9 Conclusion
External audits are used by many governments to monitor public officials and service providers and deter various forms of rent extraction. Available evidence on the effectiveness of such top-down monitoring is mixed however and interpretation of these findings is complicated in part because each study evaluates a different intervention in a different context. Yet it is of first-order importance for policymakers to know under what circumstances increased top-down monitoring is effective. We report results from a randomized policy experiment designed to test whether in the Brazilian local government context higher audit risk deters rent extraction in procurement, health service delivery and cash transfer targeting. Since audit risk increased from the same baseline level of about 5 percent to about 25 percent for all agents in the municipality, any impact heterogeneity across areas of local government activity cannot be driven by differences in the intervention itself or by differences in the broader socioeconomic or institutional context. We find substantial heterogeneity in both the level of rent-extraction and its response to increased audit risk. While there is substantial corruption in public procurement as well as widespread absenteeism among health workers, only corruption in procurement is responsive to increased audit risk. Cash transfer program fraud is almost negligible even without increased monitoring. These results suggest that increasing the likelihood of an audit alone is not sufficient to deter rent-taking if potential sanctions and the probability of sanction conditional on detection are too low. Moreover, increasing the likelihood of an audit may also be unnecessary for programs that are targeted based on easily observed individual or household characteristics.
