Investigating Corruption

Agency theory has had little to say about the control of bureaucratic corruption, perhaps the greatest agency problem that exists. I consider the role of incentive contracting for reducing corruption through the use of independent investigations, a common way to monitor corruption. In simple settings, bureaucratic corruption can be suppressed by rewarding and penalizing bureaucrats on the ndings of independent investigations. However, I also show how incentive contracting changes behavior in undesirable as well as bene cial ways. I outline three harmful behavioral responses to the use of investigations. First, many investigations are (e ciently) instigated by customer complaints. This implies that bureaucrats become excessively interested in 'keeping the customer happy', even when it is not e cient to do so. Second, bureaucrats often have private information on how cases should be handled, information which is hard to verify to investigators. I show that when o ered incentive contracts, bureaucrats have an incentive to ignore this information and 'do things by the book', o ering decisions which are more likely to be consistent with the opinions of their superiors. Finally, bureaucrats sometimes collect bribes to 'look the other way'. i.e., to ignore known transgressions. A solution to this is to o er a reward for bringing cases to light. But an implication of this is that the bureaucrat may waste resources by generating 'nuisance cases' simply to receive the bonus. In each of these cases, the harmful responses by bureaucrats may be su ciently costly that it is more e cient simply to allow corruption. The ndings are meant as a cautionary note on the how investigations should be used to combat corruption, and why corruption is likely to be a more di cult agency problem to resolve through incentive contracts than others. I am grateful to Gunnar Eskeland, Michael Raith, Tano Santos, participants at USC, the Econometric Society Summer Meetings, LACEA, and especially Kim-Sau Chung for helpful comments. This work was funded by the World Bank, NSF, and the University of Chicago. Any errors are my own.