On the Estimation of Labour Force Participation, Job Search, and Job Matching Models Using Panel Data

Abstract

In this chapter we present a class of discrete-time structural models of labour force dynamics which are estimable using the same methodology. There have been a number of such models discussed in the literature, but only a very few have been estimated in their structural form. Although the models are categorized as either fitting into labour force participation, job search, or job matching frameworks, they all share in common the reservation wage property, namely that there is a change in state only when the random wage draw moves above or below a threshold value. They key to estimation is the numerical solution of the dynamic programming problem for the reservation wage path. In the finite horizon case considered here reservation wages are time varying and all of the economic theory is contained in the restrictions on the shape of the reservation wage path; only a few parameters fully describe the entire path, although not generally in closed form. The likelihood function based on observed labour force states is also a function of the reservation wage path. Thus, the parameters which determine the reservation wage path can be estimated by iterating between the dynamic programming solution and the likelihood function.