Policy experimentation with repeated elections

When is an incumbent incentivized to experiment with risky reform policy in the presence of future elections? We study a dynamic game between two political parties with heterogeneous preferences and a voter. The voter elects a party that then chooses a policy from a status quo and two risky reform policies. Under infrequent elections, the incumbent party experiments with its preferred reform policy even if its outlook is not promising. In contrast, with overly frequent elections, the incumbent prematurely stops experimentation because an imminent election increases the danger of losing power. While infrequent elections are suboptimal, inefficiency from overly frequent elections is severe enough that voter is worse than under a dictatorship. We also show that the voter's 'incumbent advantage' behavior could lead to an efficient outcome in the frequent elections limit. Our results suggest that leaders engage in risky reform when the surrounding political environment allows them to fail.