Optimal Fines Under Announced and Surprise Inspections

This paper examines optimal fines in a regulatory framework where the regulator can choose either surprise or announced inspections to monitor a firm for compliance. The regulator  hires a supervisor who exerts unobservable monitoring effort. The firm can invest in avoidance, but it receives a fine if the supervisor finds evidence of non-compliance. Our Principal-Supervisor-Agent model focuses on the trade-off between the frequency of inspections and the magnitude of the fine.  We find that when inspections are unannounced, the optimal fine is maximal, but when they are announced, the optimal fine may be less than maximal.