supplement to vol. 22, 1990, pp. 133-154 

      Published in hardcover as Bruce J. Caldwell, ed. 
      Carl Menger and his Legacy in Economics 
      Durham, NC: Duke University Press, 1990 
      
      
      
    Austrian Capital Theory: The Early Controversies

    Roger W. Garrison* 

    I. The Austrian Vision.
    Austrian economics owes its uniqueness, in large part, to its attention to the economy's capital structure. Theorizing about the value of capital, about the time element in the structure of production, and about market mechanisms that facilitate intertemporal adjustments to the capital structure has constituted a significant part of the research agenda for the early as well as the modern Austrian school. Yet, fundamental differences emerged during the earliest developments of Austrian capital theory—differences which even today are not fully resolved.
            Menger's harsh assessment of Böhm-Bawerk's contribution is well reported in modern literature: "[T]he time will come when people will realize that Böhm-Bawerk's theory is one of the greatest errors ever committed" [Schumpeter, 1954, 847, n. 8]. Although the context in which the statement was made remains a matter for conjecture, a prevalent—and plausible—interpretation is that Böhm-Bawerk strayed too far from the subjective value theory outlined by Menger [Endres, 1987, p. 291 and passim, Kirzner, 1976, pp. 54-58, Mises, 1966, pp. 479ff, and Streissler and Weber, 1973, p. 232]. The present paper considers Menger's judgment as it applies to the treatment of the time element in the structure of capital. It is argued that the subsequent development of capital theory along formalistic lines (e.g. by Wicksell) rather than along subjectivist lines (e.g. by Mises) provides some justification for Menger's use of the superlative: "one of the greatest errors."
            Capital theory is beset with many perplexities and ambiguities. Most of the theoretical difficulties stem from the fact that capital has no natural unit of account corresponding to worker-hours of labor and acres of land. The "quantity of capital," then, has no clear meaning. If capital is reckoned in physical terms, then gauging the total quantity of it involves an insurmountable aggregation problem; if it is reckoned in value terms, then the quantity of capital becomes dependent upon its own price. Similar difficulties are associated with the notion of production time or the degree of roundaboutness of production processes. If two processes are compared strictly in terms of their respective blueprints, it may be unclear which of the two is the more roundabout; if capital values are used in gauging the comparative degrees of roundaboutness, then the comparison will depend in a critical way on the rate of interest used to calculate the capital values. Attempts to spell out the precise relationship between the rate of interest and the degree of roundaboutness are bound to run afoul of these difficulties—as was roundly demonstrated during the controversies of the 1960s over "technique reswitching" and "capital reversing."(1)
            Such perplexities and ambiguities, however, are largely if not wholly irrelevant to the early development of capital theory. What is important about the theoretical developments over the final thirty years of the nineteenth century is the new vision of capital and of a capital-using economy. Essential to this new vision were the ideas that using capital takes time, that time, in fact, is one of the dimensions of the economy's capital structure. Production time, or the degree of roundaboutness, was recognized—and highlighted—as an object of choice to be dealt with by economic theory.
            Some treatment of the time element can be found in British economics, particularly in Ricardo's discourse on machinery, and even in early French writing such as that of Turgot. But the Austrian ideas about capital and time constitute a significant break from Classical doctrine and from the corresponding vision of capital. Dominated as it was by agricultural production, Classical economics treated the time element in production as a technological datum. The very nature of agriculture dictated that the production period, the period for which wages had to be "advanced" from the capitalists to the laborers, was one year. Formal economic theory was required to take the time constraint into account, but it was not required to account for the time constraint itself. The new vision required the treatment of time as an endogenous variable in any theory of a capital-using economy. Characterizing it further requires that we speak of visions and recognize the differences between the early visionaries, particularly between Menger and Böhm-Bawerk.
            Menger's and Böhm-Bawerk's contributions can be assessed in the light of the distinction made by Ludwig Lachmann [1969, pp. 89-103 and 1978, pp. 8ff and passim] between two antithetical methods of analysis: subjectivism and formalism. For subjectivists, economic phenomena can be made intelligible only in terms of the intentions and plans of market participants; for formalists, economic magnitudes, such as inputs, outputs, and production time, can be related to one another without specific reference to the plans and actions of individuals.
            There are several related, but not perfectly synonymous methodological contrasts: causal-genetic analysis and simultaneous determinacy; market-process analysis and equilibrium theory; microeconomic analysis and macroeconomic modeling. Lachmann's distinction, however, seems appropriate for understanding the early developments in Austrian capital theory. Menger was a thoroughgoing subjectivist; Böhm-Bawerk straddled the fence between subjectivism and formalism. His formalism underlies what Menger saw as one of the greatest errors; his subjectivism allows for an interpretation thoroughly consistent with Menger's own work. 

    II. The Mengerian Formulation.
    Menger [1950, p. 152] presented his vision of a capital-using production process in terms of consumers' goods, or goods of the first order, and capital goods, or goods of the second order, third order, and higher orders. The time element of the production process was represented as a sequence of orders. Higher-order (capital) goods are transformed sequentially into goods of lower and lower order until ultimately they emerge as consumers' goods. The time element in the production process was built right into the concept of capital. The relationship between the quantity of (future) consumption and the time spent transforming goods through the several orders was made clear:

      The transformation of goods of higher order into goods of lower order takes place, as does every other process of change, in time. The times at which men will obtain command of goods of first order from goods of higher order in their present possession will be more distant the higher the order of these goods. While it is true ... that the more extensive employment of goods of higher order for the satisfaction of human needs brings about a continuous expansion in the quantities of available consumption goods, this extension is only possible if the provident activities of men are extended to more distant time periods [ibid., p. 152f].
    Menger summarizes this relationship by stating that "economizing men can ... increase the consumption of goods available to them ... but only on condition that they lengthen the periods of time over which their provident activity is to extend" [ibid., p. 153]. This statement not only captures the fundamental notion that underlies Menger's conception of capital but anticipates much of the work that was to be done in subsequent years by Böhm-Bawerk.
            In the light of modern capital controversies, it might be objected that the relationship between future consumption and the required production time holds only for some types of changes in the production process but not for others. It holds, that is, for capital "deepening" but not for capital "widening." Alternatively, it could be objected that for Menger's claims to be true, time periods cannot be reckoned in pure time units. An increase in the number of transformations that are undertaken simultaneously must be seen as involving more "time." The simple time dimension in Menger's formulation may have to be replaced with Böhm-Bawerk's more complex dimension of roundaboutness or, equivalently, with Cassel's concept of "waiting" [1903, p. 54], whose units measure both value and time. While these objections call for a clarification of the notion of production time in Menger's vision of capital, they do not constitute a wholesale condemnation of the vision. At this early stage in the development of Austrian capital theory, addressing such ambiguities is far less important than understanding the basis for the vision and the issues that the vision helped resolve.
            The reasoning that underlay Menger's concept of capital is no mystery. Nature, on its own, yields up some goods that can be consumed by man. The quantity and characteristics of these goods, however, are independent of the "wishes and needs" of the consumers. But man can, to some extent, take charge of the natural process before the consumers' goods emerge. He can exert some influence on these processes with the result that the eventual product is no longer independent of his wishes and needs. Instead, production will be determined, at least in part, by "human purposes" [Menger, 1950, p. 73f]. Clearly, the earlier in the process that man takes charge, the greater are his opportunities for influencing the final product. The earliest point at which man begins to influence the course of nature, earliest with respect to the wishes and needs he is attempting to accommodate, marks the highest order of the production process. It marks the earliest point for which there is an economic relationship between man's purposes and the resulting quantities and characteristics of the goods of the first order which are to
    serve those purposes.
            Carefully formulated in terms of purposes and "wishes and needs" (rather than in terms of physically defined inputs and outputs), Menger's arguments were exempt from the sort of criticism that was aimed at subsequent formulations. The idea, for instance, that constructing, say, a hydro-electric power station may involve less production time than building a redwood fence—owing to the many years required for the redwood tree to grow—is seen immediately to be foreign to Menger's formulation. The age of the redwood tree has no bearing whatever on the issue. Production time is meaningfully discussed only in terms of man's designs on the redwood relative to the satisfaction of his desire for a redwood fence.(2) A second sort of criticism to which Menger's conception is immune stems from the fact that physically defined inputs may continue to be used almost indefinitely. Iron mined in Roman times may be present in a modern pocket knife [Schumpeter, 1954, p. 908]. But unless it is argued that the Roman miners were motivated, at least in part, by the present-day demand for pocket knives, this physical continuium is an economic irrelevancy [See Kirzner, 1966, p. 88].
            The prehistory of production processes was excluded by Böhm-Bawerk as well as by Menger, although the grounds for exclusion were fundamentally different. Bö"hm-Bawerk [1959, vol. 2, p. 86] simply argued that production activities of the remote past are so heavily discounted that they can be safely neglected. Higher powers of the discount rate are sufficiently close to zero that, for practical purposes, they are equal to zero. But according to Menger, unless the earlier activities can be linked to the later production processes in terms of "human purposes," they should be ignored on principle. That is, in the absence of a teleological connection, even those early inputs that are not mathematically insignificant are still economically irrelevant.
            Menger's vision of the capital-using production process is important in its own right and for the understanding of subsequent developments. Two closely related aspects of his value theory, however, can be singled out as having special significance. They constitute a sharp break from the Classical cost-of-production theories and a high-water mark in the development of Austrian capital theory not to be surpassed for another two generations. First, the direction of value imputation was reversed by Menger. A consumption good is not valued on the basis of the labor and other means of production that were used to produce it. Rather the means of production are valued on the basis of the prospective value of the consumption good. Menger expresses this relationship in terms of goods of different orders. "The value of goods of higher order is always and without exception determined by the prospective value of goods of lower order in whose production they serve" [ibid., p. 150].
            Modern theorists may lament Menger's choice of terminology. Higher-order goods can be easily misinterpreted as "highly finished goods," a term used by Alfred Marshall. The terminology can clash with diagrammatic representations that show goods-in-process ascending through time from higher to lower orders. It can also clash with the fact that higher-order goods are discounted with respect to lower-order goods: Higher order means lower value. But to lament Menger's terminology is, more than likely, to overlook one of his most important insights. Consumption goods are lower, or more basic, in the logical structure of his value theory. His terminology captures the insight that the value of higher-order goods is logically dependent upon (rests upon the foundation of) the value of lower-order goods.
            The second aspect of Menger's value theory which deserves emphasis concerns the treatment of entreprenurial expectations. The notion of various orders of goods requires that time be taken into account, but Menger was careful not to introduce the time element in an overly restrictive way. His theory dealt with the valuation of higher-order goods at a point in time when the corresponding consumers' goods still lay in the future. That is, his was a forward-looking theory. In Menger's own words "The principle that the value of goods of higher order is governed, not by the value of corresponding goods of lower order of the present, but rather by the prospective value of the product, is the universally valid principle of the determination of the value of goods of higher order" [ibid., p. 151].
            Consistently taking a prospective view rather than a retrospective view is not just a matter of style. It is a means of focusing the analysis on expectations and hence on the role of the entrepreneur. Significantly, Menger excluded from his formulation any particular, or determinate, expectational scheme:
      The prospective value of goods of lower order is often—and this must be carefully observed—very different from the value that similar goods have in the present. For this reason, the value of the goods of higher order by means of which we shall have command of goods of the lower order at some future time is by no means measured by the current value of similar goods of lower order, but rather by the prospective value of the goods of lower order in whose production they serve [ibid., p. 152].
            Menger avoided the assumption of static expectation (Prospective value in his formulation is different from present value), and of any other particular type of expectations (He did not specify just what the difference is). The use of any particular or determinate expectational scheme would have had the effect, as modern theorists have recognized [e.g. Hicks, 1976], of virtually eliminating the time element from the theory. If, for instance, it is specified that expectations are static (or that they are elastic or inelastic to some specified degree), then the past, the present and the future become analytically indistinguishable. The vision of a capital-using process becomes one of a constant flow of inputs and outputs (Clarkian synchronization), or it becomes one in which changes are always anticipated and reacted to in predictable ways (equilibrium growth). In such "meta-static" visions the choice between taking a prospective view and taking a retrospective view is just a matter of style. But by giving free play to expectations, Menger maintained a meaningful distinction between the present and the future. His treatment of the time element cast the entrepreneur in a key role. The "correctness" of the valuations of higher-order goods depends critically on entrepreneurial abilities. 

    III. Böhm-Bawerk's Reformulation.
    Böhm-Bawerk's contribution to the theory of capital must be viewed with ambivalence in many respects. His lengthy and often tedious formulations, particularly his treatment of production time, give great scope for selective reading and are subject to diverse interpretations. The concept of "period of production" was seen by John B. Clark (and later by Frank Knight) to make no sense at all. It was seen by Wicksell to make a good deal of sense, but primarily from a technological point of view. For Keynes, the concept of production time was logically sound but referred to an utterly trivial aspect of the production process. And for Mises, it was the vital concept in capital theory but had meaning only when given a thoroughly subjectivist interpretation.
            Assessing the detail of Böhm-Bawerk's reformulation must be consistent with the more general assessment of his work. Is his Positive Theory a precise and definitive statement of the economic relationships that constitute capital theory, or is it a crude and skeletal outline of these relationships? Assessments can be found to support either view:

      Böhm-Bawerk's scientific work forms a uniform whole. As in a good play each line furthers the plot, so with Böhm-Bawerk every sentence is a cell in a living organism, written with a clearly outlined goal in mind.... And this integrated plan was carried out in full. Complete and perfect his lifework lies before us. There cannot be any doubt about the nature of his message.
    Alternatively:
      Böhm-Bawerk's work [was not] permitted to mature: it is essentially (not formally) a first draft whose growth into something much more perfect was arrested and never resumed. Moreover, it is doubtful whether Böhm-Bawerk's primitive technique and particularly his lack of mathematical training could have ever allowed him to attain perfection. Thus, the work, besides being very difficult to understand, bristles with inadequacies that invite criticism—for instance, as he puts it, the "production period" is next to being nonsense—and impedes the reader's progress to the core of his thought.
    These two passages provide a remarkable contrast, all the more remarkable when it is realized that both were written by one and the same Joseph A. Schumpeter.(3) A study of Böhm-Bawerk's text itself, together with a survey of the critical and interpretive work of others, suggests that the second quoted passage is closer to the truth. It will be argued, though, that it was not his lack of mathematical training that stood in the way of perfection; rather it was his failure to adhere to Menger's subjectivism.
            To treat Böhm-Bawerk's work as a valuable but rough diamond is to suggest that a detailed treatment of his argumentation may be unproductive. To focus on the simple calculations of the average period of production, or to question why the arithmetic average rather than, say, the geometric average was used is to overlook the more fundamental and more important message: Capital theory must be concerned with the time dimension in the economy's structure of production.
            Abstracting from details, Böhm-Bawerk's vision of a capital-using economy is very similar to Menger's. He depicts the essentials of a production process with a number of concentric circles [Böhm-Bawerk, 1959, vol. 2, p. 106]. Time is seen as progressing radially outward from the center of the circles. The center, then, represents the production process in its incipiency, and the outermost ring represents the ultimate fruition of the process. Menger's first, second, and higher orders become Böhm-Bawerk's first, second, and higher "maturity classes." The new terminology, like tha old, is subject to misinterpretation: The least mature capital is in the highest maturity class. (Was Böhm-Bawerk in his choice of terminology attempting to retain Menger's fundamentally subjectivist insights?)
            For the special case of the stationary state, the concentric rings have two interpretations: (1) The areas of the different rings can represent the amount of different kinds (maturity classes) of capital that exist at a given point in time, or (2) the initial inputs of the production process can be seen as radiating outward through the several maturity classes until they finally emerge at the outermost ring as consumers' goods.
            Depicting the stationary state, however, was of very little interest to Böhm-Bawerk. He briefly considers the question [ibid., p. 109]: "What is the procedure if we wish just to preserve the amount of capital in its previous magnitude?" This issue is disposed of in short order. His answer serves primarily as a stepping stone to the more important question of "What must be done, if there is to be an increase in capital?" The answer to this second question involves some sort of a change in the configuration of the concentric rings. Several types of changes are suggested, each involving the idea that real saving is achieved at the expense of the lower maturity classes (a thinning of the outer rings) and that the saving makes possible the expansion of the higher maturity classes (a padding of the inner rings) and the creation of higher maturity classes than had previously existed. Böhm-Bawerk [ibid., p. 112] even hints that in a market economy, it is the entrepreneur who brings such structural changes about and that their action is guided by changes in relative prices of capital in the various maturity classes. The most fundamental message in his discussion is clear: An increase in capital is not to be viewed as a simultaneous and equiproportional increase in each of the maturity classes. It is to be viewed as a change in the relationship between the maturity classes.
            This formulation is unobjectionable on Mengerian grounds, and it is compatible with the developments by subsequent capital theorists in the Austrian school. What is to be lamented—even by Böhm-Bawerk's followers—is that he tried to capture this important insight into the relationship between the maturity classes with a single number, the average period of production. The attempt to stipulate just how such an average period could be calculated led Böhm-Bawerk away from Menger's forward-looking vision in which (subjective) values are to be gauged by entrepreneurs. Calculating the average period of production required that the production process be formulated in terms of physically defined inputs, physically defined outputs, and the span of calendar time between them. Examples of production processes which allowed for the calculation of the average period of production required the assumption of a steady state, which virtually robbed the production period of its economic relevance. Böhm-Bawerk opened the door to criticisms of the type made by Clark (and later by Knight). It will be argued below that Böhm-Bawerk's vision of a capital-using economy can be defended against Clark's objections only by returning to the Mengerian subjectivist formulation—or, what amounts to the same thing, resorting to a subsequent reformulation, such as that of Mises. 

    IV. Clark as Counterpoint.
    The period of production appears to be calculable only when the production process is described in a very stylized manner, preferably a continual repetition of overlapping point-input/point-output processes. The archetypal case is the forest in which a fixed number of trees is cut each year, and simultaneously the same number of seedlings is planted. If the forest is characterized by a linear maturity structure, then it can be maintained in an unchanged state year after year while production and consumption (planting and cutting) proceed simultaneously. This will be recognized as Clark's most celebrated example [1924, p. 313f] of the synchronization of production and consumption.
            In Clark's example, the forest consists of twenty acres. Each year the trees on one of the twenty acres are ready for cutting. Technologically speaking, there is a time dimension to the production of wood, which can be expressed precisely and unambiguously. It takes twenty years, in his example, for a tree to mature. Clark's point, of course, is that the time element—undeniable twenty years—is totally irrelevant. The forest should simply be viewed as a perpetual source of wood. Because both planting and cutting take place each year, the inputs and outputs of the process are seen as being simultaneous. The economically relevant period of production, according to Clark, is not twenty years, but zero years. This is the synchronization view of the production process: Production and consumption are simultaneous.
            When Clark collapsed the time element out of the capital-using economy, he hedged his description of the production process: "The planting and cutting are, in a way, simultaneous" [ibid., p. 313, emphasis added]. In other similar statements he used the phrases "so to speak," "as it were," "in a sense," "virtually," etc.(4) The aspects of the production process which these phrases served to mask are the very aspects that are important in the Austrian visions of capital. Production and consumption would appear to be simultaneous only to an outside observer who had no inkling about the "wishes and needs" and the "purposes" of the individuals engaged in the production process. While Menger had grounded his theory squarely on these subjective concepts, Clark had deliberately abstracted from them.
            Böhm-Bawerk was not always careful to stress the subjective element in his capital theory, and he often departed from the subjectivist view in order to mathematize his reasoning. But his vision of a capital-using economy makes sense only when interpreted in strict Mengerian terms. To return to the overworked example, the economic significance of the planting of a seedling can be understood only in terms of present "human purposes" and future "wishes and needs." If the essential role of these subjective factors are recognized, then the time element cannot be eliminated from the theory of a capital-using economy without undermining the very basis of the analysis.
            The Mengerian view and the Clarkian view are comparable only for the special case in which future "wishes and needs" are identical with present "wishes and needs," but even then the two views themselves are not identical. One is an explanation of the production activity (Individuals plant now in order to be able to cut twenty years from now), while the other, by coincidence of the identity of the present and the future subjective factors, is descriptive (Individuals are planting and cutting simultaneously). At points in his discussion, Clark suggests a causal connection between the simultaneous planting and cutting. "[Today's cutting] is made practicable by today's planting. The tree that is just set is, then, an enabling cause of the consuming of the one that is twenty years old" [p. 313, emphasis added]. Knight was to go a step further and claim that it is today's planting that produces today's twenty-year-old tree. This is the concept of production that Hayek [1936, p. 214, n. 25] referred to as an "absurd use of words."
            In the context of the Mengerian vision, Clark's formulation must be rejected. Either his causal connection, like Knight's, is absurd, or it simply obscures the time element by assuming that the subjective factors remain fixed through time. Clearly, Stigler [1941, pp. 308-15] adopts the Clarkian vision as his basis for assessing the debate between Clark and Böhm-Bawerk over the synchronization of production and consumption. He approvingly details the activities of a stationary economy in terms of a slowly maturing forest—this time one in which the trees reach maturity in fifty years. Maintaining the forest is a technological detail; capital is permanent; it yields a perpetual income; production and consumption are simultaneous. Böhm-Bawerk's entire analysis is dismissed with the statement "We can say that any one row [of trees] takes fifty years to mature, but since there is a constant output of timber forever, there is simply no point in saying it" [ibid., p. 313].
            But if, as Menger believed, capital theory is to establish causal relationships, Stigler's statement should be turned on its head. It is the anticipation of future wants and needs that causes individuals to plant trees now. Simply to state that "there is a constant rate of output of timber forever," is to build into the theory a specific and rigid constellation of future wants and needs, anticipations, and production decisions in the guise of a technological datum. It is true, of course, that if this specific constellation materializes, then the Clarkian view will be descriptive of the associated production and consumption activities. In this particular—but, according to Böhm-Bawerk, not particularly interesting—case, we can say that individuals are producing and consuming simultaneously, but there is simply no point in saying it!
            The purpose of contrasting the Böhm-Bawerkian and the Clarkian visions of a capital-using economy is not simply to defend Böhm-Bawerk against his critics. Rather it is to call attention to the difference between formalism (Clark) and subjectivism (Menger) in the development of capital theory. The assumption of the stationary state has a different significance for each. For Clark's theory of capital, it is a technologically defined benchmark essential for theorizing about the stocks and flows that characterize a capital-using economy; for Böhm-Bawerk's theory, interpreted Mengerianly, it is an uninteresting special case. This contrast is nowhere more obvious than in the Clark-Böhm-Bawerk debate, but it is no less relevant in the assessment of Böhm-Bawerkian capital theory as formalized by Knut Wicksell.

    V. Wicksell and the Formalization of Böhm-Bawerk's Theory.
    A full assessment of Wicksell's contribution to capital theory would reveal much commonality between Austrian and Swedish theorizing. Discussions in Wicksell's Lectures [1934, vol. 1, pp. 158-66], for instance, bear a close relation to Hayek's treatment of the structure of production. The issues relevant to the present paper, however, require attention to Wicksell's formalism as it affected the Austrian-Swedish transformation of the concept of capital.
            Aspects of Böhm-Bawerk's contributions that drew criticism from Menger were, at the same time, an inspiration to Wicksell. The direction of the Swedish development can be identified in terms of Schumpeter's critical evaluation of Böhm-Bawerk's contribution cited earlier. Two claims were made in that evaluation, one that "it is doubtful whether Böhm-Bawerk's ... lack of mathematical training would have allowed him to attain perfection," the other that the concept of the "period of production" is inadequately developed and "impedes the reader's progress to the core of his thought." Schumpeter appears to be of two minds here—unless he is making the implausible judgment that the use of more sophisticated mathematical techniques would hasten the reader's progress to the core of Böhm-Bawerk's thought.
            The evaluation suggests two separate directions for development. The first claim suggests that Böhm-Bawerk's work is to be perfected by replacing the crude arithmetic examples with more sophisticated mathematics—presumably, a system of simultaneous equations. These equations would constitute a formal statement of the equilibrium relationships implied by Böhm-Bawerk's arithmetic. The second claim suggests that the essential concepts in his theory, such as the period of production and the degree of roundaboutness, could be made more intelligible by relating them to the purposes and choices of market participants. This would constitute the subjectivization of the relationships postulated by Böhm-Bawerk.
            Wicksell undertook to incorporate Böhm-Bawerk's ideas about capital-using production processes into a general equilibrium framework consistent with that of Walras's. This formalization of the theory, however, caused the essential concepts to lose economic relevance and intelligibility. If the relationship between the time element in the production process and the economic choices of individuals had become tenuous in Böhm-Bawerk's reformulation of Menger's theory, it became even more so in Wicksell's formalization of Böhm-Bawerk's. The assessment of Böhm-Bawerk's work applies a fortiori to Wicksell's. The relationships among the economic magnitudes—particularly those relationships involving the time element—make sense only when given a Mengerian interpretation.
            Not surprisingly, Stigler finds the Swedish development to be nonsense. Wicksell's entire theoretical framework is judged invalid because it contains that erroneous Austrian time element. Stigler [1941. p. 278] considers Wicksell's theory in detail only after denying, in effect, that it has any economic relevance. "[H]is technical analysis ... deserves praise on the score of elegance, but it becomes primarily a display of technique."

    VI. The Austrian-Swedish Transformation.
    Wicksell indicated a fundamental agreement with what he saw as the central thesis of the Positive Theory of Capital. "Böhm-Bawerk's main formula of the explanation of interest—that interest is an agio or a premium which arises from the exchange between present and future goods—is quite correct and appropriate" [Wicksell, 1954, p. 21]. But judging Böhm-Bawerk's presentation to be "rather clumsy" and his theory incomplete, Wicksell set about rectifying these problems by recasting the theory in a more sophisticated language. His mathematical reformulation constituted a gain in his view in terms of "simplicity and perspicuity." Wicksell also incorporated an additional variable to represent land and what he calls "rent goods." Although he considered this extension to be "pretty obvious," it was essential for achieving a "generalization" of his theory. With the additional variable he could set down five equations that would state "exactly" and "for the first time ... the relationship between the main economic factors, labour, land and capital" [ibid., p. 21]. There are further differences between the theories of Böhm-Bawerk and Wicksell which are of particular relevance to the issues of subjectivism vs. formalism. Unearthing these differences requires that we go beneath the simultaneous equations and arithmetic examples and examine the respective concepts of capital that the theories employ.
            Böhm-Bawerk's lengthy discussion of alternative capital concepts [vol. 2, pp. 10-101] constitutes a taxonomy of capital in which he identifies such categories and (overlapping) subcategories as national capital, social or productive capital, and private or acquisitive capital. For his own theoretical purposes, he was interested in a concept of capital that suited his vision of a capital-using economy. His vision makes his definitions understandable.
            Two defining statements can be found in the opening pages of Böhm-Bawerk's Positive Theory. "Capital is nothing but the sum total of intermediate products which come into existence at the individual stages of the round-about course of production" [ibid., p. 14]. And what amounts to the same thing, "We may define social capital as an aggregate of products which serve as the means of the acquisition of economic goods by society" [ibid., p. 32]. apital in its relationship to the multi-stage process of production is what Böhm-Bawerk called social or productive capital. "Private capital" includes uninvested resources, "national capital" includes capital loaned abroad, but these categories were not so closely related to his theory and hence were not of primary interest to him.
            By contrast, Wicksell's definition of capital has little or no relationship to the actual process of production. Capital in his formulation is taken to be "all interest bearing (material) goods" [Wicksell, 1954, p. 105]. His definition is meant to include all durable consumer goods as well as all means of subsistence. No attempt is made to sort out those capital goods that are used in the multi-stage production process from those that are not. Durability is the only basis for defining subcategories. "Capital in the wider sense" includes both producers' durable goods and consumers' durable goods; "Capital in the narrow sense" excludes both. "Highly durable goods" are identified as "rent goods" and are treated as equivalent to land.
            Stigler [1941, p. 272] sees Böhm-Bawerk's formulation as a "complex and artificial classification of types of capital." He sees Wicksell's formulation as incorporating a "vastly improved capital concept." This assessment reflects the rejection of the Austrian vision of a capital-using economy and the adoption of the Clark-Knight vision. But in the context of Böhm-Bawerk's vision, interpreted Mengerianly, Stigler's assessment is largely unjustified. Three separate but related issues, involving the subsistence fund, the treatment of durable goods, and the relationship between social classes and their functions, will be considered in some detail. Only on the third issue, it will be argued, does Wicksell's formulation constitute an improvement over Böhm-Bawerk's.
            (1) One alleged improvement of Wicksell's formulation over Böhm-Bawerk's involves the treatment of the subsistence fund. Stigler contrasts the Swedish with the Austrian version by saying that the former includes the "means of subsistence of laborers, even when these means are owned by entrepreneurs" [ibid.]. Contrary to the implications of Stigler's wording, Böhm-Bawerk excluded all means of subsistence from his concept of social or productive capital. These means consist of consumption goods, and Böhm-Bawerk [vol. 2, p. 71] insisted that "the concept 'means of production' is thought of as an antithesis of the concept 'consumption goods' and is intended to be just that." Restating for emphasis [ibid., p. 72], "The goods with which the working members of the community feed, warm and clothe themselves are goods of direct consumption use; they are not means of production." Again, Böhm-Bawerk's concept of capital is tailored to fit his vision of the production process. The relevant distinction is that between goods used directly by consumers and goods used in the multi-stage production process.
            Wicksell's treatment of the subsistence fund is not as clear as Stigler's discussion suggests. Two different criteria for identifying means of subsistence as capital are offered by Wicksell [1954, p. 104]. According to one, the means of subsistence "should be thought of all the time as capital until the moment when they find themselves in the possession of the consumers concerned." Clearly, Stigler's "even" should be changed to "only": Wicksell included the means of subsistence only when these means are owned by entrepreneurs. This criterion, which is completely consistent with Böhm-Bawerk's concept of private, or acquisitive, capital, could hardly be an improvement in Stigler's judgment. It commits the "classical error ... [of] making ownership a criterion of capital, when in fact this aspect is completely irrelevant" [Stigler, 1941, p. 197].
            According to the alternative criterion, the subsistence fund maintains the status of capital "up to the moment of consumption" [ibid.] Undoubtedly, it was this criterion that won Stigler's approval. It has that distinct Clark-Knight flavor. But from an Austrian perspective, this reformulation is a decided change for the worse. The category of capital goods is now so broad that it leaves no room at all for the category of consumption goods. Wicksell's formalistic approach led him to abstract from the process of production that was so emphasized by Böhm-Bawerk and to adopt a concept of capital that was more compatible with his system of equations than with the underlying economic realities. Wicksell was only a short step from rejecting altogether the distinction between capital goods and consumer goods and adopting in its place the dimensional distinction between stocks and flows. A thorough reformulation on this basis would represent a complete victory of form over substance—the ultimate consequence of the formalistic approach.
            (2) A second alleged shortcoming of Böhm-Bawerk's capital concept that was overcome by Wicksell involved the treatment of durable consumer goods. Their exclusion, according to Stigler [ibid.], constitutes a "major omission in Böhm-Bawerk's capital concept." Durable consumer goods, of course, can be considered as capital in some sense of the term, as has been recognized by theorists writing in the Austrian tradition from Menger to the contemporary Austrian theorists. Menger's view of capital, for instance, focuses on the good's "usefulness" and can be taken, as Stigler acknowledges, to include consumer durables. But capital in this very broad sense involves a distinction that is completely arbitrary if the line between capital goods and consumer goods is to be drawn short of the stock-flow distinction.
            More importantly, such a distinction is unrelated to the theoretical problems with which Böhm-Bawerk was dealing. To call a dining-room table a capital good yielding a stream of services and a table cloth a consumer good is to play word games. To call both the table and the cloth capital goods yielding streams of services is to nullify the concept of consumer goods. More to the point, neither manner of classifying table and cloth is particularly helpful in the analysis of the economy's multi-stage production processes. The problem of distinguishing between consumer goods and capital goods outside the context of a production process is not strictly related to the durability of the good. Mises [1966, p. 94] calls attention to the inherent element of arbitrariness in such distinctions:

      It is ... superfluous to enter into pedantic discussions of whether a concrete good has to be called a good of the lowest order or should rather be attributed to one of the higher orders. Whether raw coffee beans or roast coffee beans or ground coffee or coffee prepared for drinking or only coffee prepared and mixed with cream and sugar are to be called a consumers' good ready for consumption is of no importance.
    For a distinction between capital goods and consumer goods to be meaningful, it must have some relationship to the theory in which it is employed. Böhm-Bawerk's theory dealt with the process of production conceived as a sequence of maturity classes or stages of production. He was concerned with the allocation of resources among these stages and with the economic process by which resources are shifted from one stage to another. The relationship between the various stages of production can best be analyzed with a capital concept that excludes goods in the hands of their ultimate consumers—whether or not those goods are durable.
            The Austrian concept of capital was not derived from the "classical error of making ownership the criterion of capital." Rather it was designed to reflect the one-way passage of goods from the lowest stage of production (retailing) into the hands of consumers. The general lack of effective markets for second-hand consumer goods creates a relevant boundary between capital goods and consumer goods. A consumer can resell a consumers' good, say, an overcoat, but only at a great loss. To adopt this interpretation of Böhm-Bawerk's capital theory is to suggest that the is not ownership or even durability per se but remarketability. Goods that cannot be readily resold by consumers are not to be considered as part of the economy's capital structure.(5) Goods such as houses and (less so) automobiles, for which there are effective secondary markets, should be considered capital goods even from an Austrian point of view. Such remarketable goods can be shifted back into business inventories (the lowest maturity class) in response to changing supply and demand conditions. They may eventually be absorbed into even higher maturity classes, as in the case of a dwelling that is converted to a business use. For most other goods in the hands of consumers, however, the potential for such absorption into higher maturity classes is negligible.
            Replacing the criterion of durability with the criterion of remarketability refocuses our thinking about capital goods in the hands of consumers. It does not cause a great change in the goods normally thought of as capital goods. Durability, after all, is one of the characteristics that makes a good remarketable. But durability is not a sufficient characteristic. An automobile and an overcoat, for instance, may be equally durable, but the former is more plausibly categorized as a capital good than is the latter—which is to say that remarketability better describes the distinction that is typically made. There are, of course, borderline cases in distinguishing capital goods and consumer goods. There is no black-and-white distinction between goods that are remarketable and goods that are not, but the distinction does have relevance. It serves to tailor-fit the concept of capital to Austrian capital theory.
            All this is foreign to Wicksell's formulation. The need for any such distinction between capital goods and consumer goods is obscured by the formalization of the theory and the homogenization of the capital concept. A mathematical formulation is, by its very nature, ill-suited to incorporate distinctions that are based on such things as the purposes for which the good is to be used or on the notion of remarketability. It is much better suited to distinguish between dimensionally different magnitudes such as the stocks and flows of the Clark-Knight vision.
            The nature of mathematical or formalistic analysis was clearly recognized by Böhm-Bawerk in his treatment of Alfred Marshall. His critical remarks focused on Marshall's well known discussion of whether such goods as furniture and clothing are to be considered as productive capital.
      It would, to be sure, be possible and conceivable to count as income the benefits derived from the utilization of [furniture and clothing]. And it is also true that such a reckoning might appear admissible in the strictly mathematically treatment of the entire distribution problem [Böhm-Bawerk, 1959, vol. 2, p. 30].
    Böhm-Bawerk clearly saw the formalistic relationship between capital and income, but he was concerned that a strictly mathematical treatment would divert attention from the capital-using production process and the markets for the goods that are associated with the various maturity classes. The potential danger of formalism is emphasized when Böhm-Bawerk criticized the ideas of one theorist who
      could at times strangely distort [his ideas] into shapes that failed entirely to coincide with the lines of the foundation on which they were reared. I cite as an example [the] extension of the concept of social capital to include all sorts of personal traits, talents and skills. These certainly present an odd appearance, decked out as component elements of a "stock" and they were destined, like the spirits unwisely exorcized, to bedevil capital theory for many a long year [ibid., p. 22].
    These critical remarks about the notion of "human capital" were directed not at Wicksell nor even at Clark; they were directed at Adam Smith. Clearly, Böhm-Bawerk's opposition to formalism is not at all explained by his lack of mathematical training. He objected to the purely dimensional distinction whatever the mode of expression. But in defending his own formulation against the classical view, he managed to rescue substance from form only to have it relinquished once again by Wicksell.
            In sum, Böhm-Bawerk did exclude from his definition of social capital durable goods in the hand of their ultimate consumers. This definition did not involve a gross oversight, nor was it adopted and defended in ignorance of the mathematical relationship between stocks and flows. Its purpose was to direct attention to the economic relationships between the various stages of the multi-stage production process. While Wicksell's reformulation may have achieved greater "generality, simplicity, and perspicuity," it directed attention away from the production process.
            (3) The third issue to be considered stems from Böhm-Bawerk's peculiar treatment of the category of private capital and with Wicksell's corresponding generalization. It is here that Stigler's charge of artificiality has merit. But much lies behind this issue. The significance of the peculiar way in which private capital is treated derives from Böhm-Bawerk's theory of interest—or more accurately, from his particular formulation of that theory.
            Well known are his "three main grounds" for the existence of a positive rate of interest [ibid. pp. 265-89]. The first two grounds, having to do with individuals' estimates of their future incomes and with their present evaluation of future utilities were explicitly introduced as subjective factors. Böhm-Bawerk referred to these as "the factors of provision and perspective." The third ground involves subjective considerations as well in so far as it accounts for value differences as opposed to physical differences between present goods and future goods, but it is offered as a technological factor. Because of the increased productivity of roundabout production processes, present goods are preferred to future goods.
            Less well known is the differential way in which Böhm-Bawerk, adopting in this regard a very Classical viewpoint, applied the above reasoning to the social classes with which his theory dealt—the entrepreneur, or capitalist, and the worker. The productivity of roundabout processes is the concern of the capitalist alone. More significantly, the workers alone are affected by the factors of provision and perspective, the factors that make up the first and second grounds. That the capitalists are not themselves imbued with time preferences on the basis of the first two grounds is repeatedly asserted. "[I]n his subjective circumstances the capitalist as a rule places the same valuation on a sum of present goods as on an equal sum of future goods" [ibid., p. 353]. The subjective circumstances of the workers and the capitalists are vividly described in the following passage.
      The workers are in urgent need of present goods and there is little or no possibility of their accomplishing anything by working for their own account. To the last man they will prefer to sell their labor cheaply rather than not at all. But a similar situation obtains with respect to the capitalists. As far as conditions go with respect to their particular wants and the satisfaction of them, their present goods (which they would in any event save up for future use) are worth no more to them than an equal quantity of future goods. They will therefore prefer to make any purchase of labor that affords them any agio at all, even if only a small one, rather than allow their capital to lie idle [ibid., p. 354f].
    Thus, after describing at great length the subjective factors that determine the rate of interest and hence the allocation of resources among the various maturity classes of the production process, Böhm-Bawerk built into his theory a particular set of subjective evaluations, in which workers and capitalists were represented by opposite extremes. There was no relevant margin at which these evaluations could guide the process of production; the factors of provision and perspective became, instead, a part of the setting. In modern terms, the supply of labor is perfectly inelastic, and thus the level of employment is supply determined; the wage rate is indeterminate given the similarly inelastic demand conditions. The artificiality about which Stigler complained is seen to be rooted in Böhm-Bawerk's specifying, without justification, the particular evaluations and hence the particular intentions of capitalists and workers.
            Given these specifications, however, Böhm-Bawerk's treatment of private capital as distinct from social capital is understandable. Social capital includes all capital that is invested in the process of production. This excludes the (uninvested) means of subsistence whether in the hands of capitalists or workers. The means of subsistence in the hands of capitalists are private capital, nonetheless, by virtue of the capitalists' intention to invest. They cease to be capital, however, when transferred to the worker by virtue of his urgent need for present goods. True to the subjectivist perspective of Menger and the Austrian school, the means of subsistence are given or denied the status of capital on the basis of the purpose for which they are to be used. Unfortunately, the subjective factors were set in (Classical) concrete by Böhm-Bawerk's characterization of the capitalists and workers.
            The much needed generalization of Böhm-Bawerk's formulation which Wicksell achieved in this respect can be seen clearly by comparing their respective discussions of the loan market. According to Böhm-Bawerk [ibid., p. 20], "all the sellers [of credit] estimate present and future goods as just about equal in value, and all the buyers estimate the value of present goods as higher than that of future goods. In modern terms, the supply of credit is perfectly interest inelastic, and thus the volume of credit is completely supply determined. In contrast, Wicksell [1954, p. 107f., emphasis original] gives free play to subjective valuations on both sides of the market:
      [T]he interest on the loan, just like exchange value in the case of ordinary exchange, will depend on two proportions of market utility; that is to say, it will depend first on the proportion between the marginal utility of present and that of future goods for the creditor, and secondly on the proportion between the analogous marginal utilities for the debtor .... The interest which must really be paid will then fall somewhere between these two different valuations.
    This generalized formulation had the potential of keeping the subjective factors alive. But when Wicksell turned his attention from the theory of interest to the theory of capital, these factors were all but lost. Böhm-Bawerk's third ground for explaining the rate of interest, which was accepted as a purely technological consideration, became the only ground for Wicksell. It was this domination of the technological relationship in his formal theory which allowed him to dispense with Böhm-Bawerk's "complex and artificial classification of types of capital" and to adopt a "single capital concept." If the Swedish transformation appears to be a gain, it is because the subjective factors that lay behind Böhm-Bawerk's classification had been impounded by Classical prescriptions that fixed the valuations and hence the behavior of both capitalists and workers. Menger's concern with intentions and "wishes and needs" had already been eliminated. 

    VII. One of the Greatest Errors?
    Even before the turn of the century, capital theory was being developed in the direction of formalism and away from subjectivism. Böhm-Bawerk's reformulation of Menger's capital theory and Wicksell's formalization of Böhm-Bawerk's put formalism and subjectivism in direct conflict. An assessment of Wicksell's contribution from a Mengerian perspective would be the precise opposite to Stigler's. The indeterminacy of wages resulting from the dramatically differing valuations by capitalists as compared to laborers could not have troubled Menger, whose own theory of prices involved a fundamental element of indeterminacy. But Wicksell's treatment of the subsistence fund and of durable goods would have troubled Menger for reasons already suggested. The formal stock-flow distinction diverts attention from the subjective relationship between the lower-order and higher-order goods that make up the production process.
            Modern economists who prefer to theorize in terms of the dimensionally distinct stocks and flows will not see the "greatest error" as an error at all. But those who applaud the recent Austrian resurgence, who prefer subjectivism to formalism, and who believe that the time element in the economy's multi-stage structure of production is trivialized by the "average period," are likely to agree with Menger and to see Böhm-Bawerk's error as "one of the greatest."
     

    Bibliography

    Böhm-Bawerk, Eugen 1959. Capital and Interest, vol. 2 [1889]. South Holland, IL: Libertarian Press.

    Cassel, Gustav 1903. The Nature and Necessity of Interest. London: McMillan and Co., Ltd.

    Clark, John Bates 1924. The Distribution of Wealth. New York: MacMillan and Co.

    Endres, A. M. 1987. "The Origins of Böhm-Bawerk's 'Greatest Error':Theoretical Points of Separation from Menger," Journal of Institutional and Theoretical Economics, vol. 143, pp. 291-309.

    Garrison, Roger W. 1985. "A Subjectivist Theory of a Capital-Using Economy." In Gerald P. O'Driscoll, Jr. and Mario J. Rizzo, The Economics of Time and Ignorance. Oxford.

    Hayek, Friedrich A. 1936. "The Mythology of Capital," Quarterly Journal of Economics, vol. 50 (February), pp. 199-228.

    Hicks. John R. 1976. "Some Questions of Time in Economics." In Anthony M. Lang et al., eds., Evaluation, Welfare and Time in Economics. Lexington, MA: D.C. Heath and Co., pp. 135-51.

    Kirzner, Israel M. 1966. An Essay on Capital. New York: Augustus M. Kelley.

    _________, 1976. "Ludwig von Mises and the Theory of Capital and Interest." In Laurance S. Moss, ed. The Economics of Ludwig von Mises. Kansas City: Sheed and Ward, Inc.

    Lachmann, Ludwig M. 1969. "Methodological Individualism and the Market Process." In Erich Streissler et al, eds. Roads to Freedom. London: Routledge and Kegan Paul, pp. 89-103.

    ________, 1978. Macro-economic Thinking and the Market Economy. [1973] Studies in Economics No. 6, Menlo Park, CA: Institute for Humane Studies [first published in 1973].

    Menger, Carl 1950. Principles of Economics [1871]. New York: The Free Press.

    Mises, Ludwig von 1966. Human Action, 3rd rev. ed., Chicago: Henry Regnery and Co.

    Rolph, Earl R. 1980. "On Austrian Capital Theory," Economic Inquiry, vol. 18 (July) pp. 501-3.

    Rothbard, Murray N. 1970. Man, Economy, and State, 2 vols. [1963]. Los Angeles: Nash Publishing Co.

    Schumpeter, Joseph A. 1951. "Eugen von Böhm-Bawerk: 1851-1914," Ten Great Economists. Oxford: Oxford University Press, pp. 143-90.

    __________, 1954. History of Economic Analysis. Oxford: Oxford University Press.

    Stigler, Geroge J. 1941. Production and Distribution Theories. New York: MacMillan and Co.

    Streissler, Erich and Wilhelm Weber 1973. "The Menger Tradition." In John R. Hicks and Wilhelm Weber, Carl Menger and the Austrian School of Economics. Oxford: Oxford University Press, pp. 226-32.

    Wicksell, Knut 1954. Value, Capital and Rent (1893). New York: Rhineholt and Co., Inc.

    __________, 1977. Lectures in Political Economy (1934), 2 vols. New York: Augustus M. Kelley.
     

    Notes

    *The Author is Associate Professor of Economics at Auburn University, Auburn, AL 36849. Comments on earlier drafts by Greald P. O'Driscoll, Jr. (Federal Reserve Bank of Dallas), Parth Shah (Auburn University), and an anonymous referee are gratefully acknowledged. The author thanks the Ludwig von Mises Institute for financing his participation at the Menger conference.

    1. For an treatment of this controversy from a subjectivist point of view, see Garrison, Roger W. 1985. "A Subjectivist Theory of a Capital-Using Economy." In Gerald P. O'Driscoll, Jr. and Mario J. Rizzo, I. Oxford, pp. 181-84.

    2. The primacy of human purposes is often overlooked. One long-time critic of Austrian capital theory [Rolph, 1980, p. 502] called the concept of roundaboutness into question by remarking that "Not infrequently, a squirrel must be given credit for planting the proverbial acorn."

    3. The first-quoted passage was written on the occasion of Böhm-Bawerk's death [Schumpeter, 1951, p. 146]; the second-quoted passage is from Schumpeter [1954, p. 847].

    4. I am indebted to Gerald P. O'Driscoll, Jr. for calling my attention to Clark's systemic hedging.

    5. The issue of remarketability is only implicit in Böhm-Bawerk's formulation. Rothbard [1970, vol. 1, p. 320] explicitly takes account of the fact that "consumers' goods, once sold do not ordinarily re-enter the exchange nexus...."