L.S.E. Essays on Cost

Edited by: Buchanan, James M. and George F. Thirlby
(1919- )
Display paragraphs in this book containing:
First Pub. Date
New York: New York University Press
Pub. Date
Collected essays, various authors, 1934-1973. First published as a collection 1973 for the London School of Economics. Includes essays by Ronald H. Coase, Friedrich A. Hayek, Lionel Robbins, and more.

Introduction: L.S.E. cost theory in retrospect


In his paper, 'Economics and Knowledge', included in this volume, Hayek scarcely mentioned 'cost'. Nonetheless he provides indirectly the strongest argument for attempting, through the publication of this collection of essays, to focus the attention of modern economists on the elementary meaning of cost. Hayek emphasized the differences, in principle, between the equilibrium position attained by a single rational decision-maker in his own behavioural adjustments, given his preference function and the constraints that he confronts, and the equilibrium potentially attainable through the interaction of many persons. To Hayek the latter 'is not an equilibrium in the special sense in which equilibrium is regarded as a sort of optimum position'.


Despite Hayek's warning, since the 1930s, when his essay along with some of the others in this collection was written and when the L.S.E. tradition in cost theory was developed, economists have increasingly analysed equilibrium states in terms of their optimality or non-optimality properties, defined by criteria for maximizing some objective function. It is somewhat paradoxical that Robbins, whose contribution to London cost theory cannot be questioned, should also have been at least partially responsible for the drift of modern economic theory towards the mathematics of applied maximization, variously elaborated, and away from the analysis of exchange processes. In The Nature and Significance of Economic Science,*1 Robbins supplied the methodological paradigm within which modern micro-economics has been developed. Elementary textbooks throughout the world soon came to define 'economics' in terms of 'the economic problem', the allocation of scarce resources among alternative ends. So defined, the 'problem' faced by the individual on the desert island, the Crusoe so familiar to us all, is, at base, quite similar to that faced by the society or the community of persons. The paradigm was somewhat differently put, but with the same effect, by Paul A. Samuelson in his influential Foundations of Economic Analysis, when he stated:

They [meaningful theorems in diverse fields of economic affairs] proceed almost wholly from two types of very general hypotheses. The first is that conditions of equilibrium are equivalent to the maximization (minimization) of some magnitude.*2 [Italics supplied.]

The increasing conceptual quantification in economic theory was almost necessarily accompanied by increasing conceptual 'objectification'. Once the magnitude to be maximized is symbolically defined, attention is quite naturally diverted to the manipulation of the symbols and away from the initial leap into presumed objectivity itself. The increasing conceptual quantification need not have introduced confusion save for the simultaneous developments in theoretical welfare economics. Within what Hayek called the 'Pure Logic of Choice', the formal theory of utility maximization, mathematical rigour has offered aesthetic satisfaction to the sophisticated without loss of explanatory potential. More importantly, the increasingly elegant and formalistic content of general-equilibrium theory, and notably its emphasis on existence proofs and stability conditions, yields pleasure to the talented, criteria to the critical, and convictions to some who have remained unconvinced about the overall efficacy of market order.


So long as the object for discussion, and for theorizing, is either the individual decision-maker or the interactions of separate decision-makers in markets, no harm is done and perhaps some good is added by conceptual objectification. Confusion arises only when the properties of equilibrium, as defined for markets, are transferred as criteria of optimization in non-market or political settings. It is here that the critical distinction between the equilibrium of the single decision-maker and that attained through market interaction, the distinction stressed by Hayek, is absolutely essential to forestall ambiguity and analytical error. The theory of social interaction, of the mutual adjustment among the plans of separate human beings, is different in kind from the theory of planning, the maximization of some objective function by a conceptualized omniscient being. The latter is equivalent, in all respects, to the problems faced by Crusoe or by any individual decision-taker. But this is not the theory of markets, and it is artificial and basically false thinking that makes it out to be. There are properties or characteristics of equilibria in markets that seem superficially to be equivalent to those attainable by the idealized optimization carried out by the planner. But shadow prices are not market prices, and the opportunity costs that inform market decisions are not those that inform the choices of even the omniscient planner. These appear to be identical only because of the false objectification of the magnitudes in question.


This is what the great debate on socialist planning in the 1930s was all about, comment to the contrary notwithstanding. And modern economic theorists measure their own confusion by the degree to which they accept the alleged Lange victory over Mises, quite apart from the empirical record since established. The central issue in this debate should not have been the possibility or impossibility of socialist calculation. All the participants were wrong in concentrating on this. The difference in data confronted by decision-makers in different institutional settings is quite sufficient to prove that the properties of market equilibrium cannot in the nature of things be duplicated under non-market institutional structures. This is not of course to say that 'efficiency', defined in a different but legitimate planning sense, cannot be defined in an ideal-type socialist model. Of course it can. But it is a wholly different 'efficiency' framework that is involved here, informed by the marginal-value estimates of the planner and not by the participants in markets.


I think that it is legitimate to trace the sources of error to fundamental misconceptions in the theory of opportunity cost, misconceptions that the London (and Austrian) scholars were attempting to clarify, and which later I tried similarly to rectify with my little book, Cost and Choice, in 1969.*3 Unfortunately neither the London contribution nor my book seems to have exerted much effect on mainstream thinking in economic theory.


But I am getting ahead of my story. As I noted above, the increasing conceptual quantification, and objectification, of economic theory need not have sown confusion without the accompaniment of developments in theoretical welfare economics. Precisely at the time that methodologists were effective in formalizing economic theory within a more rigorous maximization framework, interest in 'market failure' rather than 'market success' was at its peak, and, with this, interest in the extension of economic theory to socialist organization became widespread. The Robbins definition of the allocation problem, with its implied emphasis on the universality of scarcity, supported such an extension. It was predictable that economists, trained professionally to analyse market equilibria, and increasingly adept at formalizing the maximization paradigm, should begin to discuss planning problems and solutions as if these required the same set of tools as those applicable to market phenomena. In retrospect it seems singularly unfortunate that the institutionalists should have lost favour precisely when their emphasis on and expertise in the functioning of organizational-institutional structures, and the impact of differing strutures on behaviour of decision-makers, might have, with some intrusion of analysis, yielded their highest marginal product in effective critical scholarship. Instead the mathematically sophisticated analytics of such scholars as Hotelling were allowed to go unchallenged despite their vulnerability in this most fundamental sense. And young economists everywhere learnt to appreciate the beauty of the mathematical models of what they called 'an economy'. Theirs was not the role of sceptic, and to question quantification and objectification itself quickly came to be the mark of eccentricity rather than excellence.


Is it any wonder that, in the idealized fully quantifiable and fully objectifiable 'economy' that commanded all attention, the market itself should come to be regarded as a 'mechanism', as an 'analogue computing device', to be legitimately treated as one among several alternative means of allocating resources, to be evaluated comparatively in terms of some criteria of accomplishment? And so it should be in such a world.


The quest for objectivity is eternal and perhaps praiseworthy, but what has modern scholarship to offer where the classical economists tried and failed? There seems little harm in speculating about the properties of an economy whose only scarce resource is a homogeneous glob of something (putty clay or little Abner's schmoos) that may be instantly convertible into any one of a large number of final goods upon which consumers place value. In this setting the cost of any one good becomes the displaced physical alternative, measurable separately in any one of the other n goods potentially available from the single homogeneous source. If a unit of good X uses up twice as much of the scarce resource as a unit of good Y, the cost of X is properly defined as 2Y, and the cost of Y as one-half X. In such a model it is meaningful to consider the planner's problem of maximizing output, defined in values or prices of goods, from the single scarce input. The norms of theoretical welfare economics can be applied directly to this purpose. The omniscient planner can solve his maximization problem quite simply by setting the prices of goods at their relative marginal costs, arbitrarily choosing one good as numeraire. As the final consumers adjust quantities demanded to the announced set of prices, the value of total output, denominated in the numeraire, will be maximized.


Nor need we limit analysis of such a conjectural economy to the planner's problems. As an alternative speculative exercise we may suppose that our homogeneous glob of scarce resource (putty clay or schmoos) is initially and arbitrarily parcelled out among persons under a private-property-rights arrangement. By assumption, the individual owners are completely indifferent as to just what set of final goods their own assigned input becomes in the transformation. These owners are motivated solely by their own desire for final goods, command over which is measured by income, denominated in some commonly agreed numeraire good. The only difference between this model and the one described earlier is that this one 'works on its own', once private-property-rights are defined and protected. The scarce resource will be allocated among uses; final goods will be 'produced'; prices will be set. The 'market' equilibrium that emerges will in this case be equivalent in all respects to the solution of the maximization problem posed for the planner in the earlier model. Prices will equal marginal costs, not because some hidden planner has now drawn on the norms of welfare economics, but because this equality is descriptive of the end of the trading process. If this equality is not satisfied, further gains from trade would be possible; potentially realizable surplus would remain unexploited. Not only can we deduce the equivalence in results between these two models on some a priori basis. We could also observe such equivalence in an objectively verifiable sense.


I do not think it a caricature to describe modern economic theory as being grounded on the two conjectural models that I have briefly sketched, and on the equivalence between their 'equilibrium-optimality' properties. Viewed in this simplistic perspective, however, the models paradoxically suggest that economic theory has advanced little, if at all, over that advanced almost two centuries earlier by the classical economists. In one respect at least, the classical writers were more honest in their efforts. They sought to explain relative prices by relative-input ratios of homogeneous labour. They fell short precisely because the deficiency in their common objective standard for measurement was revealed for all to see. This prompts the question as to why modern theorists have been so much more successful in concealing the fundamental flaw in their structure 'Camouflage by complexity' provides only a part of the answer here. The classical economists failed because their standard for measurement was demonstrably deficient, but also because their logical structure was not complete. One must read much into classical structure if any general-equilibrium theory of markets is to be discerned. They did not close the circle, and the lacunae in their essentially one-sided explanatory model provided the source for the familiar normative critique associated with Marx. The circle was completed by the subjective-value theorists, by the Marshallian synthesis, and, more explicitly, by the Walrasian theory of general equilibrium. These several contributions represent a major conceptual advance over classical economic analysis by criteria of logic and coherence. But the logical symmetry achieved in explaining the workings of the economic process was secured at a cost which is reflected by drainage of empirical, objective content. The classical economists offered us a positive-predictive theory of relative prices; this theory was falsified. But the neo-classical model contained no comparable predictive hypotheses; there was no externally measurable standard which allowed the scientist to make predictions from observable data. This post-classical theory described an interaction process and allowed the identification of certain properties of equilibrium positions. But there was nothing upon which the economist could have based objective predictions about relative-price formation.


This was surely sensed by Alfred Marshall as witnessed by his lingering adherence to classical models, and the desire for some restoration of predictive content offers a motivation for his time-period analysis. Frank Knight was also unwilling to disregard fully the classical precepts, and, despite his affinity with some of the Austrians, students of students of students of Knight continue to learn, and to learn well, the lessons of the deer and beaver. The reaction of the Austrians was quite different. They seemed quite willing to jettison the putative objective content of the classical hypotheses. The full implications of this may not have been recognized by the early Austrians, but in Mises and his followers economic theory is explicitly acknowledged to be wholly non-objective. Intellectual tidiness rather than empirical or explanatory content seemed to be the purpose of both earlier and latter-day Walrasians.


As I have suggested above, confusion emerged only when 1) theorists overlooked the absence of objective content in neoclassical and general-equilibrium analysis, and 2) when they attempted to utilize the properties of market equilibrium as norms for the optimizing solutions of problems posed in non-market institutional settings. The presence or absence of objective content assumed instrumental significance only when the planner was introduced, whether in the administration of state or public enterprises (piecemeal or in toto) or in levy of corrective taxes and/or subsidies on production in markets. The control or correction of allocation requires that norms be invoked, and these norms must come from somewhere. The presumption of modern economic theory that such norms are readily identifiable must be attributed to the acceptance of the paradigm one-resource model sketched above.


In any plausibly realistic market process, however, only prices have objective content. This being so, how can prices be settled by reference to 'costs' or to anything else? It will be useful to discuss briefly the precise relationships between prices and 'costs' in full market equilibrium. (In this treatment I shall follow closely the discussion in Cost and Choice, page 85.)


In full market equilibrium expected marginal benefit for each participant will be equal to marginal opportunity cost, both measured in terms of the person's subjective valuation. All persons confront uniform relative prices for goods; this is a necessary condition for the absence of further gains-from-trade. Since each participant is in full behavioural equilibrium, it follows that each person must also confront the same marginal cost. As a demander the individual adjusts his purchases to insure that marginal benefit equals price. Hence the anticipated marginal benefits of a good, again measured in the numeraire, are equal for all demanders. As a supplier the individual adjusts his sales to insure that anticipated opportunities forgone, marginal opportunity cost, equals price. Hence marginal opportunity cost in the numeraire is equal for all suppliers.


Prices tend to equal marginal opportunity costs in market equilibrium. But costs here are fully analogous to marginal benefits. Only prices have objective, empirical content. Neither the marginal valuations of demanders nor the marginal costs of suppliers can be employed as a basis for determining or setting prices. The reason is that both are brought into equality with prices by behavioural adjustments on both sides of the market. Prices are not brought into equality with some objectively measurable phenomena on either the demand or supply side.


The implications of this basic, and in one sense, elementary fact for applying economic theory's tools to the making of control decisions for a wholly or partially socialized institutional structure have not been fully recognized, even by those who have partially escaped the dominance of the single-resource model. To an extent the blame for this lies in the failure of the London economists, and of the latter-day Austrians, to develop a full-blown 'subjectivist economics' that commands intellectual respect while seeming to retain explanatory relevance. Mises and his followers have been too prone to accept the splendid isolation of arrogant eccentrics to divorce their teaching too sharply from mainstream interests, and too eager to launch into polemic: epistemological, methodological, ideological. Certain members of the London group, although profoundly influenced by the Austrians via both Hayek and Robbins, had the merit of maintaining more practical interest in business decision problems. But unfortunately their interest was too pedestrian to allow them to attempt the 'grand design' that might have been produced from the cost-theory foundations which they developed.


As a result, we find Hayek (and Mises even more emphatically) talking largely to the disciples of the Austrian faith, and alongside we find Coase, Edwards, Thirlby and Wiseman taking up the cudgels against orthodoxy in detailed and particularistic applications. In their later papers both Thirlby and Wiseman seemed to recognize the grander implications but both men were perhaps discouraged by their failure to secure acceptance of their particularistic arguments, discouraged to the extent that neither made the attempt to draft the 'treatise' that seemed to be required, and which still seems to offer challenge.


Perhaps the most significant L.S.E. impact on modern economics has come through an indirect application of opportunity-cost theory rather than through an undermining of basic cost conceptions. 'Marginal social cost', enthroned by Pigou as a corner-stone of applied welfare economics, was successfully challenged by R. H. Coase a quarter-century after his initial work on cost. His now-classic paper on social cost,*4 which reflects essentially the same cost theory held earlier, succeeded where the more straightforward earlier attacks on the marginal-cost pricing norm—attacks by Coase himself, by Thirlby and by Wiseman—apparently failed. Nonetheless the still-provisional success of Coase's modern challenge should be noted. As this is written, in mid-1972, the implications of Coase's attack on the Pigovian concept of social cost for the elementary textbook discussions of opportunity cost have not yet been realized. Advanced textbooks, and notably those written in what may loosely be called the 'Chicago-Virginia' tradition, devote some space to the 'Coase theorem', but the standard chapters on cost in these same textbooks remain as if the more fundamental critique in the Coase paper had never been published.*5


A primary purpose of this summary of doctrinal developments has been to emphasize the general importance of the theory of opportunity cost, and the London contributions to the development of a fundamentally correct theory which has not yet come to inform mainstream thinking in economics. The significance may, I fear, be hidden from those who glance only at the volume's title, L.S.E. Essays on Cost, and whose subjective image of 'cost' calls up carefully specified algebraic functions, sharply etched geometrical figures, or actual numbers carried to at least two decimal points in accountants' worksheets. Such an image may unfortunately be reinforced by a superficial survey of titles of some of the independent essays included here. Coase, Edwards and Thirlby, in some of the papers reprinted here, were interested in practical problems faced by business decision-makers in business administration as such. They were attempting to use economic theory in this severely practical setting, to apply opportunity-cost notions to the problems faced in everyday economic choices. In this effort the London economists did not themselves fully appreciate the uniqueness and originality of their approach. To an extent they looked on themselves as writing down, in the context of practical-problem situations, what 'everyone knew' about cost, at least everyone around L.S.E. during the period in question.


As the norms drawn from the description of competitive equilibrium came to be presented more and more as 'rules' for socialist planners, and 'marginal-cost pricing' was elevated into a paradigm for the management of public enterprise, the significance of getting the elementary confusions identified, and with this the relative importance and uniqueness of the London approach, came to be recognized. Both Thirlby and Wiseman, in the most recently published papers in this volume, recognized the depth of mainstream intellectual error, but their plaints were largely ignored. One reason perhaps lies in the fact that the critique of orthodoxy is too fundamental; to accept fully the implications of the theory of opportunity cost that is implicit in these essays requires the modern economist to throw overboard too much of his invested intellectual capital. How can we write the elementary textbooks and teach the elementary course if we cannot draw the standard cost curves? How can we carry out benefit-cost analysis and pretend that we are assisting in social decision-making? How can we say anything at all about managing nationalized public enterprises?


What is so 'revolutionary' in the theory of opportunity cost that threatens the very foundations of modern applied economics? This introductory essay is not designed to summarize the papers reprinted in the volume, and I do not propose to develop my own interpretation and application of the theory. I have done the latter in Cost and Choice. But brief elaboration of the central argument may offer some support to my assertions about significance. The basic idea is at once extremely simple and profound. Cost is inherently linked to choice. This notion did not of course originate with the economists associated with the L.S.E. in the 1930s or before or since. As students of Frank Knight learnt, elements of the correct theory of opportunity cost are found in Adam Smith's deer-and-beaver model. Even before the subjective-value revolution, Francesco Ferrara in Italy was sharply critical of classical theory on opportunity-cost grounds.*6 The opportunity-cost conception was explicitly developed by the Austrians, by the American, H. J. Davenport, and the principle could scarcely have occupied a more central place than it assumed in P. H. Wicksteed's Common Sense of Political Economy.*7 This book was independently influential at L.S.E., and it properly deserves mention here.


At the L.S.E. there was the beginning and the widening recognition of the implications of elementary opportunity-cost theory for applications of economics. Herein lies the contribution of the economists who are represented in this volume. Almost all professional economists, old and new, can provide a rough working definition of opportunity cost that is tolerably acceptable for pedagogic purposes. But very few economists, new or old, have been consistent. Almost none of them beyond the London-Austrian axis has recognized just what his own definition suggests for the application of his discipline.


Simply considered, cost is the obstacle or barrier to choice, that which must be got over before choice is made. Cost is the underside of the coin, so to speak, cost is the displaced alternative, the rejected opportunity. Cost is that which the decision-maker sacrifices or gives up when he selects one alternative rather than another. Cost consists therefore in his own evaluation of the enjoyment or utility that he anticipates having to forgo as a result of choice itself. There are specific implications to be drawn from this choice-bound definition of opportunity cost:

1. Cost must be borne exclusively by the person who makes decisions; it is not possible for this cost to be shifted to or imposed on others.
2. Cost is subjective; it exists only in the mind of the decision-maker or chooser.
3. Cost is based on anticipations; it is necessarily a forward-looking or ex ante concept.
4. Cost can never be realized because of the fact that choice is made; the alternative which is rejected can never itself be enjoyed.
5. Cost cannot be measured by someone other than the chooser since there is no way that subjective mental experience can be directly observed.
6. Cost can be dated at the moment of final decision or choice.*8


In any general theory of choice cost must be reckoned in a utility rather than in a commodity dimension. From this it follows that the opportunity cost involved in choice cannot be observed and objectified and, more importantly, it cannot be measured in such a way as to allow comparisons over wholly different choice settings. The cost faced by the utility-maximizing owner of a firm, the value that he anticipates having to forgo in choosing to produce an increment to current output, is not the cost faced by the utility-maximizing bureaucrat who manages a publicly owned firm, even if the physical aspects of the two firms are in all respects identical. As the London economists stressed, cost is that which might be avoided by not making choice. In our example the private owner could avoid the explicit incremental outlay and the incremental profit opportunity should he fail to produce the output increment. The socialist manager, by our assumptions, could avoid the same objective consequences by taking the same course of action. These consequences could be measured in monetary terms. But the opportunity cost relevant to choice-making must be translated into a utility dimension through a subjective and personal evaluation. The private owner may evaluate the objectively measurable consequences of choice quite differently from the bureaucrat, although both are utility-maximizers.


I am not suggesting that the contributors to the London tradition in cost theory fully appreciated and understood all the implications of their own conception, nor that even now they would endorse my interpretation of this conception. I suggest only that their several papers mark a beginning of such appreciation, that they reflect an early critical questioning of aspects of modern economic theory, a questioning that is more urgently needed in the 1970s than it was when they wrote.


While the contribution of the L.S.E. group of economists should be emphasized, the constructive content of their work should not be exaggerated. Taken as a whole, the London effort is largely negative in its impact. Properly interpreted, it demonstrates major flaws in the applications and extensions of economic theory. But there is little in this work which assists us in marrying 'subjectivist' and 'objectivist' economic theory. Few modern economists would be willing to go all the way with the latter-day Austrians and convert economics into a purely logical exercise. Most of us want to retain, and rightly so, positive and predictive content in the discipline, to hold fast to the genuine 'science' that seems possible. To accomplish this, however, homo economicus must be returned to scientific respectability, and economists must learn to accept that hypotheses may be falsified. Finally, and more importantly, we must try to construct meaningful, if limited, norms for decision-making in non-market institutional structures. In competitive markets prices tend to equal marginal costs, but do we want to make prices equal 'marginal costs' in non-market settings, when we fully realize that marginal costs can only be objectified by the arbitrary selection of some artificially homogenized measure? Do we really want to make one beaver exchange for only two deer when poisonous snakes abound near the beaver dams? Of course not! But how do we know that the snakes are there? Because the beaver hunters think they are?*9

Notes for this chapter

London 1932.
Paul A. Samuelson, Foundations of Economic Analysis (Cambridge, Mass. 1947), p. 5. Samuelson's Nobel lecture provides evidence that his own position has not substantially changed. See 'Maximum Principles in Analytical Economics', American Economic Review, 62 (June 1972), pp. 249-62.
Chicago 1969.
R. H. Coase, 'The Problem of Social Cost', Journal of Law and Economics, 3 (October 1960), pp. 1-44.
This summary of the impact of the London cost theory should include mention of G. L. S. Shackle. Although Shackle does not specifically present his ideas in opportunity-cost terms, his whole approach to decision is fully consistent with that developed by the London theorists. Shackle was both directly and indirectly associated closely with the London group. For Shackle's most appropriate treatment of decision, see his Decision, Order, and Time in Human Affairs, Cambridge 1961.
See my Fiscal Theory and Political Economy (Chapel Hill, North Carolina 1960), pp. 27-30 for a summary treatment. One of my own unfinished projects is a critical analysis of Ferrara's work, with a view towards making his contribution more widely known to English-language readers.
London 1910.
For a detailed discussion of each of these attributes of opportunity cost see my Cost and Choice, chapter 3.
I am indebted to my colleagues, Thomas Borcherding and Gordon Tullock for helpful comments.

Essay 2, Remarks upon certain aspects of the theory of costs

End of Notes

Return to top