Essays on the relationship between migration and labor market skill

This dissertation consists of three essays describing relationships between migration behavior and labor market skills The first essay investigates how the distribution of human capital across local labor markets evolves over time. Using a combination of U.S. data sets, I decompose generation-to-generation changes in local human capital into three factors: the previous generation's human capital, intergenerational transmission of skills from parents in the previous generation to their children, and migration of the children. I find evidence of regression to the mean of local skills at the state level and divergence of local skills at the commuting zone level. Labor market size, climate, local colleges, and taxes affect local skill measures. Skills move from urban to rural labor markets through intergenerational transmission but from rural to urban labor markets through migration. In the second essay, I assess the effect of schooling on the propensity to migrate. Consistent with a large literature, I document positive correlations between schooling and migration. However, the observed correlation between schooling and migration might reflect mechanisms other than the effect of schooling on the costs or returns to migration. So, I estimate this effect by exploiting schooling variation due to compulsory schooling laws (CSLs). I estimate negative effects of schooling on migration among those with relatively little schooling. Results from CSL policy changes in 1947 and 1957 in the U.K. provide the strongest evidence of a negative effect. Analysis with U.S. state policies yields negative point estimates of the relationship. The third essay accounts for migration behavior in the measurement of U.S. school quality through much of the twentieth century. Previously, the effect of state school inputs on labor market outcomes was studied in Card and Krueger (1992) and Heckman, Layne-Farrar, and Todd (1996). I extend their analysis in two ways. First, I correct state-level returns to schooling for selective migration, using a method from Dahl (2002). Second, I use more recent data. Higher state-level school inputs are associated with higher returns to schooling after correcting for selective migration. Hence, the results in Card and Krueger (1992) are not due to bias from selective migration