We develop an aggregate model of the presidential vote based on the appropriation of political as well as economic information by a rational voter. We argue that, depending upon historical context, information about political corruption is relevant to individual, and hence aggregate, vote choice. In preindustrial, community-oriented machine politics, the rational voter exchanged votes for particularistic benefits. As the social and political perspective shifted to a universalistic standard, information about corruption has become for him or her one of the criteria by which to evaluate the performance of the incumbent party. By including information about corruption alongside information about the economy, our model significantly improves upon conventional economic voting models in explaining post-New Deal presidential election outcomes.