A general solution method for moral hazard problems
提出一种新方法求解委托代理问题，适用于满足单调似然比性质但不符合一阶方法的场景，通过最大-最小-最大公式处理代理人行动和最优反应。
Principal-agent models are pervasive in theoretical and applied economics, but their analysis has largely been limited to the ``first-order approach'' (FOA) where incentive compatibility is replaced by a first-order condition. This paper presents a new approach to solving a wide class of principal-agent problems that satisfy the monotone likelihood ratio property but may fail to meet the requirements of the FOA. Our approach solves the problem via tackling a max-min-max formulation over agent actions, alternate best responses by the agent, and contracts.