Familiarity and Inertia in the Formation of Governing Coalitions in Parliamentary Democracies

Political scientists who set out to test theories of coalition formation in parliamentary contexts (notably Browne, de Swaan, and Taylor and Laver) found only limited evidence to support the more classical game-theoretic propositions, which predict the formation of coalitions that command a majority of seats in a parliament but are otherwise as small as possible, in some sense of the word ‘small’. As a consequence, Browne later advocated the laying aside of these size theories in favour of theories that took account of the policy preferences of potential coalition partners, and in two separate studies theories were tested that focused upon the ideological component in coalition formation. Both these studies found theories based on presumed policy preferences to perform better than size theories. A more recent study has shown that the relative performance of theories based on size and ideological considerations depends on assumptions made in conducting the research. This study employed multiple regression analysis to establish that both kinds of theory had parts to play in an explanation of formation outcomes, which were dominated sometimes by size and sometimes by ideology, depending on country and time period. In the course of the analysis an additive combination of size and ideology was found to correlate to the extent of r ≃ 0·4 with formation outcomes, producing consistently better predictions than any existing theory.