Exact Structural Inference in Optimal Job-Search Models

Abstract

This article is a study of the exact posterior distributions of parameters appearing in a stationary optimal job-search model I exploit the simple latent structure of the search model when all job offers are observed to simulate posterior distributions of structural parameters when the latent structure is imperfectly observed. These simulations enable me to show the exact, and unusual, shape of the job-search likelihood when the data are durations and accepted wages. I also develop an algorithm to resample simulated posterior distributions to impose on the model the implications of fully optimal, utility-maximizing search. The methods are illustrated using simulated data.