G.A. Cohen on Marx on Exploitation

Alexander Douglas
Genus Specious
Published in
7 min readNov 27, 2020


As I was writing on Marx on value, a few analytic political philosophers recommended this essay to me: G.A. Cohen’s “The Labor Theory of Value and the Concept of Exploitation”.

I puzzled over it for a while, but I think it must be wrong, at least on one central point. The article has many strengths, but here I want to share a criticism.

The introduction is nice and short; it reads:

This essay shows that the relationship between the labor theory of value and the concept of exploitation is one of mutual irrelevance. The labor theory of value is not a suitable basis for the charge of exploitation laid against capitalism by Marxists, and the real foundation of that charge is something much simpler which, for reasons to be stated, is widely confused with the labor theory of value.

I hope you know what the labour theory of value is. If not, you get this rough definition. The Labour Theory of Value (LTV) is:

the theory that the exchange value of a commodity is determined by the amount of labour that goes into it.

So if a fork exchanges on the market for two spoons, a fork must take twice as much labour to make as the spoon. (I’m thinking about making the little tines; I’m not a silversmith.) I’m going to shorten “exchange value” to “value”. The other sort, use value, I’ll call by its full name. This is what Marx did, and if it was good enough for him…

Marx liked the LTV, though he knew that in more complicated economies the relation between exchange-ratios and values is mitigated by other factors, though this isn’t such a serious problem as some people say. Cohen says that the LTV doesn’t help his theory of exploitation. In fact he thinks it ruins it. Cohen reads the theory of exploitation as holding, roughly, that while labour creates the full value of a commodity, capitalists keep some of that value for themselves, and that’s exploitation.

This is all pretty loose. By “keep some of that value for themselves”, I mean that capitalists make a profit. If the produced commodities sell for $1000, and they pay the workers that made the commodities $900, they thereby exploit the workers to extract a profit of $100.

Anyway, the point is that the LTV is featuring here, in the notion that labour creates the full value of a commodity.

I won’t go through Cohen’s whole argument against this. He concludes:

Now Marxists allege that the labor theory of value is required to uncover the exploitation of the wage worker, but I disagree. What is needed is not the false and irrelevant labor theory, but the mere concept of value, as defined, independently of the labor theory, in our sentence (2) [Value determines equilibrium price]. It enables us to say that, whatever may be responsible for magnitudes of value, the worker does not receive all of the value of his product.

Cohen’s point is that whatever it is that determines the equilibrium price of a commodity, we can see the worker as being exploited if she doesn’t receive the full value of that. He thinks we can express the situation of exploitation like this:

The proletarian produces the whole product, but the capitalist appropriates part of the value of the product.

This seems wrong to me. The idea of “equilibrium” is brought in by Cohen; it’s not a Marxist idea. But taking this idea of “equilibrium price”, Cohen’s view seems to be that we can understand exploitation as meaning that labour doesn’t receive the full proceeds of the commodity’s sale at its equilibrium price.

But what is its equilibrium price? The price is whatever it exchanges for. So let’s say the workers were paid $900 for a commodity that sells for $1000. Have they been paid less than the value? Well, they were paid in the past, and the commodity is sold now. The workers accepted $900 to produce it. The capitalist gets $1000 by selling it. But she gets $1000 now; they got $900 then. Rather than say that the workers are exploited, why not say that they received the full discounted value of what they produced?

Think of it like this. The workers agreed to produce a $1000 commodity for $900. So it was worth it for them to agree to hand over the commodity, at the end of production, for the $900 now. That looks like an exchange, doesn’t it? And since equilibrium price is just what a thing exchanges at, it looks like $900 was the commodity’s value before it was produced. That’s what the workers were willing to exchange it for. We have here an intertemporal equilibrium, and a time-discounted equilibrium price.

Cohen says he wants to allow for “whatever may be responsible for magnitudes of value”. So time-preference might well be part of it. But who knows what determines it? The point is, we have an intertemporal exchange and thus an intertemporal price. There is nothing to stop us matching value to this, once the LTV is out of the way.

With the LTV in the way, we can’t have those sorts of shenanigans. This isn’t because time-discounting doesn’t occur in Marx’s world, but because even if it does it isn’t relevant; there are no intertemporal exchanges. The capitalist, on Marx’s theory, buys the worker’s labour-power on the spot (Marx specifies this explicitly), at its full value, which is the wage that provides subsistence to the worker for the given production period. The use-value of this labour-power is the work then done. The capitalist’s profit comes from the fact that less than a day’s labour is enough to produce the subsistence for a full day’s labour.

There is no danger here of valuing the commodity produced by what the workers will take in wages to produce it. The commodity they have exchanged is not the finished product but rather their labour-power. For that they have received the full value. But its use-value exceeds its value. That fact is really all there is to exploitation.

Marx calls “variable capital”, v, the amount spent paying wages to workers, and “surplus value”, s, the surplus extracted by capitalists — all this measured in labour-hours. Marx sometimes writes that v corresponds to the value of labour-power, whereas s+v corresponds to its use-value. Thus, again, the difference between the use-value and the value of labour just is exploitation (therefore the rate of surplus value, also called the degree of exploitation, is expressed as s/v).

But — here is the key — the use-value is never exchanged for anything (otherwise it would paradoxically become an exchange-value). Capitalists pay for labour-power, but they simply extract use-value. You could say that the extracting is the exploiting here. But on Cohen’s reckoning, where value is just equilibrium price, there is nothing by which to determine the difference between the value and the use-value of labour power. The use-value has no “equilibrium price” at all, since it is never exchanged. You need the LTV to count the labour-hours in one and the other, to see the difference between them [§ — note below].

Also, Cohen’s main criticism of the Labour Theory, at least as applied to exploitation, is that he thinks it depends on the notion of labour creating value. He seems to think that Marx wants us to reason that capitalists exploit workers because the workers create value and the capitalists take some of that value. Put that way, it sounds more like Adam Smith’s story: the capitalist (and the rentier) muscle in and claim their share, as Joan Robinson put it. Marx criticised this for implicitly violating the law of value — that things exchange according to their value.

Marx’s critique of political economy involved preserving the law of value, by showing how every exchange occurs according to value, even exchanges between workers and capitalists. Exploitation exists within that framework. Contrary to what Cohen suggests, we don’t need the notion of value-creation to explain it. Marx uses that phrase, but you could take it out without losing any of the story. What you can’t take out is the LTV itself, unless you can find a suitable replacement [§]. Value inferred from prices won’t do for measuring the differential between use-value and value, which determines exploitation.

So Cohen’s criticism of the LTV misses. It hits at the wrong part of the theory (the idea of value-creation, rather than the differential between the use-value and value of labour-power). Nor does his alternative framing, shorn of the LTV, provide a compelling story of exploitation. Rather, it opens the door to the neoclassical picture of an intertemporal exchange at an equilibrium price.

[§] Note: Strictly speaking, you don’t necessarily need the Labour Theory of Value. You just need some way to measure the difference between the value and the use-value of labour power that doesn’t change when relative prices change. It’s the last part that’s tricky. David Ricardo had originally tried to measure the rate of surplus value in corn: the proportion of corn consumed in the production of corn. The problem, which Thomas Malthus pointed out, is that workers consume more than just corn in producing corn, so that the relative prices of commodities will affect this proportion of consumption to production. This is a problem since relative prices are meant to depend on value, part of which is surplus value, so we end up in a circle. This may be why Ricardo switched to reckoning in labour-time. Notice also that this is different from Adam Smith’s idea that estimated labour time is what governs exchange.



Alexander Douglas
Genus Specious

Lecturer in Philosophy, University of St. Andrews — personal website: https://axdouglas.com/