Incumbency and the Minimum Winning Coalition

Abstract

In order to understand the campaign behavior of candidates and parties, it is clearly important to consider their goals for the election. These goals have been the subject of many landmark works, concerned with European as well as American party systems. Looking at one type of goal-the of the election victory-William Riker postulates that social situations similar to n-person, zero-sum games with side payments, participants create coalitions just as large as they believe will ensure winning and no larger.2 American elections, Riker feels, are perceived to fall into the category of n-person zero-sum games, and therefore candidates will tend to aim for a minimum winning coalition.3 This size applies when the players have perfect information about who belongs to which coalition. In support of the principle Riker cites three instances in American history when oversized winning coalitionssuch as the one-party dominance from 1816 through the early 1820s-were eventually reduced in size. The tendency to reduce an oversized winning coalition in elections, according to Riker, has the force of political necessity behind it: the oversized coalition includes people whose support was not necessary for gaining a majority, but who nevertheless feel that they have a claim on the winnings.4 So although the principle is often assumed to apply to the coalitions