Inspection in models of adverse selection

This paper analyses the optimal inspection, tax and penalty functions, in models where the principal knows the distribution of agent types, inspections are perfect and costly, and where the actions and the types of agents are not costlessly observable by the principal. Agents and the principal have different objective functions. However, the objective functions are not completely antagonistic. The action of an agent creates an externality which affects the rest of the agents. The goal of the planner is to minimize the inefficiency created by these externalities.