2 Interpretations of Marx’s value theory

This chapter reviews critically two interpretations of Marx’s value theory, the ‘embodied labour’ views, including ‘traditional’ Marxism and Sraffian approaches, and value form theories, including those associated with Rubin and the ‘new interpretation’ developed by Duménil and Foley.

Differences between these interpretations are largely due to distinct understandings of the value relation. Whereas the former claims that value is the average labour time embodied in the commodities, the latter argues that value is the command over social labour represented by money. Different perceptions of the value relation, distinct interpretations of the role, meaning and significance of value analysis and, more broadly, divergent views about the nature and legitimacy of capitalism, help to explain why Marx’s theory has been hotly debated for over one hundred years.

This chapter does not include a comprehensive survey of interpretations of Marx’s theory of value.¹ Its aim is merely to outline the two most influential interpretations, and their best-known offshoots, and critically examine their contribution for the development of Marxian political economy. This chapter is divided into three sections. Section 1 reviews, respectively, the traditional and Sraffian interpretations of Marx. Section 2 analyses value form theories, including the Rubin tradition and the new interpretation. Section 3 draws the main conclusions from the analysis.

2.1 Embodied labour approaches

Some of the most influential readings of Marx hold that value is the labour embodied in commodities during production. These readings define abstract labour in opposition to concrete labour; it is labour in general, abstracted from the form of the activity.² Two such views are considered in this section, the traditional and the Sraffian.

2.1.1 Traditional Marxism

According to the ‘traditional’ interpretation, Marx’s theory of value is not essentially different from Ricardo’s. It may be summarised as follows:³

The main subject of the theory of value is the analysis of capitalist exploitation. The categories developed in the first three chapters of Capital 1 (commodity, value and money) are only indirectly related to this issue, because they belong to a broader set of modes of production, especially simple commodity production, where capitalist exploitation does not necessarily exist.
The concept of value is necessary for the determination of the rate of exploitation. This reading focuses upon the magnitude of value, defined as the quantity of abstract labour embodied in each commodity. The substance and form of value and the links between value and money are largely neglected.
The analysis of profit requires the determination of commodity prices, including labour power. This is done through a set of assumptions that usually includes general equilibrium (simple reproduction). Consequently, prices are only relative to a numéraire. It follows that a theory of money is unnecessary, and money is effectively a veil.
The determination of relative prices has two stages. First, it is assumed that all capitals have equal value compositions (see chapter 6), in which case the exchange ratios are determined by embodied labour alone. Second, the value compositions are allowed to vary; in this case, relative prices differ from the embodied labour ratios, but it is presumed that the latter determine the former algebraically.
The conceptual apparatus is elementary. Commodities are use values put out for sale; value is often conflated with exchange value, and the articulation between value and price is left unclear (even though they are presumed to be quantitatively comparable).
There is little concern with the distinction between levels of analysis and the interaction between tendencies, counter-tendencies and contingency. Theory arguably captures the basic tendencies of capitalism, and expects them to be translated unproblematically into empirical outcomes.

The traditional approach has important virtues, especially the focus on the mode of exploitation. This emphasis concurs with Marx’s own concerns, and it highlights some of his most distinctive contributions; it is also conducive to the critique of the structures of circulation and distribution, such as private property and the market. However, traditional Marxism suffers from two significant shortcomings. First, it disconnects the analysis of the mode of production from the circulation and distribution of the output, which grossly exaggerates their independence.⁴ Second, traditional Marxism wrongly claims that Marx’s analysis of commodities, value and money addresses a broad set of commodity modes of production, especially simple commodity production (SCP), and that his analysis of capitalism proper starts only in chapter 4 of Capital 1 (see sections 2.2.1 and 3.1 below). In this case, two sets of relative prices exist. One is based on embodied labour, and it rules pre-capitalist exchange, while the other is based on equal profitability, and it regulates capitalist exchanges (see chapter 7):

Under certain conditions which prevailed between independent small producers in pre-capitalist societies (what Marx calls “simple commodity production”) exchange of equal values was the rule. If under capitalist conditions there are other more complicated relations determining the quantitative exchange relations, this does not make an economic theory based on the determination of value by socially necessary labour inconsistent, provided there is a clear and consistent method of deriving prices from values.⁵

Presumably, the transition between these stages is a historical process, in which case the transformation between the two types of relative prices (values and prices of production) can be analysed historically as well as algebraically.⁶ This approach is wrong both logically and historically. Generalised exchange at value has never existed because, in general, products become commodities only under capitalism. Moreover, it was shown in section 1.1 that, although Marx often draws on historical studies in order to explain difficult points or trace the evolution of important categories, the only mode of production that he analyses systematically in Capital is capitalism. Hence, although commodities, value and money may have existed for millennia, Capital focuses upon their capitalist determinations only, and no systematic inferences may be drawn about their meaning and significance in other modes of production.⁷

Misperceptions such as these have contributed to the transformation of traditional Marxism into a ‘Ricardian Marxism’. However, this position is untenable (see section 8.1).⁸ Marx was at pains to distinguish his views from Ricardo’s and, in large chunks of his writings, Marx excoriates Ricardo’s methodological and other errors. In particular, Marx argues that Ricardo’s approach is insufficient because it fails to explain the relationship between money and commodities and between abstract labour and value, as well as the mode of labour and the relations of exploitation under capitalism.⁹

2.1.2 Sraffian analyses

Dissatisfaction with the shortcomings of traditional Marxism led to the development of two alternative approaches: the Sraffian (or neo-Ricardian) and value form theory (see section 2.2). The Sraffian approach is developed and explained by, among others, Pasinetti and Steedman, drawing upon works by Bortkiewicz, Dmitriev, Seton, Sraffa and Tugan-Baranowsky.¹⁰ Sraffians attempt to develop the traditional model, focusing upon the articulation between the value and the price systems.¹¹ The main features of the Sraffian approach are the following:

Only the magnitude of value is discussed in detail; its substance and form are almost completely disregarded. The analysis usually involves two sets of equations; one represents the value system, and the other the price system.
The value system is described by λ=λΑ+l=l(I-A)^-1, where λ is the (1×n) vector of commodity values, Ä is the (n×n) technical matrix and l is the (1 ×n) vector of direct labour.
The price system is described by p=(pA+wl) (1+r), where p is the (1×n) price vector, w is the wage rate, and r is the profit rate.
As the analysis is primarily concerned with the relationship between the value and price systems, money has no autonomous role and, when considered at all, it is merely a numéraire.¹²
These definitions of value and price are the basis for a wide-ranging critique of alleged inconsistencies in Marx, which leads to the conclusion that the traditional Marxist project of determining value from embodied labour is flawed. Very briefly, first, the price system has two degrees of freedom, because it has n equations, one for each commodity, but n+2 unknowns, the n prices and the wage and profit rates. Therefore, while the value system can usually be solved (provided that the matrix A is well behaved), the price system can be solved only if additional restrictions are introduced, for example the identity of the value of labour power with the value of a bundle of goods (the wage is the price of this bundle; see section 4.2), and a normalisation condition such as one of Marx’s aggregate equalities (either total prices equal total values, or total profits equal total surplus value; see chapter 7). However, the other aggregate equality is not generally possible, which is allegedly destructive for Marx’s analysis.¹³ Second, the Sraffian representation of Marx cannot distinguish labour from other inputs, in which case it cannot be argued that labour creates value and is exploited, rather than any other input, e.g., corn, iron or energy.¹⁴ Third, even if labour does create value and is exploited, the only meaningful relationship between labour and prices is through the proposition that a positive rate of exploitation is necessary and sufficient for positive profits, which has little empirical significance.¹⁵

Sraffian analyses have contributed significantly, if indirectly, to Marxian studies of the relationship between the mode of production and the structures of distribution. However, the Sraffian approach is insufficient in several respects, and its critique of Marx has been rebutted convincingly.¹⁶ In what follows, two aspects of the Sraffian critique of Marx are briefly assessed: the shortcomings of the value equation and the Sraffian inability to represent capitalist relations of production satisfactorily.

The value equation, λ=λΑ+l, states that commodity values are equal to the input values (λA) plus the living labour necessary to process them (l). Although this equation represents correctly Marx’s definition of value (see chapter 5), it is unsuitable for the calculation of commodity values. Let us see why.

For simplicity, suppose that the matrix A represents the average production technologies, however they may be determined. Suppose, also, that the vector l represents the average number of concrete labour-hours (printing, construction, assembly, etc.) necessary to transform the inputs into the output. Even under these generous assumptions, the vector l cannot be directly used to calculate the value produced because this vector measures concrete rather than abstract labour. Since these labours are qualitatively distinct, any operation across them is meaningless.¹⁷ By the same token, labour employed in distinct activities, whether or not vertically integrated, may produce distinct quantities of value per hour because of training and other differences, for example, designing and painting cars, or building and decorating new homes (see section 5.3).

Suppose, instead, that l is a vector of abstract labour.¹⁸ Although this would avoid the problems outlined above, it would still not allow the value vector to be calculated. For this assumption implies that, in order to calculate the abstract labour necessary to produce each commodity (λ), one needs to know how many hours of abstract labour are necessary to produce each commodity (l). Because it involves a tautology, the assumption that l is abstract labour does not allow the quantitative determination of value.¹⁹

These shortcomings are symptomatic of the Sraffian inability to grasp the essence of capitalist relations of production and the specific features of this mode of production (see section 1.1 and chapter 3).²⁰ The Sraffian system is such that production resembles a purely technical process, not necessarily capitalist, in which case, first, capital is merely a collection of use values rather than a social relation of production. Second, the substance of value, abstract labour, is undistinguishable from average units of concrete labour time. Finally, the social aspect of production is either assumed away or projected upon the sphere of distribution, through the rate of exploitation. In sum, a ‘social process is replaced by technical coefficients and social relations by the distribution of the product between the social classes.’²¹

The Sraffian model is not even internally consistent. It presumes that the technical relations of production are given independently of the value and price systems, and implies that, for Marx, calculation of the price vector would necessitate value magnitudes, but not the converse. Since this is not the case, value analysis is allegedly redundant. This is wrong because, first, it misrepresents Marx’s argument (see chapter 7). Second, the production structure is socially, rather than technically, determined. Under capitalism, competition determines the allocation of labour and means of production, the quantities produced and the technologies, in which case value relations are causally determinant vis-à-vis technologies and prices.²² Consequently, ‘the labour theory of value is not redundant, but rather provides the explanation of price lacking in Sraffa’s own account.’²³ In sum, Sraffian analyses cannot distinguish capitalism from other societies that equalise rates of return. As a result, they cannot explain capitalist social relations, exploitation, the distribution of income, the sources of economic data, the process of competition and, most damagingly, the price form.²⁴

2.2 Value form theories

Value form theories (VFT) were developed primarily in the 1970s, partly as a reaction against the insufficiencies of traditional Marxism and the excesses of Sraffianism.²⁵ The development of VFT was aided by the rediscovery of the works of the Soviet economist Isaak Illich Rubin (1896–1937) in the West in the early 1970s. In what follows, VFT is analysed critically through Rubin’s work. Subsequently, a contemporary approach drawing upon VFT is examined, the ‘new interpretation’ of Marx’s value theory.

2.2.1 The Rubin tradition

This interpretation of Marx’s value theory is inspired by the Soviet economist I.I. Rubin and by independent contributions from, among others, Louis Althusser, Hans-Georg Backhaus and Suzanne de Brunhoff.²⁶ This approach generally starts from the social division of labour. It claims that the essential feature of the capitalist division of labour is the commodity relation, or the production of commodities by ‘separate’, or independent, producers:

The value-form of the product of labor is not only the most abstract, but is also the most universal form, taken by the product in bourgeois production, and stamps that production as a particular species of social production, and thereby gives it its special historical character … Thus the “value form” is the most general form of the commodity economy.²⁷

The commodity features of capitalism are so important that Rubin often refers to the subject of his analysis as the ‘commodity-capitalist’ economy.²⁸ The counterpart to the independence of the producers is the need to produce a socially useful commodity or, in other words, one that is sold (the imperative to sell has been called the ‘monetary constraint’).²⁹ Because of separation and the monetary constraint, this tradition argues that commodities are produced by private and concrete labours that, at best, are potentially or only ideally abstract and social. Private and concrete labour is converted into social and abstract labour if and when its product is exchanged for money:

In a commodity economy, the labour of a separate individual, of a separate, private commodity producer, is not directly regulated by society. As such, in its concrete form, labour does not yet directly enter the social economy. Labour becomes social in a commodity economy only when it acquires the form of socially equalized labour, namely, the labour of every commodity producer becomes social only because his product is equalized with the products of all other producers … [A]bstract labour … [is] labour which was made equal through the all round equation of all the products of labour, but the equation of all the products of labour is not possible except through the assimilation of each one of them with a universal equivalent … [The] equalization of labour may take place, but only mentally and in anticipation, in the process of direct production, before the act of exchange. But in reality, it takes place through the act of exchange, through the equalization (even though it is mentally anticipated) of the product of the given labour with a definite sum of money.³⁰

The Rubin tradition has contributed in at least two important ways to the development of Marxian value analysis. First, the claim that abstract labour is social labour indirectly formed through sale is applicable to commodity economies only, and it provides the springboard for a forceful critique of ahistorical embodied labour views (see section 2.1). This critique has helped to shift the focus of Marxian studies away from the calculation of values and prices and towards the analysis of the social relations of production and their forms of appearance.

Second, this tradition has emphasised the importance of money for value analysis, because value appears only in and through price. Since money plays an essential role in commodity economies, non-monetary or general equilibrium interpretations of Marx’s theory are fundamentally wrong, the search for an unmediated expression of abstract labour is futile, and attempts to calculate embodied labour coefficients are rarely meaningful (see section 5.4).³¹ Emphasis on the importance of money has facilitated the resurgence of interest in Marxian monetary analysis (see chapter 8), and the critique of embodied labour views has opened avenues for the development of more cogent interpretations of Marx.

However, the claim that separation is the essential feature of commodity production has led the Rubin tradition to subsume capitalist relations of production under simple value relations. Consequently, in spite of its significant contribution to the analysis of value, this tradition has added little to our understanding of capital and capitalism.

Focus on the value relation implies that commodity economies are essentially a congregation of producers that, in principle, do not belong in the social division of labour. Because of separation and specialisation, the producers must sell their own goods or services in order to claim a share of the social product for their own consumption. In other words, in this type of society production is essentially for consumption, and private and concrete labour is analytically prior to social and abstract labour, which exist only ideally before sale. The equalisation, abstraction and socialisation of labour are contingent upon sale, and commodity values are determined by the value of the money for which they are exchanged. The inability to sell shows that the decision to produce was wrong, the good is useless, and the labour did not create value.³²

In contrast, in capitalist economies the essential separation is between the wage workers and the means of production, monopolised by the class of capitalists (see section 3.2).³³ Production takes place when capitalists hire workers in order to supply goods for profit. Since the performance of labour is conditioned by this social form, the output is necessarily a commodity; it has a use value, and it is a value (if the commodity is not sold its use value is not realised, and its value is destroyed; see section 5.3).³⁴ In sum, whereas the labour of independent commodity producers is relatively free of social determinations and its social character is contingent upon exchange, under capitalism the mode of labour is socially determined (see chapter 5):

Capitalism throws workers together into workplaces in increasing numbers, where their labor is collective labor. The work of a laborer employed in a plant with a thousand other workers can in no sense be described as private, nor as individual … capitalist production involves collective, cooperative labor, directly social, consciously directed and controlled—the collective and cooperative power of the working class dominated and subsumed under the authority of capital. That which is private is not labor-in-production, but the commodity that is the result of the production process.³⁵

These limitations of the Rubin tradition are largely due to the conflation between capitalist production (the systematic production of commodities for profit) with simple commodity production (the socially unregulated production of commodities by independent producers).³⁶ This is flawed both historically and theoretically:

[in] the case of individual producers who own their own means of production and … where none of the inputs used in production is bought, but all are produced within a self-contained labor process … only the final product of the labor process is a commodity. Each article of the means of production is produced in social isolation by each producer, never facing the discipline of competition. There is no social mechanism for bringing about a normal expenditure of labor time in the products that are the means of production. In such a situation, competition’s only function is to impose the rule of a uniform selling price in the market place … The only objective necessity is that his or her total labor expenditure … be sufficient to allow for the reproduction of the family. Should some producers be able to deliver their commodities with less expenditure of effort than others, the more “efficienť’ producers will enjoy a higher standard of living. This higher standard of living of some in no way pressures the less efficient to raise their efficiency.³⁷

The Rubin tradition’s sharp focus upon the value relation has contributed to important advances in Marxian value analysis. However, neglect of the wage relation and the mode of labour have limited its ability to distinguish capitalism from other modes of production. The Rubin tradition wrongly presumes that commodity exchange is the determinant aspect of capitalism, conflates money with the substance of value, and eschews the mediations that structure Marx’s value analysis. Lack of analytical depth explains its failure to illuminate important real relations identified by Marx, for example the capitalist monopoly of the means of production, the subordination of the workers in production, the social regulation of production through competition, mechanisation and deskilling, and the mediations between value and price (see chapters 5 and 7). Because of these limitations, the Rubin tradition is poorly equipped to explain the main features of capitalism and to analyse their social, economic and political consequences empirically.

2.2.2 The ‘new interpretation”

In the early 1980s Gérard Duménil and Duncan Foley independently outlined a ‘new interpretation’ (NI) of Marx’s value theory,³⁸ drawing upon works by Aglietta and Rubin.³⁹ The growing popularity of the NI among Marxists in the past two decades has helped to shift the value debate away from the relatively sterile polemics against the Sraffian critics of Marx and the highly abstract analyses of the Rubin tradition, and into more substantive issues. The contribution of the NI is largely based upon its emphasis on the net rather than gross product, and its distinctive definitions of value of money and value of labour power (see below). Drawing upon these concepts, the NI argues, first, that empirical analyses employing Marxian categories are both possible and interesting⁴⁰ and, second, that the ‘transformation problem’ is irrelevant (see chapter 7).

The NI stems from a value form interpretation of Marx, whence labour becomes abstract, and is socialised, through sales.⁴¹ Two implications follow: first, money is the immediate and exclusive expression of abstract labour and, second, the value created by (productive) labour is measured by the quantity of money for which the output is sold. This interpretation bypasses the conceptual difficulties in the relationship between individual prices and values (see chapter 5), and the pitfalls associated with the transformation problem, by remaining at the aggregate or macroeconomic level. At this level, money is essentially command over the newly performed abstract labour. There is no necessary relationship between individual prices and values, and this theory cannot discriminate between alternative price systems. This allegedly increases its generality in the light of potentially pervasive imperfect market structures.

Algebraically, the total (abstract) labour performed, lx, creates the gross product, x, but only the value of the net product, y=x – Ax, where A is the (n×n) technical matrix, l is the (1×n) vector of unit labour requirements, x is the (n×1) gross output vector, and y the (n×1) net output vector. The value of money, λ^m, is the ratio between the total labour performed and the price of the net product:

(1)

The value of money measures the quantity of labour represented by the unit of money, or the abstract labour time that adds one pound sterling (or dollar or whatever) to the value of the product.⁴² For the NI, equation (1) represents Marx’s equality between total value and total price. The newly produced money-value is allocated across the net product as the price of these commodities.

The value of labour power, V, is defined as the wage share of the national income,⁴³ and surplus value, S, is the residual (see section 4.1). If w is the hourly wage rate and wlx is the wage mass, V is the wage rate times the value of money:⁴⁴

It follows that:

where Π is total profit, in which case Marx’s equality between total surplus value and total profit also holds by definition.⁴⁵ Finally, the rate of surplus value is:

This ratio is determined when commodities are priced and wages are paid. It is unaffected by the use of wage revenues, which may include the consumption of necessities or luxuries, saving or hoarding. For the NI, this relationship shows that profit is merely redistributed surplus value (see section 4.2).

Let us consider the contribution of the NI more closely, starting with the operation in the net product. There are two ways to conceptualise the economy’s net product. In use value terms, it comprises the means of consumption and net investment, or that part of the gross output over and above that necessary to maintain the productive system, or to repeat the same pattern and level of production. In value terms, it is identical with the newly performed labour. This raises the problem of the value of the gross product, since labour creates the entire gross product but only part of its value.

The NI implies that the conventional definition of Marx’s equalities in the gross product is inconsistent because the value of the means of production is counted twice in the value of the gross product. It counts, first, as the value of the newly produced means of production and, again, as the new value of the means of production used up (see section 5.2). However, the latter does not correspond to labour actually performed either in the current period or previously; this is merely a reflection of labour carried out and value created elsewhere.⁴⁶ These insights are persuasive. However, focus on the net product may be misleading, for two reasons. First, empirically, the net product is defined over a

time period other than the turnover period of capital. Net national product, for example, is defined for a year or a quarter. In consequence, the two components of net capital value (variable capital and surplus value) are aggregated over several turnovers, and conceptually one loses sight of the fundamental aspect of circulation, which is the recapture of capital advanced through sale of commodities and the replacement of the material components of production.⁴⁷

Second, and more importantly, focus on the net product eliminates the production of the means of production (other than that required for expanded reproduction). As a result, a significant proportion of current production is rendered invisible as if it were redundant, and the largest proportion of commodity exchanges, those between the producers, vanishes as if it were inconsequential. Therefore, the use of money as capital and as means of payment, and the role of the credit system, are significantly restricted (see section 8.3).

Because of double counting of the input values in the value of the gross output, the NI defines the value of money on the net rather than the gross product. This definition of value of money is seductive for three reasons: first, it avoids the simplifying assumptions that encumber the traditional and Sraffian approaches; second, it appeals to the contemporary experience with inconvertible paper currencies and the perceived importance of the macroeconomic determinants of the value of money, especially through fiscal and monetary policy; third, it facilitates the analysis of imperfect market structures and monopoly power, which can hardly be achieved by the traditional approach.

In spite of these significant advantages, this concept of value of money is limited in two important ways. First, it is merely the ex post reflex of the relationship between (abstract, productive) labour performed and the money-value added in the period. It is known only after labour is performed, commodities are produced and priced, and the technologies are determined. In this respect, it is unrelated to the Marxian concept of value of the money-commodity, that is determined prior to circulation (see section 8.1). Second, the value of money is unable to reflect the distinct levels of complexity of the value relation, including the social relations of production and distribution, the labour performed, the relations between supply and demand, monopoly power, the quantity and velocity of money, and the credit system. Each of these factors can affect the price system in different ways, but the NI is unable to distinguish systematically between them, or to ground them analytically and explain their implications.

In short, the value of money short-circuits the real structures and relations between social labour and its representation in money in order to address the extant macroeconomic relationships. Unfortunately for the NI, these mediations inherently contain the possibility of disequilibrium and crisis. To collapse the mediated expression of value as price into the simple division of the total hours worked over the price of the total net product is to set aside the complexity of the real processes involved and to obscure the inherent potential for disequilibrium in the economy, which weakens the theory’s ability to address the very relations which it should wish to confront.⁴⁸

The NI concept of value of labour power suffers from similar shortcomings (see section 4.2). For the NI, the value of labour power is the workers’ share of the national income, which is determined by class struggle.⁴⁹ However, this definition of the value of labour power does not extend beyond one of the effects of exploitation, the inability of the workers to purchase the entire net product. This was the same aspect of exploitation which the ‘Ricardian socialist’ economists emphasised in the early nineteenth century, and this is the only aspect of exploitation considered in Sraffian analyses.⁵⁰

This notion of value of labour power can be misleading. First, it can dilute the ability of theory to explain the primary form of class conflict in capitalism, which takes place in production rather than distribution. Second, it may create the illusion that the net product is ‘shared’ between workers and capitalists at the end of each production period, or that exploitation is due to the unfair distribution of income (see chapter 4). Third, it may support the classical dichotomy between ordinary commodity values, determined by labour embodied, and the value of labour power, given by supply and demand.

In sum, there are two distinct aspects to the contribution of the NI for the development of value analysis. On the one hand, it bypasses the transformation problem (especially the spurious debate about the ‘correct’ normalisation condition), and it rightly rejects the equilibrium framework in which value theory, and the transformation in particular, was discussed in the past. These important contributions are part of a broader reconsideration of Marx’s value theory, providing the foundation for a new, critical macroeconomics. These achievements are important, and the objective is worthwhile.

On the other hand, the NI is open to criticism on several grounds. This approach has been developed in order to address the appearances directly, through a ‘Marxian macroeconomics’. However, this important objective exacts a heavy toll. The NI has little analytical ‘depth’, emphasises exchange and distribution at the expense of production, and it eliminates the mediations and the complex relationship between value and price and surplus value and profit, treating them as if they were identical. As a result, the NI becomes unable to incorporate some of Marx’s most important insights into the analysis, including technical change, accumulation, the credit system and crises, other than as exogenous accretions. These limitations are due to the internal structure of the NI, and they explain why it has been accused of tautology (because of the way in which it validates Marx’s equalities) and empiricism (because it does not highlight the structures whose development underlies value analysis).⁵¹ Therefore, it is difficult to develop the NI further without making use of arbitrariness in the choice of phenomena to be explained, the judgement of their importance and their relationship with the other features of reality.

2.3 Conclusion

The capitalist division of labour can be approached in two different ways. Most neoclassical economists and some Marxists usually adopt the point of view of circulation (exchange). From this viewpoint, the capitalist economy appears as an uncoordinated collection of competing activities, distinguished from one another by the commodities produced in each firm and their possibly distinct technologies. This approach tends to emphasise the processes that bring coherence to decentralised economies and ensure that needs are satisfied, subject to constraints. In this context, relative prices and the distribution of labour and income are highly important. The inquiry may be extended subsequently into why the ‘invisible hand’ can fail, in which case there are disproportions and crisis. These issues are worthy of detailed study and bring to light important aspects of capitalism. Unfortunately, however, they are not conducive to the analysis of the mode of production. This is a severe limitation, because the essential differences between capitalism and other modes of production stem from the relationship between the workers and the owners of means of production and the mode of labour associated with it. One of Marx’s most important claims is that, if the analysis is restricted to circulation or distribution and ignores the sphere of production, some of the most important features of capitalism remain hidden (see chapter 3).

In contrast, analyses that emphasise production at the expense of exchange sometimes impose equilibrium conditions arbitrarily, in order to focus upon the technologies of production. In this case, it becomes difficult to grasp the significance of money and the relationship between concrete and abstract labour and, more broadly, the historical limits of value analysis. The meaning of competition, technical change and capital migration, and the conflict-ridden relationship between the social classes, are correspondingly blurred (see chapter 5).

These shortcomings show that value analysis ought to consider both production and exchange, the mediations between these spheres and the different levels of analysis. It is sometimes appropriate to short-circuit certain mediations in order to focus upon certain aspects of capitalism rather than others, but this can be risky because it may become difficult to know where and how to introduce important structures or tendencies into the analysis. In this case, it may be necessary to resort to arbitrariness, or to plug into value analysis unrelated studies uncritically, which smacks of eclecticism and is rarely fruitful.