Essays on labor market

In Chapter 1, I provide a simple empirical model to estimates return to college, accounting for unobservable ability. Contrary to previous results I find that endogenous self selection generates an upward bias in OLS estimates. Corrected estimations show that Italian and US Return to college are very similar. In Chapter 2, I study the effect of labor market institutions on schooling choice. I introduce college decision in a search on-the-job model, where wage offers distribution depends on workers' innate ability and schooling level. The interaction between workers' heterogenous abilities and labor market structure endogenously determine both college wage premium and college decisions. As in the classical search on-the-job model, higher labor mobility and higher wage dispersion increase workers' probability to reach the right tail of the wage distribution. This provide agents with a larger incentive to shift wage distribution to the right by attending school. The model fits US and Italian data, two countries that present wide differences in educational achievements. Specifically, the US presents a larger college premium, higher enrollment and graduation rates in college. I find that the labor markets in US and Italy differ significantly in two dimensions. The US is characterized by i) higher job mobility and ii)higher return of innate ability. The former accounts for around 30% of the different enrollment and graduation rates. The latter rises the observed college premium in US above Italy, despite the effect of schooling on wage being higher in Italy, and can account for almost 50% of differences in educational attainment. In Chapter 3, I develop a model of voluntary union membership, where unions provide excludable employment protection to members, by supporting them in case of dismissal conflicts and enforcing higher firing costs. In Europe, where closed shops are exceptional and often unconstitutional, membership rates significantly vary across countries. When calibrated to match the features of labor markets in different European countries, the model is able to account for both heterogeneity of membership rates across Europe and the pattern of (de)unionization observed in the 90s. Specifically I find that membership rates i) are low in countries where public employment protection is stronger, and/or available to a wider fraction of workers; ii) decrease when low unemployment duration reduces the welfare loss of a dismissal.