Three essays on principal-agent problems

This dissertation is divided into three parts. The first part studies the connection and comparison between the decentralized menu contracting procedure and the centralized mechanism contracting procedure in multi-agency games with epsilon-ex post implementation. We establish the revelation principle and delegation principle for epsilon-ex post implementation. According to these principles, epsilon-ex post menu design is merely strategically equivalent to individual-based epsilon-ex post mechanism design rather than joint-based epsilon-ex post mechanism design. We show that joint-based epsilon-ex post mechanism design always weakly dominates individual-based epsilon-ex post mechanism design and epsilon-ex post menu design. We also provide economically interesting sufficient conditions for when the former strictly dominates the former two and when the former is equivalent to the latter two.;The second part investigates the counterpart in generalized multi-agency contracting games with Bayesian implementation. Bayesian menu design is merely strategically equivalent to individual-based Bayesian mechanism design rather than joint-based Bayesian mechanism design. We also provide several sets of economically interesting sufficient conditions for when joint-based BIC mechanism design can still make the principal as well off as individual-based BIC mechanism design and Bayesian menu design.;The third part (joint with Frank H. Page, Jr) provides a generalized model concerning the one-shot pure-strategy moral hazard contracting game with the inclusion of observable actions as well as unobservable actions. We present the justification of the inclusion of observable actions as contingencies in contracts in views of the influence of existing observable actions. Primarily, we propose conditions under which the solution to such a principal-agent problem exists. It is valuable for the principal to include observable actions as contingencies in contracts. Including observable actions along with the outcomes in the contract can reinforce the incentive of outcome-contingent contracts and curb the agent's opportunist tendency.