Irrational Politics

Jon Elster

Anglo-American political science is dominated by the image of rational man. Politics is the maximising of something or other: material benefits for the voter, votes for the politician. Politicians competing for the voters’ favour have to think strategically and rationally. They cannot waste effort on lost causes, nor afford to take up positions too far from the centre. Indeed, a famous theorem, which indicates the flavour of much of rational-choice political science, says that parties tend to converge towards the middle of the political spectrum for much the same reason that makes ice-cream vendors converge towards the middle of the beach. Politicians must concentrate their scarce resources to sway the marginal voter, and in doing so they have to anticipate and if possible undercut the competition from other politicians, well knowing that the latter are trying to do the same. The ‘I think that he thinks that I think …’ aspect of politics is formalised by game theory, a branch of rational-choice theory that has become increasingly important in political science. Game theory is uniquely well suited to bring out the complex interdependencies of social life, and to explain many paradoxical phenomena in terms of individual rationality. In The Presidential Election Game,[1] Steven Brams has explained the confrontation between Nixon and Justices Burger and Blackmun over the White House tapes in terms of a strategic conflict, where the actual and rationally predictable outcome was worse for both sides than another available outcome. Such paradoxes can occur because the jointly best outcome is not stable against defectors: i.e. rational players will try to go for something even better and thereby jointly bring about something worse.

The rational-choice explanation of politics offers surprising and striking insights into a wide range of phenomena. In parliamentary politics it turns out to be false that honesty pays: it has been shown that all possible voting schemes, barring dictatorship and random choice, lend themselves to strategic manipulation, so that it is possible to achieve an outcome that accords better with one’s preferences by voting against these preferences. Also it has been argued that parliamentarians will not try to maximise support for a given proposal, but rather go for a minimal winning coalition so as not to spread the spoils of victory too thinly. We know by now quite a lot about the log-rolling and wheeling-dealing side of political life, as well as about the many ways in which individual rationality may be collectively self-defeating. But there is more to politics than individuals acting out of calculated self-interest: both calculation and self-interest may be conspicuously lacking.

Consider, for a start, an alternative view of politics that is also an important tradition within the discipline. This is the view that politics supplements rather than imitates economics. The assumption of the rational-choice theorists is that voters and politicians behave like consumers and entrepreneurs respectively. Politics is like the marketplace, trading with different goods. Now it is well-known that the marketplace is not in all respects an efficient mechanism. There are many cases of market failure, notably inefficient provision of public goods and the generation of involuntary unemployment. It is not uncommon to define politics as a device for overcoming market failures: i.e. for creating the efficiency and stability that cannot be achieved by unco-ordinated individual actions. The tradition goes back at least to Hobbes, and is thus considerably older than the rational-choice model, which can be traced back to Joseph Schumpeter’s Capitalism, Socialism and Democracy of 1942. I do not want to imply that the two models are incompatible. One might argue that failures of the economic market are overcome in another market, the political arena where politicians compete for votes by proposing schemes for overcoming market failures. Still, I do not think this captures the spirit of the rational-choice theorists, nor that it is a faithful depiction of reality. If political competition turns upon the marginal voter, then the rational strategy is to offer extra benefits to some small groups rather than institutional changes that benefit all. And so politicians must be more than vote-maximising entrepreneurs if they are seen to engage in large-scale institutional change, as of course they sometimes are.

A more striking breakdown of the rational-choice model is seen in voting behaviour. The question is simple: why bother to vote? Politicians vie for the voters’ favour by appealing to their self-interest: by proposing schemes that will make them materially better-off. The voters may be assumed to have an interest in this, but why should they bother to vote when the chance of influencing the outcome is virtually nil? In an election with 1 or 100 million voters, my chance of being the pivotal voter is, for all practical purposes, zero. And so the expected utility of abstaining is certainly larger than the expected utility of voting, given that voting requires some minimal effort. If politics really were like a market, there would be a near-total abstention from voting and hence a dramatic case of market failure, but since this is not what we observe, there must be more to politics than calculated self-interest. And if voters are moved by civic virtues in deciding whether to vote, why should they suddenly leave these virtues outside the election booth when deciding how to vote?

These questions point to the need for a theory of non-rational politics, by which I mean politics that may be both less than rational and more than rational. ‘More than rational’: this means that the calculated self-interest is replaced by calculated concern for the general welfare. ‘Less than rational’: this means that the element of calculation is absent. Not all political behaviour is dominated by ends-means rationality, and not all ends are narrowly self-interested. I now want to show how the mainstream political science tradition can be enriched by incorporating elements from two outstanding recent studies of non-rational (indeed often irrational) political behaviour. The first is Paul Veyne’s Le Pain et le Cirque,[2] an analysis of civic giving in Classical Antiquity; the second Alexander Zinoviev’s The Yawning Heights,[3] a satirical and penetrating work about politics and everyday life in the Soviet Union. Veyne is, in my view, the greatest French historian now writing. I believe that his work will come to achieve the same status as Tocqueville’s on democracy in America, by which it is clearly inspired. Zinoviev’s work is only secondarily one of political science. It is a hallucinatory, and totally convincing, portrait of a society gone to rot, or perhaps one should say a society in which rot, far from being a sign of illness, has become the normal state of affairs.

The full text of this essay is only available to subscribers of the London Review of Books.

You are not logged in

[1] The Presidential Election Game by Steven Brams (Yale University Press, 1978).

[2] Le Pain et le Cirque by Paul Veyne (Le Seuil, 1976).

[3] The Yawning Heights by Alexander Zinoviev, translated by Gordon Clough (Bodley Head, 1979). Alexander Zinoviev’s new novel, The Radiant Future, will be published presently by Bodley Head.

[4] Comment on écrit l’Histoire, suivi de Foucault révolutionne l’histoire by Paul Veyne (Le Seuil, 1978).

[5] Les Hauteurs Béantes (L’Age d’Homme, 1977).

[6] Über Mehrwertige Logik (Deutscher Verlag der Wissenschaften, 1968).