Policy Gambles

This paper develops a theory of policy making, that examines the incentives for experimentation with new policies and the scrappage of adopted policies. We demonstrate that a government which cares about its reputation out of electoral concerns, takes socially ine?cient policy gambles that may result in two kinds of ine?ciencies ? Þrst, a government may ine?ciently experiment by undertaking a new policy initiative that it (and the voter) knows is unlikely to succeed, and second, the government may prefer to not learn from experience and instead persist with an adopted policy despite publicly observable evidence of its failure. Furthermore, these ine?ciencies are systematically related to the electoral cycle. Early on in its term a government is likely to enact policies that are either too conservative or too radical, while later on in its term the government is likely to show ine?cient policy persistence.