Essays on human capital heterogeneity and agglomeration
This dissertation consists of three essays exploring how human capital heterogeneity within cities enhances individual productivity. Agglomeration theory suggests productivity is driven by rapid and frequent interactions with others in spatially-constrained areas. Using formal education data from the 2011 American Community Survey, we empirically test that theory by estimating the effects of local human capital stock characteristics on individual wages. In essay one, we posit that some kinds of knowledge are harder to exchange remotely and thus workers possessing those knowledge types benefit more from close physical proximity to others. Our theoretical framework demonstrates the returns to finding a partner to exchange ideas with are heterogeneous across knowledge types. We propose agglomerative environments favor “soft skills” where creativity and informal networking are important. Our empirical results show people with non-STEM majors benefit more from locating within a city. Conversely, terminal degrees such as a J.D. or M.D. experience a smaller urban wage premium. Essay two studies the role of specialization of human capital types for individual productivity. Glaeser et al. (1992) finds local industrial specialization has a non-increasing effect on employment and wage growth. Our empirical results indicate specialization of knowledge can play an important role in promoting productivity when simultaneously controlling for a population size effect via the urban wage premium. We find STEM-related knowledge benefits greatly from local specialization of knowledge. However, the urbanization effect from city population size often exceeds the specialization effect. The third essay studies how workers in cities learn from one another in dense economic settings. Following Winters (2014), we estimate the impact of changes in the local stock of particular knowledge types on individual wages. The richness of our data allows us to estimate the productivity effects from over 400 different combinations of human capital interactions. We find most knowledge types are more productive when local STEM presence increases. The effect is strongest among workers with higher levels of educational attainment in the earlier stages of their careers. Similarly, areas such as government and psychology generate productivity gains among others. However, the lowest productivity gains occur from interactions with religious or education backgrounds.