Ulric B. and Evelyn L. Bray Social Sciences Seminar

Abstract: This paper develops identification and estimation methods for finite-horizon dynamic discrete choice models when agents' actions are unobserved to econometricians. We provide conditions under which choice probabilities and state transition rules are nonparametrically identified when there is a continuous state variable. We also consider scenarios where only discrete state variables are available. Our results extend to dynamic models with serially correlated unobserved heterogeneity. We develop sieve maximum likelihood estimation for the parameters of interest. Monte Carlo simulation results demonstrate the effectiveness of the proposed approaches.