Introduction

There are two very different logics of voting, the logic of ideology and the logic of the incumbent's performance. A citizen may vote out of ideological closeness to parties or candidates, but the citizen may also vote in an attempt to discipline and control the government. In the first case, only ideological closeness matters. In the second case, ideology does not play a role, and the voter only takes into account the government's performance. The underlying rule in each case is clear. For ideological voting, the voter follows a decision rule that is based on ideological closeness: vote for the party that minimizes ideological distance. For performance voting, the voter follows a decision rule about reelection: vote for the incumbent if the utility produced by the government's policies is higher than a certain preestablished value; otherwise, punish the incumbent.

Of course, these two logics correspond roughly to the distinction between prospective and retrospective voting (Manin, Przeworski, and Stokes 1999). As Fearon (1999) has put it, under prospective voting, elections are a matter of selecting good types, whereas under retrospective voting, elections are a matter of sanctioning poor performance. This basic distinction can also be expressed in terms of the problem that each logic of voting produces: prospective voting is associated with adverse selection, retrospective voting with moral hazard (Ferejohn 1995).

Each logic of voting has been investigated by apparently self-contained literatures. Ideological voting is studied by spatial models, whereas performance voting is studied by accountability models.