How Do Corrupt Politicians Survive in Office? Evidence from Brazilian Municipalities

I propose and test an explanation for the persistence of political corruption. I analyze the incentives of local legislators to further investigate corrupt transactions in local governments after audit results are released and to hold mayors accountable. I do this by systematically investigating changes in councilors’ assets and in the campaign revenues of tens of thousands of local legislators. I also use the randomization generated by lotteries that determine which specific municipal governments are investigated. I argue that, after an audit, incriminated mayors focus on preventing defections and blocking further investigations of malfeasance (e.g. hearings and parliamentary commissions of inquiry), which could potentially result in an impeachment. I demonstrate that the effects of audits on councilors’ personal wealth and campaign funds is conditioned by three main factors: their political affiliation, the strength of the mayor's party in the local council and on the mayor's own re-electoral incentives. Depending on the specific incentive structure provided by the local political system and their previous alliances with the mayor, legislators are able to capture a larger amount of rents and campaign funds in return for protecting the incriminated mayor from further investigations. I find that, on average, audits lead to abnormal increases in the wealth accumulation of councilors affiliated with political parties that were part of the mayor’s previous electoral coalition but were not affiliated with the mayor’s party. In municipalities where the mayor’s party needs votes from other parties to secure a majority in the municipal council, after an audit, legislators, on average, observe a higher increase in their wealth in comparison to municipalities where the mayor’s party holds a higher share of the seats in the local council. In addition, legislators in the mayor’s previous electoral coalition are, on average, immune from reductions in campaign funds after an audit in municipalities where the incriminated mayor is eligible for reelection. These findings provide evidence of how incriminated officials use the competing material interests of legislators to hinder legislative oversight and to avoid defections in their coalitions. As I show in the final section of the paper, this strategy works because legislators are not, on average, punished by the electorate for their inaction in sanctioning the mayor for corruption revelations.