From Lapdogs to Watchdogs: Random Auditor Assignment and Municipal Fiscal Performance

A fundamental question in organizational economics is how to structure organizations to better align incentives across levels. While monitoring could mitigate agency problems, it can itself be rendered ineffective if auditors are corruptible. In this paper, I evaluate the consequences of changes in the design of monitoring institutions that limit auditors' conflicts of interest. I exploit the staggered introduction of a reform that removed the control of auditors' appointment from local politicians and introduced a random assignment mechanism. I obtain four main findings. First, random matching severs auditors-mayors connections. Second, treated municipalities significantly improve their net surpluses and debt repayments, per national government objectives. Third, the fiscal improvement results from a sizeable increase in tax capacity. Fourth, treatment effects are a combination of selection, matching, and incentive effects. These findings highlight the value of auditor independence and illustrate how changes in the organizational design of the state can improve governance.