Abram Bergson's essay, "Socialist Economics," has been by far the most commonly referenced secondary source on the calculation debate. Schumpeter was to say of this essay that its description of the development of the pure theory of socialism "leaves nothing to be desired,"43 while Benjamin Ward was to refer to the socialist controversy as having "formally ended with Bergson's summary of the issues."44
Like the later Schumpeter, Bergson was unwilling to completely and uncritically adopt Lange's equilibrium interpretation of Mises' argument, but Bergson definitely leaned strongly toward that view. Concerning the questions raised by Mises "as to whether socialism can work at all, and how well," Bergson begins his article by pointing out: "By now it seems generally agreed that the argument on these questions advanced by Mises himself, at least according to one interpretation, is without much force."45 He suggests that the later emphasis of critics of socialism on the question of the possible incompatibility of planning with political freedom "has the appearance of a tactical maneuver to bolster a cause which Mises' theories have been found inadequate to sustain."46 Thus Bergson appears to accept both Lange's interpretation and refutation of Mises.
In a section entitled "The Ends," Bergson lays out some issues concerning "the formulation of a scale of values, on the basis of which the alternative uses of resources are to be evaluated." Among these are (1) the question of the extent to which either consumer sovereignty or the CPB should dictate the ends of social production, and (2) the controversy over whether individual preferences are fundamentally incommeasurable or can be meaningfully aggregated into a social welfare function which the CPB could work to optimize.47 The first of these, though important for other reasons, is irrelevant to the calculation debate, since Mises' challenge, as he explicitly noted, applies whether the ultimate preferences are those of a single dictator or of a free consumers' goods market. Likewise, the second issue over interpersonal utility comparison, while undoubtedly an important matter for economic theory (about which Mises took a very definite position), is not crucial to the calculation argument. Even an economy possessed of "milometers" enabling an accurate and meaningful cardinal measurement of everybody's preferences would still face the problem of imputing these final demands into the higher stages of production.
The next section of Bergson's article, significantly entitled "Optimum Conditions," begins to approach the calculation argument in the standard neoclassical manner.
In accord with familiar theoretic procedure, technical knowledge and tastes are taken as given____On this basis it is possible to derive from the given ends a series of conditions ("equations") which must be satis-fled if the optimum allocation is to be achieved. The optimum conditions are sufficient in number to determine the amounts of each and every sort of goods and services allocated to each and every use (the "unknowns").48
Bergson thus presents the formal case for the determinateness of a socialist equilibrium, giving some indication that this comprises a rather abstract "theoretical" argument based on some unrealistic assumptions rather than any sort of proof of practicability.
Thus, if the scale of values implied by the ends were known in complete detail (that is, if all the utility functions were known), and detailed information were available on techniques and on the stocks of resources on hand, it would be possible at least theoretically to solve this system of equations for the concrete values of all the unknowns.49
Bergson then goes on to summarize the subsequent neoclassical contributions to the debate which have been "concerned chiefly to formulate explicitly the optimum conditions., .and to develop the analysis to deal with various complexities." As with Schumpeter we have a preoccupation with equilibrium analysis, the refinement of the "optimum conditions" such as that the marginal rate of substitution for each pair of consumers' goods must be the same for all households, that "factors must be combined in a technologically optimum manner," and that the "marginal value productivity of each factor must be the same in every industry."50 A considerable literature in welfare economics — to which Bergson had made a seminal contribution—has exhaustively studied these equilibrium conditions and has undoubtedly promoted clarity in our formal depictions of both capitalist and socialist equilibria.51 The question remains, however, how producers would know what the "technologically optimum manner" of production or the "marginal value productivity of each factor" is, without a competitive capital market. Assuming technology to be "given" may be a legitimate analytical procedure within an equilibrium framework, but it crucially begs the whole question of the calculation debate.
In his section IV Bergson specifically addresses the issue of the contributions this abstract equilibrium analysis can make to the central planners' task. First, this framework might "assist the Board to formulate a conceptually satisfactory scale of values to guide the economy, one that is internally consistent and in principle at least covers the bill." Secondly, the analysis "establishes the implications of the given ends," i.e., the optimum conditions, which would be a "prerequisite for the construction of a planning scheme which might approximate the given ends in practice." It could be said that Mises' whole argument was essentially that, in practice, noncom-peting producers could not determine the complex implications of the consumer demands with respect to the more efficient applications of the means of production toward those ends. But Bergson, at least in this article, seems to grossly oversimplify the practical difficulties involved in this imputation process.
It happens that the criteria for the optimum that have been set forth are conceptually simple and, for the cases where small adjustments are possible, require for their application only facts which actually might be experienced in a given situation (the marginal rates of substitution, marginal productivities, etc.)."
Would such "facts" as marginal productivities really be "experienced" without a competitive process? Surely we cannot simply assume that actual market participants would have all the knowledge necessary for the economist's formal proof of a determinate equilibrium. That theoretical proof proceeds as if each producer initially had a complete "cookbook" of technological recipes which, in conjunction with the appropriate prices given to him by the CPB (or the Walrasian auctioneer), he uses in order to select his optimum production technique. Such a discussion may be helpful as an heuristic device in showing the interdependence of decisions in the market, but as a purported description of the decision-making process of actual producers this approach is very ill-suited and misleading.
When Bergson comments that the CPB would "adjust prices and wages from time to time in order to bring the demand and supply of goods and services into line,"53 he seems to seriously underestimate the degree to which unexpected changes permeate a modern economic system. The economy appears as if it is normally in a state of equilibrium from which it is only occasionally disturbed. But, as Hayek and Mises were frequently to emphasize, the implications of an economy's being in equilibrium show that this is a totally unrealizable situation, in which, despite the passage of time, every agent's plans are perfectly meshed with everyone else's, in which genuine uncertainty has been banished, and in which such institutions as money and even prices would be superfluous. In such a world Lange's proposed "trial and error" solution would indeed be workable, since the CPB would have ample time to alter one "price," holding the others at their equilibrium values until complete equilibrium is reached.34 As Hayek puts it, "With given and constant data such a state of equilibrium could indeed be approached by the method of trial and error. But this is far from being the situation in the real world, where constant change is the rule."55
Bergson in fact had serious doubts about the workability of Lange's solution in a continually changing world. He cited Soviet experience as indicating that on account of the economic as well as political problems there might be very real difficulties in the way of applying the trial and error Competitive Solution. Consider the matter of high-tempo industrialization, and the rapid shifts in demand and production schedules that would be associated with this process. In such a situation, the errors involved in the operation of the Competitive Solution might well be formidable; and evidently experience could not be very helpful in rectifying them.56
The Austrians in the debate, however, would contend that the "shifts in demand and production schedules" that occur in normal conditions, not just during "high-tempo industrialization," would be rapid enough to make the planners' task of locating a set of equilibrium prices by trial and error quite unmanageable. As he was to point out in a later essay, Bergson ultimately did not accept Lange's opinion that under the "competitive solution" convergence to equilibrium would be rapid, and he was to belatedly agree with Hayek that on the contrary the task of fixing prices for an entire economy, which Lange assigns to the Board, would prove formidable. Almost inevitably, the Board would find it difficult to respond quickly to continually occurring changes in supply and demand. It would also be unable to fix prices in sufficient detail to take into account almost endlessly diverse varieties of goods produced by a modern economy. Imbalances between supply and demand, therefore, might be large and persistent.57
Although this more recent assessment of the debate substantially agrees with the Austrians, Bergson's earlier and more influential discussion reflects a common neoclassical view of equilibrium as a normal situation of quantitative balance between supply and demand. By contrast Austrian economists view equilibrium as an entirely imaginary condition of complete coordination of plans where all "changes" are fully anticipated. As such, it may be a useful, even an indispensable, tool of analysis, but according to the Austrian view it is never attained in any real economy. The fact of continuous disequilibrating change renders a "trial and error" process of equilibration unworkable since we could never know which of many prices are at disequilibrium with respect to some actors' plans at any time. As Mises later remarked in this regard, "The method of trial and error is applicable in all cases in which the correct solution is recognizable as such by unmistakable marks not dependent on the method of trial and error itself."58
To his credit, Bergson in 1948 did acknowledge that in practice the CPB may encounter some severe difficulties in administering Lange's system. Among these he specifically cites three: (1) the fact that "the vast stock of detailed knowledge" necessary would be "scattered throughout the community" and the amount of this knowledge available would itself "depend on the particular administrative procedure used"; (2) it would be "physically impossible" for the Board to decide successively on all the myriads of possible alternatives; and (3) it would be difficult to control the execution of the plan once it had been formulated. 59 Each of these points was specifically argued in the debate by Hayek and will be examined here later in more detail. Bergson seems to have believed in 1948 that the so-called "competitive solution" offered by Lange, Taylor and Dickinson (particularly in Lange's "very systematic exposition") by and large overcomes all of these potential difficulties. Bergson points out the relevance to the debate of Hayek's later contributions about "The Use of Knowledge in Society," and his description of the kind of "detailed knowledge of time and place" which "by its nature cannot enter into statistics" and "cannot be conveyed to any central authority in statistical form."60 Bergson finds these remarks a "wholesome antidote" to the tendency of many socialists to view the ministry of planning as a "committee of Supermen." However, he does not accept Hayek's conclusion that these necessary limitations on the knowledge of the CPB would render the task of setting prices impracticable, since he believes that this task can be decentralized, i.e. "broken down functionally and geographically," and be guided by "general directives" issued by the CPB.61 He also dismisses Hayek's contention that the CPB would have to continually conduct detailed audits of the cost records of the individual firms, which if true would seriously impair the decentralization advantages intended in the "competitive solution," Bergson proposes that by tying managerial incentives to profits and comparing profits of different firms the Board could dispense with such detailed studies.62
Underlying this particular controversy is an ambiguity concerning the extent to which the "competitive solution" actually introduces competition into the economy. We would contend that, if genuine rivalrous competition and the market equilibrating process are not fully permitted in the "competitive solution", the planners will not be able to escape Hayek's problem of "scattered knowledge" as easily as Bergson suggests. If, on the other hand, such full-fledged competition is allowed, the resulting system will be indistinguishable from private ownership of the means of production.
A less conciliatory attempt to answer the calculation challenge has been the "Centralist Scheme" favored by Maurice Dobb. Bergson points out how this approach would seem to bring back the problem of "solving millions of equations" which the "competitive solution" was designed to resolve. Such difficulties could be reduced, however, if one can assume fixed coefficients of production under which, in effect, relative prices can be ignored, and under which there is no basis for speaking of the marginal productivity of any one factor. Bergson notes that "Leontief found it practicable" in his famous input-output study of the American economy "to assume that for broad industrial groups the production coefficients are constant."43 If fixed coefficients could be assumed, the planners would only need to devise technologically feasible allocations of resources rather than economically efficient ones. But we would contend that both the level of aggregation and the span of time for which input-output applications have been used are far too broad to aid the microeconomic production decisions with which the calculation debate was concerned.64
Finally, at the end of the article, Bergson returns to the question he had raised at the beginning: what Mises meant when he asserted the impossibility of economic calculation under socialism. Unlike many treatments of this point, he at least admits there are two possible interpretations of Mises, those of Lange and Schumpeter on the one hand, and that of Hayek on the other.
Before describing these interpretations Bergson offers to "let Mises speak for himself and selects four quotations65 from Mises' essay on "Economic Calculation in the Socialist Commonwealth," which rather forcefully make the point about the necessary relationship between private ownership and prices, and between prices and rational economic calculation. Largely because of their forceful style of expression these quotations could be seen as consistent with Lange's interpretation of Mises, but in actual content they are consistent with either interpretation. Unfortunately Bergson did not choose to select the following quotations from the same group of pages which strongly support Hayek's and explicitly contradict Lange's interpretation:
The static state can dispense with economic calculation But this is only conceptually possible a static state is impossible in real life, as our economic data are forever changing, so that the static nature of economic activity is only a theoretical assumption corresponding to no real state of affairs, however necessary it may be for our thinking and for the perfection of our knowledge of economics.66
As was argued earlier, no statement in economics stands alone. To be rendered fully intelligible it must be examined in terms of the wider conceptual framework from which it stems. It is not enough to select a few quotes,
"read innocently," to prove what Mises meant, for these words must be given meaning by relating them to the system of thought of the reader. For Lange, Schumpeter, and Bergson, rational economic calculation means calculation with equilibrium prices, but not for Mises.
Of the two interpretations Bergson considers, the one "which seems to have gained the wider currency" is Lange's: that rational calculation is "ruled out conceptually" even for a "Board of Supermen, with unlimited logical faculties, with a complete scale of values for the different consumers' goods and present and future consumption, and detailed knowledge of production techniques." If such an argument has ever been maintained by anybody, it would, as Bergson said, certainly be "easily disposed of and would have already been refuted completely by the work of Pareto and Barone. As the analysis of these writers shows, once tastes and techniques are given, the values of the means of production can be determined unambiguously by imputation without the intervention of a market process. The Board of Supermen could decide readily how to allocate resources so as to assure the optimum welfare. It would simply have to solve the equations of Pareto and Barone.67
It would be difficult to reconcile this interpretation of Mises with his specific reaction to Dickinson's similar "superman" assumption: "We do not deal with the acts of the omnipresent and omniscient Deity, but with the actions of men endowed with a human mind only. Such a mind cannot plan without economic calculation."68
Fortunately there is another interpretation of Mises, the one given by Hayek, that "Imputation is theoretically possible; but, once private ownership of the means of production has been liquidated, it cannot be accomplished in practice."59 This, too, Bergson says, can be interpreted in two ways (represented in Schumpeter's second and third questions above), "whether a planning scheme can be devised such that it can work at all" or "which is more efficient, socialism or capitalism." Concerning the former interpretation of "impracticability," Bergson is convinced that both Lange's trial and error argument and even the simple existence of the Soviet Union refute that claim. The latter issue of relative efficiency between capitalism and socialism, then, "is now the only issue outstanding."70
Yet in fact this last question was always the only issue outstanding and all the involved discussions of "conceptual" possibility and practicability "in principle" have only diverted the debate from its original and essential content.11 Mises, Hayek and Robbins never denied that socialism would be "possible" under static conditions, though they had serious doubts about the possibility of those conditions. They were denying that any procedure could be found which was efficient enough at impelling an equilibrating process to supplant the function of the capitalist-entrepreneur under private ownership of the means of production.
In 1948 Bergson viewed Hayek's arguments concerning the unworkabil-ity of Lange's solution as generally unsuccessful and this assessment has been preserved in many subsequent accounts of the debate. But just as Schumpeter reversed himself in his interpretation of Mises, Bergson was to change his mind substantively about Hayek, and in both cases it was the earlier —and in our view the less defendable—version that has been more influential in the history of thought.
In his later article, "Market Socialism Revisited,"72 Bergson argues, for example, that the "failure" on the part of Lange and his school "to establish any practical success criterion for managers" now seems to be a more serious deficiency than he had thought before, when he had said "our impression is that the question of managerial incentives would not present any serious difficulties." While earlier he had concluded that "there is no reason to suppose" that managers under public ownership "would necessarily be too venturesome or, as Hayek argues, too cautious,"73 his later opinion had substantially changed.
Given a satisfactory test of success, managerial incentives must still be related to it appropriately. Hayek argued that such a result might not be easy to achieve. In practice, managers very likely would be reluctant to take risks. This is perhaps not inevitable, but the construction of a satisfactory incentive system now appears more difficult than I envisaged it to be previously.14
Invariably the process by which an original controversy is summarized in secondary sources and digested by later thinkers until it filters its way into the textbooks is a gradual one. It was unavoidable that some of the earliest discussions of the calculation debate—of which Schumpeter's and Bergson's were the most often cited — would have had a greater impact than any more recent second-thoughts are apt to have had. Many third-level accounts of the debate were written before Schumpeter and Bergson changed their minds about Mises and Hayek, while fourth-level discussions in contemporary textbooks have often relied on these third-level discussions. Thus the standard version of the debate still retains errors—or so they are considered by this reader — which the authors themselves have long since corrected.
Was this article helpful?