This paper presents a new simulated maximum-likelihood method that rests on estimating the likelihood nonparametrically on a simulated sample. We prove that this method, which can be used on very general models, is consistent and asymptotically efficient.

Principal-agent models of moral hazard have been developed under the assumption that the principal knows the agent's risk-aversion. This paper extends the moral hazard model to the case when the agent's risk-aversion is his private information.