Optimal contracts for agents who play inspection games

In an inspection game, a monitoring agent chooses whether to monitor or not while, simultaneously, a suspect chooses whether to commit a good or a bad action. Under sensible assumptions, this game does not have a Nash equilibrium in pure strategies, only in mixed strategies. The model leaves open the question on whose behalf the monitor acts. The harm caused by the suspect's bad act does not appear in the monitor's payoffs, the reward the monitor receives (in case of monitoring and bad act are chosen) comes out of the blue, and the fine the suspect had to pay in this case simply vanishes. This projects introduces a third party, a principal, who suffers the harm, receives the fine, and pays out the reward. It analyzes contracts which consist of bonus payments and a share of the fine and tries to derive which contract structure would maximize the principal's payoff.