Ellen Meiksins Wood on Rational Choice Marxism

So I believe I want to begin to pivot, in my thinking (which probably also means in my writing on the blog), to the space of rational choice theory.  I don’t think I’m completely unaware of various widely-noted problems with rational choice theory – and I will probably worry away at some of those problems, in posts on the blog; I don’t want my thinking here to be entirely or uncritically swallowed by the rational choice paradigm.  But I also think that: 1) like it or not, rational choice theory simply is central to a lot of formal economics, and I need to get to grips with it properly, not superficially; moreover, 2) I think rational choice theory has a lot more to recommend it than many of its critics’ criticisms imply.

With that second point in mind, I want to put up some quick thoughts on Ellen Meiksins Wood’s 1989 critique of ‘rational choice Marxism’ (Wood is specifically criticising Roemer, Elster, Przeworski, and Carling).  Wood makes a lot of different moves in this essay, some of which I agree with.  Things I agree with include: she is critical of teleological theories of history; she rightly argues that many theories that transhistoricise contingently-produced contemporary categories are tacitly teleological; she rightly (by my lights), and contra Elster, argues that Marx’s remark that “Human anatomy contains a key to the anatomy of the ape” is an anti-teleological, rather than a teleological, remark; she is critical of the tendency, within a lot of ‘analytical’ (and, I would add, not just analytical) Marxism, to ‘idealise’ capitalism on the grounds that Marx’s critique can then be shown to apply even to the best-case scenario for capitalism.  I agree with all of this.

But there’s also quite a lot I disagree with in the essay. My disagreements include the following:

As is generally the case for ‘political Marxism’ (i.e. basically Wood and Brenner), I think Wood has an understanding of capitalism that is too centrally focused on the specific form of domination associated with the ‘site of production’.  Wood and Brenner tend to distinguish their position from a ‘neo-Smithian Marxism’ that (as they see it) over-emphasises the market as the key institution within capitalism.  I think this framing is often unfair to theorists working within non-’political Marxist’ traditions.  I think that’s the case both because those other Marxist theorists are often more sophisticated in their understanding of capitalism than the ‘political Marxists’ allow (for example, Brenner’s critique of Immanuel Wallerstein in my view actively misrepresents Wallerstein’s position), but also, separately, because I think Adam Smith is a great social theorist and political economist, and being a neo-Smithian is fine.  I’ll come back to this point eventually.

Wood sees Roemer’s economics as centrally motivated by the attempt to develop an alternative ethical critique of capitalism.  Maybe this is the central motive for Roemer’s work – but Roemer’s major contribution, as I see it, lies in his effort to wrestle Marxist political-economic categories into an analytic framework that isn’t completely embarrassing as economics.  Roemer’s critique of (or, perhaps better, radical circumscription of the claims made by) the labour theory of value is mainly driven by the fact that (as I’ve blogged about before) the ‘classical’ labour theory of value simply makes no sense as economics.  Now, if you’re a historian this probably isn’t as big a problem as it is if you’re an economist, and Marx’s major contributions can be seen as lying elsewhere – but for those of us who are interested in doing good economics, Roemer’s project makes a lot of sense; Wood’s argument that there’s no ‘there’ there seems very misguided to me.

But the main thing I wanted to talk about in this post is one of Wood’s more general critiques of rational choice theory (here writing about Przeworski, but the point is broader). As Wood puts it:

Since for the purposes of the rational choice model capitalism exists only as a ‘game’ between consenting adults, a game in which the stakes are no more compelling than a desire for ‘optimization’, there is no ‘logical place’ in his argument for capitalism as a compulsive and coercive system, a ruthless logic of process.

I think there are two defensible claims that can be made in this space.  First, as I said above, the tradition of Marxist economics of which Roemer is one representative has a tendency to construct ‘ideal’ models of capitalism, which aspire to show that, even in a capitalist system that is minimally coercive, exploitation still takes place.  This ‘ideal capitalism’ approach is certainly defensible, but it naturally has a tendency to sideline the analysis of coercive dynamics.  Second, there is a much broader tradition within mainstream economics that takes free and equal, mutually beneficial exchange as its paradigmatic economic interaction.  (This is the approach mocked by Marx as inhabiting a world ruled by “Freedom, Equality, Property, and Bentham”.)  In this tradition there is a very widespread tendency to move from the formal representation of economic interactions in terms of individual optimisation, to the Panglossian claim that in an economy composed of such interactions, every interaction must be mutually beneficial for every party.  This move, similarly, has a tendency to elide coercion, and arguably with less theoretical warrant than Roemer-style idealisation.

So to the extent that Wood is criticising these phenomena, I think her critique has some force.  But Wood makes much broader claims than can be justified by these points.  In the passage I quoted, Wood seems to be arguing that rational choice theory is incapable of capturing coercion or compulsion at all – and this is flatly wrong.

There are, I think, implicitly two separate arguments in this space, in the passage I quoted.  One is basically just a play on words: because game theory represents economic life as a “game”, Wood seems to be suggesting, it is treating economic interactions as trivial and low-stakes.  One can, along these lines, regret the fact that game theory is called ‘game theory’, but I think it’s extremely clear that as a substantive critique of the analytic framework, this is nonsense.  We’re just talking about a theoretical toolkit here – the toolkit can be used to analyse matters of life and death, and is used in that way all the time.

The second implicit argument, I think, is really a cluster of points: that rational choice theory is about choice, and therefore cannot capture coercion; and that because agents are engaged in ‘optimisation’, they are understood as necessarily coming out of interactions ‘ahead’ in some sense.  As I said above, I do think a lot of ‘mainstream’ political economy makes the move that Wood is here tacitly highlighting – but I also think it’s clear that this move is not ‘baked in’ to the rational choice analytic framework at all.

Let me take this point in two stages.  First, the issue of choice.  Of course it’s true that ‘choice’ contrasts with ‘coercion’, and therefore to some extent these are opposed concepts.  But I think it’s a big mistake to therefore conclude that frameworks which emphasise choice have difficulty tackling coercion.  The reason is the following: there is no human scenario in which agents completely lack choice.  Even if I am being ordered to act in some way with a gun to my head, I have the choice to refuse to act in that way – and to be shot.  Even if I am imprisoned, I have some choice about what actions to take in my confinement.  Of course, these are scenarios of sharply limited choice – they are scenarios of coercion.  But the coercive dimension of these scenarios is captured by the range of choices available to me – the way in which other social actors’ actions influence my choice set.  The choice is still present.

This point applies to rational choice theory also.  The scenario of violent coercion can perfectly well be modelled within a rational choice framework – one social actor’s threat of violence transforms the set of options between which another social actor must choose.  The second social actor then optimises within the new choice options.  Nothing about this is contrary to the rational choice framework – the issue of whether a choice is coerced or not is simply orthogonal to the use of the analytic toolkit.

A similar point applies to the question of ‘optimisation’.  The fact that a social actor chooses what they take to be the most welfare-enhancing (or whatever) action among the available options, thereby optimising their expected utility (or whatever), really doesn’t speak at all, in the abstract, to whether they are better off or worse off, out the other side of this interaction.  They are better off relative to the other available options, as they see it.  But again, they may be worse off relative to the prior state – it all depends on the scenario in question.

These are basic points, but I think it’s clear from various remarks Wood makes across her piece that Wood does not accept them.  Of course, these aren’t Wood’s only critiques of rational choice theory – she makes many other points in the essay, some of which I’ve mentioned briefly above, some of which I haven’t touched on at all.  But I think it’s important to be clear that this specific argument – the idea that a rational choice approach to political economy has no place “for capitalism as a compulsive and coercive system” – is straightforwardly false.

I had thought about writing more on Wood’s essay, and issues arising from it, but I think this is a good stopping point for now.