Saturday, September 04, 2010

Rational choice marxism?

Responding to my post on the imperialism of market reason, Henry at Crooked Timber points out a 'disjunction' between my generally hostile attitude to explanatory paradigms based on 'rational choice', game theory and methodological individualism, and my later endorsement of a book by Erik Olin Wright, himself a supporter of 'rational choice marxism', along with Jon Elster, Alan Carlin, Adam Przeworski, and John Roemer (indeed, Roemer far more than Wright is committed to 'rational choice' as the microfoundational basis for marxist theories of class). Henry's main point is to suggest that there is a perfectly serviceable left-wing version of rational choice theory, and that it provides good models of exploitation and unequal exchange. I can well see that rational choice theory is compatible with a wide range of liberalisms, from right to left. What I want to address is whether rational choice models can add anything of value specifically to marxist theories of history, society and social change.

Before going any further, it's worth clearing up a few points. Firstly, when Henry complains about my "standard-issue dismissal of economic notions of rationality as a kind of imperialism", he is making this seem far more controversial than it actually is. The 'economics imperialism' that is spoken of is the tendency for marginalist precepts to become the basis not merely of economic theory, but increasingly of social sciences covering everything from crime, romance, and statecraft to urban geography and drug addiction. And the public choice theory that I wrote about is a fairly typical example, taking market-based precepts about human behaviour that was designed to explain marketplace behaviour and applying them to a non-market terrain where an 'implicit market' is just assumed. To describe this as imperialism is not to 'dismiss' market-based notions of rationality. I am unaware of any circumstances in which one would 'dismiss' imperialism. Rather, it is to acknowledge their apparent power, capacity and flexibility as explanatory principles. If one dismisses market rationality it is on other grounds, about which more in a moment.

Secondly, when he says that "there is an unexplained jump" in my original post "from the description of microfoundations to more specific claims about public choice economics", it is only fair to point out that these microfoundations - individualism, rational self-interest and exchange - do provide the basis for the specific claims of public choice theorists. Whether or not Henry is right that the right-wing bias lies in the models used by public choice economists rather than the microfoundations, which we will come back to, there is no inexplicable 'jump' from foundations to explanations. And my rejection of the microfoundations as essentially circular and reductionist does not depend on the rejection of public choice theory. One the contrary, rejecting the microfoundations of market-based 'rationality' simply provides the basis for a more thoroughoing dissent from public choice theory than someone committed to methodological individualism and game-theoretic models could concur with. Thirdly, I don't think I need to spell this out, but just to mention that it is possible to appreciate Wright's attempt to offer a rigorously marxist interpretation of class without agreeing with his estimation of rational choice - I will say more about this later. Lastly, in passing, I'm also not convinced that the problem of how social democrats "win elections where the working class is not actually in the majority" is actually a real problem. This observation implies that the model of class and exploitation that Henry is working with is incommensurable with the marxist model.

The marxist theory of history
So, these quibbles stated, let me also state what I understand to be the basis of the marxist theory of class, exploitation and modes of production, so that we can see whether rational choice models can add anything valuable to it. The basis of Marx's theory is not the individual of liberal theory but the human society, and the fundamental aspect of any society is how they produce sufficient material goods through the alteration of 'nature' to survive and reproduce themselves in their given state. Marx's contention is that in pre-class ('primitive communist') societies, there was no surplus available to support an elite or 'ruling class', so egalitarian relations prevailed. But at a certain level of the development of the forces of production, a small surplus became available, and thus it became possible for a ruling class to arise. More than that, since the further development of the means of production requires individual sacrifice and privation that no individual would willingly consent to, the emergence of a ruling class is optimal for the further development of productive forces and thus of mankind's ability to transform nature for its own benefit. For a class society to emerge, it was then necessary for a group to have the capacity and opportunity to sieze control of the surplus and of the means of physical and ideological coercion. Thus begins the sequence of class societies that, hopefully, is finally terminated with the supercession of capitalism.

There is more. The level of surplus, and the forces of production will allow only a certain range of relations of production. The forms of exploitation that arise at any given point are thus limited by the available means of production, such that - for example - slave-based systems would be unfeasible in an era of computerised technology, because of the level of culture and education required to work those systems. At a certain point in the development of the forces of production, the further development of these forces is basically incompatible with existing relations of production. Assuming that there is an agency sufficient to carry through a revolutionary transformation of society - a big assumption to make - the forces of production will continue to be developed within new relations of production. The agency necessary to achieve this change will be a class, or perhaps an alliance between fractions of different classes, with the relevant ideological, organisational and material capacities.

The earliest forms of class society, with a small surplus and only rudimentary forms of technology, were either tributary or slave-based. The larger and more centralised they were, the better they tended to mobilise productive forces, satisfy human needs, and develop culture and technology - the ancient and medieval empires are a case in point. But these tended to reach their limits when further overland expansion become impossible, or cost the empire more than it added in surplus. The tributary empires and feudal societies could develop the productive forces to a certain reasonable high degree, with complex global commercial networks, remarkable technologies, new developments in mathematics, a thriving vernacular literature, and the emergence of the rudiments of science. But they tended to hit a ceiling. The empires collapsed and disintegrated, the feudal societies were prey to conquest, diseases wiping out whole populations, mass famines, and political instability.

Marxists legitimately differ on how the transition to capitalism took place, but I would put it like this: the English ruling class, experimenting with various property forms in response to feudal crisis and class struggles, alit upon what became the capitalist form, most basically in the transition from freehold tenancies to leasehold tenancies. In this new form of exploitation, instead of being bonded to the land and to the feudal ruler, the labourers were 'free', and their access to the means of production, and thus to a share of the social product, dependent on their availability for exploitation. Instead of the surplus increasing through more overland expansion, which meant war, or through a brutal ratcheting up of tribute extraction, which could lead to insurrection, surplus was increased as a result of the imperative to continually improve the productivity of labour, and enhance the means of production. This proved optimal for the further 'improvement' of the land, which outstripped feudal methods in productivity considerably. The capitalist aristocrats and bourgeois asserted themselves as England's new ruling class by deposing the monarch and destroying the power of the feudal aristocracy. By virtue of the way in which England's new productive relations could develop the forces of production, its ruling class went from being a clan of nepotistic barbarians ruling over some backwater into an emerging global power, laying down capitalist property relations in Ireland and North America, and eventually leading a global transformation of social property relations. Subsequently, through revolutions, nationalist wars, colonial tyranny, subsequent anti-colonial rebellions, and various putative 'revolutions from above', capitalist relations were established globally.

In theory, as the market-based coercion of capitalism produces a mega surplus, or at least productive forces sufficient to potentially produce a mega surplus, it also removes the necessity of coercion and exploitation for the further development of the means of production. If there is enough of a surplus to feed, clothe, house and educate everyone, then there is no need for the surplus to be concentrated among a few. In fact, at a certain point, this begins to hold back the further development of the means of production, as concentrating ownership and fundamental decision-making in the hands of a few necessarily squanders the potential expertise, talent and ingenuity of the great mass of humanity.

By virtue of the same theory, the agency that is exploited - the working class, who must form the majority in any developed capitalist society - has both an interest in abolishing class relations and a unique capacity to do so because of their centrality to production, their numbers, their ability to organise and the fact that they are socialised in the workplace, the very site of their exploitation. Through the exploitation process, they are united with other groups of workers who are similarly exploited, and form a potentially powerful agency for limiting and resisting that exploitation. Through the same process, they also acquire the expertise and skills necessary to challenge the authority of line managers, and assert their own 'common sense' solutions to production dilemmas. Working class composition, cohesion and power with respect to the employers varies, but there is in this the possibility that the different fractions of the working class could be convoked into a revolutionary agency that is capable of taking control of the means of production and asserting egalitarian productive relations.

To summarise. The character of a society is most basically determined by the available forces of production, and the relations of production (ie between producers and appropriators). This combination of factors is known as the 'mode of production'. The mechanism of exploitation in each case involves the coercive appropriation of a surplus by a ruling class. Change from one mode of production to another takes place when the existing relations of production impede the further development of the forces of production, and a class or class alliance with an interest in social transformation and the capacity to carry it out revolutionises the society. The necessity for class relations only persisted for as long as the surplus was relatively modest. Capitalism, providing the potential for superabundance and convoking a potentially revolutionary agency through its very mechanism of exploitation, offers us a way out of exploitative relations. Such is the marxist theory of history, and the revolutionary socialist faith, as I understand it.

Rational choice explanations for exploitation and class societies
Now we can look at what 'rational choice' marxists think are the essential microfoundations of this theory. According to rational choice marxism, the different modes of class rule and exploitation develop from the microfoundations of individuals pursuing their rational self-interest, that interest determined by the individual's location within the class structure. Class relations emerge out of exchanges between differently endowed individuals in a competitive setting. Where the relevant assets that make such exchanges advantageous or not are distributed unequally, the result is necessarily exploitative.

Roemer's approach has it that someone is an exploiter when a more equal distribution of certain resources would leave them worse off, and exploited when a more equal distribution of the same resources would leave them better off. The resource in question is relative to the mode of exploitation and the hypothetical alternative. In the feudal mode of production, the asset that is most inequitably distributed is the freedom to trade one's product - bonded labourers enjoying this asset least of all. This is exploitative in comparison to the hypothetical alternative of capitalism. In this circumstance, the bonded labourer enters into an unequal exchange with the feudal lord in a rationally self-interested way as her alternatives are vagabondage, or death. Most people, rationally seeking to improve their mastery over nature and the material resources available to them, would enter into that exploitative relationship. Similarly, in the capitalist mode of production, the asset that is unequally distributed is access to the means of production. The hypothetical alternative, against which this is deemed exploitative, is socialism. The worker enters into an exploitative relationship with the employer because the alternative of privation in conditions of unemployment, beggary or self-employment is less preferrable. The capitalist likewise takes advantage of this exploitative relationship because of a preference for greater material rewards.

Thus, rational choice theory would appear to have offered a model of exploitation without recourse to the labour theory of value. Even if no transfer of surplus labour to the employer has taken place, there has still been exploitation. Exploitation is a matter of asset distribution, and relative advantage or disadvantage, not of productive relations. This is obviously quite different from the account outlined above, where the class relationship is a social and not an individual relationship. Rational choice marxism, by starting with the individualist, game theoretic precepts that it does, it does not merely fill in missing micro-foundations. It ends with a different - I would say radically different - theory of history, class and exploitation.

But does it provide a superior theory to the classical marxist account? I would say that it fails on a number of grounds. Firstly, by positing differently endowed individuals as the microfoundations of macro-analysis, 'rational choice' marxism smuggles the wider social relations into the unit of individual transaction. It becomes circular, rather than informative, because the very microfoundation that is supposed to explain larger structural facts such as exploitation, class and the structural capacities that flow from these actually contains these embedded within it. So it turns out that the individual is not a self-sufficient explanatory unit constituting social structure, but rather that her capacities and assets are themselves constituted by larger historical, structural and relational factors.

Secondly, each actor is supposed to participate in the exploitative relation in order to maximise utility, but utility is too recursive a notion to be of much explanatory use here. Utility is maximised wherever an individual satisfies her most preferred preferences (assuming that individuals have a well-ordered hierarchy of preferences, a convenient fiction for the purposes of the theory). Whatever an individual is doing, it is assumed that she is maximising utility. If a person commits a crime that is likely to reduce her material well-being we could argue, pace Becker, that she is satisfying a preference for risk. Or that she is satisfying sadistic, or masochistic urges. Or that, quite simply, she is fulfilling a preference to murder which she ranks above other preferences for material well-being. Whatever - the point is that there is no form of behaviour that could not be described as utility-maximising.

If we are trying to explain social phenomena, however, rather than simply redescribe it in market-based language, we have to do better than this. And here methodological collectivism is unavoidable. To explain why a capitalist acts the way she does, it is not sufficient to refer to individual preferences. The imperatives of capital accumulation exert themselves through the competitive pressures of the capitalist market, an historically produced social structure which has forced individuals to become dependent on the market for their survival and reproduction. For the rational choice model to work, capitalist imperatives must be naturalised - they must be simply assumed to be among the existing preferences of capitalists, which they are able to exercise on account of their particular endowments.

Thirdly, the account of exploitation explains the transaction as a freely chosen exchange by individuals pursuing optimising strategies with the endowments available to them. The class position they end up in will depend on the endowments and whether they choose to work for themselves, hire themselves out, or hire others to work for them. But as we have said, these endowments contain the social relations they are supposed to explain, notably under capitalism the capitalist's control of the means of production, and the worker's separation from the means of production. Free choice is severely restricted here, since for the worker there are few alternatives to surrendering a portion of her labour to the owner of the means of production in exchange for access to the means of survival. Some can become lone traders, criminals, beggars, etc. But the opportunities for each within the system are not widespread, so that the majority must hire themselves out if they are to survive. Of necessity, many people with the relevant endowments, and the appropriate order of preferences, would not be able to exercise those alternative choices mentioned. Further, there are elaborate social mechanisms which construct people's choices, from education, with its various selective and grading mechanisms, and the full range of state and private sector bureaus intended to direct people into this or that line of employment, to culture, which plays a vital role in propagating the values that ensure the reproduction of the system.

Indeed, the combined mechanisms of compulsion - which are central to the classical marxist explanation of exploitation - while not completely obviating individual choice, do so severely limit choice as to make it questionable whether grounding the exploitation mechanism in individuals freely pursuing optimal utility-maximising strategies is a sensible way to proceed. But for rational choice marxism, a free exchange has to be the starting point for explaining exploitation, because its model breaks society down into free and roughly equal individuals. And because it explains exploitation in this individualistic way, it has difficulty explaining class cohesion, social conflict, and particularly the transition from one mode of production to another. For example, in the transition to capitalism there is no question of capitalist social relations simply, mysteriously emerging as an option that individuals could take up by, for example, becoming waged labourers (as in Roemer's imagined model). Capitalist social property relations emerged through class struggle and coercion, and were consolidated in England at least by means of regicide. One mode of production displaced the other not through the aggregate rational choices of individual aristocrats and peasants, gradually converting to capitalism, but through collective interest and action.

The problems with rational choice marxism, not exhausted here, are basically reducible to the problems with rational choice theory itself - the circularity of its explanations, its dogmatic assertion that the individual is the proper foundation of social analysis, and its vulgar, reductionist model of agency. Rational choice marxism's attempt to reinscribe the marxist theory of history and class struggle in the terms of methodological individualism has to be judged a failure. There remains the question of whether the precepts market 'rationality' are commensurable with marxist theory.

Embedded normative assumptions
Henry's most basic assertion is that rational choice models don't contain inbuilt normative assumptions, or theses about the social world that would lead to them having a right-wing bias. It is certainly possible for rational choice theorists to resist right-wing interpretations, but to do so it would be necessary to acknowledge a few things that Henry does not. First of all, to return to one of the previous quibbles, the origin of rational choice theory is precisely in the Cold War state and the right-wing free market theology it produced, which Henry claims are somehow neatly separable from it.

Secondly, the founding assumptions of rational choice theory are directly imported from classical liberalism. The idea that social action ultimately takes place at the level of the individual, that people are utility maximisers, and that such maximisation optimally takes place in the process of exchange is obviously derived from liberal, contractarian political economy. This is inescapably normative, and unavoidably contains propositions about the social world - most obviously that, whatever mode of production prevails, social life is composed from an aggregate of self-sufficient individual units with varying endowments and attributes engaging in repeated acts of exchange in a competitive setting. Thus, social and temporally specific aspects of behaviour under capitalism are, through their insertion into a set of ahistorical claims about human nature, naturalised and universalised. Characteristic of Marx's approach, by contrast, is that he does not assume that the logic of capitalism results from individual propensities.

It is obvious enough that rational choice theory is not necessarily right-wing, since there are plenty of left-wing exponents of it, but nor is it avalent. And insofar as there have been attempts, largely in the retreat from Althusserian structuralism, to reconcile rational choice with marxism, the tendency has been for one to be adulterated rather than complemented and strengthened by the other.

Since my citation of E O Wright triggered Henry's reflections, I will use him as a case in point. Wright is an 'analytic marxist' with an interest in rational choice theory. But the more closely he cleaves to marxism, the less weight he attributes to rational choice precepts. In 'Marxism and methodological individualism', an essay written with Elliott Sober and Andrew Levine, he argues that what is distinctive about methodological individualism - its commitment to explaining social phenomena by attributing primary explanatory power to individuals, such that it recognises no irreducible explanatory power for the properties and relations of aggregate social entities - is incompatible with marxism (and bad social science to boot). He is right. But methodological individualism is a fundamental commitment of rational choice theory since its inception. It is a commitment shared by Henry's favoured game theorist, Tom Slee. And it is a commitment made by most who choose to call themselves rational choice marxists. To that extent, inasmuch as Wright is (perhaps was) a far more orthodox marxist than his rational choice peers, he is a far less enthusiastic rational choice theorist.

For my part, while I think it is correct to attribute some explanatory power to individual relations and actors, I don't think the market-based models of rationality are very useful for assessing individual motives and behaviour. They can possibly sustain left-liberal interpretations, but I maintain that they will still be bad left-liberal interpretations. And I certainly don't think they can provide 'microfoundations' for a marxist theory of society and history.