S C Littlechild*

Thirty years ago, Jim Buchanan elucidated and developed the subjective nature of opportunity cost and its relation to choice. Ten years later, Jack Wiseman and I were hoping to write a textbook embodying a thorough-going subjectivist approach. Jim kindly offered to host our work at the Center for Public Choice, VPI, in Blacksburg in the summer of 1980. This was an invaluable opportunity: Jim’s own understanding of and sympathy for the approach and his openness to discussion, as well as the active intellectual milieu and unfettered working conditions at the Center, made this an ideal environment. No matter how early we arrived in the office there would be a typed note from Jim on our desks, reflecting his even earlier morning thoughts on the discussions of the previous day.

Regrettably Jack and I did not succeed in producing the subjectivist textbook. There were many difficulties. For example, Jack would propose that we could not use market demand and supply curves because they did not exist in anyone’s head. I would counter that we could not realistically produce an economics textbook that did not contain such demand and supply curves. We would argue furiously for a week. The next week we would each concede, after due reflection, that the other was right. Then we would argue for another week from the opposite points of view. By the third week we were each back to our original positions. And so on. We wrote several drafts, and Jim kindly organised a small seminar to discuss them, but I’m afraid we never did get past chapter four. But we did later publish a paper on Cost, Choice and Political Economy as exemplified by the "real" Robinson Crusoe.¹ The material for that paper was first collected during our stay at the Center, and Jack drafted the later version of the paper during a subsequent stay at the Center, by then moved to George Mason University, in Spring 1985. Both Jack and I were always grateful to Jim for his support, friendship and stimulating companionship.

Chapter 2 of Buchanan’s monograph Cost and Choice² traced the contributions of the economists at the London School of Economics. In an Appendix to that chapter Buchanan commented that Shackle’s treatment of the decision process was consistent with the LSE doctrine although Shackle surprisingly did not make the linkage between his own work and that of his LSE counterparts. This stimulated some subsequent discussion as to whether or not Shackle’s work was in fact consistent with the LSE opportunity cost tradition. The present note describes and evaluates some of this discussion and seeks to clarify the issue by giving Shackle’s own view on this matter, as expressed in Shackle’s correspondence now available for study³ and as set out in a hitherto unpublished note annexed to this paper. The paper concludes by recording an exchange of views between Buchanan and Shackle on subjective economics.

Buchanan comments as follows on Shackle.

To substantiate the first claim Buchanan gives excerpts from Shackle’s Decision, Order, and Time in Human Affairs⁵. These illustrate Shackle’s emphasis that the outcomes of actions between which a decision-maker chooses are things imagined by the decision-maker; that the consequences relevant for the decision-maker’s choice are the experiences of the decision-maker, rather than experiences coming from outside; and that it is imagined outcomes at the time of decision, rather than actual outcomes observed later, that are relevant. In a later paper Buchanan asserts that

It was Buchanan’s claim concerning the analytical relationship between Shackle’s ideas and those of the LSE group that attracted interest. In the very first issue of the Austrian Economics Newsletter, Richard Ebeling took Buchanan’s claim a stage further, and asserted in effect that Shackle did think of cost in opportunity cost terms. He wrote

In the next issue of the Newsletter Lawrence White explicitly rejected this suggestion, and claimed that "Shackle does not endorse the concept of opportunity cost". Rather, "It is [the] distressing prospect of loss which Shackle identifies as the cost accompanying choice." White's brief comment may usefully be reprinted in full.

I found White’s argument intriguing but not convincing, though I considered that Shackle himself seemed to be inconsistent on usage. I drafted a note to the Newsletter which I did not subsequently publish but which I discussed in correspondence with White⁹. He maintained and expanded on his view; I agreed (or thought I did) with his detailed arguments, but suggested that Shackle’s approach could nonetheless be interpreted so as to fit into and generalise the LSE framework. Briefly, my suggestion was that the opportunity cost of a chosen action A was the value of the pair of focus outcomes associated with the rejected action B. I thought (erroneously) that White and I agreed as to the senses in which Shackle was or was not consistent with the LSE tradition.

In July 1978 the University of Birmingham had awarded Shackle an honorary degree, and I took the opportunity to put my proposition to Shackle himself. He professed shock that he should be thought to reject opportunity cost and seemed to think his approach consistent with it. But as I confessed in my letter to White, "You must understand, however, that Shackle invariably agrees with everyone, and one has to seek amongst delicate shades of enthusiasm to ascertain his real view."

So on 25 September 1978 I invited Shackle to consider the matter explicitly. On 28 September he wrote back saying "On the question of opportunity cost, it is evident that I must really pay undistracted attention to this and sort out my thoughts by writing a short piece (for my own benefit)." Given his interest I copied to him White’s letter to me and summarised my own interpretation.

On 7 November Shackle wrote back saying "I spent three weeks trying a variety of attacks on the question whether and how opportunity cost can be expressed in terms of focus-outcomes. In the end I merely convinced myself that your paper says everything. It seems to me that Lawrence White only differs from your view in leaving out the necessary second stage of comparison. I think his contribution has been very valuable in raising the whole question." He enclosed "the draft into which I sorted my ideas". Shackle’s key conclusions are perhaps as follows.

I sent White a copy of Shackle’s manuscript, and a couple of months later he wrote to Shackle, expressing surprise that his initial statement in the Newsletter was considered controversial. "I was under the impression that you had explicitly broken with the opportunity cost tradition on the grounds that a chancy opportunity cannot be assigned a single value ex ante…. Instead an uncertain prospect must be assigned an ordered pair of epitomizing values (focus gain, focus loss)."¹² He also expressed concern about the implications for the concept of the neutral outcome.

In his reply to White, Shackle expressed very great admiration for a copy of a paper on entrepreneurial price adjustment that White had sent,¹³and suggested an additional reference to his own work. He continued, "I must in coming days enter upon a period of most intensive exertion, in order to come to terms with your searching criticism regarding opportunity cost. I must look at the whole matter again from the start."

In fact, Shackle did not write further to White on this topic. But three years later, he was asked about this issue by Alex Shand, who was then writing a monograph on subjective economics. His reply shows that he had not changed his view.

Buchanan was right and perceptive to observe that many of Shackle’s statements on decision were consistent with the subjective cost discussions of other LSE economists. This was a novel and stimulating claim, which predated some of Shackle’s more explicit writings in this vein. For example, it was not until four years after Cost and Choice, and after Buchanan had written his paper on "LSE Cost Theory in Retrospect", that Shackle’s argument was published that

This quotation further tends to confirm Buchanan’s suggestion about the acceptance of the subjective concept of cost in London and the lack of acceptance elsewhere.

How far "Shackle was both directly and indirectly associated closely with the London group" is less clear. Hayek was Shackle’s Ph.D. supervisor at LSE, and Shackle was very familiar with Hayek’s economic thinking, though he later decided to work on Keynes’ ideas instead of Hayek’s. And, as Buchanan notes, Thirlby acknowledged the influence of Shackle and indeed notes that Shackle read and commented on one of his papers in draft.¹⁶ But Shackle was very much a loner, not associated closely with any group. He often wrote of the impact that particular economists and ideas had on him, but I am not aware that he wrote of his London colleagues in this way.

The Shackle papers at Cambridge shed further light on the issue.¹⁷ Shackle reviewed a couple of Robbins’ books and there is evidence that he read carefully a 1938 paper by Robbins on methodology (and mysteriously deemed it "mispelt"). There was one exchange of letters with Robbins, in December 1968, associated with Robbins Festschrift, in which Shackle refers to Robbins’ influence and to his lectures on the history of economic thought. Shackle also corresponded with Thirlby in December 1951, commenting on one of the latter’s papers. This seems to be the reference acknowledged by Thirlby above, and there does not seem to be any further correspondence. Wiseman wrote to Shackle in October 1952, about "the probability bogey" and "excludable and non-excludable hypotheses", and sent him a copy of his article on uncertainty, costs and economic planning (reprinted in Buchanan and Thirlby LSE Essays on Cost). Correspondence between these two continued until 1990, as well as meetings. And if Wiseman did not raise the subject of subjective opportunity cost at some stage, he was not the Jack Wiseman that Jim and I knew. There is no evidence that Shackle corresponded with the other members that Buchanan identifies as in the LSE group, such as Coase and Edwards.

Given the deliberation with which Shackle approached every book and article, the scope of his reading programme was necessarily limited. For example, Shackle himself says that he did not read anything by Popper until much later [than it was written]. If this was true despite the importance and topicality of Popper at the time, and given the focus of Shackle’s interests on business cycles and related issues, my initial inclination was to be surprised if he had read much by the LSE writers on cost. But the evidence now suggests that he must have been familiar with the thinking of Thirlby and Wiseman in 1951, and so I conclude that Buchanan was right in his judgement.

Buchanan suggested that "in his general treatment of cost itself, Shackle reverts to orthodoxy" but did not give sources to substantiate this claim. I drew White’s attention to Shackle’s Expectation, Enterprise and Profit, which in some respects is orthodox, but he quite reasonably pointed out that "this textbook is, among Shackle’s writings, the least expressive of Shackle’s own views."¹⁸

At the time of writing (1978), White was right to point out that Shackle did not endorse the concept of opportunity cost, at least in his published writings, and that it was the distressing prospect of loss that Shackle identified as the cost accompanying choice. That was what interested Shackle, and what he focused on.

Furthermore, it would seem that Shackle had not at that time fully considered the meaning and implications of opportunity cost within his own system of thought. But I think it was going too far to claim that Shackle actually rejected the notion of opportunity cost. White cites Shackle’s exposition of the doctrine of opportunity cost (what he calls displacement cost) in Epistemics and Economics (pp. 132-3), and suggests that Shackle was doing so in order to discuss its inconsistency with the deterministic overtones of standard microeconomic theory. I am not confident that I know what Shackle’s purpose was in that particular passage. He seems to be saying that opportunity cost and freedom of choice, and more generally rational conduct, can only be made consistent with a deterministic world if the relevant data are somehow provided to decision-makers (and that economists had failed to explain how this was done). But whichever interpretation is taken, it does not seem to be a rejection of opportunity cost; rather, it seems to be a rejection of standard microeconomic theory. And if it is not an explicit endorsement, it perhaps is an implicit endorsement of opportunity cost, and of the associated reality of choice in a world where knowledge is necessarily incomplete, if the alternative option is to accept standard microeconomic theory.

When Shackle’s attention was drawn to the issue, he concluded that his own approach was not inconsistent with the concept of opportunity cost. (Nor, in effect, with Ebeling’s characterisation.) He argued, essentially, that when uncertainty was fully acknowledged, and what he liked to call a skein of possible outcomes had to be envisaged for any action, then the concept of opportunity cost had to be defined more generally than economists had done previously. It had to be expressed in terms of the value associated with a foregone skein of envisaged possible outcomes rather than with a single envisaged outcome.

Because Shackle came to this question only after developing his own apparatus of decision-making, he did not express it in quite this way. He started from the proposition that each skein could be reduced to a pair of outcomes, namely the focus gain and focus loss. For an action to be attractive its focus gain had to be more persuasive than its focus loss. And for it to be more attractive than all other actions the excess of this persuasive power had to be greater than for any other action. In this latter sense the concept of opportunity cost was taken on board, but only as one of two hurdles that a chosen action had to overcome.

White’s main objection was that Shackle had rejected opportunity cost because he denied the existence of well defined (i.e. single valued) opportunities, and instead proposed that actions be characterised in terms of a pair of focus-outcomes. If focus gain and loss are measured in relation to a neutral outcome, which may be hypothetical and not necessarily in the decision-maker’s choice set, and if decisions are made on the basis of these concepts, White suggests that this is inconsistent with straight opportunity cost doctrine which ranks alternatives only relative to one another. I would again argue that "rejected" and "inconsistent" are too strong here. The criterion of focus gain exceeding focus loss is just one part of the test for a chosen action. Even Shackle’s early expositions accept the need to choose between alternative available actions, on the basis of what he called the gambler indifference map.

White also expressed concern that Shackle was, in effect, reducing courses of action to single values at the final stage of decision, by characterising each action in terms of the numerical difference between focus gain and loss, and then comparing these differences and choosing the action with the greater difference. But this is not what I interpret Shackle to mean. He uses phrases like "ascendancy" and "persuasive influence" that mean more than simple subtraction of numerical amounts: they take account also of the degree of interest or excitement or concern generated by a each particular outcome and of its likelihood or plausibility. For an individual to choose an action, or enterprise, "because its focus gain more decisively overpowers its focus loss than is the case with any of its rivals" seems to me consistent with "retaining in the whole act of decision a formal role for uncertainty".¹⁹

I see force in several of White’s concerns about the concept and definition of the neutral outcome. But they seem to me a problem for the concept and definition of a neutral outcome, rather than for the concept of opportunity cost within Shackle’s world of uncertainty. I am not sure that the analysis or exposition of focus outcomes needs to refer to neutral outcome at all. Furthermore, Shackle’s more precise propositions concerning reduction to and choice between focus outcomes seem to me essentially empirical ones and less convincing than the more fundamental concepts of the skein of possible outcomes and the interrelated relevance of attractiveness and subjective belief.²⁰ I am sure that Shackle would have approved of empirical work in psychology to try to learn more about how decisions actually are taken.

This paper began by quoting Buchanan’s 1969 comments on Shackle’s work on subjective cost and choice. A dozen years later Buchanan again cited Shackle on subjective economics, in a very supportive way. For example, towards the end of his paper he wrote

Having examined Buchanan’s views on Shackle, it is of interest briefly to explore Shackle’s views on Buchanan. He reviewed very favourably Buchanan’s Selected Essays on Fiscal Theory and Political Economy²², particularly his study of the Italian tradition with which Shackle was especially familiar.

The annotations on Shackle’s copy of the book enable us to see some of the thoughts that did not make it into print.²³ For example, he thought that at one point in his paper on the Italian tradition Buchanan fell into the same fallacy – consideration of partial effects – that he had been arguing against. But against this there are numerous passages marked with approval for quotation, one marked "a good paragraph", another "excellent" and in one case "a brilliant passage".

The annotations also reveal, in a way that the review does not, Shackle’s sympathy for Buchanan’s emphasis on subjectivism (though I’m not sure that either uses that actual term) in the context of welfare economics. For example, Buchanan writes

This passage gets Shackle’s "excellent" rating.

The availability of the Shackle collection of letters at Cambridge now enables us to learn Shackle’s views on Buchanan’s later work. He was tremendously impressed and pleased with the latter’s paper in the Mises symposium just cited. In his letter to Shand that covered opportunity cost (11 November 1982) he also commented, "I think James Buchanan is nowadays more pleased with my views than he was in Cost and Choice since he was kind to me in the paper he wrote for the Ludwig von Mises symposium last year." In the present Festschrift it is appropriate and pleasing to end with the tribute that Shackle wrote to Buchanan himself.

* Honorary Professor, University of Birmingham School of Business. I am grateful to Lawrence White for helpful comments on this paper; and to Ms Kathleen Cann, Manuscripts Department, Cambridge University Library and Mr Adrian R Allan, University Archivist, and his colleagues at Sydney Jones Library, Liverpool University, for assistance in accessing the material referred to below.

G.L.S. Shackle

22 October 1978

Let us refer to a scheme of actions directed to some self-consistent unified effect, which actions and individual might envisage taking in his early time-to-come with the means at his disposal, as an enterprise. With these same means, any one but only one of a plurality of rival distinct enterprises will seem open to him. How will he decide amongst these enterprises?

If he looks upon such an act of choice as having in its making some element of uncause, so that its character could not have been inferred from any knowledge, however complete, of the circumstances and thoughts of the individual antecedent to its making, then he is obliged in logic to recognize that the sequel of any enterprise cannot be uniquely foreknown. For this sequel will be shaped in some degree by the choices which he and others will make in time-to-come, and the character of these choices is not capable of being inferred. Nonetheless, his knowledge of natural principles, of human organization and of the posture of things in his present, including the qualities and quantities of his own resources, will seem to set bounds to the diversity of the sequels which, up to some moment in the course of his business of choice, he can imagine and judge possible for some specified enterprise. These sequels will be capable of being ordered by him according to their greater or less desiredness. If all of them are judged by him equally possible (equally free, in his thought, from fatal obstacles discernible by him in the light of his knowledge) only the most and the least desired will concern him. For all of the imagined sequels are rivals, the coming-true of one would exclude the coming-true of any of the others, there is no meaning in any summing or averaging of these mutual exclusives, what matters to the individual is what according to his judgement can follow his election of a particular enterprise, at best and at worst. These are what I mean by the focus-outcomes of the enterprise.

The concept of focus-outcomes takes a sharper edge if we apply it to some concrete and particular field of decision, since this makes it easier to treat the desiredness of a conceived sequel of action as a scalar variable. If this field is that of enterprise in the commercial sense, the setting-up and equipping of a business, then the relevant aspect of any sequel is the trading profit derived from the working of the business, that is to say, the excess, in each unit calendar interval, of the sale-proceeds of product over the expense of operation. As an element in decision, such a stream of trading profits is a supposition, something conceived in thought, one of many or indefinitely many rivals. In order to give it a scalar value, its instalments must each be discounted at that market rate of interest on loans which applies to the length of deferment of the particular instalment. The sum of these discounted or 'present' values will be the spot cash equivalent of the supposed stream of trading profits when the latter is considered in abstraction from its epistemic standing. Instead of a skein of conceived sequels the individual can now consider the set of their respective discounted values. The extremes of this set will provide the focus-outcomes. Thus we can respresent each enterprise which the individual envisages as open to him to choose, as a point in a Cartesian frame.

In this frame let the west-east axis serve for focus-loss and the south-north axis for focus-gain. Points in the north-east quadrant will stand for enterprises to which the individual ascribes focus-gain and focus-loss both of them numerically greater than zero. A point in the north-west quadrant would stand for an enterprise to which the individual adjudged a focus-loss numerically less than zero. That is to say, the enterprise would be assigned two focus-outcomes, both of them gains. No use need be made of the south-east or south-west quadrants. In choosing amongst several enterprise, each requiring the investment of one and the same value of resources, the individual will disregard any enterprise whose focus-gain is smaller, and whose focus-loss is numerically larger, than those of one of the other enterprises which he has in mind. His relevant comparisons, that is to say, will be only between such points as lie north-east and south-west of each other.

The expression opportunity-cost names the idea that when faced which a number of mutual exclusives amongst which he is free to choose, the individual in electing one of them is sacrificing the next-best of the choosables, and that it is this sacrifice which for him is the cost of the choice he actually makes. The question which concerns me is what meaning is to be given to opportunity-cost when choice amongst enterprises, in the more general or in some more special sense, is examined in terms of focus-outcomes.

The notion of opportunity-cost expresses an extremely general aspect of choice. The sacrifice of a second-best is part of the essence of any act of choice. We may still, however, ask in what that sacrifice precisely consists. It must consist in that psychic experience, inferior only to the one he has elected, which was within his reach instead. If we represent both the elected and the forgone enterprise by means of pairs of focus outcomes, the psychic experience in each case is the enjoyment by anticipation of that course of things represented by the focus-gain, combined with the distress of contemplating that course of things represented by the focus-loss.

May we not say that when our particular individual envisages that enterprise, amongst those which his own thought has created and judged possible, and has offered for his choice, which appeals most to him, he does so because its focus-gain more decisively overpowers its focus-loss than is the case with any of its rivals? What he sacrifices, as the cost of choosing the best enterprise, is this lesser ascendancy of the focus-gain of the second-best over ist focus-loss. What, then, will express completeley the formal necessary conditions for his choosing one particular enterprise out of a number of rivals? The chosen enterprise must have a focus-gain more persuasive than its focus-loss, and this excess of influence must be greater than that which is to be found in any rival. Two stages of comparison are needed: focus-gain with focus-loss in each of the two rivals, and excess of persuasive power in one case with excess of persuasive power in the other. And at last, can we say that one of these stages and not the other is opportunity-cost?

The matter can be penetrated a little further. We can step back, as it were, from the notion of opportunity-cost and look upon it as one necessary element in an act of choice, but not one having a right, in the nature of things, to be treated as the only one that matters. Cannot cost be a more complex idea than that of a single scalar quantity? However, we might be tempted to seek to restore simplicity by asking, if the excess of persuasive influence of focus-gain over focus-loss in one enterprise is greater than in another, how much greater is it? Here also, we need not be intimidated by traditions of exactness. Can we deny meaning to such expressions as 'distinctly greater', 'decisively greater', excitingly greater' which the individual might read from a private scale built up by his own experience and intuition. These may be deemed to hang between qualitative and quantitative assertion. But choice between enterprises is not a question of small differences. One exceedingly complex whole is almost of necessity discretely, not continuously, different from another. Men do not fall in love according to a calculation of marginal advantage. Thus I think we may eschew any resort to the invention of a (metrically defined) surface whose arguments would be the focus-gain and focus-loss variables represented on the axes of our Cartesian diagram. A private scale can be imagined to make such a surface conceivable. But what purpose would be served?