Spousal Labor Supply and Unemployment: Estimates from a Household Search Model

The majority of job search models assume decision making takes place on the individual level, while in practice many people coordinate their labor market activity with a spouse. In this paper, I develop a job search model in which married couples jointly decide both how much to save and whether to accept or reject new job offers based on their current assets, wages, and unemployment insurance (UI) benefits. I am able to account for the endogeneity in these key variables through exclusion restrictions, variation in state UI laws and unemployment rates, theoretical restrictions on the hazards out of and into employment, and incorporation of various forms of unobserved heterogeneity in my model. I estimate my model with the 2004 and 2008 Panels of the Survey of Income and Program Participation using a modified version of a newly developed Bayesian Markov chain Monte Carlo algorithm, which I enhance to handle the sizable number of state variables in my model. My policy simulations show that both assets and spousal employment affect a household's responsiveness to changes in UI benefits, and that an estimated 75%-88% of the effect of UI on job search behavior is due to the liquidity-constraint effect, while only 12%-25% of the effect is due to the moral hazard effect.