Abstract
A formal model of electoral behavior is developed under the assumption that candidates have policy preferences as well as an interest in winning per se. This model is shown to have an equilibrium in a k-issue space when there are two candidates. The implications of this model are compared to the implications of the Downsian-type model where candidates are interested only in winning. Testable propositions are derived via the use of comparative statics. The results of recent studies are shown to coincide with the synthesis model but not the pure Downsian model.The theoretical model bridges the gap between formal theory and empirical research and unifies a variety of seemingly unrelated studies.