Dynamic decision making in frictional labor markets

This dissertation contains three essays relating to the dynamic decision making of workers in labor markets with search frictions. In the first essay I determine the optimal unemployment insurance (UI) replacement rate in an equilibrium matching model in which workers are able to self-insure against unemployment risks by accumulating precautionary wealth. The optimal replacement rate is between 34 and 40 percent and balances the benefits of higher wages and better insurance provision with the cost of longer unemployment durations. I find that the optimal replacement rate is increasing in workers' dislike for work as measured by the Frisch elasticity of labor supply and in workers' degree of impatience as measured by their personal discount rate.; The second essay explores the role of retraining the unemployed in economies that are subject to structural change. I develop an equilibrium labor market model of structural change in which low-skilled workers may enroll in retraining programs in order to acquire industry specific skills. I use this model to investigate to what degree differences in retraining rates can explain the different evolutions of unemployment in the U.S. and Europe since the 1970s. I document that retraining rates in the U.S. are between two and five times higher than in Europe, suggesting that retraining opportunities are more widely available and more flexible in the U.S. I find that economies with better retraining institutions adjust better to structural change and experience lower unemployment, higher productivity, and higher aggregate output than economies with poor retraining institutions.; In the third essay I develop a framework to analyze the effects of policy on the intensive and extensive margins of labor supply. I extend a simple equilibrium matching model to allow workers and firms to bargain over working hours. Furthermore, I examine how well this model can explain differences in hours per worker and unemployment rates across educational groups in the U.S. and between OECD countries. I find that while the model is reasonably successful at explaining differences in hours of work, it is less successful at explaining differences in unemployment rates.