilibrium in the form of saying that the whole system of debits and credits must be in equilibrium.” (281) In Keynes's theory “the rate of interest on long-term bonds is taken to stand for the whole gamut of rates and yields, on securities of all kinds, that are established on the market. As soon as one begins to ask questions about the structure of these rates, it becomes apparent that the choice between money and bonds is only one of the many possible choices between forms of asset-holding into which similar considerations of liquidity enter.” (283) From these considerations the following picture emerges: “There is a maximum to all rates of interest, set by the expected rate of return on real investment (I simplify by the assumption that there is just one rate of return); there is a minimum set by the rate of interest paid by the bank.... All other rates of interest (those paid by firms to savers, and those paid by firms to the bank) must lie, in equilibrium, between these limits. Where they will lie will be determined by a balance of liquidity considerations in the balance-sheets of lenders and borrowers respectively.” (286) Within this gap there is a place for financial intermediaries. “The financial intermediary can prosper if it can make use of specialized knowledge about the prospects of particular kinds of real investment; so that it can make advances to firms ... which the bank would not know were sound investments; and if it can acquire resources which enable it to make these financial investments at a less loss of liquidity than they would entail upon the private saver.” But while such action will reduce the gap, it can never close it altogether, a fact which has certain obvious implications for monetary policy in a Radcliffian world.

In the last chapter, “The Production Function,” Professor Hicks examines Mrs. Robinson's famous criticism of this elusive notion.6 Here at last technical progress is introduced, though it occurs discontinuously. “There are inventions (let us say) in 1900, 1910, 1920; in 1909 and 1919 the economy has settled into a stationary state.” (295)

Technical progress requires a transmutation of the capital stock. “Can one treat the supply of capital as fixed, when capital has been transmuted according to our particular rule? The answer is that one can.” (297) In fact “so long as we are only concerned with the comparison of equilibrium positions, the production function (or a production function) gets through.” (298) But our mentor adds significantly: “How much use it is, when it has to be put into this sophisticated form, may indeed be questioned.” He then admits “that the rate of profit on new investment is raised, while the profit that is earned on past investment may be lowered.” (301) In other words, technical progress entails capital losses on specific resources.

This is certainly a matter of great importance. It is to be regretted that it is only mentioned in the concluding pages of the book. After all, technical progress is not the only cause of capital losses and gains. Any disappointment of expectations concerning the use of specific resources has the same effect. Why, the reader cannot help asking himself, did not Sir John tell us that before? Would it not have been useful, at least on the Traverse, to know that the transmutation of the capital stock will be affected by such losses and gains? Should we not also have been told that, together with the classical notion of a homogeneous capital stock, we must abandon the corresponding notion of the uniform rate of profit?

4.

We return to the Traverse. Chapter XVI is, in a sense, the pivot of the book. It is here that we have to decide whether the notion of Growth Equilibrium is a tenable conception. The problem is posed early on in the chapter: “But let us now suppose that the Harrod difficulty has been got over: that a suitable change in the overall propensity to save, for whatever reason, has occurred—will that be the end of the trouble?” (185) Our author has told us as early as on page 17 that, if the equilibrium assumption is to be justified, we must be able to assert the existence of a tendency to equilibrium, and that it must be a strong tendency. Can we assert this for the Traverse from an old growth path to a new?

The problem of the Traverse consists essentially in the need for a time interval to elapse before the new equilibrium path is reached, because the transmutation of the capital stock, the change of its mode of composition from that appropriate to the old to that appropriate to the new conditions, takes time. But if any of our conditions of equilibrium, which include expectations and wealth distribution, changes during the interval, the final equilibrium will be modified. This is an old and familiar problem which Edgeworth and Walras saw clearly and, within their stationary framework, attempted to solve by means of “recontract.” Sir John spurns these “artificial arrangements.” (54) But how does he tackle the problem?

Recontract is out of the question and a suspension of all business dealings during the Traverse hardly feasible. The transmutation of the stock obviously requires firm commitments. We are thus driven to the conclusion that, so far from being able to assert a tendency to it, we do not even know what the new equilibrium will be like until we get there—if ever we do. Nor are we entitled to speak of a transmutation of the stock since we are unable to specify the terminus ad quem beforehand. To speak of an “adjustment to new conditions” is positively misleading when we do not know what they are.

How does our author avoid these conclusions? He tells us that, when prices have to change, “a corresponding Fixprice policy would presumably imply that prices are adapted at once (or sought to be adapted at once) to the new equilibrium.” (196) Nothing is said here about what would happen with flexible prices, but here, too, the system would evidently have to adapt itself to the new set of equilibrium prices at once, if malinvestment and the adoption of disequilibrium techniques are to be avoided. But this could only happen miracle and hardly permits us to assert a strong tendency to equilibrium.

What lessons are we to draw from this disconcerting experience?

In the first place, we must realise that our discomfiture is due to a misguided attempt to use the equilibrium concept in fields far away from its natural habitat. With the household and the firm equilibrium makes very good sense as here it is something actually aimed at. Interindividual equilibrium already raises issues concerning mutual knowledge which have never been properly appreciated or fully discussed. But in the Marshallian type of commodity market with flexible prices it still has a clear meaning. To extend the concept to the economic system as a whole was a bold venture, but Walras and Pareto showed that, in a stationary state, it could still be done. But to extend it even further, to an economic system in motion, would appear to lie beyond the range of the feasible.

Secondly, therefore, we must consider the possibility of a retreat to a more congenial terrain. Two positions can now be seen to have become untenable.

On the one hand, once we acknowledge, with our author, the inadequacy of all mechanical rules about human reaction to change, we also have to acknowledge the autonomy of expectations at every point of time, because this autonomy is the true cause of that inadequacy. But with this all possibility of an equilibrium over time, based upon convergent expectations, vanishes. For real expectations always diverge. This simple fact appears to destroy the, even theoretical, possibility of a determinable time path of economic processes. All this, however, does not invalidate the possibility of equilibrium at a point of time, an equilibrium in which each price reflects a balance of contemporary expectations.

On the other hand, there can be no such thing as a dynamic macro-economic equilibrium. For outside the stationary state there is, in general, owing to the ubiquity of “lags,” no market mechanism to bring the divergent expectations of all individuals within the same economic system into simultaneous consistency with each other. Nor is there any reason why the quantities of the various capital goods held in different sectors should necessarily be such as to earn their owners an actual, let alone expected, uniform rate of profit.

We are thus forced back to a micro-economic version of the Temporary Economic Equilibrium at which Professor Hicks and Lindahl tried their hands in 1939. We have to assume a market, an intemporal market which of course permits of forward transactions, on which individuals express their expectations, with a resulting equilibrium price reflecting a balance of such expectations.

This may seem a poor “optimum” for equilibrium analysis. But we may draw some comfort from at least two qualifications (there may be more) which we may permit ourselves to make to the rule about the necessary micro-economic character of our markets.

In the first place, there is, in a market economy, a Stock Exchange, a market for future yield streams, in which expectations are brought into consistency every day and a price reflecting the balance of such expectations is struck. And since the Stock Exchange is also, in every reasonably developed economy, the central market for existing capital goods, or titles to them, we can say that expectations pertaining to the whole economy are here coordinated without a necessary lag. In fact, if the classical notion of a uniform rate of profit, the corollary of the assumption of capital homogeneity, is to retain any significance at all in the real world, it is only on the Stock Exchange, where a uniform rate of yield is produced every day by the price changes of existing assets, that we can really speak of it.

Secondly, once we recognise, with our author, the heterogeneity of all capital, we must also recognise that existing capital combinations have to be dissolved from time to time, as expectations change. Existing capital combinations will thus have to be “re-shuffled,” at intervals which may, but need not, coincide with those between our “market days,” by the discarding of some and the purchase of other existing capital goods, such as buildings, equipment, ships, etc. This secondhand market for certain kinds of capital goods provides another link between various sectors of the economy. But here of course there will be lags.

Lastly, we should remember that equilibrium analysis, and indeed all formal analysis couched in terms of functional relationships, is neither the beginning nor the end of economic theory. When confronted with a disequilibrium situation, we certainly have to assume that each individual seeks to attain a (flow and stock) equilibrium. But these individual equilibria may not be compatible with one another and therefore be unattainable. Economists will have to learn to live with, and give an intelligible account of, circumstances which have no determinate outcome.

5.

In the opening section of this review article we described Sir John Hicks as a great mediator of economic thought, a most successful broker of ideas. In reality of course there is no broker, however successful he may be, to whom it does not happen, from time to time, that a deal falls through. Similarly, we find at least one conflict inherent in modern economic thought which our mediator has been unable to appease. We shall hardly be surprised that it first comes to our notice in the historical chapters of the first part, and even less that it fully comes to the surface in Chapter XVI. The question at issue is that of the compatibility of subjective attitudes (tastes, expectations) with the requirements of modern formal analysis in the shape of models. The elements of our models, parameters and variables, must be, at least in principle, objective and measurable entities. But are subjective attitudes?

This problem has existed, in one form or another, at least since the “marginal revolution” of the 1870s in which human preferences were acknowledged to be the ultimate basis on which the economic edifice rests. To trace it in the work of the major neo-classical writers would be a fascinating task. All we can do here is to make a few comments on the way in which it affects the present work.

Quantifiability is not, as has often been thought, the root of the matter. The outcome of the long discussion on cardinal versus ordinal utility showed that tastes qualify for inclusion in our models provided they can be ordered; it is unnecessary for them to be quantifiable in any cardinal sense. There seems to be no reason why the same should not apply to expectations.

The root of the matter is the autonomy of the human mind: men can and will change their tastes and expectations for no objectively ascertainable reason. Pareto saw this problem, as he saw so many others, far more clearly than most of his contemporaries. He insisted that the individual, having once recorded his preferences for us, “having left us this photograph of his tastes,” as he put it, must disappear from the analytical scene and worry us no further with the unpredictable acts of his mind.7 Whether he realised equally clearly that, by making this postulate, he also limited the validity of his whole system to the conditions of a stationary state, in which alone today's photographs will still be valid tomorrow, it is hard to say. But we may safely assume that he would have been willing to pay that price.

But around 1930, just about the time when our author joined the staff of the London School of Economics, expectations arrived on the scene. And expectations, since in a stationary state they are in any case without significance, cannot be disposed of in the Paretian fashion. The assumption of their continuous convergence, made in all the familiar growth models, is simply an attempt to sterilize them, as Professor Hicks sees clearly.

While constant tastes over a period of time are at least conceivable, expectations cannot remain constant as soon as they diverge, since some of them must turn out to be wrong sooner or later, hence be revised, though we can say very little about the mode of their revision. While therefore expectations cannot be constants, we must not treat them as variables either. They are clearly not dependent variables as they do not “depend” on any observable events. But if we try to treat them as “exogenous” data, we soon find that they will “take over” and “swallow up” most of our other data. This is the real lesson of the story of the Traverse. Divergent expectations, prompting transactions at non-equilibrium prices, will themselves affect the composition of the capital stock as well as the interindividual distribution of resources.

We must therefore conclude that expectations, and other subjective elements, constitute an alien body within the organism of formal model analysis. The conflict remains unresolved. Marshall was uneasily aware of it. Pareto saw it, drew his sword and cut the Gordian knot, but, alas, knew nothing of expectations. Our mediator, for once, has been unable to mediate in a conflict of the existence of which he is clearly aware. This of course is hardly his fault. Sir John Hicks has failed to do what cannot be done. It remains a tribute to the qualities of this remarkable book that for one brief moment, in Chapter XVI, a reader could bring himself to imagine that he might do it.

NOTES

[[1]]John Hicks, Capital and Growth (Oxford: Clarendon Press, 1965).

[[2]]See, e.g., F. H. Hahn and R. C. O. Matthews, “The Theory of Economic Growth: A Survey,” Economic Journal 74 (December 1964): 799–902.

[[3]]“We were such ‘good Europeans’ in London that it was Cambridge that seemed ‘foreign.’” Hicks, The Theory of Wages, 2d ed. (London: Macmillan & Co., 1963), p. 306.

[[4]]It is true that in the autobiographical sketch added to the 2d edition of Theory of Wages (1963) he says that in 1936 “I was (I think I may say) an almost whole-hearted Keynesian.” (310)

Perhaps we should not take the author of the famous “bootstraps” argument at his word. Perhaps, as often happens, the enthusiasm of 1936 soon turned into the skepticism of 1939.

[[5]]Cf. L. M. Lachmann, Capital and Its Structure (London: London School of Economics and Political Science, 1956), p. 64.

[[6]]Joan Robinson, “The Production Function and the Theory of Capital,” Review of Economic Studies 21(1953–4): 81–106.

[[7]]“L'individu peut disparaitre pourvu qu'il nous laisse cette photographie de ses goῦts.” V. Pareto, Manuel d'Economie Politique, 2d ed. (Paris, 1927), p.170.