Capital theory is notorious for being perhaps the most controversial area in economics. This has been so ever since the very inception of systematic economic analysis. Much of the interest in the theory of capital lies in the fact that it holds the key to the explanation of profits. Since the notion of ‘capital’ is at the centre of an inquiry into the laws of production and distribution in a capitalist economy, controversies in the theory of capital are reflected in virtually all other parts of economic analysis.

We can distinguish between debates within different traditions of economic analysis and debates between them. In what follows our concern will be mainly with the latter. At the cost of severe simplification, the various traditions in the theory of capital and distribution may be divided into two principal groups, one rooted in the surplus approach of the classical economists from Adam Smith to Ricardo and the other in the demand and supply approach of the early marginalist economists. The so-called ‘Cambridge controversies’ (cf. Harcourt 1969), triggered off by a seminal paper by Joan Robinson (1953), consisted essentially in a confrontation of these two radically different traditions. The debate is still continuing. Currently, the discussion focuses on some of the neoclassical authors’ claim that the classical theory, as it was reformulated by Sraffa (1960), is a ‘special case’ of modern general equilibrium theory. We shall come back to this questionable proposition towards the end of the entry.

The Surplus Approach

The general method underlying the classical economists’ approach to the theory of capital and distribution was that of ‘normal’ or ‘long-period’ positions. These were conceived as centres around which the economy is assumed to gravitate, given the competitive tendency towards a uniform rate of profit. Because of the assumed gravitation of ‘market values’ to the ‘normal’ levels of the distributive and price variables, the former were given little attention only, being governed by temporary and accidental causes, a proper scientific analysis of which was considered neither necessary nor possible. Emphasis was on the persistent or non-temporary causes shaping the economy. Accordingly, the investigation of the permanent effects of changes in the dominant causes was carried out by means of comparisons between ‘normal’ positions of the economic system.

The development of a satisfactory theory to determine the general rate of profit was thus the main concern of the classical economists. As regards the content of this theory, profits were explained in terms of the surplus product left after making allowance for the requirements of reproduction, which were conceived inclusive of the wages of labour (Ricardo 1817, vol. 1, p. 95). As Sraffa (1951, 1960) emphasized, the determination of the social surplus implied taking as data (i) the system of production in use, characterized, as it is, by the dominant technical conditions of production of the various commodities and the size and composition of the social product; and (ii) the ruling real wage rate(s). In accordance with the underlying ‘normal’ position the capital stock was assumed to be so adjusted to ‘effectual demand’ (Adam Smith) that a normal rate of utilization of its component parts would be realized and a uniform rate of return on its supply price obtained. Thus the classical authors separated the determination of profits and prices from that of quantities. The latter were considered as determined in another part of the theory i.e. the analysis of accumulation and economic and social development.

The rate of profit was defined by the ratio between social surplus and social capital, i.e. two aggregates of heterogeneous commodities. Thus the classical theory had to face the problem of value. Ricardo’s ingenious device to solve this problem consisted in relating the exchange values of the commodities to the quantities of labour directly and indirectly necessary to produce them. This led to the first formulation of one of the key concepts in the theory of capital ever since – the inverse relationship between the real wage and the rate of profit (Ricardo, vol. 8, p. 194).

It was not until Marx that additional important steps in the development of the surplus approach were taken. In particular, in Marx the analytical role of the ‘labour theory of value’ in the determination of the general rate of profit was brought into sharp relief. According to him the explanation of profits in terms of the surplus approach would have been trapped in circular reasoning if the value expression of either aggregate (surplus and capital) were to depend on the rate of profit. The measurement of both aggregates in terms of labour values, which themselves were seen to be independent of distribution, was considered a device to circumvent this danger and provide a non-circular determination of the rate of profit, \( r=s/\left(c+v\right) \), where r is the general rate of profit, s the ‘surplus value’, c the value of the means of production or ‘constant capital’, and v the wages advanced or ‘variable capital’. A central message of Marx’s Capital reads that the rate of profit is positive if and only if there is ‘exploitation of workers’, i.e. there is a positive ‘surplus value’.

In Marx’s opinion it was only after the rate of profit had been determined that the problem of normal prices, or ‘prices of production’ as he called them, could be tackled. Marx dealt with it in terms of a multisectoral analysis of the production of commodities by means of commodities; the deviations of relative prices from labour values are systematically traced back to sectoral differences in the ‘organic composition of capital’, i.e. the proportion of ‘constant’ to ‘variable’ capital (cf. the so-called ‘transformation’ of values into prices of production; Marx 1959, Part II).

Yet Marx did not fully succeed in overcoming the analytical difficulties encountered by the classical economists in the theory of capital and distribution. He was particularly wrong in assuming that the determination of the rate of profit is logically prior to that of normal prices. Given the system of production and the real wage the rate of profit and prices can be determined only simultaneously. This was first demonstrated by Bortkiewicz (1907). For a rigorous and comprehensive formulation of the classical surplus approach see Sraffa (1960), whose contribution will be dealt with in more detail below.

The Neoclassical Approach

The abandonment of the classical approach and the development of a radically different theory, which came to predominance in the wake of the so-called ‘marginalist revolution’ in the latter part of the 19th century, was motivated (apart from ideological reasons ever present in debates in capital theory) by the deficiencies of the received (labour) theory of value. Since the new theory was to be an alternative to the classical theory, it had to be an alternative theory about the same thing, in particular the normal rate of profit. Consequently, the early neoclassical economists, including, for example, Jevons (1871), Walras (1874), Böhm-Bawerk (1889), Wicksell (1893, 1901), and Clark (1899), adopted fundamentally the same method of analysis: the concept of ‘long-period equilibrium’ is the neoclassical adaptation of the classical concept of normal positions.

The basic novelty of the new theory consisted in the following. While the surplus approach conceived the real wage as determined prior to profits (and rent), in the neoclassical approach all kinds of incomes were explained simultaneously and symmetrically in terms of the ‘opposing forces’ of supply and demand in regard to the services of the respective ‘factors of production’, labour and ‘capital’ (and land). It was the seemingly coherent foundation of these notions in terms of functional relationships between the price of a service (or good) and the quantity supplied or demanded elaborated by the neoclassical theory that greatly contributed to the latter’s success.

As regards the supply side of the neoclassical treatment of capital, careful scrutiny shows that its advocates, with the notable exception of Walras (at least until the fourth edition of the Eléments), were well aware of the fact that in order to be consistent with the concept of a long-period equilibrium the capital equipment of the economy could not be conceived as a set of given physical amounts of produced means of production. The ‘quantity of capital’ in given supply rather had to be expressed in value terms, allowing it to assume the physical ‘form’ best suited to the other data of the theory, i.e. the technical conditions of production and the preferences of agents. For, if the capital endowment is given in kind only a short-period equilibrium, characterized by differential rates of return on the supply prices of the various capital goods, could be established by the forces constituting demand and supply. However, under conditions of free competition, which would enforce a tendency towards a uniform rate of profit, such an equilibrium could not be considered a ‘full equilibrium’ (Hicks 1932, p. 20).

Thus the formidable problem for the neoclassical approach in attempting the determination of the general rate of profit consisted in the necessity of establishing the notion of a market for ‘capital’, the quantity of which could be expressed independently of the ‘price of its service’, i.e. the rate of profit. If such a market could be shown to exist, profits could be explained analogously to wages (and other distributive variables) and a theoretical edifice erected on the universal applicability of the principle of demand and supply.

Now, the plausibility of the supply and demand approach to the problem of distribution was felt to hinge upon the demonstration of the existence of a unique and stable equilibrium in the market for ‘capital’. (On the importance of uniqueness and stability see, for example, Marshall, 8th edn, 1920, p. 665n.) With the ‘quantity of capital’ in given supply, this, in turn, implied that a monotonically decreasing demand function for capital in terms of the rate of profit had to be established (see Fig. 1). This inverse relationship was arrived at by the neoclassical theorists through the introduction of two kinds of substitutability between ‘capital’ and labour: substitutability in consumption and in production. According to the former concept a rise in the rate of profit relatively to the wage rate would increase the price of those commodities, whose production is relatively ‘capital intensive’, compared to those in which relatively little ‘capital’ per worker is employed. This would generally prompt consumers to shift their demand in favour of a higher proportion of the cheapened commodities, i.e. the ‘labour intensive’ ones. According to the latter concept a rise in the rate of interest (and thus profits) relatively to wages would make cost-minimizing entrepreneurs in the different industries of the economy employ more of the relatively cheapened factor of production, i.e. labour. Hence, through both routes ‘capital’ would become substitutable for labour and for any given quantity of labour employed a decreasing demand schedule for capital would obtain. In Fig. 1 the demand schedule DD′ corresponding to the full employment level of labour L* (determined simultaneously in the labour market) together with the supply schedule SS′ would then ensure a unique and stable equilibrium E with an equilibrium rate of profit r*. Accordingly, the division of the product between wages and profits is expressed in terms of the ‘scarcity of factors of production’, including ‘capital’ conceived as a value magnitude that is considered independent of the rate of profit.

figure 32figure 32

Capital Theory: Debates, Fig. 1

Let us now briefly look more closely at some of the characteristic features of neoclassical capital theory and point out differences between the main versions in which it was presented.

To define ‘capital’ as an amount of value requires the specification of the standard of value in which it was to be measured. A rather common procedure was to express capital in terms of consumption goods or, more precisely, to conceive of it as a ‘subsistence fund’ in support of the ‘original’ factors of production, labour and land, during the period of production extending from the initial expenditure of the services of these factors to the completion of consumption goods. This notion corresponded to the view that capital resulted from the investment of past savings, which, in turn, implied ‘abstention’ from consumption. Thus it appeared to be natural to measure ‘capital’ in terms of some composite unit of consumption goods. However, there was a second dimension of capital contemplated by these authors: the time for which capital is invested in a process of production. The idea was that capital can be increased either by using more of it or by lengthening the period of time for which it is invested.

The first author to use time as a single measure of capital was Jevons (1871). The gist of his argument consisted in the concept of a ‘production function’ y = f(T), where output per unit of labour, y, is ‘some continuous function of the time elapsing between the expenditure of labour and the enjoyment of results, T; this function is assumed to exhibit diminishing returns (1871, pp. 240–41). Jevons showed that in equilibrium r = f′(T)/f(T).

Jevons’s contribution was the starting point of the Austrian theory of capital and interest with Böhm-Bawerk and Wicksell as its main representatives. Böhm-Bawerk’s concern was with establishing a temporal version of the demand and supply approach. This involved the appropriate reformulation of the data of the theory. The central elements of his analysis were the concepts of ‘time preference’ and the ‘average period of production’, used in describing consumer preferences and technical alternatives, respectively. As in Jevons social capital was conceived as a subsistence fund and was seen to permit the adoption of more productive but also more ‘roundabout’, i.e. time-consuming, methods of production. It was to the concept of the ‘average period of production’ that the marginal productivity condition was applied in the determination of the rate of interest.

Among the older neoclassical economists it was perhaps Wicksell who understood best the difficulties related to the problem of a unified treatment of capital in terms of the demand and supply approach. In particular, Wicksell was critical of attempts to work with the value of capital as a factor of production alongside the physically specified factors of labour and land in the production function of single commodities. This implied ‘arguing in a circle’ ([1901] 1934, p. 149), since capital and the rate of interest enter as a cost in the production of capital goods themselves. Hence the value of the capital goods inserted in the production function depends on the rate of interest and will change with it. Moreover, Wicksell expressed doubts as to the possibility of providing a sufficiently general definition of the ‘average period of production’ that could be used to represent capital in a way that is not threatened by this kind of circularity. In the Lectures he tried to overcome these difficulties by introducing production functions in terms of dated services of the ‘original’ factors labour and land.

While Wicksell shared Böhm-Bawerk’s procedure of conceiving the ‘capital endowment’ of the economy as a value magnitude, he become increasingly sceptical whether it was admissible to identify it with some unspecified stock of subsistence goods, which, in turn, was seen to provide some measure of ‘real’ capital. With capital as a value magnitude Wicksell showed that the rate of interest is generally not equal to the marginal productivity of ‘capital’. This discrepancy is due to the revaluation of the capital stock entailed by a change in distribution. The phenomenon is known as the ‘Wicksell effect’ and was regarded by Joan Robinson as the key to a criticism of the marginal productivity theory of income distribution.

Authors like J.B. Clark and Marshall appear to have been less aware of the fact that the conditions of production of single commodities cannot be defined in terms of production functions that include ‘capital’ among the factors of production. Obviously, the criticism levelled against these versions applies also to the concept of the ‘aggregate production function’, which boomed in the late 1950s and throughout the 1960s in conjunction with neoclassical growth theory.

Alternative views of the fundamentals of capital theory were expressed in a controversy between Böhm-Bawerk and J.B. Clark around the turn of this century (cf. in particular Böhm-Bawerk 1906–7; Clark 1907). Böhm-Bawerk criticized Clark’s attempt to differentiate between ‘true capital’, a permanent abiding fund of productive wealth, and ‘concrete capital goods’, each of which is destructible and has to be destroyed in order to serve its productive purpose; in Böhm-Bawerk’s view this is ‘dark, mystical rhetoric’. Furthermore, Böhm-Bawerk refuted Clark’s claim that no concept of ‘waiting’ or ‘abstinence’ is needed to explain interest in stationary equilibrium. Without some concept of time preference, and thus a theory of saving, the determination of the rate of interest is left hanging in the air.

Irving Fisher (1930) extended general equilibrium theory to intertemporal choices. However, he proceeded as if there were a single composite commodity to be produced and consumed at different dates. In his discussion of the theory of interest all prices, wages and rents are assumed to be fixed. Hence the interrelationship between the rate of interest, prices and the remaining distribution variables is set aside. The ‘investment opportunities’ available to an individual and to society as a whole are summarized in intertemporal production possibility frontiers. Due to the assumption of diminishing returns Fisher arrived at a decreasing demand function for saving with respect to the rate of interest. As Keynes noted, this is equivalent to his ‘marginal efficiency of capital’ schedule (Keynes 1936, p. 140). Because of ‘impatience’ the supply of saving is considered to be positively related to the rate of interest. The market equilibrium between the supply of, and the demand for, saving gives the rate of interest, which is equal to the marginal rate of return over the cost of the marginal increase in the capital stock. (For an attempt to generalize Fisher’s rate of return approach see Solow 1967. For a critique of Fisher and Solow see Pasinetti 1969; Eatwell 1976).

The 1930s brought a further controversy on the theory of capital (cf. Kaldor 1937). This was triggered off by a series of articles by F.H. Knight (e.g. Knight 1934), in which he launched an attack on the concept of the ‘period of production’ revived a few years earlier by Hayek, among others. In particular, Knight argued that there is no need to refer to a ‘quantity of capital’ and that therefore the ‘vicious circle’ disappears. The rate of interest could be ascertained with reference to the instantaneous rate of investment and the present value of the additional stream of future income generated by it, However, Knight’s proposed solution to the problem of circularity in terms of a ‘theory of capital without capital’ is illusory, since if the accusation of circularity applies at all (because the value of capital goods cannot be ascertained independently of the rate of interest), it applies both to the stock variable ‘capital’ and the corresponding flow variable ‘investment’.

Finally, some recent attempts to revive and reformulate basic elements of the doctrines of the older neoclassical and Austrian authors should be noted, in particular: Weizsäcker (1971), Hicks (1973), and Faber (1979) on the Austrian theory, Morishima (1977) on Walras, and Hirshleifer (1970) and Dougherty (1980) on Fisher. (For a critical assessmenty of the older theories see especially Garegnani 1960).

The Recent Critique of Neoclassical Theory

Sraffa (1960) deserves the credit for having elaborated a consistent formulation of the classical surplus approach to the problem of capital and distribution. His analysis provided the fundamental basis for a critique of the prevalent neoclassical theory during the so-called ‘Cambridge controversies in the theory of capital’ (see Harcourt 1969; Kurz 1985).

Sraffa starts from a given system of production in use in which commodities are produced by means of commodities. If wages are assumed to be paid at the end of the uniform production period, then, in the case of single-product industries (i.e. circulating capital only) and with gross outputs of the different products all measured in physical terms and made equal to unity by choice of units, we have the price system

$$ p=\left(1+r\right) Ap+ wl, $$

where p is the column vector of normal prices, A is the square matrix of material inputs, l is the vector of direct labour inputs and w is the wage rate. Under certain economically meaningful conditions, for any given feasible wage rate in terms of a given standard the above equation yields a unique and strictly positive price vector in terms of the standard and a unique and non-negative value of the rate of profit. The investigation of the ‘effects’ of variations in one of the distribution variables on the other one and on the prices of commodities, assuming that the methods of production remain unchanged, yields the following results. First, the system possesses a finite maximum rate of profits R > 0 corresponding to a zero wage rate. Second, the vector of prices in terms of the wage rate pw (prices in terms of quantities of labour commanded) is positive and rises monotonically for 0 ≤ r < R, tending to infinity as r approaches R. Third, at the maximum level of wages corresponding to r = 0 relative prices are in proportion to their labour costs, while at r > 0 relative prices generally deviate from relative labour costs and vary with changes in r (or w); it is only in the special case of uniform ‘proportions’ of labour to means of production in all industries that prices are proportional to ‘labour values’ for all levels of r (w). (For a discussion of joint production, fixed capital and land, see Pasinetti 1980.)

While earlier authors were of the opinion that the capital-labour or capital-output ratios of the different industries could be brought into a ranking that is independent of distribution, this is generally not possible: ‘the price of a product…may rise or it may fall, or it may even alternate in rising and falling, relative to its means of production’ (Sraffa 1960, p. 15). This result destroys the foundation of those versions of the traditional theory that attempted to define the conditions of production in terms of production functions with ‘capital’ as a factor. Moreover, as regards the concept of the ‘capital endowment’ of the economy conceived as a value magnitude, the same ‘real’ capital may assume different values depending on the level of r. Sraffa concludes that these findings ‘cannot be reconciled with any notion of capital as a measurable quantity independent of distribution and prices’ (1960, p. 38).

Samuelson (1962), in an attempt to counter Joan Robinson’s (1953) attack on the aggregate production function, claimed that even in cases with heterogeneous capital goods some rationalization can be provided for the validity of simple neoclassical ‘parables’ which assume there is a single homogeneous factor called ‘capital’, the marginal product of which equals the rate of interest. But, alas, Samuelson based his defence of traditional theory in terms of the construction of a ‘surrogate production function’ on the assumption of equal input proportions (cf. 1962, pp. 196–7). By this token the ‘real’ economy with heterogeneous goods was turned into the ‘imaginary’ economy with a homogeneous output, i.e. the ‘surrogate production function’ was nothing more than the infamous aggregate production function. (For a critique of Samuelson's approach see particularly Garegnani 1970).

Implicit in the above system of price equations is the inverse relationship between the wage and the rate of profit, or wage curve, of the given system of production, w = w(r). We may now turn to the hypothesis that for one or several industries alternative technical methods are available for the production of the corresponding commodity. The technology of the economic system as a whole will then be represented by a series of alternative techniques obtained from all the possible combinations of methods of production for the various commodities. Expressing w and p in terms of a commodity produced in all the alternative systems, we obtain as many different wage curves as there are alternative techniques. In Fig. 2 it is assumed that only two techniques, α and β exist. Clearly, at any level of the wage rate (or rate of profit), enterpreneurs will choose the cost-minimizing system of production. It can be shown that, whichever the system initially in use, the tendency of producers to switch to the cheaper system will bring them to the one giving the highest rate of profit (wage rate), whereas systems giving the same r for the same w will be indifferent and can coexist. Thus, in the example of Fig. 2, in the two intervals 0 < w < w1 and w2 < w ≤ Wα technique α will be chosen, while in the interval w1 < w ≤ W2 technique β turns out to be superior; at the two switch points P and Q both techniques are equiprofitable. It follows that with a choice of technique the relationship between w and r, or wage frontier, will be represented by the outermost segments or envelope of the intersecting wage curves.

figure 33figure 33

Capital Theory: Debates, Fig. 2

Figure 2 shows that the same technique (α) may be the most profitable of a number of techniques at more than one level of the wage rate even though other techniques (here β) are more profitable at wage rates in between. The implication of this possibility of the reswitching of techniques is that the direction of change of the input proportions cannot be related unambiguously to changes of the so-called ‘factor prices’. The central element of the neoclassical explanation of distribution in terms of supply and demand is thus revealed as defective. This element consisted in the proposition that a rise of r must decrease the ‘quantity of capital’ relative to labour in the production of a commodity because of the assumed substitutability in production and consumption. The demonstration that a rise in r may lead to the adoption of the more ‘capital intensive’ of two techniques clearly destroys the neoclassical concept of substitution in production. Moreover, since a rise in r may cheapen some of the commodities, the production of which at a lower level of r was characterized by a relatively high ‘capital intensity’, the substitution among consumption goods contemplated by the traditional theory of consumer demand may result in a higher, as well as in a lower, ‘capital intensity’. It follows that the principle of substitution in consumption cannot offset the breakdown of the principle of substitution in production. Finally, it is worth mentioning that reswitching is not necessary for capital-reversing cf. Symposium 1966, p. 516).

The negative implication of reverse capital deepening for traditional theory can be illustrated by means of the example of Fig. 3, in which the value of capital corresponding to the full employment level of labour is plotted against the rate of profit. Obviously, if with traditional analysis we conceived the curve KK′ as the ‘demand curve’ for capital, which, together with the corresponding ‘supply curve’ SS′, is taken to determine the ‘equilibrium’ level of r, we would have to conclude that this equilibrium, although unique, is unstable. With free competition and perfectly flexible distributive variables a deviation of r from r* would lead to the complete extinction of one of the two income categories. According to the critics of traditional theory, the finding that the quantity of a factor demanded need not be related to the price of the factor service in the conventional, inverse manner demonstrates the failure of the supply and demand approach to the explanation of normal distribution, prices and quantities.

figure 34figure 34

Capital Theory: Debates, Fig. 3

Neoclassical Responses

Neoclassical economists tried to counter the attack in various ways. At first it was claimed that reswitching is impossible. When this claim was shown conclusively to be false (cf. Symposium 1966), doubts were raised as to its empirical importance (see, for example, Ferguson 1969), thereby insinuating that neoclassical theory was a simplified picture of reality, the basic correctness of which would not be endangered by ‘exceptions’ of the kind analysed in the capital debate. Other advocates of the neoclassical approach were conscious of how defective the attempt was to play down the importance of reswitching and capital-reversing using the ‘empirical’ route. Since the phenomenon was irrefutable it had to be absorbed and shown to be compatible with the more sophisticated versions of the dominant theory.

Perhaps the first move in this direction was made by Bruno et al. (1966), who drew an analogy between reswitching and the long-known possibility of the existence of multiple internal rates of return. However, whereas the latter phenomenon is a discovery within the partial, ‘fixed-price’ framework of microeconomic theory of investment, reswitching presupposes a total, general framework. Moreover, we are not told how traditional theory was both able to cope with reswitching and yet preserve its basic structure.

A more interesting challenge came from authors such as Bliss (1975) and Hahn (1982). They contended that because of its concern with a uniform rate of profit Sraffa’s analysis can be considered a ‘special case’ of general equilibrium theory. According to these authors the criticism of traditional neoclassical capital theory implicit in Sraffa is correct but has no bearing upon modern general equilibrium theory. Since in the latter the distribution of income is explained in terms of given physical endowments of agents, there is no need to find a scalar representation of the capital stock. The uniformity of profit rates is taken to be ‘a very special state of the economy’ (Hahn 1982, p. 363) which, for given preferences and production sets, presupposes a particular composition of initial endowments. In general, there will be as many own rates of return as there are different assets in the endowment set.

The first thing to be noticed is that the preservation of the basic supply and demand approach to the explanation of prices, distribution and quantities in modern general equilibrium theory is effectuated at the cost of the abandonment of the traditional long-period method. As we have seen, this method was shared by all ‘forerunners’ of this theory, including, most notably, Walras and von Neumann (1936). Indeed, the change in the notion of equilibrium involved expresses a fundamental break with the analytical method used by all economic theory up to the 1930s, when partly because of a growing perception among neoclassical economists that the whole approach was threatened by the difficulties concerning the notion of capital a drastic methodological reorientation was advocated (cf. Garegnani 1976; Milgate 1979). Most influential in this move away from the traditional method was apparently Hicks’s Value and Capital (1939; second edition 1946). Interestingly enough, Hicks himself appears to have become increasingly sceptical as to the usefulness of the ‘temporary equilibrium method’ then suggested by him (see, for example, Hicks 1965, pp. 73–4).

The second observation concerns Hahn’s attempt to interpret Sraffa’s analysis as a special case of general equilibrium theory. Since the latter takes as data (i) the preferences of consumers, (ii) the technical conditions of production, and (iii) the physical endowments, Hahn’s view necessarily leads to the question of which constellation of these data is compatible with a uniform rate of profit. Clearly, to superimpose the latter specification on an ordinary general equilibrium system would render it over-determined, as some of the older neoclassical authors were well aware of. Hence, following the interpretation under consideration, (i), (ii) or (iii) cannot be taken as independent variables. Now it is Hahn’s contention that at the basis of Sraffa’s price equations there must be a special proportion between the initial endowments; i.e. (iii) is tacitly assumed to be specified accordingly. However, as we have seen there is no evidence in support of this presupposition. The surplus approach does not require given endowments of produced means of production in order to determine distribution and normal prices. In fact, looking at classical analysis as a whole the quantities of the capital goods available may be considered as dependent rather than independent variables. In analysing the problem of value, capital and distribution the classical economists took the capital stocks installed in the different industries as exactly adjusted to given outputs, such that the latter could be produced at minimum costs. The tendency towards normal capital utilization and a uniform rate of profit was seen to be the outcome of the working of the persistent forces of the system reflected in the competitive decisions of producers.

Since the opinion entertained by Hahn that Sraffa’s analysis can be subsumed as a ‘special case’ under modern neoclassical theory has to be rejected, the question remains, which of the two is the more powerful instrument of analysis. There does not seem to exist a ready-made answer at present. The following remarks on the dominant neoclassical theory must suffice.

Obviously, to take the capital endowment as given in kind implies that only ‘short-period’ equilibria can be determined. Because firms ‘prefer more profit to less’ (Hahn 1982, p. 354) the size and composition of the capital stock will rapidly change. Thus, major factors which general equilibrium theory envisages as determining prices and quantities are themselves subject to quick changes. This, in turn, makes it difficult to distinguish them from those accidental and temporary factors, which, at any given moment of time, prevent the economy from settling in the position of equilibrium. More important, the fast variation in relative prices necessitates the consideration of the influence of future states of the world on the present situation.

This can be approached in two different ways. First, if there were complete futures markets the analysis could be carried out in terms of the concept of intertemporal equilibrium. However, the assumption that all intertemporal and all contingent markets exist, which has the effect of collapsing the future into the present, can be rejected on grounds of realism and economic reasoning (see, for example, Bliss 1975, pp. 48 and 61). Moreover, there is the following conceptual problem (see Schefold 1985). If in equilibrium some of the capital stocks turn out to be in excess supply these stocks assume zero prices. This possibility appears to indicate that the expectations entrepreneurs held in the past when deciding to build up the present capital stocks are not realized. Hence, strictly speaking we are faced with a disequilibrium situation because otherwise the wrong stocks could not have accumulated. Therefore, the problem arises how the past or, more exactly, possible discrepancies between expectations and facts influence the future.

Since the notion of intertemporal equilibrium cannot be sustained the theory is ultimately referred back to the introduction of individual price expectations concerning future deliveries of commodities for which no present markets exist. This leads to the temporary equilibrium version of modern neoclassical theory. The basic weakness of the theories of temporary equilibrium concerns the necessarily arbitrary choice of hypotheses about individual price expectations. Indeed, as Burmeister stresses, ‘all too often “nearly anything can happen” is the only possible unaqualified conclusion’ (Burmeister 1980, p. 215). Moreover, the stability properties of this kind of equilibrium are unclear, since small perturbations caused by accidental factors may entail changes in expectations, which define that very equilibrium.

The danger of lapsing into empty formalism and of depriving the theory of clear-cut results was of course recognized by several supply and demand theorists and considered a fundamental weakness. In view of it some of them were prepared to dispence with the alleged generality of general equilibrium theory and return to some version of traditional neoclassical analysis. After the recent debate in capital theory this involved ruling out reswitching and other ‘perverse’, i.e. non-conventional, phenomena in terms of sufficiently bold assumptions about available techniques. It comes as no surprise that given these assumptions the central neoclassical postulate of the inverse relation between the capital-labour ratio and the rate of profit should re-emerge as ‘one of the most powerful theorems in economic theory’ (Sato 1974, p. 355). However, in order to be clear about this move it deserves to be stressed that it was motivated, as one author expressly admits, by the fact that ‘regular economies’ have ‘desirable properties’ (Burmeister 1980, p. 124).

See Also