The spatial model of electoral competition first proposed by Anthony Downs and subsequently extended by many authors is a core part of formal political theory. It has been and is currently used to study a wide variety of electoral processes and political institutions and its properties under many alternative conditions are now well known. All of this work, however, maintains the assumption that voters' preferences are exogenously given and can be treated as fixed while studying the behavior of competing parties and candidates. This makes all the resulting predictions conditional on this assumption. Empirical studies of voter preferences, by contrast, have connected changes in preferences to the platforms and actions of competing parties and candidates. The proposed research will connect these two literatures by developing a model of electoral competition that makes preferences endogenous, meaning that they co-evolve with party platforms during the election process. The model will then be explored to ascertain its implications for the existence of stable outcomes, for the ability to predict these outcomes based on initial conditions and assumptions about party behavior, and for its dynamic properties. The assumption of fixed preferences can be treated as a special case of this general model.