Partially binding platforms: The effects of campaign promises in political competition

Most past models of political competition do not stress the role of campaign platforms: they assume either that candidates cannot betray their platforms, or that candidates can betray their platform without any cost. My dissertation examines electoral competition when campaign platforms are partially binding. A candidate who implements a policy different from his platform must pay a "cost of betrayal" that increases in the size of the discrepancy. My dissertation is thus the first to consider platforms as a commitment device. The first chapter shows that under partially binding platforms, the median-voter result does not hold, and it examines some comparative statics. The second chapter predicts that when voters do not know the candidates' ideal policies, an extreme candidate whose position is far from the median policy may win with higher probability than a moderate candidate. The third chapter shows that, with asymmetric candidates and partially binding platforms, one party wins with higher probability: this contrasts with most past models of asymmetric candidates, which predict ties.