Search, matching and heterogeneity[Recherche, appariement et hétérogénéité]

This thesis consists of three chapters that study frictional markets. The first chapter asks the question of what are the sources of labor income shocks, with a special focus on the scarring effects of recessions. I develop and estimate a dynamic frictional model of the labor market with heterogeneous workers and firms. The economic contribution of the first chapter is to show that sorting — the degree of complementarity between firms and workers — is a key component of idiosyncratic labor income risk. A technical contribution is to show that, while the determination of wage is a priori complex in a dynamic search model with heterogeneity, an efficient and robust algorithm exists. The second chapter explores to what extent a localized drop in commuting costs may lead to an increase in local employment. This chapter makes use of a discontinuity introduced by a French reform in September 2015 in the Paris metropolitan area. I find that cities that enjoyed a decrease in commuting costs experienced an increase in local employment. While the first two chapters analyze the labor market, the last chapter focuses on another key frictional market: the housing market. Little is known on the rental market because there are no comprehensive datasets recording rental agreements. To circumvent this issue, I collected data on rental ads in the Paris metropolitan area using web scraping techniques for a period of three months. I show that the rental housing market is well described by a directed search model. However, a non-negligible proportion of landlords use a two-step pricing approach when setting the rent, which raises interesting welfare and modeling questions.