From International Socialism 2:4, (Spring 1979).
Transcribed by Christian Høgsbjerg.
Marked up by Einde O’ Callaghan for the Encyclopaedia of Trotskyism On-Line (ETOL).
In the first part of this article I argued that the Labour Theory of Value is the necessary foundation for understanding Capitalism. The object of Marx’s analysis is not some lifeless abstraction called an economy but an historically specific mode of production. To defend the theory is to defend the theoretical premises of revolutionary socialism. However few of us may have read Capital its message has become part of the everyday vocabulary of Marxists. It is not dogmatic to suggest that those who want to abandon the core propositions of that work today are, like so many in the past, really seeking to alter our understanding of the deeply contradictory nature of Capitalism.
It should go without saying that the most important task today is to build on the work discussed in the first part of the article. Capitalism is the most dynamic mode of production in history and its forms have radically changed. But the essentials of the theory remain applicable today. The sale of labour-power and the extraction of surplus-value from the working class, remain the prerequisites of the survival of the system. The enormous growth of the forces of production has not lessened the compulsion of each unit of capital to maximise the ‘productivity’ of the labour force. Technical change still serves not to reduce the length of the working day but to continually recreate a reserve army of labour. The interdependence of the different national economies, the world division of labour, the internationalisation of the forces of production, continue in contradiction with the division of the world into nation states, jealous of their sovereignty and escalating their rivalry in the face of the world-crisis.
Marx’s analysis of the contradiction between the ‘private’ character of the relations of production, and the ‘social’ character of the productive forces, between the division of the system into ‘autonomous’ units and their mutual relations sustained only through the exchange of things, remains acutely relevant, however much the forms of that contradiction may have changed. The working class, those who labour under the command of Capital, the human living force which sustains the system, is still the arbiter of the future course of history.
The Neo-Ricardian critics of Marx are not concerned with these questions. Their focus is on some narrowly selected appearances of the system. They argue that the labour theory of value is either unnecessary to, or incapable of, explaining those appearances, the relative prices of different commodities and the rate of profit.  Ian Steedman makes no attempt to show how his ‘Sraffa based analysis’ could generate the understanding of the actual history of Capitalism we derive from Marx’s work. If his criticisms of the logic of the theory of value are correct today they would have been just as correct at the time Capital was written.
Originally this part of the article dealt mainly with the Transformation Question of how values are related to prices, and the general differences of method and perspective between Marx and his critics. It has since been forcefully brought home to me that these are not the decisive matters. They are both easily resolved in Marx’s favour, and of little immediate political concern. Marx’s theory of the tendency of the rate of profit to fall, however, is rejected even by many who want to retain at least some sort of Labour Theory of Value. Andrew Glyn, John Harrison and Sue Himmelweit all fall into the category of those who retain the theory of value as a ‘sociological’ theory of exploitation, or abstract labour, whilst adopting an essentially ‘neo-Ricardian’ method for looking at the rate of profit of the dynamics of the system.  The bulk of this article is therefore now taken up by this question.
On the one hand Marx takes for granted that labour is the basis for all hitherto existing societies. Labour is not only human activity but the relationship of man to nature and to other human beings, the production of the goods necessary to the survival of any society. This is a firmly materialist foundation. It goes hand in hand with an insistence upon the decisive importance of the relations between the producers and non-producers, between the ruling and the subordinate classes in any society. It governs Marx’s understanding of history as the succession of different modes of production.
On the other hand, Marx is concerned not with labour and exploitation in general but with the specific form taken is the form of value, the form taken by labour in a society of commodity exchange.
Most critics of Marx follow Smith and Ricardo and see the Labour Theory of Value as nothing other than a theory of relative prices. Their confusion in the face of the first chapter of Capital is therefore understandable. In moving so rapidly from the exchange relation between two commodities to the labour which they both contain Marx appears to simply assume what he is supposed to be proving. But what is at issue is not the actual prices of these two commodities but an exposure of their hidden social character. This rests not just on labour, but labour of a specific sort, abstract labour. It presupposes a division of society into private property-owners who can only relate to each other through exchange, in the competition of the market.
Smith and Ricardo had, unlike their more vulgar successors, understood the materialist basis of society. They had reduced the prices of individual commodities to the labour (direct and indirect) required for their production. They had failed, Marx repeatedly observes, to examine the form of value, to see exchange as the expression of specific social relations of production. 
The abstractions of the classical economists served to invoke features of society in general or the laws of nature. They took the existence of exchange and prices for granted. The abstract categories of Marx, of which Value is the most important, capture the essential social relations of Capitalism, the real structure of this mode of production concealed beneath the fluctuations of a market economy.
Marx begins by discussing the most general features of a society of commodity production, abstract labour, exchange-value, money, but proceeds rapidly to an analysis of the whole in which these relations find their fullest extension, Capital. As Rosdolsky has shown, on the basis of the Grundrisse, in the first two volumes of Capital Marx is concerned with ‘Capital-in-General’.  This involves an abstraction not from competition or the existence of many Capitals as even Rosdolsky at times suggests. Rather it involves abstracting from certain effects of competition, those deriving from the differences between Capitals when rates of profit are equalised, and the mass of surplus-value distributed in the course of exchange. Nor does it involve ignoring that this is a monetary economy. On the contrary Capital is analysed as but a developed form of Value, and in terms of the generation of surplus-value. These are measured in quantities of labour-time but both suppose that production starts and finishes with the purchase of one set of commodities and the sale of another.
’To the extent that we are considering here, as a relation distinct from that of value and money, capital is capital-in-general, i.e. the incarnation of the qualities which distinguish value as capital from value as pure value or as money. Value, money, circulation etc., prices etc., are presupposed as is labour etc. But we are still concerned neither with a particular form of Capital, nor with an individual capital as distinct from other individual capitals etc. 
Having firmly established the nature of Capital as a sum of value dependent upon the sale of labour-power and the performance of labour for its further expansion, Marx proceeds to a thorough discussion of Capitalist production. The relationship between the productivity of labour, the increase of surplus-value and the accumulation of Capital are the central concerns here. In Volume 2 Marx returns to the realm of exchange, only to consider it now in terms of the circulation of Capital. The analysis of the circuit of Capital shows how a given mass of Capital-value must continually pass through periods of production in which its value expands, and periods of circulation in which the mass of commodities is turned into money and back again into means of production and labour-power. The essential point of this is to show both the necessity of exchange, that profits are never made until the commodities have been sold — and that this conceals the fact that the mass of value and surplus-value cannot increase in the course of buying and selling alone.
It is on this basis that Marx comes to discuss the effects of the formation of equal rates of profit and prices of production. Yet this controversial matter is but a preliminary, a stage at which the mass of profit is still regarded as equal to the mass of surplus-value. Yet to be discussed are the further transfer of portions of this mass as interest and rent. Marx never does reach the stage of discussing the effects of international trade, the State and taxation, trade cycles and, not least, the concrete eruption of crises. At each stage the argument is becoming more complex and the actual determinants of particular prices multiplying. The abstract categories of value are not abolished in the course of this development but are continually presupposed. The surface forms can only be grasped on their basis.
The critics of Marx have usually followed Ricardo and identified value and exchange-value. Value becomes merely a relative price, at best distinguishable from the variety of market-prices by virtue of its stability, its function as the equilibrium point around which actual prices fluctuate. If this point is not proportional to the labour contained in the commodity; if, for example, it can as easily be explained in terms of the costs of production and the rate of profit, the labour theory of value in this version collapses. The structure of Marx’s work, the manner in which he contrasts the hidden social reality of value to the fetishistic appearances of the system, the exchange of things, is just ignored. The notion of value as an ‘intrinsic’ property of the commodity becomes metaphysical, an invocation of unseen essences to explain facts which can as readily be explained without them. 
If the word ‘value’ is not identified with price it is reduced to the actual performance of labour-time as in Ian Steedman’s use of the term ‘value quantities’. But this does not express Marx’s concept adequately either. Value is the ‘socially-necessary’ labour-time contained in a commodity produced for sale. It is the product of abstract labour, labour of a specific social character. There are three levels, to put it simply, in Capital, not two, each level containing its physical and social dimensions: the performance of labour (concrete and abstract labour); the commodity produced (use-value and value); and the commodity or money received in exchange (here it is the quantity of the use-value which serves as the measure of exchange-value or price, concealing the relations of labour behind the exchange of things). Value expresses the fact that the commodity has to be both produced and sold. The distinctions may appear pedantic but they are essential to a clear understanding of Marx’s theory of value.
Ian Steedman and his associates have no time for any of this. On the one hand; the labour theory of value is assessed only as a possible means of deriving relative prices and the rate of profit.  On the other, Steedman’s ‘value-quantities’ are explicitly derived from physical relations, or the technical coefficients of production.  The quantities of labour-time are treated as physical data without any reference to the social relations of Capitalism. The result, however much he may deny it, is that his ‘values’ will exist in any society in which labour is actually expended. This is a return to the natural laws of political economy. This conception, like Ricardo’s, completely fails to locate the prices and rate of profit which it invokes in the only context in which they can be understood, the nature of Capital. Ricardo at least had the merit of retaining labour as the basis of his theory of value despite the difficulties he faced in accounting for a whole number of phenomena on the surface of society. Sraffa resolves Ricardo’s difficulties only by abandoning all reference to labour-time. The ‘physical qualities’ picture in this respect is a regression beyond Ricardo, to the Physiocrats of the 18th century.
With this the question of how prices are related to values is posed. In supposing that Capitalists move between industries in search of the maximum rate of profit on the mass of their Capital, Marx has to consider both how a general rate of profit is formed for the economy as a whole and also the effects of this on the formation of what Marx calls ‘prices of production’, and how these deviate systematically from the values of commodities. Prices of production are themselves only the notional equilibrium points around which market-prices will fluctuate in response to shifts in supply and demand. They in turn will be modified by other influences not taken into account at this stage.
Capitalists do not set prices on the basis of a direct measurement of labour-time. They assess the costs of a unit commodity and add a mark-up for the profit-margin they consider obtainable. In a world of fully mobile Capital, without monopolies, these margins on costs will tend to be equalised across industries. Marx’s prices of production are based on these simple facts. They are given by the formula (C+V)(1+r) where C and V are Constant and Variable Capital respectively, the money spent on means of production and wages, and r is the average rate of profit. When Organic Compositions of Capital (C/V) are unequal, these prices will not be proportional to values.
These variations provide the basis for attempts to present Capital as an additional factor of production capable of creating value. This is how it appears to the individual Capitalist as both parts of Capital play an equal role in securing his profits. This led Adam Smith to abandon the labour theory of ‘value’ altogether in favour of the simplistic idea that prices were composed of the revenues earned by the different factors engaged in production How these revenues were themselves determined he could not really say.
Marx cut the knot which had strangled classical political economy. The premises of his solution have already been outlined. In the aggregate neither total value nor total surplus-value could be affected by the exchange of commodities. Their distribution among particular capitals would, however, be a function of the ratios at which commodities exchange for money. Money is the expression of abstract social labour, and particular sums of money are claims to a share in the total labour of society. Variations in prices will alter these claims but not the total to be distributed. If some prices of production are higher than what have been termed ‘direct prices’ (prices proportional to values), others can only be lower. 
The average rate of profit was measured by Marx in value terms, as the ratio of total surplus-value to total Capital. Equal rates of profit will involve a flow of surplus-value from labour-intensive to capital-intensive industries, or between Capitals of low and high Organic Compositions. This is the simple effect of the fact that prices are above ‘values’ in the latter, and below in the former. When Organic Compositions are equal, prices will be proportional to values, as a result of movements of Capital which involve no direct comparison of rewards to labour. In any event changes in the productivity of labour will be the most important causes of shifts in prices of production. Values are the underlying regulators of prices.  It is the conclusion which most strikingly establishes the manner in which the laws of value work behind the backs of the producers themselves.  None of this has been challenged in the subsequent debate.
This stems from the indisputable fact that, as he himself recognised, Marx’s solution was incomplete.  In his numerical examples Marx assumed that the costs on which the rate of profit was calculated were equivalent to the values contained in the commodities concerned. In reality alterations in the prices of means of production and means of subsistence (assuming a given real wage) will in turn alter the costs on which prices of production are based, producing a whole series of further effects.
This problem has been solved. The mathematics of the question are not in doubt although much argument rages about several minor variations. The reader is referred to two recent articles by Gerstein and Shaikh for definitive accounts of the manner in which both a complete set of relative prices and a corresponding rate of profit can be derived from a specification of the labour which enters directly and indirectly into each commodity.  There are only two serious problems posed by these solutions.
Firstly Marx’s two ‘invariance’ conditions. Total Value equal to Total Price and Total Surplus-Value equal to Total Profit will, except under restrictive conditions, be incompatible. This is easily resolved. Total Value expresses the total direct and indirect labour contained in the mass of commodities. If the value of money is unaltered, this entails that it is the first condition which must be retained. This is because we are dealing here only with relative prices and money cannot have a relative price (changes in its value will affect all prices equally). 
Changes in total profits as a result of the Transformation reflect the fact that profits will be affected by changes in relative prices, rising or falling with overall movements in the costs of production. Since such movements must be balanced by changes in the prices of the goods received by the Capitalists, the actual physical quantities of those goods will not be altered.
Secondly Marx’s formula for the rate of profit in value terms (S/(C+V)) will normally be incorrect. These alterations to the rate can be explained, however, in the same way as the alterations in total profits with which they go hand in hand. Although Steedman and his associates make great play with this, it is in fact just another calculable deviation. Marx’s account of the underlying determinants of the rate of profit is not affected. 
Steedman argues that there is no need to transform values into prices because relative prices can perfectly well be explained without them. This ‘redundancy’ argument collapses complete once the structure of Marx’s work is understood. There is only one serious question, given the centrality of quantities of socially necessary labour-time to the functioning of Capitalism, the question of how values and prices are related. Sraffa has shown that the same prices and rate of profit as those in a complete transformation of values can be derived directly from a specification of the different physical conditions of production of commodities (their technical coefficients) and the real wage. In purely formal terms this may even be simpler than what Steedman terms the ‘detour’ through values. But we are not dealing with purely formal matters. Sraffa’s models just do not capture the complex structure of Capitalism. Their deficiencies can be crudely summarised as follows.
Firstly they give us only relative prices, expressed not in terms of money but of ratios between things. Money itself cannot be fitter into Sraffa’s input-output physical tables.  It can only be understood as the measure of value, the commodity which represents total social labour. This demands the use of those aggregates such as total value and total price which have no meaning for the Neo-Ricardians.
Secondly a logical ‘derivation’ is not a causal explanation. The derivation of prices from physical conditions of production tells us something negative about supply and demand theories. But the physical data contain within themselves two sets of information which Marx clearly separates, quantities of labour-time and differing ratios of dead to living labour, or compositions of Capital. Such a separation is not unnecessary but essential to understanding the causes at work.
Thirdly labour is represented in the equations only by the means of subsistence which the workers consume, which in turn appear as inputs into production on a par with means of production. An increase in the productivity of labour will be represented by a reduction in the amount of ‘wage-goods’ required for the production of a given output of commodity. Instead of explaining the source of Capital’s ability, as value, to expand itself, it remains an unexplained mystery.
Fourthly input-output models of the physical interdependence of use-values can deal only with the technical relations between industries. The basic social unit of Capital, however, can be a firm, or even a State, but never an industry. This, and the absence of money, means that the essential social character of the competitive relations between Capitals, both within and between industries does not appear. Exchange relations become the mere expression of technical coefficients, not the necessary result of a social division of labour between private units of production which only relate to each other through the unplanned chaos of the market.
Finally Sraffa’s models are purely static exercises in equilibrium analysis. This, like several of the above, is a general feature of Transformation operations whether they start from values or not. In this lifeless, frozen world it doesn’t make any difference whether one substitutes quantities of things for quantities of human labour. This simply shows that one cannot squeeze the content of the theory of value into the material required for equilibrium models of relative price determination.
Sraffa himself cannot be entirely blamed for the use to which his work has been put. He makes no pretentious claims for it and clearly specifies the very restrictive assumptions with which he is working. But his is a world in which we have rates of profit but no Capitals, prices but no money, production but no labour. The consequences of mistaking this for reality are serious and will be seen further in the next section.
This is the question which produces the sharpest confrontation of Marx and the Neo-Ricardians. Can the rate of profit only fall (excluding questions of cyclical shifts in demand) if either real wages rise or raw material costs rise as a result of declining productivity? Or will it fall as an expression of the most far-reaching contradictions of Capitalism, of the character of Capitalist Accumulation, the replacement of living by dead labour, the rise in the Organic Composition of Capital, and most basically because of the fact that labour-time remains the measure of value, the basis for exchange, despite the enormous growth of the forces of production? The former was the position of Ricardo, and is held today not only by Steedman but also by many who claim to adhere to the Labour Theory of Value. It provides the basis for an explanation of the current world crisis solely in terms of a distributional struggle between classes.
Marx’s argument explaining the tendency of the rate of profit to fall, followed directly from his understanding of the Labour Theory of Value and the Accumulation of Capital. The basis for it is to be found in his account of the latter in Part 7 of Capital I, although the Law itself is not discussed until Volume 3, after the account of the formation of a general rate of profit for the economy as a whole.
Whereas Ricardo identified the rate of profit with the rate of surplus-value (S/V) Marx shows that the rate depends also upon the size of the Constant Capital employed in the economy. In value terms this is captured by the formula (S/(C+V)). In price terms the rate of profit expresses the ratio of the mass of profits to the mass of money-capital spent by the Capitalist in purchasing means of production and labour-power.
The accumulation of Capital can proceed without altering the Organic Composition of Capital (C/V or the ratio of Constant to Variable Capital), as long as reserves of labour-power are available. Generally, however, it will involve technical change. The Capitalists will not just extend the scale of production but seek to cheaper costs by introducing new sorts of machinery. Increases in the productivity of labour will go hand in hand with the replacement of living labour by dead labour, workers by machines. This process will generally increase the rate of surplus-value, as the proportion of ‘necessary’ labour-time or the value of labour-power falls (assuming that the ‘real wages’ of the workers do not rise). But this increase in the rate of surplus-value goes hand in hand with a tendency to reduce the labour-force and thus the sole source of surplus-value. It also involves a rise in the Organic Composition of Capital, an increase in the mass of Constant Capital tending to bring down the rate of profit. This is a perfect example of what Marx meant by the contradictory nature of Capitalist development.
If accumulation proceeds rapidly enough, the workers displaced by the machines will find jobs elsewhere, the whole process continually displacing labour from one sphere of production and making possible an expansion without technical change in other sectors. The mass of surplus-value will rise, and thus the mass of profit, but the rate of profit may still decline if the Constant Capital rises even faster. If accumulation does not occur, the pressure on the rate of profit will be reduced from this direction but the reserve army of labour will rapidly rise in the face of continual technical change. The mass of surplus-value must decline as a result and the system will enter into crisis.
It is vital to grasp that the rise in the ratio of dead to living labour (or C/(V+S) which in many ways captures the important relationship better than just C/V) involves a rise in the one portion of Total Value which can rise without limit. Given that accumulation is the accumulation of surplus-value, initially as money-Capital and then in the form of means of production and labour-power purchased on the market; and given that reserves of labour will have strict limits, three possibilities follow. Firstly, the accumulation regularly of even a fixed share of surplus-value will lead to a rise in the value of the mass of Capital; or secondly, much of this accumulated Capital is wiped out, losing its value, which, regardless of how it occurs, must be seen as a loss by the Capitalists concerned; or, thirdly mechanisms ensure that Capital is never accumulated faster than the expansion of the available labour-force. These require further consideration, for we are not dealing just with abstract possibilities.
It is also necessary to see that Marx’s analysis of the effects of increases in productivity is radically different from orthodox or even ‘common-sense’ accounts. The latter see an increase in productivity as increasing directly the mass of physical use-values and therefore the mass of value. Marx sees the mass of new value produced in any given period of time as given by the time and effort expended by the labour-force. Increases in productivity entail a devaluation of each commodity (a decline in the amount of socially-necessary labour-time they contain). In price terms (assuming that the value of money is unchanged) the sum of prices will not change if the sum of value does not alter.  As output rises the price of each commodity must fall as indeed occurred in the pre-inflationary age of 19th century Capitalism.
Productivity changes cannot therefore directly increase the mass of surplus-value. Relative Surplus-Value will rise with the rise in S/V as a result of increased productivity but only indirectly, as a result of the fall in the value of labour-power. (In a world of falling prices this entails that money-wages are forced down whilst real wages are constant. In a world of inflation the same effect occurs if money-wages are stable while money-prices rise, but this requires a fall in the value of money which can legitimately be ignored without affecting the validity of the argument.) 
Ricardo saw the decline in the rate of profit as stemming from diminishing returns in agriculture or a decline in the productivity of labour leading to a rise in the price of corn. This in turn led to a rise in wages forcing down the rate of profit. This made the whole process dependent upon a law of nature and not something inherent to the character of Capitalism. Marx rejects this position firmly both on factual grounds and because he sees that
’The progressive tendency of the general rate of profit to fall is, therefore, just an expression peculiar to the capitalist mode of production of the progressive development of the social productivity of labour’ (Capital III, p: 213) …
’The rate of profit does not fall because labour becomes less productive but because it becomes more productive. Both the rise in the rate of surplus-value and the fall in the rate of profit are but specific forms through which growing productivity of labour is expressed under Capitalism.’ (ibid., p. 240)
But how does this work in practice? The mechanisms can be complicated but it is essential to see that the dependence of the mass of profits upon the surplus labour-time performed by workers is not apparent to the individual capitalist. The result is that the rational actions of each Capitalist in seeking to increase profits can have the effect of reducing the rate of profit for the Capitalist Class as a whole. This can be seen in two stages:
What though of the rise in the rate of surplus-value? This will certainly offset the effects of a rise in the OCC. But there are clear limits to this. The higher the rate of surplus-value, the higher the proportion of the working-day devoted to surplus labour-time. As the system ages, as the rate of surplus-value rises, the scope for further expansion in the mass of surplus-value is limited.  It is on this basis that Marx argues that the tendency for the rate of profit to fall will gradually become more marked over time. Other factors, listed by Marx as Counteracting Tendencies, may also prevail for periods of time.  But if the OCC continues to rise the underlying pressure on the rate of profit will steadily increase. It is the question of whether the OCC actually does rise in the long-run that has provoked most controversy.
The critics of Marx almost all share one feature in common. They effectively abstract from the context in which alone the law can be understood, the Accumulation of Capital. Operating with the static equilibrium models of technical change typical of bourgeois economics, they treat it in terms of a choice between ‘available techniques’ as if it was a matter of choosing a curtain design. The competitive compulsion to accumulate on the one hand and the manner in which the labour-process is altered on the other, both disappear. The actual criticisms though are varied. They will be taken in order of seriousness, the most trivial first.
(1) Price Rate of Profit not equal to Value Rate of Profit. This stems from the Transformation Debate and has already been considered. Such deviations do not undermine Marx’s location of the basic determinants of the rate of profit. Nor has any evidence been provided to show that in reality such deviations will be significant. If anything such deviations will reinforce any fall in the rate of profit. 
(2) Wage-Profit Frontiers.  This argument can be made to look very impressive, with diagrams etc., but is in reality almost unbelievably crude. It states that if Capitalists follow profit-maximising policies then for any given level of the real wage, there will be a maximum rate of profit possible at which Capitalists will operate. They will never introduce techniques which give less than this maximum rate.
Now this assumes that Capitalists are able to perceive the effects of their actions, and that there can be no contradiction between the interests of the individual capitalist and the Capitalist class as a whole. Have these people really read Capital?
And even if Capitalists could see the effects of their actions the argument still requires a further assumption—that they act together, that they are effectively a single Capitalist, or that there is no competition within the industry, even in the world-market. This, as noted earlier, is the assumption implicit in the model of Sraffa from which the argument derives. In reality there can be techniques which lower costs of production but, when generalised, also lower the rate of profit. Competition will ensure that such techniques are introduced.
This mistake stems from the misuse of equilibrium models to deal with what are inherently disequilibrium situations. The wage-profit frontier shows that as wages change the rate of profit will change and that new techniques may be introduced as a result. It isolates distribution changes as the sole cause of technical change, abstracting from all other forces, i.e. all those discussed earlier.
(3) Okishio’s Theorem. Faced with such objections to the last argument economists such as Himmelweit and Harrison resort to an obscure piece of mathematical reasoning known as Okishio’s Theorem. Okishio assumes that Capitalists only introduce techniques which lower the costs of production. Unlike the ‘profit-maximisation’ assumption this is perfectly correct. On this basis, however, Okishio argues that any innovation which lowers the costs of production can only increase the overall rate of profit.
This is a more serious argument. But the mathematics seem to exclude a possibility which is perfectly plausible in Marx’s framework. This requires looking at the economy as a whole. The total Value or the total product in physical terms can be broken down into C+V+S or their use-value equivalents. Okishio agrees that these price-changes will have further effects. If they arise in the capital goods or the wage goods sectors they will lower the price of C or V and thereby increase the rate of profit. If they occur in Dept. III they will not affect it at all. 
In Himmelweit’s version of the argument the mistake is as follows. Starting from a given set of physical product ratios she sees only the price-effects of changes in labour productivity. But, unlike Transformation exercises, neither the Total Value nor the physical ratios between Departments can be taken as given in this case. The rise in labour productivity is not disembodied but will occur as a result of a rise in the quantity of machinery or Constant Capital in the economy as a whole. As Shaikh has pointed out, Okishio treats all Capital as only Circulating Capital and thus as a matter of raw materials and wages.  He abstracts from fixed Capital or machinery and thus from the fact that such increases of labour productivity have to be purchased at a ‘cost’ to the Capitalist.
This, unlike the rest, is the genuine problem which is often used to give a verbal plausibility to the argument of the Okishio theorem. The fact that a rise in the OCC leading to a fall in the rate of profit can occur does not mean that it will occur. Increases in the productivity of labour can mean a cheapening of the means of production and this can offset the rise in the Value of Constant Capital as a whole. Levels of mechanisation can increase the ‘technical’ composition of Capital, but not the value ratios which are decisive for the rate of profit. Marx himself recognises (indeed clearly formulates) the question but states that
’... with the increasing productivity of labour, the mass of the means of production consumed diminishes. Their value therefore rises absolutely but not in proportion to the mass.’ (Capital I, p. 774)
There is no satisfactory discussion of this question in the literature. The following is but a sketch of an answer to the critics who seize on this as their last resort.
The question is often wrongly posed, in terms of ‘capital-saving’ and labour-saving innovations. As Yaffe correctly insists the notion of ‘capital-saving innovation’ taken by itself is ideological’.  If labour is the sole source of value, a cheapening of the value of machinery must itself involve increases in the productivity of labour. These will normally themselves require a rise in the Constant Capital of Department I industries making machinery. This in turn is why Marx is quite correct to argue that it is the rise in the OCC which is the basic tendency and the cheapening of Constant Capital only a ‘counteracting influence’.
It may be objected that the new machines which increase the productivity of labour in Department I can also be cheaper than the machines which they replace. Scientific inventions may occur which make this possible.  Marx is supposed to have ignored this yet we find in Capital itself:
’It the productivity of labour has increased in the place where these instruments of labour are constructed (and it does develop continually, owing to the uninterrupted advance of science and technology), the old machines, tools, apparatus etc. will be replaced by more efficient (and considering their increased efficiency) cheaper ones ... science and technology give capital a power of expansion which is independent of the given magnitude of the capital actually functioning. They react at the same time on that part of the original capital which has entered the stage of its renewal. This in passing into its new shape incorporates, free of charge (my emphasis) the social advances made while its old shape was being used up.’ (Capital I, pp. 753–54)
The important point is that such inventions involve no cost to the Capitalist. No accumulation of surplus-value as Capital has occurred. The replacement of existing Capital and the accumulation of new Capital must be clearly distinguished even if they can occur together in reality. If existing Capital is replaced more cheaply, extra funds will be released for further accumulation which will restore the mass of Capital-value to its original level.
Only if either a) such funds released are used for consumption or hoarded rather than assimilated, or b) the cheaper machinery forces a depreciation of fixed Capital of an older variety still; employed by other Capitalists in the same industry (see below), will such innovations lower the Organic Composition of Capital. To the extent moreover that accumulation continues the tendency for the OCC to rise will reappear. Only in periods of stagnation when no net investment is taking place will such ‘intensive’ innovations simply replace more expensive by cheaper machines (microprocessors today perhaps). But in such periods they will go hand in hand with a rapid rise in unemployment and an absolute fall in the production of surplus-value.
However, there is ample evidence to suggest that Capitalists seeking to innovate are usually confronted not with cheaper but with more expensive systems of machinery. It is precisely the enormous scale and accompanying costs of the most up-to-date Steel and Chemical plants which make them unobtainable by the smaller and weaker Capitals in those industries, and demand ever-greater centralisation of Capital. The rise in the mass of profits which goes with seizing a greater share of the market will compensate the larger Capitals for any decline in the rate of profit.
To refer to the contemporary world again, silicon-chip based mini-computers etc may prove to be cheaper than the machines they replace (although living labour employed will also fall dramatically, raising the OCC from the other side). But re-equipping whole car-plants with new automated machinery and robots is still seen as prohibitively expensive by most manufacturers.
Simple numerical examples establish that if the productivity of labour increases uniformly in all Departments, changes in the ‘technical composition’ of Capital (the ratio of Means of Production to the Net Product) will be matched by proportional if smaller changes in the ‘value composition’ (C/(V+S)). The OCC will therefore rise unless either a) the technical composition does not rise (as in the cases discussed in ii) above or b) productivity increases in Department I are greater than average. Even in the latter case the increases in productivity in Department I alone must be sufficient to offset the accumulation of Constant Capital in the economy as a whole. This, however, will require a depreciation of fixed Capital already in existence which will have further severe implications (see below).
No algebraic proof can be provided to show that the OCC must rise in the abstract world of disembodied technical possibilities. The critics are forced to argue that the supposed rise in the ‘technical composition’ itself involves an assumption that technical change has a ‘labour-saving’ bias. It does. In a completely automated world there would certainly be no point in using machines to produce the same quantity of goods, (or more corn to produce a given output of corn in the Ricardian model). But we are in a world in which labour-time remains the measure of value and in which the whole drive towards automation expresses the tendency of Capital to free itself from a dependence on living labour, regardless of whether wages are rising or not. Marx’s whole analysis of machinery, spelled out in great detail , showed that this accumulation of means of production was the great historical achievement of Capitalism. That it results in a fall in the rate of profit and crises is simply the expression of the conflict between this growth in the forces of production and the social relations of Capitalism.
If the question is looked at concretely, instead of with arbitrarily chosen examples, and technical change located in its actual historical context, the basic validity of Marx’s argument is confirmed. On the one hand there are several pieces of empirical evidence which show that Capital-Output ratios (a very approximate index, based upon admittedly bourgeois statistics) rise in periods of slump, such as the interwar period, or in war. 
On the other hand there is Marx’s own discussion of the mechanisms of accumulation. The decline in the reserve army of labour as expansion proceeds entails that Capital can be accumulated only by raising the Organic Composition of Capital (Capital, Ch. 25 sections 1 and 2). Capitalists will feel confident about introducing techniques which greatly raise both the scale and capital-intensity of production. Whole new systems of machinery will emerge. The cheapening of unit commodity costs will go hand in hand with a greatly enlarged scale of production. As the new methods spread markets in particular industries will be glutted, prices will fall, and the least productive Capitals will be forced out of business. The mass of profit will increase but the rise in OCC will put pressure on the rate of profit. Only a quite remarkable series of ‘Capital-saving innovations’ could offset this tendency. Despite the great increase in productivity, pressures to increase the rate of exploitation will intensify. Speed-ups, continuous shift systems and attempts to hold down wages will all accompany the new machinery. To the extent that the working-class resists successfully they can accelerate the emergent crisis. They can scarcely be considered its cause simply for refusing to be its sacrificial victims.
Periods of slump, however, will involve the destruction of Capital-values, as a simple consequence of bankruptcies on the one hand, the scrapping of old machinery in the face of its rapid depreciation on the other. Marx refers to this repeatedly in his chapter on the Internal Contradictions of the Law  (of the fall in the rate of profit). At the same time, the accumulation of new Capital will slow down, and innovation increasingly take the form of cheapening machinery rather than extending the scale in production. A full discussion would however require looking at the mechanisms by which crises occur. There is no space for this.
We are not arguing here for some mechanical breakdown theory. The tendency may underlie the regular trade cycle but a major crisis can serve to restore the rate of profit—at an enormous cost. It is in the course of crises that prices are forced down to the level of the most productive capitals. This is but an expression of the manner in which the laws of value work only ‘in and through anarchy’. The forces of competition, so abstractly invoked in Steedman’s models, assume a savage and bloody form, with costs which are scarcely recognised in the harmonious, non-contradictory world of equilibrium.
The system continues to age and with it the historical forms of crisis. The contradictions are resolved, at a severe cost, only to be reproduced at a new and higher levels. The concentration and centralisation of Capitals is accelerated in each major crisis with the elimination of the smaller and weaker firms. The crisis itself end what Marx calls ‘the over-production of capital’ giving breathing room to the Capitals which remain, in which they can expand. But as the units of Capital get larger the consequences of their destruction become more severe. Within each national economy States intervene to sustain the credit-system and prop up the weaker Capitals. Inflation serves to conceal the underlying value relation; even more, the limits set to total surplus-value by surplus labour time obscured by the rise in gross money-profits. In this unprecedented era of stagflation we need to look more than ever at the new forms taken by the current world-crisis.
This requires that we develop Marx’s categories not abandon them. To this work Steedman and company have nothing to contribute, except insofar as they force us to consider what Marx was really up to in writing Capital.
Both parts of this article owe much to conversations with Derek Howl and Jim Kincaid, and a challenge posed by Dave Turner.
1. I. Steedman, Marx after Sraffa (NLB 1977). Steedman shifts back and forth throughout his book between the ‘redundancy’ and the ‘inconsistency’ argument. See the careful discussion in P. Armstrong, A. Glyn and J. Harrison, In Defence of Value (Capital and Class 5, Summer 1978).
2. See the verbal argument in J. Harrison, Marxist Economics for Socialists (Pluto 1978), pp. 115–17, and mathematical versions in S. Himmelweit, The Continuing Saga of the Falling Rate of Profit (Bulletin of the Conference of Socialist Economists – henceforth CSEB – No. 9, Autumn 1974). In their reply to Steedman, Glyn et al. defend the use of ‘measurements’ in terms of labour-time on ‘pragmatic’ grounds e.g. to clarify the nature of exploitation; Himmelweit and Mohun, The Anomalies of Capital (Capital and Class 6, Autumn 1978) reject any quantitative approach in favour of an odd notion of ‘Abstract Labour’. While they correctly point out the basic differences of Marx and Ricardo their view loses the ‘materialist’ aspect of Marx’s theory. In both case the final result is that Sraffian models are still invoked to handle the movements of prices and the rate of profit, and the Neo-Ricardian offensive can continue unscathed.
3. Capital I, p. 174. (All references to Penguin ed.)
4. R. Rosdolsky, The Making of Marx’s Capital (Pluto 1977), pp. 41–53 esp.
5. Marx, Grundrisse (Penguin 1973), p. 310.
6. Joan Robinson, Economic Philosophy (Penguin 1964), p. 39, remarks echoed in Steedman’s talk of ‘obscurantism’ (op. cit., p. 21). This sort of attitude tends to go with a distaste for Hegelian elements in Marx, and for the notion of ‘contradiction’ in particular.
7. This is seen clearly in Steedman’s article in the New Statesman, 5 January 1979. The editorial preface to this illustrates perfectly how Steedman’s work is being used, stating ‘that the “labour theory” of profit and exploitation has long seemed like a superstition, shutting off the Marxists as a formidable but self-absorbed tribe’.
8. Steedman, op. cit., esp. Chapter 3, and pp. 64–65. See the pointed criticism in B. Fine and L. Harris, Controversial Issues in Marxist Economic Theory (Socialist Register 1976). Steedman’s method of calculating values by simultaneous equations also leads him, not unwillingly, into such absurdities as ‘negative values’. The difficulties of determining quantities of labour-time are a recurrent theme in Steedman reinforcing his main arguments. Most of them stem from the attempt to impose methods of calculation whose relevance is only given by the mathematical constructions at stake. Different methods of calculation produce different results (cf. Armstrong, Glyn and Harrison, op. cit.). Others such as Joint Products (two or more products produced by the same labour-process) express a real indeterminacy which will be resolved by the Capitalists in practice alone. (Capital I, p. 135).
9. The term ‘direct prices’ is taken from an article by A. Shaikh, Marx’s Theory of Value and the “Transformation Problem”, to be found in a collection The Subtle Anatomy of Capitalism, ed. J. Schwarz (Goodyear Pub. Co., California 1977). This unfortunately inaccessible work is by far the best treatment of the Transformation question. A similar idea termed ‘Simple Prices’ is found in D. Yaffe, Value and Price in Marx’s Capital (Revolutionary Communist, No. 1). Yaffe, however, refuses to recognise that there are any problems with Marx’s solution, following Rosdolsky in this respect, in the belief that any attempt to ‘Transform the Inputs’ entails abandoning values altogether.
10. Capital III (L&W 1971), p. 166 & p. 206.
11. Capital I, pp. 166–67.
12. Capital III, p. 161 and pp. 164–65. But he comments ‘However this always resolves itself into one commodity receiving too little of the surplus-value while another receives too much, so that the deviations from value which are embodied in the prices of production compensate one another. Under Capitalist production, the general law acts as the prevailing tendency only in a very complicated and approximate manner as a never ascertainable average of ceaseless “fluctuations”.’ This, while imprecise, is essentially correct.
13. Shaikh, op. cit., and I. Gerstein, Production, Circulation and Value (Economy and Society, 1976, Vol. 5, pp. 243–91). Shaikh presents an alternative method of calculation, the ‘Iterative Solution’ which he strongly argues reveals the actual movements better than the method of Simultaneous Equations preferred by Gerstein and most other commentators. The final results are the same. Gerstein draws on the article by F. Seton, The “Transformation Problem”, in The Economics of Marx, ed. M. Howard and J. King (Penguin 1976), which provides the definitive mathematical version of the question to which Steedman adds nothing. Seton, a bourgeois economist, concluded that “the logical Superstructure” of Marx’s work was “sound enough” whilst rejecting the basic “doctrine”.
14. Von Bortkewiecz confused matters totally by including money in his equations, in which respect, he has been followed by Steedman. This is quite wrong. Sweezy, The Theory of Capitalist Development (MR 1968), does the same thing and then concludes that differences of Total Value and Total Price etc. result simply from moving from a Value unit of account to a Money Unit of Account. This has got nothing to do with it, is in fact clearly left aside by Marx (he had dealt with that question in Vol. 1). G. Hodgson makes the same very confusing mistake in his Marxian Epistemology and the “Transformation Problem” (Economy and Society, Vol. 3, 1974). Yaffe (op. cit.) was the first person to point this out correctly in the recent literature.
15. Sraffa’s own Standard Commodity amounts to an exact specification of an average industry for which the rate of profit will not be affected by the Transformation (see the account in A. Medio, Profit and Surplus-Value: Appearance and Reality in Capitalist Production, in E. Hunt and J. Schwarz (eds.) A Critique of Political Economy (Penguin 1972) ,which argues for putting Sraffa into Marx rather than vice-versa). Sraffa’s formula for the rate of profit (r = Rmax(I−w)) isolates precisely the two factors, the Organic Composition of Capital, and the rate of surplus-value, with which Marx determines the rate of profit. Sraffa’s account however is placed in the distorted context of Ricardo’s search for an invariable measure of value, a commodity whose value would not change as wages and profits altered. For Marx’s comments see Theories of Surplus Value, vol. 3, pp. 133–35 where he calls Ricardo’s problem ‘spurious’. Also see Shaikh (op. cit.) for why a method which takes Marx’s own solution as a ‘first step’ towards a complete answer is preferable.
16. Sraffa simply takes an arbitrarily chosen commodity out of his equations as a unit of account (equal to one) to give him absolute numbers rather than just relative prices (ratios of one thing to another). Money is not an arbitrary commodity, and cannot simply be added onto the equations.
17. See the first part of my article, International Socialism 2:3.
18. Strictly speaking if money wages are stable and money-prices are stable but output rising, the value of labour-power will be falling, and the rate of exploitation increasing. If we follow the strict logic of Marx’s argument, a stable price-level implies that the value of money is falling (if productivity rises and values fall commodity-prices must fall if the value of money is unchanged.) Nobody seems to have discussed the implications of this, not least for the manner in which the underlying value-relations are concealed by movements in money-prices. The neglect of this question undermines attempts to apply value-categories ‘immediately’ to the concrete dynamics of the system.
19. This matter is fully and clearly discussed in both R. Rosdolsky (op. cit., pp. 398–411) and R. Meek, The falling rate of profit, in his Economics and Ideologv and Other Essays (Chapman & Hall 1967). Both are well worth reading.
20. Marx lists several such ‘Counteracting Influences’ in Capital III, Ch. XIV, such as an increasing intensity of labour, the cheapening of Constant Capital, over-population, foreign trade etc. This is only a list however, bearing the marks of an uncompleted work. The question must be looked at historically, taking into account those forces which have in practice diverted large quantises of surplus-value from the Accumulation of Capital, notably Imperialism, Wars, and the Permanent Arms apparatuses of the post-war period.
21. Steedman, op cit., ch. 9 uses this, as well as several other arguments suggesting that he does not find this objection on its own very conclusive. According to Pete Binns, Garcia (in a Spanish revolutionary journal, Debate, 1978) has analysed the effects of actual variations in organic compositions of capital between industries and concluded that they are likely to accentuate any fall in the rate of profit, if they have any effect at all.
22. These can be found in Steedman, op. cit., and Himmelweit (CSEB, No. 9, 1974). They stem from Sraffa’s discussion of the effects of changes in distribution upon the choice of techniques. In this purely static framework they have a valid use, and have served to expose inconsistencies in marginalist economics. But dealing only with industries not firms, and thus abstracting, from the real processes of competition, they are useless for dealing with the effects of accumulation in which wage-changes are the ‘dependent, not the independent variable’. (Capital I, p. 770)
23. The conclusion that changes in the conditions of production in Dept. III producing luxury goods cannot affect the overall rate of profit, corresponds to Sraffa’s argument about non-basic goods, and the famous ‘Von Bortkewiecz corollary’. But the fact that one can drop the equation for Dept. III from an equilibrium model and still obtain a ‘rate of profit’ says nothing at all about what happens in the real world. It is astonishing how readily such an idea, which conflicts with the essence of Marx’s analysis of surplus-value, has been accepted. In an equilibrium model, the conditions of production in Dept. III will be given by a specification of those in Depts. I & II simply by the assumption that the total surplus of goods in the first two departments is sold to Dept. III and that the latter purchases none from elsewhere. The fact that changes in the price of luxury goods cannot affect the overall rate of profit then enables you to drop Dept. III from the equations.
Once money, firms and disequilibrium are introduced the argument collapses. If conditions of production ‘worsen’ in Dept. III causing a fall in the rate of profit firms or money-capital will depart and the price of luxury goods will rise restoring the rate of profit. This change in price will not affect the overall rate of profit. That is as far as the Neo-Ricardians can go. But what of the drop in demand for goods from other Depts. which stems from this contraction of Dept. III? And what of the firms which left the industry? These will enter the other Departments and lower the rate of profit there as a result. These changes in the conditions of production or the technical data in Depts. I & II as a result of changes in Dept. III are excluded by definition from the Sraffian models. The final effect, however, will be to lower the overall rate of profit.
24. A. Shaikh, Political Economy and Capitalism: notes on Dobb’s theory of crisis (Cambridge Journal of Economics, 1978, No. 2). Shaikh makes an important distinction between the profit margin on the costs which enter into the commodity, and the rate of profit calculated on the whole fixed Capital engaged in production. This in turn leads into a series of complications arising from different turnover periods of Capital which Marx discusses in Volume 2 of Capital.
25. D. Yaffe, The Marxian Theory of Crisis, Capital and the State (Economy and Society, Vol. 2 No. 2, 1973). The fact that much of this article is an attack on the theoretical tradition represented by this Journal should not detract from the important and correct points made by Yaffe on the basic tendency of the rate of profit to fall.
26. B. Schefold, Different Forms of Technical Progress (Economic Journal, Dec. 1976, No. 86) calls this case ‘innovation’ as distinct from the ‘mechanisation’ upon which Marx focussed, which will tend to lower the ‘maximum rate of profit’. Schefold is interesting as a Sraffian who rejects Marx’s basic thesis but seems to allow that rises in the Organic Composition of Capital can lower the rate of profit independently of rising wages. This at least permits a sensible discussion of the question which the ‘a priori’ rejections do not.
27. cf. Theories of Surplus Value, vol. III, p. 366.
28. Capital-Output ratios do have to be handled carefully. Physical measures involving heterogeneous commodities are especially problematic as distinct from value or price measures. But the latter will be affected by relative price changes, levels of demand etc. Allowing for these the evidence from several sources shows a remarkably similar pattern. See J. Gillman, The Falling Rate of Profit (Dobson 1957) for American data which show a steady rise in the Organic Composition up until 1920 and, allowing for shifts in capacity utilisation, a steady fall from 1920 to 1950. Hodgson (op. cit.) makes great play with the latter but as Cogoy correctly observes it is absurd to isolate such a fall from the context of a massive world-slump and a world-war. M. Cogoy, The Fall in the Rate of Profit and the Theory of Accumulation (CSEB, II.7, Winter 1973). More recent evidence suggests a steady rise again in Capital-Output ratios since 1950. See for example on Britain the survey Measures of Real Profitability in the Bank of England Quarterly Bulletin (Vol. 18 No. 4, Dec. 1978).
29. Capital III, Chap. XV passim.
Last updated: 16.5.2012