Spatial job search and commuting distances

Abstract

An empirical model for spatial job search is developed and estimated. The model allows for heterogeneous jobs. It can be extended to incorporate the effects of housing market search induced by acceptance of a job. In order to deal with unobserved heterogeneity among workers, the reservation utility level is treated as a random variable in the empirical work. The estimation procedure uses simulation in order to evaluate an otherwise cumbersome integral in the likelihood function. The model is estimated on a sample of married or cohabiting female workers in the Netherlands and takes into account wage rates, commuting distances and working hours as relevant job characteristics. Estimation results suggest, among other things, that workers are on average willing to accept an hourly wage rate that is 0.12 Dutch guilder lower in order to avoid one additional kilometre of commuting, that there are important spatial elements in the way vacant jobs are offered to workers and that workers in their early 30s have the highest arrival rates of job offers.