Elections, political competition and bank failure

We exploit exogenous variation in the scheduling of gubernatorial elections to study the timing of bank failure in the US. Using hazard analysis, we show that bank failure is about 45% less likely in the year leading up to an election. Political control (i.e., lack of competition) can explain all of this average election year fall in the hazard rate. In particular, we show that the reduction in hazard rate doubles in magnitude for banks operating in states where the governor has simultaneous control of the upper and lower houses of the state legislature (i.e., complete control) heading into an election.