Counterfactual Analysis for Structural Dynamic Discrete Choice Models

This paper studies the identification and estimation of counterfactual outcomes in structural dynamic discrete choice models. The authors show that while observed discrete choice data identify utility differences, many policy-relevant counterfactuals remain only partially identified. The paper develops methods for deriving bounds on counterfactual outcomes under mild model restrictions and introduces computational tools and inference procedures suitable for high-dimensional settings. Through simulations and an empirical application to firms’ export decisions, the study demonstrates how partial identification methods can yield informative counterfactual analysis while clarifying the role of modeling assumptions in structural estimation.