Authors
Abstract
This paper develops a method to analyze policy reforms in environments with discrete choice, such as occupational choice or default. Computing transition paths in these settings is computationally challenging, particularly in models with substantial heterogeneity and many endogenous states. We extend perturbation methods to handle discrete choice by appropriately tracking both intensive-margin changes conditional on discrete choices that are relatively small and extensive-margin changes resulting from a switch in a discrete choice that are relatively large. The method is fast, scalable, and efficient, providing good initial estimates for global solution methods. We demonstrate our method by analyzing optimal business taxation in a model with occupational choice between paid- and self-employment.