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

PART THREE

THE MARKET PROCESS

Some Notes on Economic Thought, 1933–1953

In commenting on the thought of an epoch immediately after its end, the commentator faces a task similar to that of the biographer of a contemporary. However intimate his acquaintance with his subject may have been, however copious the sources he can tap, sources which may no longer be available twenty or thirty years hence, he stands to lose by the lack of historical perspective. We all know that a biography written after fifty years will in many respects be different from one written soon after a man's death.

The problems of historical perspective are notoriously complex and intricate. No doubt, as time goes by author and readers gain a clearer view of their subject by being able to see it at a distance, but at the same time it becomes more and more difficult for them to appreciate the social climate, no longer their own, which prompted the actions of the men in whom they are interested.

There is of course no ready recipe for commenting on the recent past and not looking foolish in fifty years' time. But, unable as we are to forecast what future historians will have to say on our subject, we should probably not go far wrong if we—

first, endeavour to discount those events the influence of which is already visibly vanishing, i.e., clear our minds of what can already be seen to have been purely ephemeral; and

second, devote our effort primarily to discerning the major underlying trends of our epoch which will also shape the future, unless all of them are reversed or interrupted, which is unlikely.

Reprinted from South African Journal of Economics 22 (March 1954).

It goes without saying that reasons of space impose severe limitations on our endeavour. Of course there can be no question of our attempting anything even approaching a reasonably complete account of the ideas and discussions of the past twenty-one years. Nor is this all. All we can do here is to emphasize what appear to us to have been the “critical points” in the economic thought of our period. This means not merely that a good deal will have to be left out, but that the selection of these critical points for discussion in these pages will be highly subjective. The reader must bear in mind that, were somebody else to write this commentary, his selection of topics for discussion as well as his emphasis on the various topics selected, would necessarily differ from ours. In the present context this is inevitable, but it is in our view no serious sacrifice. All history is interpretation. The reader of what follows will be in a position to compare our interpretation of the thought of the period with his own, and thus to judge for himself.

As seen from the close proximity of 1954, three major events seem to characterise the economic thought of the past twenty-one years: the rise of the Keynesian economics, the evolution of various theories of mixed market forms, like monopolistic and imperfect competition, and the new developments in Welfare Economics associated with the names of Professor Hicks and Mr. Kaldor and their critics.1

Very little need be said here about the new Welfare Economics. In spite of its impressive name and the ingenuity shown by many of its protagonists, the subject matter is somewhat remote from reality. To be sure, this whole body of thought has been evolved ostensibly as a code to guide policy. But it is hard to see how in the world as it is it could ever be brought into operation. Its central concept, the “social welfare function,” is not exactly a plaything for politicians. And all policy after all is made by politicians. In reality, as every newspaper reader knows, politicians pursue power, not welfare. In fact, one eminent welfare economist has candidly admitted that “our arrangements may perhaps be more properly described as constituting a discussion of a theory of rational behaviour rather than a complete theory of the state: for we are very little concerned with what a government does in fact do in any particular case, and in no case have we considered the ethical question of what a state should do.”2

In the field of economic thought the rise of the Keynesian theory of employment and incomes was undoubtedly the most dramatic, as it was the most widely discussed event of the past twenty-one years. The products of both Keynesian and anti-Keynesian literature have by now reached mountain-size. To do justice to even a few of the problems raised is for us clearly impossible. To survey and assess the new doctrine, even were we to confine ourselves to the most hotly debated issues, would require a frame of discussion of at least the size of a book. Fortunately there is here no need for such an endeavour, as Professor Hutt will, elsewhere in this volume, deal with what is probably the most critical issue in the Keynesian doctrine, viz., the relationship between the scale of prices and the income level. But a few brief comments on the significance of the Keynesian economics as a whole will not be out of place.

If we look at it simply as a theoretical model, the Keynesian system is sound enough. It is consistent in the sense that, if we grant the premises, the conclusions will follow: the “level of incomes and employment” will be determined by the well-known determinants. The real issue is precisely whether the premises can be granted: to what extent they reflect reality. In schumpeter's words, the realism of Keynes's “vision,” not the logical consistency of his system is at issue.

It has sometimes been said that the Keynesian economics, so far from providing us with a “General Theory,” reflects in its assumptions, explicit and implicit, the conditions of the Great Depression of 1929–1933 under the influence of which Keynes wrote his book. This is at best a half truth. It does less than justice to the great architect of the Allied economic war effort to whom we all owe so much, and to the man who devoted so much penetrating thought to the problems of the post-war world. Moreover, in “How to Pay for the War” (1940) Keynes showed with his usual brilliance how the “multiplier” technique can be used to describe inflationary processes. And in general we need not doubt that conditions of full and “over-full” employment, as we found them in the war and post-war years, lend themselves to description in Keynesian terms just as much as conditions of general unemployment do.

The truth appears to be that for the Keynesian model there lies the other limit of its validity. The Keynesian economics is an economics of extreme situations: it fits the circumstances of war and post-war inflation with the universal shortage of labour and material resources just as much as it did the world of the early 1930s with almost universal unemployment and “excess capacity.” In other words, the Keynesian model fits reasonably well any world in which we find the various classes of factors of production in approximately similar conditions, and where they therefore can be treated as thought they were homogeneous. In such a world the actual heterogeneity of factors may often be disregarded with impunity. It is here, but only here, that the famous “macro-economic” method works satisfactorily.

But by the same token our model can tell us little about what we may regard as the normal situation of a progressive economy. Where there is unemployment in some industries and labour shortage in others, where shortage of equipment in some rapidly expanding sectors coincides with excess capacity in others, the macro-economic notions are of little use. In such circumstances a “point of full employment” which we could hope to reach, but not to overshoot, by applying the familiar nostrums, does not exist. The assumption of universal homogeneity breaks down. Economists have to look round for other tools.

When we now turn to the theories of mixed market forms, of monopolistic and imperfect competition, to apply there, as we did in the Keynesian case, our twin tests of internal consistency and correspondence to reality, we see a very different picture. For one thing, the singleness of analytical purpose, the unity of structural design, which are such fascinating features of the Keynesian system, are here lacking. The theories of competition were not all cast in one mould. As a result we witness Professor Chamberlin loudly disclaiming an intellectual affiliation which Mrs. Robinson protests does exist.3

On the other hand, most of the attacks made on the new theories on the grounds of lack of realism have been defeated with surprising ease. In staving off what, for a time, looked like the most dangerous of these attacks, the onslaught of the “full-cost pricing” enthusiasts,4 the defenders have all shown considerable dexterity and usually a much better understanding of the actual circumstances in which business action has to be taken, in particular in the multi-product firm, than their opponents, for all their vaunted realism, could show.5

There are, nevertheless, some ominous cracks in the doctrinal edifice. Recently both Professor Chamberlin and Mrs. Robinson found it necessary to revisit the scenes of their earlier triumphs, a visit which, at least on Mrs. Robinson's part, seems to have led to considerable heart-searching, while Mr. Harrod has now submitted a revised version of the theory of imperfect competition.6 No major structural alterations were found necessary, but there seems to be a common tendency to reassess the part of the marginal revenue curve which twenty years ago was widely regarded as the very linchpin of the new doctrines. While Professor Chamberlin dismisses it as “a piece of pure technique unrelated to the central problem,”7 Mr. Harrod bases his rejection of the doctrine of excess capacity on a distinction between long-period and short-period marginal revenue of which, according to him, only the first determines price and output under imperfect competition.

But the most interesting problems in the theory of mixed market forms arise in connection with the question whether, to what extent, and, if at all, in what sequence the various market forms can be said to succeed each other in time. In this context the “inevitability of monopoly,” or perhaps oligopoly, calls for particular attention. But these are questions to which it will be better to return after we have explored the wider issues of which they form part.

Thus far we have dealt with problems which loomed large in the discussions of the past two decades and occupied most of the literature. We must now turn to those wider issues which, though not recognised at the time, and even now perhaps barely visible, were in fact implied in, and underlay the questions which were currently discussed. But before we set out to plumb the depth of the stream of economic thought we have to deal with one issue which cannot be thus easily classified: a problem some aspects of which appeared on the surface and were widely discussed, but which had roots and ramifications that have not been laid bare. Throughout our two decades we notice a growing feeling of dissatisfaction with the traditional equilibrium methods of neoclassical economics, and a strong desire to make economic analysis “more dynamic.”

Equilibrium analysis was felt to be unrealistic. In reality, we were told, equilibria are hardly ever found. In this form, to be sure, the criticism need not be taken too seriously. No theoretical model of course can ever provide a completely adequate picture of reality. The merits of a particular model have to be judged by comparison with those of another model, actual or potential, not by comparison with “reality” which is, and must always remain, beyond our theoretical grasp. The common sense case for the equilibrium method is that if we wish to survey a constellation of diverse forces, the easiest method of doing so is to perform the mental experiment of imagining that state of affairs which would be reached when all these forces have unfolded all their implications. This is certainly much simpler than to have to go through the laborious business of describing and classifying each force separately. The method can, however, be applied only if, first, the unfolding of the forces can take place without interference from outside and, second, the mode of interaction of our forces is known and can be predicted. The first condition, usually stated in the familiar ceteris paribus terms, is of course simply a fundamental postulate of all scientific method. But the second condition raises an issue peculiar to the social sciences. Our forces after all reflect human action prompted by knowledge. The second condition therefore means that the individuals acting will during the process of interacting which leads to equilibrium, not acquire new knowledge: otherwise their actions cannot be predicted. There are many cases (arbitrage is an obvious example) in which the process of interaction is so swift that the second condition will be approximately fulfilled, but there are others where it is not. The real objection to the equilibrium method is that it must ignore the process by which men acquire and digest new knowledge about each others' needs and resources. But during our period the problem was rarely seen in this light, except by Professor Hayek whose penetrating studies of these problems broke much new ground and opened up entirely new vistas.8

The wish to give the prominent ideas of the time a more dynamic colour than that with which they made their original appearance, was strong throughout the period under review.9 In particular, Keynes's vision of the capitalistic economy bogged down in a morass of permanent unemployment clearly called for a theory of economic development to which the master himself had only contributed a few bare fragments. From Mr. Harrod's first “Essay in Dynamic Theory”10 through his later contribution “Towards a Dynamic Economics” to Mrs. Robinson's “Generalisation of the General Theory”11 there have been many attempts to “dynamise” the Keynesian doctrine. If none of these attempts has been very successful, this was, on the face of it, due to the fact that the model employed, that of an expanding economy, was somewhat too simple, just Cassel's “uniformly progressive economy” brought up to date to match the Keynesian background of the times, the “society making less than full use of its human and material resources.” But we need only probe a little beneath the surface to see that the real reason for their discomfiture was the neglect of the problems of time and knowledge.

This is not to say that the rôle of time in economics was neglected during our period. It certainly was not,12 but its implications were. Mrs. Robinson in a retrospective mood, has confessed: “In my opinion, the greatest weakness of the Economics of Imperfect Competition is one which it shares with the class of economic theory to which it belongs—the failure to deal with time.”13 As a generalisation about neoclassical economics this is hardly fair comment. Marshall after all had a good deal to say about time and its economic effects. But while time as a dimension of economic phenomena was by no means unknown to economists before 1933, its true economic significance was but tardily recognised.

Time brings change, and change brings the need for adjustment to new conditions. But a ready response to this need cannot be taken for granted. In a society based on division of labour men have to know each other's needs and resources in order to achieve their aims.

In a stationary economy, in a world in which to-day is as yesterday was and to-morrow will be like to-day, the question how men got the knowledge by which they live, offers no particular problem. We need ask it no more than we need ask, in general, how the stationary economy came to be stationary. Here it is not unreasonable to assume, as classical and neoclassical economists did, that all men have the knowledge requisite to go about their daily business.

But in a changing world the question cannot be eschewed. Here change implies that part of yesterday's knowledge is to-day no longer up to date. Men have to fight a running battle with the forces of change and ignorance, since every day that passes turns former knowledge into present ignorance. Here the economic problem begins to consist largely, if not exclusively, in “catching up” with the stream of change. He will be master who understands better, and more quickly, than the next man what recent change “means” in terms of needs and resources. Moreover, there now emerges the task of guessing accurately to-day what to-morrow's change will bring. It becomes clearly impossible to assume that new knowledge is acquired by everybody with the same speed with which conditions change, or even that, if there is a lag, it will be the same for all people. Change brings the need for adjustment to new conditions, but few people will at first understand what these new conditions are, or what they require, and the few who do profit at the expense of the others. (The typical reaction of the saving public to the secular inflation of our age provides ample illustrations for this.) Time thus entails changes in knowledge and its distribution, and thus also changes in the resources of the various individuals, a conclusion hardly congenial to equalitarians.14

This problem, which any serious attempt to bring time into economic theory has to face, has as a rule been hitherto ignored. The “dynamic models” of Messrs. Harrod and Hicks are prominent examples of this tendency, while the most ingenious attempt so far made to evade the problem openly, by assuming “perfect foresight,” was soon seen to entail too many absurdities to find ready acceptance. Yet, during our period, again and again the problem came to the surface. This fact was reflected in the growing interest in expectations.

It would of course be quite wrong to think that expectations did not exist for economists before 1933. No economist who had to deal with concrete problems could ever permit himself to forget that in an uncertain world men base their actions not on what is, but on what they think will be. It remains true nevertheless that the introduction of expectations into economic theory was one of the major events of our period. We believe that future historians of economic thought will rank it as the outstanding event of our period.

We must first briefly outline the position as it existed in 1933. As early as 1912 Schumpeter15 drew the distinction between the “entrepreneur,” the man who has the mental power to imagine that to-morrow will be different from to-day and who is able to act accordingly, and the “static individual” who lacks this power and can only adapt himself to existing circumstances. Professor Knight, by a different route, reached virtually the same conclusion, viz., that in an uncertain world uncertainty-bearing becomes a function of specialists. In Sweden, by the late 1920s, the pupils of Wicksell had encountered the problem, and Professor Myrdal wrote the first book explicitly devoted to it.16 Even in England Keynes, though probably unwittingly, had introduced expectations in 1930 when he discussed the influence of the “bullishness” and “bearishness” of the public.17

It remains true that only in the General Theory were expectations officially introduced into Anglo-Saxon economics. It is to be regretted that it was done in such a haphazard fashion. Thus the marginal efficiency of capital and liquidity preference are expectational magnitudes, but the all-important marginal propensity to consume is not, though it is hard to see why consumers' decisions should not be influenced by expectations of future prices. Moreover, in a world in which most durable, and semi-durable consumer goods, from television sets to clothes, can be bought on credit, consumers' expenditure is not limited to current income, and the consumer is in a position not really different from that of those who make investment decisions. It is difficult to avoid the impression that Keynes introduced expectations whenever it suited his argument, and left them out when it did not. Furthermore, in his Chapter 12 on “The State of Long-Term Expectation,” the famous diatribe against the Stock Exchange, it becomes painfully evident that Keynes failed to grasp the nature of the problem posed by the existence of inconsistent expectations. Instead of studying the process by which men in a market exchange knowledge with each other and thus gradually reduce the degree of inconsistency by their actions, he roundly condemned the most sensitive institution for the exchange of knowledge the market economy has ever produced!

It cannot be said that the theory of expectations has made much progress since Keynes wrote. To be sure, we now have a set of impressive-looking tools of analysis. The Hicksian “elasticity of expectations,”18 Dr. Lange's “practical range,”19 and Professor Shackle's “potential surprise function”20 all testify to the large amount of ingenuity that has been devoted to the subject during our period. If, for all the efforts made, the results have been rather meagre, the reason has to be sought in the mechanistic nature of the tools and the theories in which they are employed. None of these theories came to grapple with the central fact of a dynamic world: the human acts of interpretation by which men try to keep abreast of the changes in needs and resources. All these authors disregard the fact that man casts the material of his knowledge in the mould of expectations.

The dissatisfaction with the shortcomings of the equilibrium method mentioned earlier gave, during our period, rise to the first experiments with a new method of analysis which has come to be known as “Swedish Process Analysis.” The common sense of this new method is, briefly, that while each individual, producer or consumer, at the moment at which he makes a plan, may reasonably be expected to co-ordinate his resources in such a way as to use them to his best advantage (so that “he is in equilibrium”), these various plans need not, and probably will not, be consistent with each other. Hence, from time to time, these plans will have to be revised in the light of the new knowledge prompted by their failure. In other words, Process Analysis takes account of the fact that in a changing world men only gradually and imperfectly acquire knowledge about each other's needs and resources.

The new method made its first appearance in the Anglo-Saxon world in 1937 in Professor Lundberg's Studies in the Theory of Economic Expansion. Its rationale was lucidly explained by Professor Lindahl in 1939.21 It was used with dexterity by Professor Hicks in Parts III and IV of Value and Capital. Though it has not been without its critics,22 it was perhaps one of the most hopeful departures of our period.23 In the postwar years it proved most useful in the study of processes of inflation, open or suppressed.24

Dynamics has also invaded the theory of market forms during our period. With oligopoly this became inevitable as soon as it was realised that, for better or worse, oligopolists have to act on what they expect their rivals to do in the future. But here again, the real issue goes much deeper, and certainly passes the precincts of oligopoly. As long as competition was regarded as the principal market form, with monopoly as an exception, it was sufficient to ask what peculiar circumstances caused monopoly. But in the last two decades we have learned that most actual market forms are hybrids of monopoly and competition. The question arises now whether all these various market forms have to be regarded as alternative, though permanent types of market organisation, or as successive stages of a process. If the latter, we have to ask what is the typical sequence of this process, and also whether there is only one such type of process or whether there are several.

The problem finds its crudest expression in the neo-Marxists' assertion that “competitive Capitalism” is inevitably followed by “Monopoly Capitalism.” But even outside the orbit of Marxism the problem is important enough to merit discussion. It is one of those issues which the discussions of the last two decades have raised without giving a conclusive answer to them.

Thus far the problem was as a rule discussed in the context of Increasing Returns. It has always been known that perfect competition is incompatible with increasing returns. This fact of course provided the original starting point for the Economics of Imperfect Competition. But do increasing returns necessarily lead in the end to monopoly or oligopoly? At the end of our period we find the problem by no means solved. Mr. Harrod thinks that “increasing returns are compatible with any kind of imperfect competition, but not with perfect competition.”25 Mrs. Robinson, on the other hand, has arrived at the conclusion that: “The chief cause of monopoly (in a broad sense) is obviously competition. Firms are constantly striving to expand, and some must be more successful than others.”26

The inevitability of oligopoly is here inferred from the existence of increasing returns. To the extent to which the latter are due to “technical indivisibilities” the argument is plausible enough: the bigger firm has the advantage over the smaller firm. But, as Mr. Harrod has shown, increasing returns are also often a function of time. And time, as we saw, entails the diffusion of knowledge. It is hard to see why the knowledge acquired by one firm during the course of its expansion should for ever remain its exclusive possession, unless we assume that each firm's position at any moment is of such a unique character that no one else can learn from it anything to his profit, an assumption which would of course destroy most generalisations in our field and, in any case, make competition impossible.

It is, however, possible to feel that the whole discussion rests on a fundamental misconception of the nature of competition. Almost invariably it has been assumed that competition, perfect or otherwise, is one market form among others. In the discussion just mentioned the question at issue was merely whether it was a “stable” market form. In reality, however, as Professor Hayek put it, “competition is by its nature a dynamic process whose essential characteristics are assumed away by the assumptions underlying static analysis.”27

In other words, competition is not a market form, but the very process by which one market form evolves into another. And this process is identical with the spreading of knowledge, not only from producers to consumers, but also from producers to their rivals. The “state of perfect competition” which in the last two decades has so often been made to serve as the standard model of the text books is, if at all, conceivable only as the end-product of this process of competition. For a situation in which all consumers are completely indifferent between the products of the various sellers must be a situation in which each consumer knows already all there is to be known about all goods on the market, and has nothing further to learn from it. On the other hand, all new knowledge, technical or otherwise, is at first necessarily the possession of a few on whom it will probably confer a temporary monopoly position. Gradually, as the new knowledge is tested in the workshop as well as in the market, more and more people come to know about it, and thus the spreading knowledge of it gradually undermines the erstwhile monopoly. In the course of progress we may expect that as one “wave of knowledge” reaches the periphery of the system, becomes “common knowledge,” a new wave will emanate from somewhere else, and the process starts all over again. This, we need not doubt, is the real meaning of Schumpeter's “process of creative destruction.”

A “state of perfect competition” in the text book sense would require therefore that this process has come to an end. In other words, it denotes a state of stagnation. In reality knowledge is always unequally distributed though at every moment forces are operating to widen its distribution. There is no reason to believe that these forces cease to operate under oligopoly. In order to understand what happens in a market it is not sufficient to count the number of sellers. What one has to establish is the existing degree of differentiation of knowledge, and whether and why it has recently increased or decreased.

As a final example of the misinterpretation of market forces likely to occur when elements of the competitive process are forced into the Procrustean bed of static analysis, we may choose the notion of Product Differentiation which has occupied a prominent place in the discussions of our period. Product differentiation is usually conceived as the result of deliberate attempts by entrepreneurs to protect themselves against the forces of competition. They are supposed to do this by spreading misleading information, by advertising and other means, among consumers who have no means of obtaining better knowledge. No doubt, if we look at a market at a given moment, we may often get this impression: but it is nevertheless likely to be a misleading impression. When set against the background of the process of economic progress, the assertion that product differentiation is practised by wily producers on an unsuspecting public appears absurd. Quality improvement is one of the hallmarks of economic progress. It is clearly impossible without product differentiation. Can anybody imagine how the aeroplanes, motor-cars, typewriters, etc., of fifty years ago could have evolved into their present forms without product differentiation? The view of product differentiation here criticised thus appears to fall into the class of illegitimate generalisations.

We need not doubt that a producer will often attempt to hide a particular bit of information from the public, and for a time he may well succeed. But in this case he has to pay the penalty of not being able to utilise his own knowledge by testing it, and to improve it by utilising it. Sooner or later new waves of knowledge will sweep over him. The process of diffusion of knowledge is inherent in a society of specialists who exchange goods and services with each other. It is a concomitant of the division of labour. Even politicians cannot stop it altogether, though they may well slow it down.

NOTES

Methodological Individualism and the Market Economy

1.

For over a century and a half, from David Hume to Gustav Cassel, the defenders of the market economy were able to draw intellectual strength no less than moral comfort from the existence of a body of economic thought which supported their cause and which appeared to show that interference with the free play of market forces would, at least in the long run, do more harm than good and prove ultimately self-defeating. During this period an attitude favourable to “interventionism” almost invariably went together with an attitude critical of the doctrines of classical economics. In the Methodenstreit, Schmoller appears to have felt that what his opponents were really defending was not so much a methodological point of view as the principle of the market economy—”Das Manchestertum.“

In the course of this century all this has changed. Today economic theory, encapsulated in an artificial world of “perfect competition,” coherent plans, and instantaneous adjustments to change, has come to rest so heavily on the notion of equilibrium, embodied in a system of simultaneous equations, that the significance of its conclusions to the real world is more than dubious. In a sense it is easy to explain what has happened. The notion of equilibrium which makes very good sense when confined to individual agents, like household and firm, is less easily applied to the description of human interaction. It still has its uses when applied to a very simple type of market, such as Marshall's corn market. But “equilibrium of the industry” is a difficult concept to handle. Equilibrium of the “economic system” is a notion remote from reality, though Walras and Pareto showed its logical consistency. Equilibrium of an economic system in motion, “equilibrium growth,” borders on absurdity. What has happened is that a notion which makes good sense in the description of human plans, within the universe of action controlled by one mind, has illegitimately been extended to a sphere where it has, and can have, no meaning. A formalistic methodology which uses concepts without a proper understanding of their true meaning and natural limits is apt to defeat its own ends and bound to lead us to absurd conclusions.

Reprinted from Roads to Freedom:Essays in Honour of Friedrich A. von Hayek, ed. Erich Streissler et al. (London: Routledge & Kegun Paul, 1969), pp. 89–104.

Professors Mises and Hayek have taken a prominent part in emphasizing the implications of this unfortunate state of affairs. They have both underlined the shortcomings of the notion of equilibrium when employed out of context. Mises in 19401 described this notion as “an auxiliary makeshift employed by the logical economists as a limiting notion, the definition of a state of affairs in which there is no longer any action and the market process has come to a standstill.... A superficial analogy is spun out too long, that is all.”2

Hayek has twice dealt with the same problem. In Chapter II of The Pure Theory of Capital he pointed out why capital problems cannot be discussed within the framework of traditional stationary equilibrium theory.3 And in “The Meaning of Competition” we were told that “competition is by its nature a dynamic process whose essential characteristics are assumed away by the assumptions underlying static analysis.”4

Today the defenders of the market economy are finding themselves in a difficult position. The arsenal of economic thought, which served their fathers so well, no longer provides what they need. In fact it now often happens that what it has to supply proves more useful in the hands of their enemies than it does in their own. Their enemies will hardly fail to point out, for instance, that actual market competition, as distinct from “perfect competition,” is bound to fall short of the high ideal of “Pareto Optimality,” an equilibrium notion which occupies a prominent place in modern “welfare economics,” another spurious offshoot of contemporary economic thought.

In these circumstances upholders of the market economy are confronted with two tasks which are as unenviable as they are inevitable. They must, in the first place, be ready to turn themselves into stern and unbending critics of the economic doctrines currently in fashion, ever ready to point out the aridity of their conclusions, the unreality of their assumptions, the artificial nature of their procedure. Secondly, and even more important, they must henceforth be able to forge their own weapons. What follows in this paper is offered as a modest contribution toward the achievement of these aims.

2.

The fundamental question, i.e. in what form we should conceive of the market economy, once we have rejected the general equilibrium of the economic system, has already been answered by Mises and Hayek: The market is a process of continuous change, not a state of rest. It is also clear that what keeps this process in continuous motion is the occurrence of unexpected change as well as the inconsistency of human plans. Both are necessary conditions, since without the recurrence of the first, in a stationary world, it is likely that plans would gradually become consistent as men come to learn more and more about their environment. The recurrence of unexpected change by itself, on the other hand, would not suffice to generate a continuous process, since the elements of the system might respond to each change by a finite process of adjustment to it. We would then have an “open system” on which external change impinges in the form of “random shocks” each of which the system, possibly with variable time lags, contrives to “absorb.” But the existence of human action consciously designed to produce certain effects, prompted by expectations which may, and often do, fail, makes it impossible to look at the market process in this way. Conscious action oriented to a certain state of the market cannot possibly be conceived as a “random event.” Nor is the inconsistency of the plans of different agents, without which there can be no competition, to be regarded in this manner without doing violence to the facts. For such plans have to be drawn up and carried out with great care if they are to have a chance of success. To speak here of “random shocks” would mean to profess ignorance where we have knowledge.

We now have to consider the significance of these facts for the methodology of the social sciences. It seems to us that they provide the justification for “methodological individualism” and the “compositive method.”5

Let us retrace our steps. We have rejected the conception of the market economy as a closed system in a state of equilibrium, or at least with an inherent tendency towards it. We are unable to conceive of it as an open system on which random shocks impinge from “outside.” Mere outside shocks without the inconsistency of plans would not necessarily generate a continuous process, certainly not the market process with which we are all familiar. This requires the inconsistency of plans prompted by divergent expectations, an inevitable concomitant of human action in an uncertain world. But in these plans the future as image affects the present as action in a way which makes nonsense of the notion of “random events.” Hence, if we wish to explain the nature of the forces which propel the market process, we have to explain the nature of the relationship between action geared to the future and plans embodying a mental picture of the future.

The case for methodological individualism, for the method which seeks to explain human action in terms of plans conceived before action is actually taken, thus rests on a positive as well as a negative reason. The negative reason is, of course, that an event designed to take place in a certain situation, but not otherwise, cannot be regarded as a random event. The positive reason, on the other hand, is that in the study of human action we are able to achieve something which must for ever remain beyond the purview of the natural sciences, viz. to make events intelligible by explaining them in terms of the plans which guide action.

The scope of this principle of explanation is, of course, much wider than the area of significant action in a market economy. Needless to say, the fact that plans often fail and hardly ever are completely successful, provides no argument at all against our postulate. In fact it is only by comparing the outcome of action with the plan which guided it that we are able to judge success, another achievement which is beyond the reach of the natural sciences. The alternative principle of explanation is, of course that of “response to stimulus.” It is perhaps unnecessary to stress that the kind of entrepreneurial action mainly responsible for keeping the market process in motion, i.e. innovation and the formation and dissolution of specific capital combinations, does not lend itself to this type of explanation. Spontaneous mental action is not a “response” to anything pre-existent. Neither is it a random event. One might think otherwise of the process in the course of which, in a market economy, large numbers of producers are “learning by doing,” and gradually find out more and more efficient, and cheaper, methods of producing goods, or ways of improving the quality of their products. Here, a formalist would speak of “adding a time dimension to the production function.” But in reality this process is no more a response to stimulus than is spontaneous action in the form of innovation. The process is part and parcel of the general process of competition in the course of which even those who were unsuccessful in improving their own methods of production can benefit by adopting those of their more successful rivals. In any case, the continuous nature of the process reflects continuous acts of human will and effort, and emulation of the successful is here just as important as in the process by which innovations are diffused.

The method which explains human action in terms of plans, constituted by mental acts and linking an imagined future to an active present, has two aspects of which one is forward-looking while the other is backward-looking.

What Hayek has called the “Compositive Method”6 denotes the forward-looking aspect. Here we start with the plans of the individuals, those mental schemes in which purposes, means and obstacles are welded together into a whole and, as it were, projected on a screen. We then ask whether the plans made by different individuals are consistent with one another. If so, the conditions of success do exist, a “general equilibrium” is possible, though in reality, of course, for a large number of reasons it may never actually be reached. If not, inconsistency of plans is bound to generate further changes. In this case we have to argue from the divergence of plans to their disappointment and hence to their revision. But while we can say that disappointed expectations will lead to a revision of plans, we never can tell what new expectations the acting individual will substitute for those which were frustrated by the course of events. It may be impossible to use a durable capital good for the purpose for which it was designed. That may happen for a large number of reasons. It will then have to be turned to “second best” purpose. But what this will be depends on the new expectations of its owner at the moment of the turning decision, and about that we can say nothing.

But we can also employ the method in the reverse order. Instead of asking what are the implications of a number of plans simultaneously carried out, we can reverse the procedure and ask what constellation of plans has given rise to an existing situation. This is the real meaning of the method of Verstehen, which is also, of course, the historical method. There appears to be no reason why the theoretical social sciences, when they pursue their enquiries into the typical causes of typical social phenomena, should not make use of it.

Methodological individualism, then, in its backward-looking form, means simply that we shall not be satisfied with any type of explanation of social phenomena which does not lead us ultimately to a human plan. This entails that explanations couched in terms of so-called “behaviour variables” are not satisfactory explanations of human conduct. We have it on Hayek's own authority that the main task of the theory of capital is to explain why existing capital goods are used in the way they are. But we may also enquire how the existing capital structure came into existence, i.e. in the pursuit of which plans the existing capital resources came to assume their present form. In fact, it is hardly possible to explain present use without answering these questions. But this means that we analyse an observed phenomenon in terms of the plans in the pursuit of which it came into existence. This is the obverse of the compositive method.

Such analysis of observed phenomena in terms of pre-existent plans has nothing to do with psychology. We are here concerned with purposes, not with motives, with plans, not with the psychic processes which give rise to them, with acts of our conscious minds, not with what lies behind them. As soon as our thoughts have assumed the firm outline of a plan and we have taken the decision to carry it out over a definite period of future time, we have reached a point outside the realm of psychology, a point which we can use either as the starting point or as the final goal of our enquiry. In the former case we make use of it as the starting point of the application of the compositive method, in the latter as the final point to which we carry the method of Verstehen. In neither case are we trespassing on the domain of psychology.

3.

We must now make an attempt to look at our principle of explanation (hereafter referred to as Subjectivism)7 in the perspective of the history of economic thought.

Hayek has given it as his view that “it is probably no exaggeration to say that every important advance in economic theory during the last hundred years was a further step in the consistent application of subjectivism.”8 Naturally one thinks of marginal utility and expectations. But in exactly the same way as in writing the history of a realm an historian would not be entitled to confine himself to reciting the triumphs of its kings, soldiers, and statesmen, but must also deal with the vicissitudes they faced and the failures they suffered, the historian of thought has to record the defeats as well as the triumphs of subjectivism.

It seems to us that, from the point of view of methodology, the history of economic thought of the last 100 years has to be seen as a continuous struggle between subjectivism and its opponent (hereafter referred to as Formalism). In this long drawn-out battle success has by no means always been on the side of the subjectivists. They confronted a formidable foe with whose general character we are already familiar. The same late classical formalism which, as we saw, has brought about the alienation of modern economic theory from the market economy, is also responsible for the vicissitudes of subjectivism. Acts of the mind do not fit easily into the formal apparatus of a body of thought the main purpose of which is to produce a closed system within which it is possible to assign numerical values to a large number of magnitudes. But plans are products of mental activity which is oriented no less to an imagined future than to an experienced present. No wonder there were difficulties.

The story of the “subjective revolution” of the 1870s offers an instructive example of the vicissitudes which befell subjectivism. Its main thrust was directed against the classical theory of labour value. To the Ricardians value was a kind of economic “substance,” a property common to all economic goods. The subjectivists were able to show that value is not a property inherent in goods, but constitutes a relationship between an appraising mind and the object appraised, a manifestation of mental activity. But most of the fruits of their victory were subsequently lost when neoclassical economics contrived to “absorb” subjective utility within the framework of its formal apparatus. “Tastes” were embodied in its system as a class of “data,” a status which they came to share with resources and technical knowledge. Naturally the successful counter-revolution of neoclassical formalism raised problems of its own. Tastes can, and often will, change in an unpredictable fashion. Whenever this happens, the other elements of the system, i.e. the dependent variables, must adjust themselves accordingly. To be able to speak at all of “the system having an inherent tendency towards equilibrium,” we should therefore have to assume that the velocity with which the other elements adjust themselves to changes in tastes is always so high that no new change will occur before a full adjustment to the previous change has taken place. It is difficult to imagine such circumstances.

From our point of view it is most important to realize that formalism, by assuming all tastes to be “given,” whether in the form of utility functions or of indifference curves, is in fact evading the whole problem of how plans are made, a problem which is of crucial significance to subjectivism. The indifference curves which are imputed to consumers are in reality comprehensive lists of alternative plans to be put into operation if and when opportunity offers. In other words, what is really assumed here is that individuals never need make actual plans, because from the start they are equipped with such a large number of alternative plans that all contingencies are covered! The question how these lists of alternative plans ever came into existence is then ruled out of order as falling outside the sphere of economic questions! The whole purpose of the subjectivist revolt, which was to show that prices and quantities are the indirect results of the decision-making acts of millions of individuals who are renewing or revising their plans every day, is thus thwarted. Consumers' preferences, separated from the mental acts which daily shape and modify them, are turned into independent variables of a system in which there is no scope for planning and plan revision. Spontaneous action has been transformed into a response to stimulus. The formalists are able to claim that they have incorporated into their system the contribution of subjectivism, albeit in an emasculated form.9 Robertson's famous bon mot on Keynes's theory of interest fully applies to the formalist theory of consumers' action: “The organ which secretes it has been amputated, and yet it somehow still exists—a grin without a cat.”10

When we turn to expectations, our second instance of subjectivist success during the last 100 years, we see a very different picture.

In the first place, the problem of expectations did not make its appearance on the stage of economic thought in one thrust, as did marginal utility between 1871 and 1874, but rather by gradual infiltration. As a result it is virtually impossible to date its appearance. If we were to set the date e.g. in 1930 (“bullishness” and “bearishness” in Keynes's Treatise), we should be ignoring the fact that the problem was clearly foreshadowed in the work of Schumpeter and Knight, as well as in the early writings of Lindahl and Myrdal in the 1920s. But on the other hand, before 1930, at least in Anglo-Saxon economics, the problem was hardly recognized at all. It remains true that it began to make its impact in the 1930s.

From our point of view, the crucial significance of the emergence of expectations as a problem rests in the fact that, by contrast to what happened to utility, they have thus far proved refractory to all attempts to incorporate them into the formal apparatus of the late classical economics of our time. The reason is not far to seek. Expectations refer to processes of change. (In the stationary world of Walras-Paretian equilibrium they are in any case of no significance.) It is hard to see how they can be treated as elements of a system. They are not constants, since they are bound to change, while tastes can at least be conceived of as constants. Expectations, that is, always refer to a future point of time which we approach more closely as time passes. But neither can they be treated as variables. We cannot regard them as dependent variables since we cannot specify any mechanism of response. Different men's expectations will react differently to the occurrence of the same event. And if we regard them as independent variables, very little will be left of the rest of the system. Changes in expectations would then come to overshadow all other causes of change.11 J. Schumpeter12 and E. Lundberg13 saw this very clearly already in the 1930s and reacted with characteristic vigour.

This does not mean that, if we compress our period of decision-making to a point of time, to “market day equilibrium,” expectations could not be used and regarded as data. In this case they clearly can, but any conception of equilibrium over time, of “moving equilibrium,” is incompatible with changing expectations. It is therefore hardly surprising that most of the authors of those macro-economic growth models which have gained prominence in recent years, such as Sir Roy Harrod and Joan Robinson, have on the whole preferred to keep the problem of expectations at arm's length. Only G. L. S. Shackle has been a vigorous and indefatigable exponent and student of its implications.

It is of some interest to cast a cursory glance at Keynes's contribution in the perspective of the continuous struggle between subjectivism and formalism. Fundamentally, Keynes was a subjectivist, aware of the contrast between the variability of expectations and the determinateness required of any formal system, such as his own short-term equilibrium model.14 He mocked at long-period equilibrium (“In the long run we are all dead”), but then had to use what Marshallian tools lay most readily at hand for the purpose of giving unity to his thought. So he cast it in the mould of a short-period equilibrium system. Moreover, the General Theory was largely written as a polemic against what Keynes regarded as the neo-classical orthodoxy of his day. Since his argument relied so heavily on expectations, the polemical effect would certainly have been marred had the contrast between the rather indistinct character of the expectations he used to support his argument and the ostensible rigour of his model been too clearly revealed. In these circumstances he found himself compelled somewhat to “underplay” the significance of expectations. He introduced them where he needed them for his immediate purpose, as e.g., in the theory of investment and in liquidity preference theory, but left them out where he did not, as in multiplier theory.

But, when seen in the historical perspective which concerns us here, Keynes certainly was on the side of the subjectivists. As Professor Shackle has said so well:

The whole spirit of Keynes' book insists on the unfathomable subtlety, complexity and mutability of the influences which bear upon the decision to invest. To build a self-contained dynamic model would have been, for him, to contradict the very essence of what he was trying to say, namely, that it is uncertainty, the feeling of a helpless inability to know with assurance how a given course of action will turn out, that inhibits enterprise and the giving of full employment.15

No wonder that his successors found themselves somewhat embarrassed when they attempted to distil macro-economic growth models from his work.

Within the confines of this paper we are unable to do more than record a few episodes of the great struggle mentioned. But one such episode of recent years, which constitutes quite a remarkable success of subjectivism, should not go unrecorded.

In 1965 Sir John Hicks, who for many years had been one of the foremost exponents of formal analysis and one of its most skilful practitioners, appears to have changed sides. In an attempt to define the limits of the static method, which is of course the method of formalism, he showed that this method is incompatible with the existence of any planned action. “In statics there is no planning; mere repetition of what has been done before does not need to be planned. It is accordingly possible, in static theory, to treat the single period as a closed system, the working of which can be examined without reference to anything that goes on outside it (in the temporal sense). But this is not possible in dynamics.”16

The implications of this passage are far-reaching and intimately concern the matters pursued in this paper. Sir John has not only made clear why it is that expectations, which must transcend the single period, cannot be fitted into any model which employs the static method. He has at the same time shown within what narrow limits the instruments of formalism can be at all usefully employed. And in doing so he has opened up a vast new area for economic research, an area which is of paramount importance to us. For the world “outside the single period,” the world in which men have to act with a sense of the future and a memory of the past, the world of action and not merely of reaction, this world is none other than the realm of the market economy.

4.

At the end of our first section we promised to make a contribution to the arsenal of the market economy. The reader may well be wondering how far the methodological reflections presented in our second and third sections can be said to have furthered this cause. But what we in fact have done is to lay the ground for an attempt to cast what we hope will be new light on two notable features of the market economy which are all too often misunderstood—and not only by its critics.

The first of these is the Stock Exchange, perhaps the most characteristic of all the institutions of the market economy. In fact it is hardly an exaggeration to say that without a Stock Exchange there can be no market economy. What really distinguishes the latter from a socialist economy is not the size of the “private sector” of the economy, but the ability of the individual freely to buy and sell shares in the material resources of production. Their inability to exercise their ingenuity in this respect is perhaps the most important disability suffered by the citizens of socialist societies, however large their incomes might be, however wide the range of choice of consumption goods that may be available to them.

In the traditional view the chief function of the Stock Exchange is to serve as a channel through which savings flow before they become transformed into additions to the capital stock. Keynes taught us to regard the apportioning of the flow of savings to various investments as a function subsidiary to the constant turnover of an existing stock of securities prompted by divergent expectations. Thus, seeing the importance of expectations in asset markets, and disliking the implications of what he saw, he launched his famous diatribe on the Stock Exchange as a “casino.”

The Stock Exchange consists of a series of markets for assets, i.e., future yield streams. In each market supply and demand are brought into equality every market day. Demand and supply reflect the divergent expectations of buyers and sellers concerning future yields. Transactions take place between those whose expectations diverge from the current market price. Since as much must be bought as is sold, we may say that the equilibrium price in an asset market reflects the “balance of expectations.” As without divergence of expectations there can be no market at all, we can say that this divergence provides the substrate upon which the market price rests.

Since all assets traded on a Stock Exchange are substitutes, albeit imperfect substitutes, for one another, these markets form a “system.” And as equilibrium is attained simultaneously in each market which forms part of it, our system is free of those problems which in the Walrasian system are apt to arise when equilibrium is reached in some markets before it is attained in others.

In this way the market economy accomplishes daily a consistent, because simultaneous, valuation of all its major productive assets. The practical importance of this fact is that it makes possible, whether in the form of “take-over bids” or otherwise, the transfer of the control of material resources from pessimists to optimists, i.e. to those who believe they can make better use of them than others can. Critics of the market economy who scoff at the continuous and often violent day-to-day fluctuations of share prices, have failed to notice that an equilibrium price which rests on a balance of expectations is bound to be flexible since it must change every time the substrate of this balance changes. For precisely the same reason for which equilibrium in an asset market is reached so smoothly and speedily, it cannot last longer than one day. For expectations rest on imperfect knowledge, and not even a day can pass without a change in the mode of diffusion of knowledge.

The methodological significance of these fats, which is of interest to us here, even transcends their practical importance for the market economy, great as this is. For we are now able to see that the market process in asset markets has a more restricted function than is the case in commodity markets. In the latter, as we said above, the market process is kept in continuous motion by the occurrence of unexpected change as well as the incoherence of human plans. But in asset markets, in which equilibrium is established every day, human plans are made coherent every day. Here the lapse of time between market days serves only to diffuse new knowledge and facilitate the re-orientation of expectations. It does not have to serve to display the inconsistency of, for instance, production plans, which is what must happen between “market days” in commodity markets if such are to exist. Equilibrium in asset markets, as in the Marshallisn corn market, makes sense because it is confined to the exchange of existing stocks. Where these conditions do not exist, as in a flow market, and a fortiori in the relations between such markets, it makes no sense, and all there exists in fact is the continuous market process.

The formalists, in extending the equilibrium concept from asset markets, where it makes sense, to the Walrasian system of commodity markets, where it does not, have not only rendered a poor service to economic thought. They have rendered an even poorer service to the market economy by blurring one of its distinctive features. But in doing so, they have unwittingly provided the friends of the market economy with an instructive lesson that they must henceforth forge their own weapons.

A second feature of the market economy, with which we shall deal even more briefly here, is the fact that quantities produced and prices paid apparently depend on the distribution of wealth. We are, for instance, often told that “the Invisible Hand will only maximize total social utility provided the state intervenes so as to make the initial distribution of dollar votes ethically proper.“17 We shall refrain from comment on the ethical propriety of such statements. But it is perhaps clear that the nature of the market process, which is a continuous process that cannot be interrupted, has here been misconceived. There is, of course, no such thing as an “initial distribution” before the market process starts. The distribution of wealth in terms of asset values at any point of time is the cumulative result of the market process of the past. In the asset markets, the sources of income streams are revalued every day in accordance with the prevailing balance of expectations, giving capital gains to some, inflicting capital losses upon others. What reason is there to believe that interference with this market process is any less detrimental than interference with the production and exchange of goods and services? Those who believe that such a reason does exist (and most of our contemporary “welfare economists” do!) must assume that asset holders, like Ricardian landlords, somehow stand outside all market processes and “get rich in their sleep.” Nothing we have said about differences in the modus operandi of the market process, in asset and commodity markets respectively, can impair the validity of the simple truth that all these processes form part of an integrated whole.18

NOTES

Economics as a Social Science1

In attempting to outline the main characteristics of Economics, I shall maintain a triple thesis:

• (1) that Economics is a Science,
• (2) that it is a Social Science,
• (3) that it is an Analytical Social Science.

By saying that Economics is a Science I mean that economists endeavour to establish systematic generâlisations about observable phenomena.

The real nature of truth, the ultimate grounds of human existence, the universal criteria of the Good and the Beautiful, are the province of the philosopher, not of the scientist. For this very reason the economist, as an economist, must refrain from making value-judgements. He is concerned with the World as it is, not with the World as it ought to be. About what ought to be men will always disagree. Arguments of this kind cannot be settled by an appeal to reason and experience. For each value-judgement presupposes another value-judgement of a higher order, and thus cannot be sustained without an appeal to the ultimate grounds of human existence. Every discussion of a value problem inevitably leads to a metaphysical problem, the kind of problem the scientist has to eschew.

The object of Economic Science is Human Action, a class of observable phenomena. But before we can begin to study its characteristics we have to meet a possible objection. Can there be a science which is not “deterministic”? If not, how can a science of human action be reconciled with our consciousness of a Free Will? I believe this objection can be met, but the problem is undoubtedly a serious one. In answering it we must, for the reasons just given, keep clear of the philosophical depths of the Free Will problem. I shall merely assume that, where human affairs are concerned, Free Will is a useful hypothesis which has not hitherto been invalidated. It is, in fact, as hypothesis which is universally accepted, even by those who profess to disbelieve in it. For how otherwise could they take part in discussions without regarding themselves as mere human gramophones emitting strange but irrelevant noises, and how could they ever hope to “convince” anybody else?

Reprinted from South African Journal of Economics 18 (September 1950).

Fortunately there is a way out of our dilemma. It lies in the distinction between means and ends. In choosing ends we are free. Choice indeed is a manifestation of Free Will. But there are some ends which are incompatible, “having one's cake and eating it” is a significant example in the economic field. Here, by making one choice, we eliminate the possibility of another. Moreover, the means at our disposal are almost always limited, and this sets further limits to our choice, whilst there are few, if any, economic problems in the Land of Cockaigne. But it is important to understand that where these means are of a kind to leave us no choice, no economic problem exists either. The nature of economic activity lies in that we have some choice, to the extent to which the means at our disposal have alternative uses. In this way the freedom of choice and the determinacy imposed on us by our limited resources can be reconciled. “Economics is the Science which studies human behaviour as a relationship between ends and scarce means which have alternative uses” according to Professor Robbins's well-known definition.2

It remains to clarify the difference between technical and economic problems. Technical problems can also be stated in terms of means and ends, but they only arise where we have one end and more than one means. How to produce gold is therefore a technical problem; whether to produce it at all, or to devote our resources to other ends, is essentially an economic one. It is the possibility of choice which makes it so.

The second part of my triple thesis contends that Economics is a Social Science as distinct form the natural sciences. I hasten to stress that the criterion of distinction does not lie in the nature of the objects studied. It would be wrong to think that “Man” constitutes a field of study intrinsically separate from “Nature.”

Natural sciences of Man can and do exist: Anatomy and Physiology are obvious examples. But it is a materialistic fallacy to believe that the material nature of the objects studied determines the fields of the various sciences. The pursuit of knowledge consists in asking a series of questions, the answers to which we try to relate to each other. And it is the relationships between the problems thus raised, and sometimes solved, and not any relationship between material objects, which constitute the field of each science.3

The difference between natural and social sciences therefore lies in the nature of the questions they ask. I have defined the method of Economics in terms of means and ends. But means and ends have no measurable “material” existence. They are categories of the mind. The angle from which economists approach their problems assumes the form of a general classification into means and ends. To them all economic phenomena have, in the first place, to be interpreted as manifestations of the human mind, of decisions to seek certain ends with given means.

This means that among the observable phenomena which economists, like other scientists, attempt to relate to each other, human decisions play a most prominent part. In fact the business of the economist consists in very little else but asking what human choices have caused a given phenomenon, say a change in price, or output, or employment.

But behind these choices we must not go. Why female fashions change more rapidly than male fashions, why more people prefer the music of Irving Berlin to that of Stravinsky than the other way round, is no business of ours. The economic consequences (implications, if you like) of the choice once it is made, not the psychological causes, belong to the province of Economics.4

The difference between natural and social sciences may further be illustrated by the different part played in both by notions which originally were commonly used in both.

The concept of “Purpose,” for example, has long been discarded by the older natural sciences like physics, and has now even been expunged from biology. Yet, it remains an indispensable tool of the social sciences. Where human action is concerned, a purely behaviouristic approach can answer none of our questions. It certainly cannot explain, i.e. make intelligible, a single human act, let alone a complex series of acts of production and exchange.

The same applies to the “continuity of environment.” In the natural sciences this is an axiom. The “uniformity of Nature” has long been recognised as the logical basis of inductive inference. Natura non facit saltum. But in the social sciences where, of course, we also have to assume some continuity of environment, it is not an axiom. Its logical basis is here an assumption about purposes, and such an assumption may, in a concrete case, be falsified. Whether we turn on the wireless, post a letter, or wait for a train, in each case our conduct is guided by an implicit assumption that the purposes in the pursuit of which men yesterday operated the social environment in which we live, will continue to inspire them to-day. The probability we assign to such assumptions is evidently something entirely different from that with which we expect the moon to rise to-night. A general strike, for instance, would upset our assumptions in the former case, while Nature, broadly speaking, does not go on strike.

The idea of Causality falls into the same class of notions discarded by modern natural science, but which the social sciences must retain. The very “anthropomorphic connotations” which make the concept so suspect in the eyes of modern scientists eager to purge their terminology of anything not “observable,” make it valuable to us. After all, we are concerned with the “anthropomorphic.” For us it is not true that all we can observe are the uniformities of sequence between “events.” Our object of study is the pattern of relationships between decisions made and the practical carrying out of these decisions, the co-ordination of means and ends. For us the category “means and ends” is logically prior to any observation we make, and the phenomena we observe are to us not just “events” in themselves meaningless. For us our observations fall at once into two distinct classes, human decisions and all other social phenomena. And inasmuch as decisions have to be made before they can be carried out and have consequences, we are entitled to regard them as social causes.

That the making of decisions, the co-ordination of means and ends, takes the form of mental processes does not, of course, mean that it is not “observable.” Without the assumption that we and our fellow-men, broadly speaking, “know what we are doing,” there could not only be no social science, there could be no social life. The social scientist, we may conclude, not merely describes but explains social phenomena by reducing them to acts of the mind. We may therefore say that the “causes” of these phenomena are our choices, co-ordinated in the form of plans.

These plans may, of course, fail. Very few things ever go according to plan. In war, for example, nothing ever does, not even for the victorious side. It remains true none the less that the outcome of a war is the cumulative result of the conflicting plans of both belligerents. Could anybody describe the course of a war otherwise than in terms of the rival plans successively adopted, failures though they all were? We may therefore conclude that cause and effect as well as means and ends are fundamental categories of the social sciences.

Compared with the natural sciences the social sciences are in some respects inferior, in others superior. Their inferiority rests in their inability to predict and control. As human action is governed by choice, and choice is free, there can be no prediction of our actions. All attempts to smuggle in predictability by the large back-door labelled “the Law of large numbers” are bound to fail since human events lack the quality of “randomness” essential for this purpose. The essence of social life consists in that men get to know about each other and modify their conduct in accordance with such a knowledge. Human action, directed by knowledge gained in that process of intercommunication which is the very texture of society, can never be regarded as “random.”

Nor can there be “control” in the full scientific sense of the word. A zoologist making a breeding experiment with guineapigs need have little fear that the guinea-pigs, knowing that they are being watched, will change their breeding habits. But a politician experimenting with taxation or import control measures will very soon find that the objects of his experiments are guided in their action by inspired guesses about how long he will stay in office.

I said there can be no control in the full scientific sense of the word. I do not wish to be misunderstood. Nothing is farther from my mind than to deny the possibility of social control. The essential point is that social control requires the willing co-operation of those whose actions are to be controlled. And co-operation, like every other type of action, requires a continuous effort of the human will. Without it control cannot succeed.5

I do not wish to deny the possibility of what I would call “negative prediction,” based on inconsistency. If an economist observes a government trying, at one and the same time, to reduce the cost of living and to create an export surplus, he can predict that one of these actions will be a failure. To uncover such inconsistencies and to warn the public that what the politicians propose to do cannot be done, is, in all countries, perhaps the most important public duty of economists in our time.

With such meagre and unimpressive contributions to human progress to their credit, wherein lies the superiority of the social sciences? In the fact that they can go beyond mere description and correlation, and render the social world intelligible by reducing the phenomena of human action to that irreducible final cause: human choice. The natural sciences, after all, adopted their present-day methods after centuries spent in a vain search for ultimate causes, not out of strength but out of despair. I can see no cogent reason why we, who are in a more fortunate position, should follow their lead. We shall never know why a rose smells as it does, but I can see no insurmountable obstacle to our knowing why a perfume, say Chanel No. 5, smells as it does. In the second case we can ask the creators what they had in mind; in the first we cannot. In the social sciences the quest for final causes is a meaningful enterprise, and in this lies their superiority.

The third part of my thesis is that economics is an analytical social science, as distinct from the descriptive social sciences, or History. In a moment I shall return to the very important relationship between Economics and History. Before doing so, however, I must stress that the only successful method of the analytical social sciences is the “compositive” method.

The modus operandi of all sciences consists in analysing complex phenomena into their elements. Where not causation but correlation is the type of relationship under examination, its degree may provide the standard of comparison. But where causation is our quest, the elements of our analysis must be the causes of the phenomenon observed. Only where we can account for all the necessary and sufficient conditions can we claim to have grasped all the elements of the problem.

The logical character of the relationship between a phenomenon and its elements raises, of course, a number of crucial issues with which I need not deal here. But at least the most fundamental aspect of this problem requires some comment. The question has been asked, with what right we apply the logic of our minds to the external phenomena of nature. There are, of course, a number of answers to this question, not all of them consistent in themselves, few consistent with each other. But where Human Action is concerned, fortunately the matter is much simpler. For the Logic with which we think is also the Logic with which we act. As Professor Mises has put it: “Human Action stems from the same source as human reasoning. Action and reason are congeneric and homogeneous; they may even be called two different aspects of the same thing. That reason has the power to make clear through pure ratiocination the essential features of action is a consequence of the fact that action is an offshoot of reason.”6

The Logic of Action is essentially a Logic of Success. We start by imagining a desired state of affairs as aim of our action, and call its achievement “success.” We then proceed to eliminate all those courses of action which, in the situation as we see it, would be inconsistent with this achievement. What remains is the course of action we take.

It is not hard to guess that I shall be accused of excessive methodological rationalism. “Where,” I shall probably be asked, “is there any room left for the non-rational aspects of behaviour, for custom and habit, for the overwhelming force of passion, and the all-pervasive influence of human inertia?” The answer to this objection is that by making choice, whatever it is and however motivated, the starting point of our analysis we have already taken care of these objections. Not the psychological causes of human decisions, but their logical consequences form the subject-matter of the analytical social sciences.

The number of hours worked in a community is, as a rule, fixed by custom, but it certainly has an economic effect: it determines the magnitude of output. The remuneration of officials is everywhere determined outside the marketplace; it remains true none the less that it has an effect on the supply and demand of their services. The most ardent traditionalist who regards it as his main vacation in life to maintain an existing way of life and social order, must seek to make this order “work”; otherwise it will not survive.

I now turn to the relationship between the analytical social sciences and History. Perhaps a Professor of Economics and Economic History may crave the indulgence of his audience if, on an occasion like this, he spends a few minutes pondering the correct relationship between the two halves of his function. But something more serious is here involved. The relationship between the analytical social sciences and History encompasses, in the social field, the problems of “theory” and “fact,” or, to be more precise, the whole set of problems which concern the relationship between the formal-logical apparatus of a science and its empirical material.

At once we are confronted with a dilemma. If Economic Science and Economic History both deal with the same empirical phenomena, is not one of them superfluous? Is there anything the one could tell us which the other could not? In trying to solve this dilemma it has been said that History deals with facts, Theory with inductive generalizations from these facts. If this were so, History could not be regarded as a science, for the mere accumulation of facts is, of course, not a scientific, but a pre-scientific scientific activity. But it is readily seen that this view of the matter is quite wrong.

It is plainly impossible to write the simplest village chronicle, let alone a biography, or the history of wars and revolutions on a purely behaviouristic basis, without an attempt at causal explanation, that is to say, without referring to ends sought and means employed. This simply follows from the fact that all History deals with Human Action which cannot be rendered intelligible otherwise. It is also hardly an accident that the method I have described, the method of explaining social phenomena in terms of human decisions, possibly rival and conflicting decisions, was originally developed in the writing of History. Not in the sense that its logical character and implications were at all clearly realised, they were not, but for the simple reason that History cannot be written otherwise.

So we seem to be thrown back on our dilemma. If Theory and History both aim at causal explanation, is one of them superfluous? The answer has to be sought in the methodological principle I mentioned earlier in this address. It is not the nature of our empirical material, but the nature of the questions we ask of our material, that determines the boundaries between sciences. But do not the theorist and the historian both ask causal questions of their material? They do, but the questions of the one presuppose the answers to those of the other.

The work of the historian consists largely, though not exclusively, in applying the broad generalizations of theory to concrete facts. The relationship between the analytical social sciences and History is, broadly speaking, the same as that between pure and applied science. Whether the historian ascribes the vicissitudes of the British economy of the postwar period to “suppressed inflation” (teruggedrongen inflasie) or to Full Employment, in either case his historical judgement involves the valid existence of some general theory linking money and employment. Whether he sees the chief cause of the French Revolution in the stubborn blindness of a ruling class which failed to make concessions when there was still time to make them, or whether he sees it in the equally disastrous blindness of a rising professional class (to wit, the modern professional politician of legal extraction), neither explanation would make sense without a theory about the relationship between social stratification and political power.

In the language of modern Logic, the function of the historian is to “fill in” the descriptive signs between the logical signs, to tell us what ends by what means men in a given situation pursued. In applying the general means-ends category to concrete historical facts new problems are encountered. The applied scientist knows many woes that are undreamt of in the philosophy of pure science. The general “scarce-means-multiple-ends” scheme, for example, works well enough where we have to deal with the action of one man, as in a biography, or an organized group, say, a company, a party, a nation. It works less well where the situation we study is the result of the complex interplay of a large number of social forces. Like every other scientist the historian dislikes having to handle too many variables. And the temptation to treat as constant what one knows not to be a constant is often very strong. In the worst cases this takes the form of seeking an explanation of phenomena observed by “personifying” the forces whose very modus operandi should be explained, and ascribing means and ends to such pseudo-characters, for example, if the evolution of the modern form of the joint-stock company is “explained” as an “indispensable tool of Capitalism.” This, of course, is not History but mythology, somewhat reminiscent of the Olympian interventions in he struggles of the Homeric heroes whenever the author is at a loss to account for their actions. Explanations of events in the history of a group in terms of the Hegelian “group spirit,” or the “culture patterns” currently in anthropological fashion fall into the same class of pseudo-explanations.

Some economic historians explain almost everything that happened between 1815 and 1914 as either the result, or at least a concomitant, of the “process of industrialization.” Here we postulate a given change, or rather, a given process of continuous change, as a quasi-external “cause,” and assume that everything that happens constitutes a “response” of the social group concerned to the initial and continuous “stimulus.” One can hardly grudge a working scientist an attempt to reduce the number of his independent variables to manageable proportions, but at the same time it is possible to feel that the cases in which the results of this method will be an unqualified success will be few. It is, for example, obvious that the process of industrialization in Britain, Germany, and the United States produced, besides a number of similar, also certain highly significant dissimilar results which it is also the task of the historian to explain. All this reminds us of a passage from Tocqueville:

M. de la Fayette has said somewhere in his memoirs that the exaggerated system of general causes provides wonderful comfort to mediocre politicians. I would add that it also does this admirably for mediocre historians. It always provides them with some really good reason which, in the most difficult part of their book, will promptly get them out of all trouble, and encourages the weakness or laziness of the mind, while all the time paying homage to its profundity,

I trust these remarks will not be construed as a criticism of historical method. They are not. The historical method, as outlined above, is the only method that enables us to understand complex social phenomena. My remarks were prompted by a desire to see some of the applications of this method improved, not to see it replaced by another method. I cannot help feeling that to the working historian notions like “Industrialization” or “Colonization” offer a frame of reference which is too wide and therefore can explain very little. I am pleading for a narrowing of the frame of reference used for the explanation of certain concrete events, not for a narrowing of the scope of historical method.

In fact, there are fields of study in which the historical method if it were used more widely, might be used to great advantage. In recent years, in all countries, large quantities of statistical figures have been turned out by official and semiofficial bodies, by research institutions and ad hoc agencies. While it is always useful to know more facts, it is undeniable that from the point of view of gaining knowledge, the results have, on the whole, been rather meagre and often disappointing. The reason for this lies in the simple fact that statistical figures merely depict certain aspects of historical events, and these events, to become accessible to our minds, require an interpretation of the statistical picture. Without such an interpretation statistics have no meaning. By themselves they tell no story. From all the excellent statistical information about economic conditions in Europe which the Economic Commission for Europe has recently put at our disposal, we would yet fail to learn the most important single fact about Europe's economy in the post-war years: that in every country outside the British Isles and Scandinavia the attempts of the interventionists to impose and maintain a “controlled economy” have failed.

The so-called “Trade Cycle” offers another instance in which the historical method might be more widely employed. Almost from the time the ups and downs of modern economic life began to attract attention, economists have shown themselves eager “to explain it all,” to grasp the essence of the phenomenon by various devices. Theorists tried to catch the elusive ghost by tying him up with long deductive chains derived from a few general assumptions. But as they could hardly ever agree on which assumptions to start from, their quest failed to succeed. “Empirical” economists, their positivistic faith undimmed by logical reasoning, sat poring over innumerable series of production, price, and employment figures, waiting patiently for a moment of inspiration that would show them what was cause and what effect.

To-day it is becoming more and more clear that these ups and downs do not conform to a single invariant pattern. There is no such thing as a Trade Cycle in the sense of a periodically recurrent movement of a given number of variables.7 Unlike the celestial bodies the modern body-economic does not obey the laws of uniform rotation. Each economic crisis has to be studied as an historical event. But it is now also possible to see that the effort spent on the construction of so many theoretical models was by no means in vain. The inconsistency of the various models disappears once we realize that each historical crisis was due to a different configuration of circumstances. And to find its proper model of explanation for each crisis is essentially the task of the economic historian.

We may therefore conclude that the spheres of History and the analytical social sciences, so far from overlapping, are actually complementary. The historian endeavours to render his narrative intelligible by means of causal imputation. But if I study the causes of an event E, I can meaningfully attribute it to, say, factors A and B only if I have some prior general knowledge which makes me think that the class of events to which A and B belong may, in general, generate events of the class to which E belongs. If, on the other hand, two other factors, C and D, belong to a class of which there is no reason to believe that in any circumstances they could give rise to events of the class E, we shall refuse them causal status a priori, before even beginning the study of the facts. If we hear it suggested that a nation was ruined by the incompetence of its rulers, all we can do is turn to the facts. It is a plausible hypothesis, for in general it is possible for incompetence to have such results. But a hypothesis attributing the ruin of a nation to a lack of matrimonial virtues on the part of its rulers need not even be investigated.

The chief task of the analytical social scientist is to tell the historians what factors will not bear a causal imputation. The general analytical schemes of theory furthermore provide the historian with alternatives of explanation. But the actual choice of the alternative, the act of causal imputation itself, is very much the historian's own. It requires that specific understanding of a concrete situation, that ability to weigh each element of it in accordance with its proper significance, for which no general theory, however broadly conceived and elegantly formulated, can offer a substitute. On the other hand, all causal imputation has to depend on broad and general frames of reference describing connections between classes of events. It is the task of the analytical social sciences to provide, in the social sphere, such frames of reference and build a system out of them, for the historian and for all of us.

You may have noticed that in the later part of this address the concept of Economics became almost imperceptibly fused with that of analytical Social Science. This is as it ought to be. I trust I shall not be thought guilty of “economic Imperialism” if I claim that Economics has more nearly approached the ideal of a closed theoretical system in which all propositions are linked to each other and the number of fundamental hypotheses reduced to a bare minimum, than any other social science. This can hardly be an accident. No doubt such an achievement was easier for a science which deals with a sphere of life in which conduct has to be rational, on penalty of bankruptcy, and which can thus use the Logic of Action as the logical cement of its own edifice. But although possibly more difficult elsewhere, I do not think it is an achievement entirely beyond the power of other social sciences. After all, it is not merely in business life that failure carries extreme penalties. If Economics studies the implications of consumers' choice and business decisions, I can at least imagine a Political Sociology which applies the same method to voters' choice and political decisions. The fundamental principle that inconsistent action cannot succeed, that feasible plans must at least be free of inherent contradictions, applies wherever and whenever men strive for success.

At the beginning of this address I paid homage to my eminent predecessors. Let me, in this concluding passage, cast a glance into the future. My distinguished successor in this Chair, who fifty years hence may perhaps address a similar audience on “The Social Sciences in the Twentieth Century,” will undoubtedly have much richer material to draw upon. But I venture to doubt whether he will find it necessary to modify in their essence, to add much to or to detract from, the few logical principles of Social Science I have set before you to-day.

NOTES

Ludwing von Mises and the Market Process

1.

In the thickening gloom of our age, an age of declining standards, rampant inflation, and egalitarian ideology, it is perhaps too much to hope that the realm of economic thought alone will remain unscathed and at leat this province of the human mind escape invasion by our contemporary follies. In fact, what we find to-day is very much what one might have expected. We see a few thinkers engaged in a valiant but desperate struggle to defend and strengthen the great tradition they have inherited. The large majority of economists have to-day adopted an arid formalism as their style of thought, an approach which requires them to treat the manifestations of the human mind in house-hold and market as purely formal entities, on par with material resources. Not surprisingly, the adherents of this style of thought have come to find the mathematical language a congenial medium in which to give expression to their thoughts.

They are fond of referring to themselves as “neoclassical” economists. This label is, however, rather misleading. The classical economists, in their great day, were concerned with human action of a certain type, the forms it takes in varying circumstances and the results it is likely to produce. They took the market economy of their time as object of their thought and asked why it was what it was. Gradually they built up a formal apparatus of thought in order to deal with these problems.

The “neoclassical” economists of our time have taken over, developed and considerably refined this apparatus of thought. But in doing so they have taken the shadow of the formal apparatus for the substance of the real subject matter. It will not surprise us to learn that when confronted with real problems, such as the permanent inflation of our time, neoclassical economics has nothing to say. “Late classical formalism” appears to us a much better designation of the style of thought currently in fashion in these quarters.

Reprinted from Friedrich A. Hayek, ed., Toward Liberty: Essays in Honor of Ludwig von Mises, 2 vols. (Menlo Park, Calif.: Institute for Humane Studies, 1971), 2:38–32.

A prominent economist of this school has recently told us, “Until the econometricians have the answer for us, placing, reliance upon neoclassical economic theory is a matter of faith.” What a faith! Economics is by no means exclusively concerned with what happens, but also with what might have happened, with the alternatives of choice which presented themselves to the minds of the decision-makers. In fact, it is in terms of these alternatives alone that the decisions can be rendered intelligible, which is after all the main purpose of a social science. Statistics, as Mises has often explained, merely record what happened over a certain period of time. They cannot tell us what might have happened had circumstances been different.

Thirty years ago Mises warned us of the futility of late classical formalism. Characteristically he thrust his blade into his opponents' weakest spot. He showed the inadequacy of the main tool of the formalists, the notion of equilibrium. “They merely mark out an imaginary situation in which the market process would cease to operate. The mathematical economists disregard the whole theoretical elucidation of the market process and evasively amuse themselves with an auxiliary notion employed in its context and devoid of any sense when used outside of this context.”1 And he added, “A superficial analogy is spun out too long, that is all.”

In voicing these strictures Mises gave pointed expression to that opposition to the work of the school of Lausanne in general, and its fundamental concept, the notion of equilibrium, in particular, which has for long been a characteristic feature of the whole Austrian school. From Menger's letters to Walras to the work of Hans Mayer and Leo Illy a succession of Austrian writers have expressed their distrust of the Lausanne approach and criticised the theory of general equilibrium. Schumpeter is the obvious exception, but in the sense relevant to our problem, as in several other senses, he may be said not really to have belonged to the “inner core” of the Austrian school. Mises, by contrast, established his claim to this title by his rejection of the equilibrium concept and thus showed himself to stand firmly in the true line of the Austrian succession. But he did not confine himself to criticism of the work of the school of Lausanne. He took an important step forward. He replaced the notion of equilibrium by the concept of the Market Process. We shall have more to say later on about this fundamental concept and its significance within the structure of Mises's thought. But there is another matter to which we must turn first.

In the 30 years which have now elapsed since Mises made his attack on the late classical formalism of our age and its notion of equilibrium a certain re-orientation of modern economic thought has taken place. Less is heard to-day of what Mises called the “evenly rotating economy” (Kreislauf) as the framework of the equilibrium concept. Instead the notion of “growth equilibrium” or “steady state growth” has come to acquire a place of prominence in contemporary thought. We shall therefore have to ask ourselves whether, and how far, this metamorphosis of the notion of equilibrium has affected the validity of Mises's criticism of 30 years ago.

In this essay we set ourselves two tasks: in the first place, to examine the question whether the new notion of equilibrium growth may be regarded as exempt from the criticism of the old variety of static equilibrium which Mises has presented. In the second place, Mises's hints about the Market Process as an alternative to equilibrium as a fundamental concept will have to be worked out more fully. We shall have to ask what are the conditions of the continuous existence of such a process. we shall also have to ask, what is, within the framework of the market process as a whole, the status of those equilibrating forces which tend to produce at least partial adjustments.

2.

In this section we propose to show that the new notion of “growth equilibrium” which has come into fashion in the last quarter of a century is even more inadequate than was the older version which Mises so trenchantly criticised. Though the new variety acquired fame and came into fashion as a feature of the Harrod-Domar model of economic growth, its origin has to be sought in Cassel's work in the second decade of this century. Cassel was critical of Wicksell's work, and in particular of the latter's attempts to analyse dynamic processes in terms of concepts, such as the “natural rate of interest,” which can be given little meaning outside an unchanging world. He realised that economic processes in an industrial society subject to continuous change could not possibly be analysed with the help of such instruments of thought. But he remained enough of a Walrasian to want to retain the notion of general equilibrium and the static method. So he proposed the “uniformly progressive economy,” the model of an economy in which output of all goods and services increases at a uniform rate all over the system while relative prices and the relative marginal products of the factors of production remain unaffected. Thus our economic system can remain in a state of general equilibrium all the time while output, population and the stock of capital grow steadily. We now have equilibrium persisting in a world of steady change. The static method remains applicable to a world which is not stationary. In a sense we might say that here we have another type of an “evenly rotating economy,” only that the economic system as a whole achieves motion while it is rotating. Horrod and Domar, when they worked out their model, appear to have been unaware of Cassel's contribution.2

It is noteworthy that the protagonists of modern growth theories appear to believe that their models bear at least some resemblence to reality. Professor Solow asks, “What are the broad facts about the growth of advanced industrial economies that a well-told model must be capable of reproducing?” and, following Kaldor, then proceeds to state six “stylized facts.” The first of them is according to him: “Real output per man (or per man-hour) grows at a more or less constant rate over fairly long periods of time. There are short-run fluctuations, of course, and even changes from one quarter-century to another. But at least there is no clear systematic tendency for the rate of increase of productivity in this sense to accelerate or to slow down. If, in addition, labour input ... grows at a steady rate, so will aggregate output....” The second is stated as “the stock of real capital, crudely measured, (our italics) grows at a more or less constant rate exceeding the rate of growth of labour input.”3

That some fascinating games can be played with “macro-economic” aggregates, and the size of the capital stock in particular, is not a new discovery. When Cassel presented his model, at a time when macro-economics had not been thought of, he had to stress the need for a uniform rate of progress in all sectors. In our age this implication is conveniently forgotten together with the Cassellian original.

If the equilibrium of a stationary economy is an unsatisfactory tool of analysis for an industrial economy, growth equlibrium of the kind we described above is readily seen to be even less satisfactory. When real incomes per head increase, income recipients do not spend them in the same proportion as before. They will begin to buy some goods which previously had been entirely beyond their reach, buy more of some other goods, but less than in proportion to their higher incomes, and may actually reduce their consumption of some other goods they have come to regards as “inferior.” The pattern of relative demand will certainly change. For the pattern of relative supplies to adjust itself instantaneously we at once have to assume that producers foresaw this change correctly as well as the time pattern of the change. We also have to assume that costs are constant over the relevant ranges of output in all industries affected and that wage rates do not change, otherwise relative prices will change. Such assumptions about constant costs and wages when relative output changes must be regarded as being already somewhat unrealistic. But the degree of lack of realism inherent in such assumptions pales into insignificance when compared with that of perfect foresight on the part of the producers without which we can have no instantaneous adjustments of supply to demand. In fact it is this assumption of perfect foresight that deprives the model of growth equilibrium of any resemblance to the market processes of the real world.

Yet, without such foresight the adjustment of supply to changes in demand will certainly be delayed, and during the delay there will be disequilibrim in the markets affected. If any transaction take place during the period of disequilibrium (and, in a continuous market, how could this fail to happen?) the conditions of our moving equilibrium will be changed for the very same reasons for which Edgeworth the Walras had to introduce “re-contract” to safeguard the determinate character of their final equilibrium position. To out knowledge, however, none of the many economists who have presented to us equilibrium growth models in recent years has attached the condition of re-contract for transactions during periods of disequilibrium. They have all, of course, assumed continuous and uninterruped existance of equilibrium. It is this which, without instantaneous adjustments of supply to changes in demand, is impossible.

Similar problems arise in connection with the composition of the stock of capital. The maintenance of a constant capital-output ratio (whatever this vague notion may mean and imply) is, of course, not a sufficient condition of the maintenance of general equilibrium in a growing economic system. The actual composition of the capital stock in terms of the various capital resources must be appropriate to the composition of total output demanded. The capital stock must contain no single item which its owner would not wish to replace by a replica, if he suddently lost it by accident, otherwise the stock cannot be in equilibrium. Such changes in demand for consumer goods as we discussed above must therefore be at once accompanied by a corresponding change in the composition of the capital stock, otherwise this stock cannot retain its equilibrium composition and we confront a new source of disequilibrium. Of course, so long as we regard all capital as homogeneous the problem does not arise. As soon as we face the fact that most durable capital goods, even if not actually specific to the uses for which they were originally designed, have at least a limited range of versatility, the continuous maintenance of the equilibrium composition of the capital stock in a world in which relative demand and technology are bound to change in quite unpredictable fashion, emerges as a serious problem.

It is instructive to look at the whole problem from the point of view of the convergence of expectations. A society in which economic progress occurs is part of an uncertain world. Nobody knows the future. In a stationary world it is possible to appeal to the constancy of the “data” and the continuous recurrence of events to justify the belief that all members of such a society will sooner or later become familiar with them and their expectations will converge on the recurrent pattern of events. In an uncertain world this is impossible. Experience shows that different people will entertain widely divergent expectations. This will be so not merely because some men are, by temperament, optimists and others pessimists. Differences in knowledge are here often of fundamental importance. The diffusion of new knowledge is not a uniform and not often a continuous process. Some sources of knowledge are only available to some, but not to others, while the ability to make use of new knowledge is most unequally distributed among men.

For all these reasons expectations in an uncertain world are bound to diverge. But divergent expectations cannot all be fulfilled. Some are bound to be disappointed. The plans based upon them will fail. Some plans will be even more successful than their makers had expected. In either case the planners will not be in equilibrium over time. At the end of the period they will wish they had pursued different plans, and this will apply to those whose plans failed as well as to those whose plans succeeded better than expected. They will thus have revise their plans in the light of an unsatisfactory experience. But continuous equilibrium requires continuous success of plans. We have to conclude therefore that in an uncertain world in which expectations diverge and the plans based upon them cannot be consistent with one another the particular type of dynamic equilibrium known as “growth equilibrium” is impossible.

3.

Mises rejects the notion of equilibrium and proposes to replace it by that of the Market Process. In following him we confront a number of difficulties. Not the least of them stems from a fact of history which none of us can eschew. The ascendancy which the school of Lausanne has gained in this century has created a situation in which for most of us it has become difficult even to conceive of a world without equilibrium. It nowadays requires quite an effort to do so. So much of what we have learnt and thought seems to depend on it that without it we appear to be drifting helplessy on an uncharted sea without a possibility of taking our bearings. But the inadequacy of the Lausanne notion of general equilibrium has been established. We have to tackle the uncomfortable task of substituting for it something else, something at once more akin to reality and more congenial to praxeological thought.

Fortunately we have Mises's work to guide us in this task. In ridding our minds of the domination of the equilibrium notion the market process presents itself as a better alternative. Perhaps such a conception came more naturally to somebody who shaped his fundamental conceptions in the Vienna of the first decade of this century, the decade in which the reputation of the Austrian school was at its peak.4 No doubt the young Mises, imbibing the “pure atmosphere” of the school of Vienna, not as yet contaminated by alien particles, found himself able to conceptualize, with little effort, the essence of the market economy in the form of the market process. For us, as we explained, an effort is here required. We should make a start by looking at different meanings of the notion of equilibrium.

First of all, we have to note that what has happened to the notion of equilibrium is that the economists of Lausanne and their successors to-day have stretched the meaning of equilibrium to such an extent that a notion, in its original meaning useful and indeed indispensable, has been applied far outside the borders of its natural habitat.

The Austrians were concerned, in the first place, with the individual in household and business. There is no doubt that here equilibrium has a clear meaning and real significance. Men really aim at bringing their various actions into consistency. Here a tendency towards equilibrium is not only a necessary concept of praxeology, but also a fact of experience. It is part of the logic inherent in human action. Interindividual equilibrium, such as that on a simple market, like Böhm-Bawerk's horse market, already raises problems but still makes sense. “Equilibrium of an industry” à la Marshall is already more precarious. “Equilibrium of the economic system as a whole,” as Walras and Pareto conceived of it, is certainly open to Mises's strictures. “Growth Equilibrium,” as we have tried to show, the equilibrium of a system in motion, is simply a mis-conception.

The vice of formalism is precisely this, that various phenomena which have no substance in common are pressed into the same conceptual form and then treated as identical. Because equilibrating forces operate successfully in the individual sphere of action, we must take it for granted, so the formalists tell us, that they will also do so outside it. From Walras to Samuelson we find the same manner of reasoning, the same arbitrary assumptions, the same unwarranted conclusions.

What, then, are we to do? If, with Mises, we adopt the Market Process as our fundamental Ordnungsbegriff, how much of equilibrium can we embody in it? We suggest that we envisage a world in which millions of individuals attempt to reach their individual equilibria, but in which a general equilibrium that would embrace all of these is never reached. The Market Process derives its rationale from, and has its place in, a world in which general equilibrium is impossible. But to deny the significance of general equilibrium is not to deny the existence of equilibrating forces. It is merely to demand that we must not lose sight of the forces of disequilibrum and make a comprehensive assessment of all the forces operating in the light of our general knowledge about the formation and dissemination of human knowledge.

If, with Mises, we reject the notion of general equilibrium, but, on the other hand, do not deny the operation of equilibrating forces in markets and between markets, we naturally have to account for those disequilibrating forces which prevent equilibrium from being reached. In other words, to explain the continuous nature of the market process is the same thing as to explain the superior strength of the forces of disequilibrium.

The market process is kept in permanent motion, and equilibrating forces are being checked, by the occurrence of unexpected change and the inconsistency of human plans. Both are necessary, but neither is a sufficient condition. Without the recurrence of the first, i.e. in a stationary world, it is indeed likely that plans would gradually become consistent as men came to learn more and more about their environment including one another's plans. Without the inconsistency of plans promoted by divergent expectations, on the other hand, it is at least possible that all individuals would respond to exogenous change in such a manner that general equilibrium can really be established. A good deal would here, of course, depend on the speed of such adjustments. Where this is high, each adjustment may have been completed before the next unexpected change occurs. What however, will in reality frustrate the equilibrating forces is the divergence of expectations inevitable in an uncertain world, and its corollary, the inconsistency of plans. Such inconsistency is a permanent characteristic of a world in which unexpected change is expected to recur.

Within the general framework of the market process, prompted by the two permanent forces whose modus operandi we have just attempted to describe, equilibrating adjustments in individual markets, both price and quantity adjustments, will, of course, take place. The equilibrating forces will be found to do their work. But we can never be sure that the spill-over effects which an equilibrating adjustment in one market has on other markets will always be in an equilibrating direction. They may well go in the other direction. Equilibrium in one market may be upset when the repercussions of the equilibrating adjustments in other markets reach it. There is therefore no reason why the effects of such inter-market repercussions must always on balance be equilibrating. But our inability to assess the net result of this interplay of equilibrating forces in different markets does not amount to the discovery of another permanent force which keeps the market process in motion. It is a process within the market process.

We have never been able to understand why in the discussion on Keynes's so-called “under-employment equilibrium” some economists, opposed to Keynesian teaching, should have regarded it as either necessary or desirable to argue that in a market economy the market process, if only left unhampered, would “in the end” tend to bring about full employment. In the light of the considerations presented above such a conclusion appears unwarranted. If the outcome of the contest between equilibrating and disequilibrating forces is at best uncertain, why should it be less so in the case of the labour markets, affected as they are by a variety of factors, many of them noneconomic? If we have good reason not to believe in the generality of equilibrium, why should we want to assert that in the labour market alone equilibrium will always come about in the end? The cause of the market economy is not served by such assertions which a deeper understanding of the market process and the complex play of forces on which it rests will show to be fallacious. We have to learn to live with unemployment as with other types of disequilibrium.

4.

It may be useful to elucidate the ideas presented above on market process and equilibrium by restating them in terms of the diffusion of information, somewhat in the manner in which Leijonhufvud has recently interpreted some ideas of Keynes.

We pointed out above that a good deal always depends on the speed of the adjustments following disequilibrium. Where these are made rapidly, equilibrium may be reached before the next unexpected change occurs. Most economists agree that the market is an agent for the diffusion of information, but we may well doubt whether this can be at all regarded as a rapid process. Equilibrium theory, in order to affirm the existence of a strong tendency towards it, has to assume that correct information about equilibrium prices and quantities is readily distilled from market happenings and available to all participants. Otherwise there can be no immediate adjustment. With slow adjustments a good deal may happen in the meantime before equilibrium is reached.

In reality, of course, information will spread slowly because not all participants have the same ability to assess the informative significance of the events they observe. But even apart from this fact, which in any case prevents equal knowledge by all market participants, we have to take note of two further facts which in reality cannot but impede the diffusion of information.

Firstly, nobody can be certain whether an event he has observed constitutes a “real change” or a random fluctuation. He has to wait for confirmation and this takes time. Secondly, nobody knows for how long the information provided by a market event will remain relevant to his plans. In a changing world information which is relevant knowledge to-day may have become obsolete by to-morrow. These two facts, pulling the individual in opposite directions, account for the divergence of expectations.

We thus have to conclude that the diffusion of information does indeed form an indispensable part of the market process and by itself constitutes an equilibrating force. But it is in reality bound to be a rather slow process, likely to be hampered by the divergence of expectations and overtaken by unexpected events.

Mises, as a critic of equilibrium theory and exponent of the Austrian tradition, assumed the rôle of an innovator when he presented his conception of the Market Process as an alternative. It is, however, noteworthy how slowly and gradually the Austrian school evolved these fundamental concepts which serve to unify economic action in society.

In the Walrasian system the notion of equilibrium is employed as a formal device to unify economic action on the three levels of individual, market, and system. This unification is apparently accomplished at one stroke on all three levels. Hence the formal elegance and architectonic unity which have so fascinated many of our contemporaries. But, as we saw, poverty of content is here the price to be paid for elegance of form. While we learn something useful about what governs and unifies individual action, we merely learn a few half-truths about the forces operating in the system as a whole.

The Austrian school presents a very different picture. Here conceptualization and unification are often painfully slow. Even on the level of the individual it took half a century and was not achieved until Schönfeld's Wirtschaftsrechnung of 1924. In the development of Mises's thought as we said above, the idea of the market process was probably conceived 60 years ago, but it was not formulated until the 1930s.

But the slow progress has now brought its reward. We are now able to gain an insight into the complex nature of the forces operating, in particular between markets, which was never dreamt of in the halls of the palace on the shore of the Lake of Geneva.

Mises has provided his disciples with an instrument of thought which promises to be of superb power. In years to come it will be for them to prove their worth by handling it with care and adroitness.

NOTES

[[1]]N. Kaldor, “Welfare Propositions of Economic and Interpersonal Comparisons of Utility,” Economic Journal 49 (September 1939): 549–52; J. R. Hicks, “The Foundations of Welfare Economics,” Economic Journal 49 (December 1939): 696–712. For criticism of the Hicks-Kaldor view see I. M. D. Little, A Critique of Welfare Economics (London: Clarendon Press, 1950), especially Chapters VI and VII.

[[2]]W. J. Baumol, Welfare Economics and the Theory of the State (Cambridge: Harvard University Press, 1952), p. 140.

[[3]]“I have never been able to grasp the nature of the distinction between imperfect and monopolistic competition to which Professor Chamberlin attaches so much importance.... It appears to me that where we dealt with the same question, in our respective books, and made the same assumptions we reached the same results (errors and omissions excepted). When we dealt with different questions we naturally made different assumptions” (Joan Robinson, “Imperfect Competition Revisited,” Economic Journal 63 [September 1953]: 579n).

[[4]]See P. W. S. Andrews, Manufacturing Business, 1949.

[[5]]E. A. G. Robinson: “The Pricing of Manufactured Products,” Economic Journal 60 (December 1950): 771–80; and “The Pricing of Manufactured Products and the Case against Imperfect Competition,” Economic Journal (June 1951): 429–33; R. F. Harrod, Economic Essays (London, 1953), pp. 157–74; E. H. Chamberlin, “‘Full Cost’ and Monopolistic Competition,” Economic Journal (June 1952): 318–25.

[[6]]E. H. Chamberlin, “Monopolistic Competition Revisited,” Economica 18 (November 1951): 343–62; Joan Robinson, “Imperfect Competition Revisited,” Economic Journal (September 1953): 579–93; R. F. Harrod, “The Theory of Imperfect Competition Revised,” in Economic Essays, pp. 139–87.

[[7]]Economic Journal 62 (June 1952): 321.

[[8]]“Economics and Knowledge” and “The Use of Knowledge in Society” reprinted in Individualism and Economic Order (London: Routledge & Kegan Paul, 1949), pp. 33–56 and pp. 77–91.

[[9]]For fairly obvious reasons this feeling found its strongest expression in business cycle theory. But as Professor Schumann deals with this field elsewhere in this volume we can neglect it here.

[[10]]R. Harrod, “An Essay in Dynamic Theory,” Economic Journal 49 (March 1939): 14–33.

[[11]]Joan Robinson, The Rate of Interest and Other Essays (London: Macmillan & Co., 1952), pp. 69–142.

[[12]]See e.g. P. N. Rosenstein-Rodan, “The Role of Time in Economic Theory,” Economica 1 (February 1934) 77–97.

[[13]]“Imperfect Competition Revisited,” Economic Journal 63 (September 1953): 579–93.

[[14]]On this whole problem see Ludwig von Mises, Human Action: A Treatise on Economics (New Haven: Yale University Press, 1949), especially pp. 308–11, and pp. 580–83. This important book has so far not met with the attention it deserves.

[[15]]In the original German version of his Theory of Economic Development [English ed., Cambridge: Harvard Economic Studies Series, 1934].

[[16]]Gunnar Myrdal, Prisbildningsproblemet och föränderligheten (Uppsala: Almquist & Wiksells, 1927).

[[17]]Treatise on Money, vol. 1, Chapters 10 and 15.

[[18]]J. R. Hicks, Value and Capital (Oxford: Clarendon Press, 1939), p. 205.

[[19]]Oscar Lange, Price Flexibility and Emloyment (Bloomington, Ind.: Principia Press, 1944), p. 30.

[[20]]G. L. S. Shackle, Expectation in Economics (Cambridge: Cambridge University Press, 1949), p. 4.

[[21]]Studies in the Theory of Money and Capital (London: Allen & Unwin, 1939), Part 1.

[[22]]See for instance A. P. Lerner, Essays in Economic Analysis, pp. 215–241.

[[23]]See also Karl Bode, “Plan Analysis and Process Analysis,” American Economic Review 33 (June 1943): 348–54.

[[24]]ee Bent Hansen, A study in the Theory of inflation (London: Allen & Unwin, 1951).

[[25]]Economic Essays, p. 186.

[[26]]“Imperfect Competition Revisited,” p. 592.

[[27]]Individualism and Economic Order, p. 94.

[[1]]In the original German edition of Human Action.

[[2]]L. von Mises, Human Action (New Haven: Yale University Press, 1949), p. 352.

[[3]]F. A. Hayek, The Pure Theory of Capital (London: Routledge & Kegan Paul, 1941), p. 14.

[[4]]Individualism and Economic Order (London: Routledge & Kegan Paul, 1949), p. 94.

[[5]]F. A. Hayek, The Counter-Revolution of Science (Glencoe, III.: Free Press of Glencoe, 1955), pp. 38–39.

[[6]]The Counter-Revolution of Science, loc. cit., pp. 39, 212.

[[7]]The Counter-Revolution of Science, loc. cit., p. 38.

[[8]]Ibid., p. 31.

[[9]]The Austrians alone stubbornly resisted this trend and retained a wholesome distrust of indifference curves in particular and the whole Lausanne approach in general. Unfortunately, they never were able to show, with the cogency their case required, the incompatibility between the idea of planned action, the very core of Austrian economic thought, and an analytical model which knows no action, but only reaction.

See Hans Mayer, Der Erhenntniswert der funktionellen Preistheorien (Vienna, 1932).

See also Ludwig M. Lachmann, “Die geistesgeschichtliche Bedeutung der österreichischen Schule in der Volkswirtschaftslehre,” Zeitschrift für Nationalökonomie 26 (January 1966): 158–62.

[[10]]D. H. Robertson, Essays in Monetary Theory (London and New York, 1948), p. 25.

[[11]]This is, of course, what happened to Keynes's liquidity preference theory.

[[12]]Business Cycles, 2 vols. (New York and London, 1939), 1:140.

[[13]]Studies in the Theory of Economic Expansion (London: Allen & Unwin, 1937), p. 175.

[[14]]“There is an arresting contrast between the method and the meaning of Keynes' book. The method is the analysis of equilibrium, the endeavour to account for men's actions as a rational, calculated and logically justifiable response to circumstances which in all relevant essentials they thoroughly know. The meaning is that such rationality is in the nature of things impossible and baseless, because men confront an unknown and unknowable future.” G. L. S. Shackle, A Scheme of Economic Teory (Cambridge: Cambridge University Press, 1965), p. 44.

[[15]]Ibid., p. 98.

[[16]]John Hicks, Capital and Growth (Oxford: Oxford University Press, 1965), p. 32.

[[17]]Paul A. Samuelson, Collected Scientific Papers, 3 vols. (Cambrige, Mass.: M.I.T. Press, 1966), 2:1410. (Italics in the original.)

[[18]]“There is no distributional process apart from the production and exchange processes of the market; hence the very concept of ‘distribution’ becomes meaningless on the free market. Since ‘distribution’ is simply the result of the free exchange process, and since this process benefits all participants on the market and increases social utility, it follows directly that the ‘distributional’ results of the free market also increase social utility.” Murray N. Rothbard, “Toward a Reconstruction of Utility and Welfare Economics,” in On Freedom and Free Enterprise, ed. Mary Sennholz (New York: D. Van Nostrand, 1956), p. 251.

[[1]]An Inaugural Lecture given at the University of the Witwatersrand on 19th April, 1950. The chair was taken by the Vice-Chancellor Dr. H. R Raikes.

[[2]]Lionel Robbins, An Essay on the Nature and Significance of Economic Science (London: Macmillan & Co., 1962), p.16.

[[3]]Max Weber, Gesammelte Aufsäze zur Wissenschaftslehre (Tübingen: J. C. B. Mohr, 1922), p.166.

[[4]]This delimitation is not arbitrary. It simply follows the natural frontier of our conscious thought, which is also that of Logic. Human Action controlled by the mind has a logical structure and can thus be “understood.” Those subconscious processes, on the other hand, which precede the choice of purpose and the decision to act, lack this natural structure. They are to us “external phenomena,” essentially structure-less, just like any other event we happen to observe.

[[5]]Forty years ago it was said of a highly civilized country in Europe that there the government was so universally loved and respected, that it was quite sufficient for them to say that they did not want a certain thing to be done, and everybody would start doing it!

[[6]]L. v. Mises, Human Action (New Haven: Yale University Press, 1949), p. 39.

[[7]]So much is now more or less generally agreed. Dr. J. R. Hicks, in his recent Contribution to the Theory of the Trade Cycle (Oxford: Clarendon Press, 1950), although he still speaks of “cycles,” makes it clear that he does not mean uniform sequences of identical constellations.

“In the real world we shall not to expect to find such uniformity; and in consequence we ought not to expect that actual cycles will repeat each other at all closely. Certainly the cycles of reality do not repeat each other; they have, at the most, a family likeness. Thus, in order to explain the facts, we do not want to assume uniformity in conditions; what we want is a theory which allows variation in conditions, but still leaves us with a cycle of the same basic character.... One of the things which give different cycles their different histories (our italics L.M.L.) may thus be found in a change in the investment coefficient. It looks very likely that variations in others of the fundamental conditions may explain other varieties of cyclical experience” (108–9).

Evidently this is a task for historian.

[[1]]Human Action (New Haven: Yale University Press, 1949), p. 352.

[[2]]Theoretische Sozialökonomie (Leipzig, 1918), I. Kapitel, para. 6.

[[3]]R. M. Solow, Growth Theory: An Exposition (Oxford: Oxford University Press, 1970), p. 2.

[[4]]“These years, during which Böhm-Bawerk, Wieser and Philippovich were teaching at Vienna, were the period of the school's greatest fame.” F. A. von Hayek, “Economic Thought: The Austrian School,” in International Encyclopedia of the Social Sciences, 4:461.