Divided we reform? : evidence from US welfare policies

Divided government is often thought of as causing legislative deadlock. I investigate the link between divided government and economic reforms using a novel data set on welfare reforms in US states between 1978 and 2010. Panel data regressions show that under divided government a US state is around 25% more likely to adopt a welfare reform than under unified government. An analysis of close elections providing quasi-random variation in the form of government and other robustness checks confirm this counter-intuitive finding. The empirical evidence is consistent with an explanation based on policy competition between governor, senate, and house.