A Simple and Robust Estimator for Discount Factors in Optimal Stopping Dynamic Discrete Choice Models

We propose a simple and robust two-step estimator for discount factors in a class of dynamic discrete choice models. The estimator follows from constructive identification results, including a new identification result for a general, time separable discount function. The estimator is derived as the solution to well-behaved sample moment conditions which are linear in the discount factors and are independent of the utility function. The estimator is therefore easy to implement, computationally light, and in contrast to existing estimators, robust to biases from finite sample approximations to the unknown utility function. We apply the estimator to data on mortgage defaults under an identifying assumption of time homogeneity of the utility function. We compare the performance of the proposed estimator to alternative two-step estimators that jointly estimate the discount factor and the utility function. The results show that our proposed estimator's robustness to finite sample approximation bias and its computational ease do not necessarily come at a material expense of precision