The Concise Oxford Dictionary of Politics (223 page)

(1) Broadly, any study of politics using the methods and characteristic assumptions of economics. The methods are deductive and rely heavily on differential calculus because they depend on the marginal principle. The marginal principle stresses that changes in one quantity (say, propensity to vote for the incumbent party) depend on changes in another (say, the level of unemployment last month). A fully-fledged public choice application to politics would form a theoretical model, deduce its consequences, and then test them on observed behaviour. Most actual applications are less ambitious. The characteristic assumptions of economics are: that individuals, not groups or societies, are the appropriate unit of analysis; that tastes are taken as given; that people make choices under scarcity; that they would always rather have more than less, but that their preferences reflect diminishing marginal substitutability between any two goods. The last condition means that, faced with a fixed budget to split between goods A and B, the consumer will substitute more and more Bs in exchange for one A, the more As she already has.
Well-known work in this spirit includes:

(a) the median voter theorem of Duncan
Black
and its many derivatives;
(b) analysis of the logic of collective action (M. Olson , 1965), which builds on the
prisoners' dilemma
to explore why any interest groups exist (left to themselves, rational political actors would almost always leave the job of lobbying for somebody else to do), and which sort of interest groups are likely to be stronger than their relative weight in the population would warrant;
(c) the ‘political business cycle’ literature which tries to predict the popularity of parties from the state of the economy; and
(d) the properties of actual and potential voting systems, and the rational behaviour of political actors given that a particular voting system exists (this last shades off into
social choice
).

(2) More specifically, a school of writers founded in Virginia by J. M. Buchanan (Nobel laureate in economics, 1986) and Gordon Tullock in the 1960s. Their most important work is
The Calculus of Consent
(1962). In the
social contract
tradition, this argues that only a constitution with unanimous support is legitimate: they regard such a constitution as embodying the
Pareto
condition into politics because nobody would accept it unless he or she thought he or she would be at least as well off with it as without. The Virginia school are suspicious of governments, because they argue that political actors are no less likely to be driven by selfish motives than economic actors; therefore it is inconsistent to suppose simultaneously that the economy is driven by self-interested actions and the polity by altruistic ones. Recent Virginian research has concentrated on the alleged oversupply of
bureaucracy
in modern democracies, and on
rent-seeking
. Buchanan and Tullock are personally identified with the libertarian right, and many of their ideas have been adopted by right-wing politicians ( see
Thatcherism
), but nothing in the central ideas of public choice leads necessarily to right-wing conclusions.