Bayesian inference in models based on equilibrium search theory

摘要

The equilibrium search model represents a substantial advance over previous models of job search, but its empirical performance fails in some crucial ways. Hence, flexible extensions are called for. In this paper, we consider three such extensions inareas where Bayesian methods can be used to great advantage. These are: (i) a specification where Bayesian priors are used to center a model over the restrictions implied by economic theory; (ii) a model involving a nonlinear production function where aclosed form expression for the likelihood function does not exist; and (iii) a model which allows for unobserved heterogeneity. We show how posterior simulation methods can be used for empirical analysis for all three models. The paper includes an empirical exercise involving the school-to-work transitions of Canadian and US school leavers.