would end by sharply reducing the rate of surplus-value, is by accumulating a larger and larger part of capital in the form of fixed constant capital – i.e. substituting machinery for living labour). Moreover, empirical evidence overwhelmingly confirms that branches of production which are more labour-intensive than others do not normally realize a higher rate of profit.
So the conclusion Marx draws is the following: in a fully developed and normally functioning capitalist mode of production, each industrial branch does not receive directly the surplus-value produced by the wage-labour it employs. It only receives a fraction of all surplus-value produced, proportional to the fraction it represents of all capital expended. Surplus-value in a given bourgeois society (country) as a whole is redistributed. This results in an average rate of profit more or less applicable to each branch of capital. Branches of production which have an organic composition of capital below the social average (i.e. which employ more labour, spend more variable capital, in relation to total capital spent) do not realize part of the surplus-value produced by ‘their’ wage-labourers. This part of surplus-value is transferred to those branches of industry where the organic composition of capital is above the social average (i.e. which spend a larger proportion of total capital on equipment and raw material, a smaller proportion on wages, than the social average). Only those branches of industry whose individual organic composition of capital is identical to the social average realize all the surplus-value produced by the wage-labour they employ, without transferring any portion of it to other branches or receiving any fraction of surplus-value produced in other branches. As a result, each capital receives a part of the total surplus-value produced by productive labour which is proportional to its own part in total social capital. This is the material basis of the common interest of all owners of capital in the exploitation of labour – which thereby takes the form of a collective class exploitation (competition between many capitals only deciding the way in which this total mass is redistributed between the capitalists).
This process of equalization of the rate of profit raises three series of problems. What is its relation to the labour theory of value in general? What are the concrete mechanisms which allow equalization of the rate of profit to occur in real life? What is the ‘technical’ solution to the problem of transformation of values into prices of production (capital outlays, i.e. production costs, going into the output of each commodity + average profit multiplied by these outlays)? The first two problems have provoked relatively less controversy than the third, probably because of their more ‘abstract’ character. They are, however, of the highest importance for the inner cohesion of Marxist economic theory. Marx’s treatment of them, moreover, shows his dialectical method at its most mature.
Briefly, with respect to the first, Marx argues that as value in the last analysis is a social not an individual category, those branches of industry which have an organic composition of capital below the social average objectively waste social labour from the point of view of capitalist society as a whole (i.e. from the point of view of ‘equality’ of commodity-owners).9 Therefore, the market does not return to their owners all the value effectively created during the process of production in these branches. Inversely, those branches of industry which have an above-average organic composition of capital, i.e. an above-average social productivity of labour, objectively economize socially necessary labour. Their owners are rewarded for this by the market, which attributes to them a higher proportion of all surplus-value produced than that which is directly produced by the wage-labourers they employ.
Various objections have been raised to this solution. Is productivity of labour comparable in different branches of output, inasmuch as these do not produce goods that are interchangeable? This difficulty can be resolved dynamically, i.e. by comparing the different rates of increase in productivity of labour in different branches of output over time. More generally, the specific organic composition of capital in each branch of production, which constantly changes as a result of these different changes in the productivity of labour, can be considered as a general index, a means of measurement, of social productivity of labour.10 In a capitalist market economy, with its constant revolutions in the techniques of production, its constant shifts in demand from one commodity to another, its constant flux of capital investment from one branch to another, this assumption is both theoretically tenable and empirically verifiable.
But is there not a basic contradiction between considering all labour effectively expended in the process of production of each branch of production as value-producing, and at the same time explaining the transfers of value (surplus-value) between different branches as a function of objective waste or economy of social labour?11 I do not believe so. What we have here, on the contrary, is a demonstration of the unique way in which social labour and private labour are combined and interrelated under capitalism, i.e. under generalized commodity production.
For Marx, the problem of value as an embodiment of abstract human labour is not a problem of measurement, of numéraire, but a problem of essence.12 Each community has at its disposal a given total labour capacity (a total number of producers effectively engaged in productive labour, multiplied by the socially accepted average of annual work-days and daily work-hours). This potential is an objective category, in a given country and for a given stretch of time (for purposes of simplification, we can take the work-year as the basic time-framework). From it flows the total value produced during a year (in so far as part of this labour potential has not been idle, for reasons independent of its will). Again, this is an objective social category: the total number of labour-hours effectively produced in the course of the process of production. The category of ‘socially necessary labour’, which treats some of these labour-hours as ‘wasted’ and hence not accounted for from a social point of view, only implies redistribution of value inside each branch of production, except in cases of monopoly.13
If we extend the same reasoning to the economy as a whole, nothing changes. All labour actually expended in the process of production has been value-producing. It cannot be made larger or smaller by anything which occurs outside the actual sphere of production. The problem of compensation on the market for labour expenditure is one of distribution, not one of production. Thus it is perfectly possible that actually expended private labour in a given branch, at the average rate of productivity of that branch, is socially necessary labour and has really produced value, while at the same time the owners of the commodities in which it is embodied do not receive full compensation on the market for all that embodied value, or receive a counter-value higher than the amount of value embodied in their commodities.
This dialectical unity-and-contradiction between, on the one hand, private labour effectively expended in production and effectively value-producing and, on the other hand, socially compensated value is mediated through the understanding that total value is equal to total prices of production (i.e. represents an equal sum of labour-hours, or labour-weeks, or labour-years: an equal total labour potential). What is modified on the market, i.e. what the Volume 3 notions of ‘objective waste’ and ‘objective economy’ of social labour represent when different branches of production are compared (in contrast to the notions of ‘waste’ and ‘economy’ of quantities of social labour inside each separate branch of industry, studied in Volume 1), is exclusively a problem of (re)distribution of value, not one of production of value.
The second question regarding equalization of the rate of profit between different branches of industry is how this operates in practice. In order to understand this, we should start from the assumption that this equalization is always a tendency, never a permanent reality. If we start from the actual realization of the total mass of surplus-value produced in each branch of production by the capitalists operating in that branch, a much higher rate of profit will occur in those branches of production which have a lower organic composition of capital and spend a larger proportion of capital outlays on wages than in those which have a higher organic composition of capital and spend a larger proportion of total capital outlays on equipment and raw materials. All things remaining equal (which means, above all, not assuming for the moment any changes in the distribution of total demand for different use-values produced by different branches of output), such an above-average rate of profit will attract additional capital in these branches. This will increase production (supply) above social demand, which will precipitate a decline in prices, which will precipitate a decline in the rate of profit. Inversely, in those branches of production where the average organic composition of capital is above-average, hence the ‘initial’ rate of profit below average, capital will be withdrawn; production will decline, till it falls below social demand; prices will rise; profits will rise, until they reach the socially average rate of profit.
In other words, it is the flux and reflux of capital between different branches of production, from those with lower rates of profit to those with higher rates of profit, which is the driving force behind equalization of the rate of profit. This flux and reflux of capital between different branches of production is indeed the main way in which capital accumulation (growth) occurs in actual life, i.e. as an uneven process, all branches never growing at exactly the same rhythm and over the same span of time. Equalization of the rates of profit indeed presupposes their relative inequality. It is a process which constantly realizes itself by negating itself. Anybody who studies the real history of capitalist branches of industry, mining and transport may easily confirm this view.
This uneven process does not necessarily presuppose that it starts with great unevenness in the rates of profit between various branches, nor that higher rates of profit each time coincide with greater labour intensity in given branches of industry. Indeed, it would be sufficient to assume a single initial situation of that kind to make the process perfectly logical and coherent with the given analysis.14 In fact, very early in the history of modern industrial capitalism, the average rate of profit is a known entity (bank credit and the stock exchange playing a not unimportant role in establishing this).15 The real process is, therefore, not so much one of capital flowing from branches with below-average to branches with above-average rates of profit. The real process is usually one of firms looking for surplus profits over and above the known average rate of profit, essentially through revolutionary innovations (which might imply creating completely new branches of industry). The average rate of profit is constantly shaken and reestablished by the reactions which this constant revolution in the technique of production and the organization of labour provokes. Each firm trying to maximize its own rate of profit contributes, independently of its wishes and designs, to the tendential equalization of the rate of profit.
If we abandon the initial simplifying assumption of a stable structure of demand in a given time-span, we only have to introduce additional mediations; the result remains substantially the same. If, in regard to branches of industry with below-average organic composition of capital, there is additionally an above-average increase in social demand for their output, prices will decline less rapidly in spite of the influx of additional capital and the ensuing increase of production.16 But this will only attract even more additional capital, until equalization of the rate of profit finally occurs. Inversely (and this occurs more often), if branches of industry with below-average organic composition are relatively ‘older’ branches suffering from relative decline of total demand, the influx of additional capital and the ensuing increase of output will lead more rapidly to a decline of prices and profits, and to the final equalization of the rate of profit. It is not necessary to repeat, for those branches which witness an outflow of capital because of initially lower rates of profit, the reasoning for the combination of fluctuations in final demand with the process of equalization of the rate of profit. It is an obvious counterpart of the analysis just developed.
It is the third category of problems raised by the equalization of the rates of profit between different branches of production which has provoked most argument: that concerning the ‘technical’ problem of the transformation of values into prices of production for each specific commodity (or group of commodities), i.e. the problem of how one can ‘technically’ prove the operation of the law of value under conditions of competition of capitals between different branches of output. This can be divided into two main bodies of argument, which I shall refer to as the feedback controversy and the monetary confusion.
The feedback controversy arises from the fact that, in the way in which Marx solves the transformation of values into prices of production in Chapter 9 of Volume 3, apparently only the values of currently produced commodities (outputs) are being ‘transformed’ and not the values of ‘input-commodities’. Ever since the Prussian statistician Ladislaus von Bortkiewicz first raised this objection,17 a constant stream of authors – some claiming to be Marxists, others obviously adhering to other economic doctrines or at any rate other theories of value – have repeated this assertion about a basic flaw in Marx’s reasoning.18
This ‘flaw’ seems, at first sight, all the more evident in that Marx himself appeared to be aware of it. Again and again, the following passage from Chapter 9 has been quoted: ‘The development given above also involves a modification in the determination of a commodity’s cost price. It was originally assumed that the cost price of a commodity equalled the value of the commodities consumed in its production. But for the buyer of a commodity, it is the price of production that constitutes its cost price, and can thus enter into forming the price of another commodity. As the price of production of a commodity can diverge from its value, so the cost price of a commodity, in which the price of production of other commodities is involved, can also stand above or below the portion of its total value that is formed by the value of the means of production going into it. It is necessary to bear in mind this modified significance of the cost price, and therefore to bear in mind too that if the cost price of a commodity is equated with the value of the means of production used up in producing it, it is always possible to go wrong.’19
However, this quotation from Marx should not be made to say more than it does. It says only that if one uses value calculations in inputs and prices-of-production calculations in outputs, then one is likely to arrive at numerically erroneous conclusions. This is rather obvious, since the whole analysis precisely concerns the deviation of prices of production from values. But the extract cited does not imply that prices of production of inputs should be calculated within the same time-span as prices of production of outputs. Such an interpretation is even explicitly rejected in a passage which immediately follows that quoted by von Bortkiewicz and so many others: ‘Our present investigation does not require us to go into further detail on this point. It still remains correct that the cost price of commodities is always smaller than their value. For even if a commodity’s cost price may diverge from the value of the means of production consumed in it, this error in the past is a matter of indifference to the capitalist. The cost price of the commodity is a given precondition, independent of his, the capitalist’s, production, while the result of his production is a commodity that contains surplus-value, and therefore an excess value over and above its cost price’20 (my italics).
And even more clearly: ‘For all the great changes that constantly occur in the actual rates of profit in particular spheres of production (as we shall later show), a genuine change in the general rate of profit, one not simply brought about by exceptional economic events, is the final outcome of a whole series of protracted oscillations, which require a good deal of time before they are consolidated and balanced out to produce a change in the general rate. In all periods shorter than this, therefore, and even then leaving aside fluctuations in market prices, a change in prices of production is always to be explained prima facie by an actual change in commodity values, i.e. by a change in the total sum of labour-time needed to produce the commodities’21 (my italics).
In other words, inputs in current cycles of production are data, which are given at the start of that cycle, and do not have a feedback effect on the equalization of the rates of profit in various branches of production during that cycle. It is sufficient to assume that they are likewise calculated in prices of production and not in values, but that these prices of production result from equalization of rates of profit during the previous cycle of production, for any inconsistency to disappear.
Such an assumption eliminates the logical inconsistency of which von Bortkiewicz and his followers accuse Marx, between supposedly calculating inputs in the form of values and outputs in the form of prices of production. But is it compatible with what we know about the actual operation of capital movements in a given time-span (a year, for example)? Could it not, for instance, be argued that raw-material prices fluctuate constantly, changing many times during one year: hence one may assume that, where this is the case, feedback effects do indeed occur; and that the final equalization of the rate of profit is not only a function of redistribution of surplus-value between branches of production whose commodities can be considered only as industrial outputs, but should include, at least with regard to raw materials, part of the inputs as participating in the current (annual) redistribution of surplus-value between various branches?
This objection, however, is not a valid one. I repeat, prices of production of raw materials, like all other inputs bought by capitalists currently occupied in production, are unchangeable data. They cannot vary through ups or downs of current production of surplus-value, or current changes in the organic composition of capital occurring during a given year. The capitalists have to pay a given price for them, which does not change a posteriori as a function of what is occurring during a given year in the field of final surplus-value redistribution. They are results of the equalization of the rate of profit which occurred during the previous period. Even if one were to assume that capitalists buy their raw materials currently and not only at the beginning of the year, and even if one were to eliminate all existing stocks of previously produced raw materials to explain the origin of these current purchases, the argument would still hold.
The formation of prices of production, i.e. the calculation of the average rate of profit, is not a constantly moving process. It is linked to the overall realization of surplus-value of all (most) of the commodities currently produced. That is why a minimum time-span must be assumed before one may speak of a new average rate of profit replacing a previous one. Even the assumption of such an annual change is probably an exaggeration, rather than an underestimate. Therefore, one has to assume that currently purchased raw materials on a quarterly or even monthly basis do not fundamentally change the prices of production (average rate of profit), as resulting from the capital movements which had occurred during the previous year. One should, of course, not confuse the formation of prices of production – which result from a redistribution of the total surplus-value produced for society as a whole – with current fluctuations of market prices, which Marx explicitly excludes from the study of prices of production, as is clearly stated in the passage cited above.
The reason for this relative rigidity of prices of production (of average rates of profit in a given country) is linked to the very nature of the processes of which the equalization of rates of profit is a result: the determination of the total mass of surplus-value (surplus labour) produced; and the fluxes and refluxes of capital (large-scale capital movements) between various branches of production, determining changes and differences in the organic composition of capital both of productive sectors as a whole and of each productive sector taken separately. It is clear that such overall social movements cannot vary from quarter to quarter, let alone from month to month. The relative indivisibility of fixed capital alone is a formidable obstacle to such broad movements under advanced capitalist conditions, except in the case of radical devalorization of capital under conditions of severe crisis. Therefore, not only is Marx theoretically consistent when he assumes prices of production of inputs resulting from equalization movements in different time-spans (during different years) from prices of production of outputs. This also corresponds much more closely to the real, empirically verifiable operation of the capitalist system as we know it than does the opposite assumption of von Bortkiewicz and his followers.
Numerous attempts have been made both to extend von Bortkiewicz’s critique of Marx’s solution to the transformation problem, and to provide an alternative solution to that proposed by von Bortkiewicz himself. J. Winternitz sought to formulate one in which total prices of production would still equal total value. More recently, Anwar Shaikh has proposed yet another solution, using the ‘iterative method’ rather than that of simultaneous equations.22 However, mathematical models cannot, in and of themselves, ‘solve’ theoretical problems. They can only formalize interrelations previously understood as such, whose nature and implications have to be grasped before a meaningful formalization can take place. Unfortunately, many authors of such models operate by silently assuming correlations which have not been previously proved or empirically tested. Their equations lead to conclusions which are, of course, mathematically consistent, but may nevertheless be theoretically wrong: i.e. which do not correspond to a meaningful representation of the problem supposedly to be solved.
In the ‘Okishio theorem’, for instance, the author puts fixed capital between brackets altogether, in order to arrive at conclusions regarding the trend of the rate of profit. But if one postulates that precisely the growth of fixed capital is one of the main – if not the main – determinant of the tendency of the rate of profit to decline, then this theorem does not prove anything.23 Similarly, in the von Bortkiewicz ‘solution’ of the transformation problem (accepted by Paul Sweezy, Piero Sraffa, F. Seton and many others), besides uniform profits for all products (not all branches of industry or even firms, which is quite another story), it is assumed that only those equations are needed for a solution which involve commodities entering into the production of other commodities. It is logical that, under these circumstances, the organic composition of department III (whose commodities do not enter the reproduction process) does not influence the average rate of profit.24 But this tells us nothing either about department III in Marx’s analysis, where such a distinction is explicitly excluded, or especially about what happens in the really functioning capitalist economy, i.e. in real life. To say that the organic composition of the armaments industry, including its size, is immaterial to the real rate of profit of a real capitalist economy is quite untenable – especially if one takes a look at the size of that department in, say, 1943 in Germany or 1944 in the U.S.A.
A second line of attack on Marx’s solution of the transformation problem has involved a confusion between prices of production and market prices, and more generally the introduction into the problem of questions concerning the expression of values as prices, i.e. money. Sweezy, in particular, has been guilty of such a confusion, in the way he has taken over von Bortkiewicz’s critique.25 Others, like Ian Steedman recently, have followed in his footsteps.26
Marx himself, however, makes crystal clear that prices of production do not concern market prices, i.e. values (or prices of production) expressed in money terms. The very title of Chapter 9 specifies this, referring as it does to the transformation of values of commodities into prices of production. Values are quantities of labour, and have nothing to do with money prices as such. The equalization of the rate of profit between different branches of production occurs through the transfer of quantities of surplus-value from one branch to another. Again, quantities of surplus-value are quantities of labour (surplus labour) and not quantities of money. At the end of the last passage cited from Volume 3, there follows a sentence which I deliberately omitted but will now quote – a sentence which again eliminates all doubt as to the non-inclusion of monetary questions in the transformation problem: ‘We are not referring here, of course, to a mere change in the monetary expression of these values.’27 If the problem does not concern changes in the monetary expression of values, it ipso facto does not concern changes in the monetary expression of prices of production either.
In Chapter 10, immediately following that in which he gives his solution to the transformation problem, Marx does indeed introduce market prices, and the influence of competition, etc. upon them. But he there clearly and explicitly distinguishes fluctuations of market prices and of monetary expressions of value (prices of production) from fluctuations in the average rate of profit which determine fluctuations of prices of production.28
Behind this confusion, there lies an insufficient understanding of the nature of Marx’s theory of money. Marx considers money (gold) as a special commodity having its own ‘intrinsic’ value. It is only for this reason that it can serve as a general equivalent for the exchange-value of all other commodities. It immediately follows that fluctuations of market prices (monetary prices, expressions of value in money) may always be the result of a dual movement: the changes in the value of a commodity and the changes in the value of the money-commodity, gold. But changes in the intrinsic value of the money-commodity have identical effects on the market prices of all other commodities, i.e. cannot change their mutual exchange relations (their mutual ‘relative prices’). Paper money does not alter anything in this respect. Inflation of paper money only means that an increasing amount of paper dollars, paper pounds, etc. represents the same quantity (e.g. one ounce) of the money-commodity, gold. What is true for the money expression of value is likewise true for the money expression of prices of production, as they concern only a redistribution of quantities of surplus-value between different branches of production.
The ‘inputs’ in the reproduction tables could only be treated as inputs in really occurring capitalist production (i.e. in real life) if they were expressed in market prices, and not in prices of production: for capitalists obviously buy raw materials, machines, buildings, etc. at market prices. So the problem would be how to ‘transform’ values, not into prices of production, but into market prices; or, in two successive stages of transformation, values into prices of production and the latter into market prices. This final stage, of course, would have to involve real monetary problems: specifically, the interrelationship between the average value of commodities and the average value of gold. What is really involved in this controversy is whether the ‘transformation problem’ concerns the immediate move from essence to appearance, in other words to the process of production and circulation in day-to-day reality, or whether – as I would strongly maintain – it is only a mediating link in the process of cognition, which does not yet deal with immediately verifiable, empirical data, i.e. market prices.
The diagram on page 30 will help to elucidate the relations between Marx’s various concepts of value, market value, price of production and market price, which are often rather confusing.
An excellent overall critique of the von Bortkiewicz/Sraffa ‘corrections’ of the way Marx deals with the relation between prices of production and values has been furnished by Pierre Salama.29 It has, among other qualities, the merit of revealing a series of underlying theoretical assumptions of which the authors themselves are not always aware. It shows that a further disaggregation of the von Bortkiewicz system – in other words, the application to von Bortkiewicz of some of the criticisms he himself directs at Marx (for example, it is evident that in the aggregate of department I, those means of production which are exclusively used for the production of commodities in department III will have a different status) – leads unavoidably to the elimination of all value calculations and, therefore, of exploitation itself from the system. I do not want to imply that Salama, Farjoun and others have definitively resolved all the difficulties raised by the ‘transformation problem’: there is clearly still room for further discussion and research. But neither have von Bortkiewicz, Seton and Sraffa ‘definitively’ proved Marx wrong.
From his definition of the average rate of profit as the sum total of surplus-value produced during the process of production divided by the sum total of capital, Marx derives the central ‘law of motion’ of the capitalist mode of production. Since that part of capital which alone leads to the production of surplus-value (variable capital, used to buy labour-power) tends to become a smaller and smaller part of total capital, because of the fundamentally labour-saving tendency of technical progress – the gradual substitution of dead labour (machinery) for living labour – and because of the gradual increase of the value of raw materials in that of total output: since, in other words, the organic composition