Keynes, bottom left, at the signing of the Anglo-American loan agreement in 1945. Photo: wikimedia commons Keynes, bottom left, at the signing of the Anglo-American loan agreement in 1945. Photo: wikimedia commons

In the fourth part of our series, Dominic Alexander looks at the problems with Keynesian critiques of Marx’s labour theory of value 

The Limits of Keynesianism series:
Part One: John Maynard Keynes and Orthodox Economics
Part Two: The Assumptions Keynes Makes
Part Three: Marx, Keynes and the Analysis of the Trade Cycle
Part Four: The Keynesian Attack on the Labour Theory of Value
Part Five: can the working class advance within capitalism?

In Keynesian thinking, there are arguments in favour of more egalitarian societies, and a preference for higher shares of social wealth to be directed towards the mass of people. Nonetheless, its theory never abandons the perspective of capital. Analytical individualism is so ingrained that even a consciously left-wing Keynesian economist like Joan Robinson misinterprets Marx at several points in her critique, because she assumes the perspective of a single capitalist to be the starting point, where Marx’s analysis takes the social whole as prior. Alongside this is the refusal to understand value as a social relation, insisting upon a reductive materialist analysis. With these two prior assumptions at work, Keynesians like Robinson or heterodox economists like Steve Keen leap to dismiss the labour theory of value.

Robinson attacked the very concept of value, arguing that it was just metaphysical, a ‘mysterious emanation’ in Marxism that ‘was still somehow lurking in relative prices’.[1] Yet there is a difference between a quality that cannot be apprehended as a concrete substance and something that has no real existence. There are different analogies that can explain this, but one is human consciousness. So long as we do not invoke Descartes’ ghost in the machine, consciousness is a phenomenon of the material world, yet there are no particles of consciousness. It does not exist in this or that neuron as a tangible unit. Rather, it is the creation of the totality of brain activity; it is an emergent property that depends upon movement, or process, to come into a very real existence.

Value is very similar, except that it exists at the level of total social interactions, rather than neuronal ones. In a parallel way, value can only exist if the volume and complexity of economic interactions reaches a certain level in a society. Thus, in medieval economies the level of exchange was at a relatively low level and intensity (in the sense that production for exchange represented a low proportion of household activity). While the presumption that labour had something to do with the value of a commodity was present, it was not strong enough as a social relation to overcome other influences on price. Value was not, consequently, a dominant relation over the processes of production.[2]

In a different analogy, David Harvey uses gravity to explain:

As such it [value] is, like gravity, an immaterial but objective force. I cannot dissect a shirt and find atoms of value in it any more than I can dissect a stone and find atoms of gravity. Both are immaterial relations that have objective material consequences.[3]

There are many very real phenomena that cannot be isolated as discrete physical objects, and social relationships are almost a category on their own of these real but ‘immaterial’ forces; there is no substance in which we could locate love, solidarity, or class conflict.[4] However, if we follow the arguments of economists like Joan Robinson, we should dismiss their existence, since each case is like that of value; ‘It has no operational content. It is just a word.’[5]

Marx against reductionism

Dismissals of Marx’s labour theory of value often assume his conception to be identical with that of his predecessor, the classical economist David Ricardo (who was no friend of the working class). Marx took Ricardo’s theory but altered it in subtle but crucial ways, so these criticisms actually fall at the first hurdle.[6] One type of argument used against the labour theory of value is to take some substance with rarity value and claim that its high price as against other commodities shows that the labour inherent in it cannot therefore be the measure of its value.

One example that has been used by way of illustration is the case of ambergris (whale vomit). A chance find of a pile of ambergris upon the beach can be sold very lucratively to the perfume industry, and yet the person who finds it has expended very little labour, while walking upon the beach, relative to the amount of exchange value involved. The flaw in this reasoning, and many others like it, is that it assumes that value is defined through individual cases. This is not so; value is a social relationship and is defined through the social processes involved.

It is not the individual labour in a particular commodity item that is the measure of value, it is the total socially necessary labour time that defines the value contained in commodities of a particular kind. It is not one person walking upon a beach, but the average social labour taken. That is to say, what counts is the total hours spent by the total number of people, walking along all the available beaches, that has to occur before one person strikes it lucky. The case of ambergris actually bears out Marx’s theory so long as it is read, as it needs to be, in the social frame, not the individual one. As Harvey says:

Marx defines value as socially necessary labour time. The labour time I spend on making goods for others to buy and use is a social relation.[7]

This is important in a methodological way. It defines the difference between Marx’s dialectical version of materialism, and the standard positivist version almost universally espoused in academia:

Physical materialism, particularly in its empiricist garb, tends not to recognise things or processes that cannot be physically documented and directly measured.[8]

A real science of economics must not take a reductive materialist approach if it is to understand capitalism. Capitalism needs to be seen in terms of a series of social relations which create particular dynamics. Positivists want all facts to lie upon the surface, to be directly observable and countable, and yet, as anyone at all acquainted with modern physics will grasp, reality just does not work like that. What Marx’s analysis achieves is to lay bare the often counter-intuitive results of the dynamics of capitalism, which are hidden from plain sight.

For Robinson, wedded to the capitalist’s pragmatic perspective, this is precisely what is wrong with the theory of value; it does not correspond to common sense; ‘the academics can score a point against Marx, who always reckoned in terms of average cost, because in this connection the principle of marginal cost, or rather cost at the margin, corresponds to common sense.’[9] Yet, Robinson fails to notice that just as common sense does not necessarily correspond to the real workings of the world, the perspective of the individual capitalist does not correspond to the social dynamics of the whole. In this respect, Keynes’ thinking was superior to his supposedly more radical ally, as when he was able to perceive the counter-intuitively negative impact of individual saving on the whole economy.[10]

The dynamics of the social whole

So, again, in Marx, value is determined by the socially average amount of labour needed to produce a commodity, but how much a capitalist will produce and at what price that commodity will be sold depends upon a range of factors.[11] At a social level, the profit returning upon commodities will be determined by the ratio of labour and capital in their production, which is known as the organic composition of capital (the greater the amount of machinery and other capital investment, the higher is the organic composition). This does not mean that a particular factory relying upon more human labour than machinery will receive a higher rate of profit than a technologically more advanced one, even though the former factory is more productive of surplus value than the latter.

This is because the more advanced industries, with a higher organic composition of capital, tend overall to capture a higher share of the socially available surplus.[12] The dynamic of capitalism is contradictory here, with the industries which produce more surplus value capturing less of the socially available surplus than those with a higher organic composition of capital. This is even though the latter, because they use less labour relative to capital, or have a higher labour productivity, produce relatively smaller quantities of surplus value. The tendency over time is therefore, at the social level, for a smaller proportion of value to be produced than capital advanced, leading to a tendency for the rate of profit to fall.

Critics of Marx simply do not follow this line of argument, insisting that the ‘facts’ must indicate that everything, capital, labour, machinery and raw inputs, must all contribute to value. The failure to accept that the analysis must proceed from the perspective of the social whole is the problem. Robinson, for example, insists on evaluating the concept of value according to whether an individual capitalist can use it to determine the price at which particular commodities should be sold, showing to her satisfaction that a capitalist will not rationally sell commodities on the basis of value, but rather of marginal utility.[13] In arguments such as this, Robinson entirely misses the purpose of Marx’s analysis because of the profound methodological differences between her positivist-Keynesianism and the dialectical Marxist approach.

Most of Robinson’s arguments against Marx founder because she refuses to accept the concept of value according to Marx’s usage; the critique therefore only proves Robinson’s own assumptions to be erroneous, not Marx’s.[14] The degree to which Keynesians tend to import their own assumptions into the discussion of Marx is revealed acutely where Robinson absurdly claims that Marx’s argument depends upon Say’s Law at points, when Marx is well-known to have held Say in some contempt.[15]

More fundamentally, Robinson, at the start of her argument on the labour theory of value, refers to labour time without clarifying the role of labour power. The distinction is absolutely crucial to Marx’s revision of Ricardo’s version of the theory. Labour power is a qualitative term, in which the social relations are revealed; it is the potential for productive power, whose realisation depends upon a struggle between labour and capital over how much value the latter can extract from the former, in return for a given quantity of wages.[16] The difference between labour time and labour power also explains why a machine or an animal cannot be the source of extra value; both can only release the value embodied in their reproduction, whereas only human labour can, firstly, transform past labour into a new form, and secondly, be made to create more value than is required to reproduce itself.[17]

Reducing Marx

The core of Robinson’s critique of Marx’s labour theory of value, and Steve Keen’s following Robinson, comes down to what is known as the transformation problem.[18] This is a highly technical issue centred on only one section of Capital, so it is rather disingenuous of both Robinson and Keen not to acknowledge this.[19] It has been pointed out that this section, even though Engels included it in volume three, was written in 1864/5, before the publication of volume one in 1867.[20] The assumptions made in it cannot, therefore, be taken to represent the entirety of Marx’s labour theory of value.

Unfortunately, this is precisely what the Keynesian critics do, holding it as representative of the entire argument. Robinson and Keen claim, on the basis of the passage in volume three, that Marx holds the rate of surplus value to be constant across time and different industries, and apply this to the logic of the whole theory. That, however, is just an analytical limiting assumption made at this particular point in order to make a set of calculations possible.[21] Elsewhere, since the rate of surplus value is a measure of labour productivity, Marx allows it to vary. Any reading of the rest of Capital would show Marx’s analysis works through the consequences of the variations of the rate of surplus value. Marx’s transformation tables in volume three were not meant to do the work that the Keynesians imputed to them, and the critique has therefore been founded on false premises. Whatever the technical problems involved, the issue is certainly not the definitive disproof of the labour theory that Marx’s critics have presented it to be.

Robinson insists that Marx’s analysis depends upon a constant rate of surplus value in order to debunk the argument about the falling rate of profit.[22] Her discovery of fatal ‘inconsistencies’ in Marx simply ignores how the analysis is built through holding some factors constant at certain points, in order to explore the resultant dynamics. Re-introducing the held factors then produces a dialectical understanding of the movements of different tendencies. In spite of Robinson’s assumption, the impact of a rise in the rate of surplus value has, in fact, contradictory effects.

On the one hand, if new processes or technology make labour more productive, that can cheapen the commodities needed for the reproduction of labour. The resultant lower wage costs can have the outcome of a boost to the rate of profit. However, a generalised raising in the organic composition of capital, also a result of rising labour productivity, will depress the rate of profit. Which tendency within the dialectical movement becomes dominant depends, for example, on the extent of generalisation of new technology through different sectors of industry. Over time, nonetheless, the tendency for the rate of profit to fall is likely to re-assert itself. Robinson’s reading of the whole argument insists instead on a one-to-one offset, reducing complex dynamics of real economic relations to an abstract, linear movement up or down in magnitudes.

There is a constant tendency among non-Marxist economists to attempt to reduce the qualitative relations through which Marx analyses the dynamics of capitalism, such as value or labour power, to quantitative measurements. Yet, this is neither the goal of Marx, nor is it the function of the concept of value. Paul Mattick dismissed the ‘transformation problem’ some time ago explaining that:

Of course, the “transformation” is only a way of saying that although everything in the exchange process occurs in terms of prices, the latter are nevertheless determined by value relations of which the producers are not aware. This determination of price by value cannot be established empirically; it can only be deduced from the fact that all commodities are produced of labour … There is no observable “transformation” of values into prices; and the value concept has meaning only with regard to total social capital.[23]

Both Robinson and Keen find Marx frustrating and irrelevant because value does not automatically convert to prices. What is the point of an analysis which cannot predict particular prices?[24] A commitment to the reductive materialist outlook is necessary for a capitalist perspective, but that is not the same as it embodying truths about social relations. In fact, Marx’s employment of irreducible, qualitative concepts reveals his analysis of capitalism’s drive towards alienation as a pervasive and unfolding structure of capitalist social relations. This is the history and the future of capitalism.

By contrast, in Keynes, capitalism has no complex history, only a quantitative, linear trajectory towards an increase in the availability of capital. Shorn of a history embedded in social relations, the abstract capitalism of Keynes’ analysis is far more one-dimensional in its dynamics than Marx’s. Thus, Keynes expected the increase in capital over time to mean that it would become cheaper, as the ahistorical law of supply and demand would indicate, and therefore society would benefit. Instead, quite different results accrue upon the mounting overproduction of capital in a global economy where the proportion of value being produced is declining relative to capital. The result is an economy where speculative booms and crashes are driving forces, and where the post-war Keynesian consensus broke down in the course of the 1970s.

Real consequences of philosophical differences

The methodological and philosophical differences between Marxism and Keynesianism are not, therefore, just abstract issues, but go to Keynesianism’s inability to perceive that all economic questions are really questions about social power. The policy choice that favours capital or labour will not be determined by an objective, rational judgment based on a wider economic good, but on the social power of the two contending camps. The absence of value theory also allows Keynesians to imagine that a balance can be found between the interests of capital and labour. This can lead Keynesians in government to demand sacrifices from labour for the sake of that balance. Value theory allows Marxism to see how crises of profitability have and will lead capital towards the most vicious possible of assaults on labour in order to restore profitability.

None of this means that Keynesian policies are not useful to the working class, or that a reformist government basing its economic programme on some form of Keynesianism should not receive support from the radical left. It does however mean recognising that Keynesian economics of themselves are not likely to lead even to partial victories for labour; the ability of such policies to deliver improvements and ameliorations to capitalism has depended historically on the willingness of capital to make concessions. When conditions have changed for capital, as they did in the 1970s, there have been concerted campaigns from the supporters of capital to shift the burden onto labour.

Keynesianism was thus also unable to deal with the collapse of its standard model in the 1970s, with stagflation leading to a newly determined ruling-class attempt to reverse the gains of post-war reformism. In essence, this was about the crisis of profitability.[25] If then, Keynesianism is a double-edged sword at least, does this mean that the capitalist economy can never be bent towards the interests of the proletariat? The history of labour struggles suggest that what matters is the organised power of the working class.



[1] Robinson, Economic Philosophy, p.36.

[2] For a brilliant exposition of this, and the emergence of the law of value from a pre-capitalist to a capitalist economy, see Octavio Columbo, ‘Simple Commodity Production and Value Theory in Late Feudalism’, in Studies on Pre-Capitalist Modes of Production, eds. Laura da Graca and Andrea Zingarelli (Leiden/Chicago 2015/16), pp.237-67.

[3] Harvey, Marx, Capital, p.5.

[4] To paraphrase E. P. Thompson, The Making of the English Working Class (London 1963), p.8.

[5] Robinson, Economic Philosophy, p.47.

[6] Harvey, Marx, Capital, p.144.

[7] Ibid. p.5.

[8] Ibid. p.5.

[9] Robinson, An Essay on Marxian Economics (London 1942), p.xxi.

[10] See above, part 1, p.4.

[11] Robinson, Marxian Economics, pp.ix-x fails to distinguish the social tendency for prices to average around values, from the value to price proportions of particular commodities.

[12] Harvey, Marx, Capital, pp.34-5.

[13] Robinson, Marxian Economics, pp.26-7.

[14] Steve Keen’s attack on Marx’s labour theory of value fails on many of the same grounds as does Robinson’s, not least because he proceeds from Robinson’s assumptions; Keen, Debunking Economics, p.423 and p.427, appears to simplify Robinson’s argument, producing the entirely erroneous claim that Marx always supposed the rate of surplus value to be constant; see Robinson, Marxian Economics, p.38, where she refers to it as the rate of exploitation. Keen in general is enthusiastic about the ‘neo-Ricardian’ economist Piero Sraffa, whose theories, Harvey explains, were used ‘to destroy the prevailing (non-dialectical) notion of value theory in Marx’, Companion to Marx’s Capital, vol. 2, p.331, note 9. The version of value theory being defended here is the dialectical one.

[15] Robinson, Marxian Economics, p.85; see above n.25, and Harvey, Companion to Capital, vol. 2, p.40.

[16] Robinson, Marxian Economics, p.xii.

[17] The failure to understand this seems to lie behind Steve Keen’s rejection of labour as the source of surplus value; Debunking Economics, p.442. What both Keen and Robinson also fail to notice is that the labour theory of value is about the production of surplus value within capitalism, not the existence of nature’s ‘free gifts’, for example; see John Bellamy Foster, Marx’s Ecology: Materialism and Nature (New York 2000), particularly p.167. 

[18] For the most recent, and very effective, Marxist rebuttal of the transformation problem, discussed at the general level, see Michael Roberts, Marx 200 – A review of Marx’s economics 200 years after his birth ( 2018), pp.81-5.

[19] Robinson, Marxian Economics, pp.36-8.

[20] See Geert Reuten, ‘The Productive Powers of Labour and the Redundant Transformation to Prices of Production: A Marx-immanent Critique and Reconstruction’, Historical Materialism 25.3 (2017), pp.3-35; p.3; Reuten offers an interesting solution to the transformation problem itself, within Marx’s own terms.

[21] Robinson, Marxian Economics, p.ix, finds that the ‘constant rate of exploitation in Volume III [of Capital] is not explained’, noticing therefore that in other parts of the argument the rate of surplus value does change, but not appreciating the function of holding it constant for the purpose of the overall analysis.

[22] Ibid. p.36.

[23] Paul Mattick, Marx and Keynes: The Limits of the Mixed Economy (London 1971), p.43. Mattick goes onto explain how values turn into prices ‘by way of competition, by the search for profits and extra-profits which constitutes the capitalist contribution and reaction to the increasing productivity of labor’. For a full explanation of this whole issue, see ibid. ch.4, ‘Value and Price’, pp.40-50.

[24] Keen, Debunking Economics, p.429; Keen demands that Marx’s argument can achieve a multiplicity of conclusions without varying the holding assumptions, and also assumes that the point of the transformation tables is to predict prices from values, in real terms, which was clearly not Marx’s purpose in this fragment.

[25] See Roberts, The Long Depression, chapter 4, pp.59-64.

Dominic Alexander

Dominic Alexander is a member of Counterfire, for which he is the book review editor. He is a longstanding activist in north London. He is a historian whose work includes the book Saints and Animals in the Middle Ages (2008), a social history of medieval wonder tales, and articles on London’s first revolutionary, William Longbeard, and the revolt of 1196, in Viator 48:3 (2017), and Science and Society 84:3 (July 2020). He is also the author of the Counterfire books, The Limits of Keynesianism (2018) and Trotsky in the Bronze Age (2020).

Tagged under: