Computation of Moral-Hazard Problems

We study computational aspects of moral-hazard problems. We consider deterministic contracts as well as contracts with action and/or compensation lotteries, and formulate each case as a mathematical program with equilibrium constraints (MPEC). We investigate and compare solution properties of the MPEC approach with the linear programming (LP) approach with lotteries. We propose a hybrid procedure that combines the best features of the both. The hybrid procedure obtains a solution that is, if not globally optimal, at least as good as an LP solution. It also preserves the fast local convergence property by applying the SQP algorithm to MPECs. Numerical examples show that the hybrid procedure outperforms the LP approach in both computational time and solution quality in term of the optimal objective value