Essays on Human Capital: Re-skilling, College Major Choice, and the Incidence of Local Labor Market Shocks

Abstract

This dissertation consists of three chapters that study individual investment in humancapital and the incidence of economic shocks. I use administrative data to answer these questions, developing and applying structural methods to analyze individual choices in combination with a careful consideration of research design. Chapter 1 studies the efficacy of using late-in-life educational investments to adapt to economic change. I focus on the Great Recession, an important shock that was accompanied by large increases in adult community college enrollment. I use administrative panel data on earnings and enrollment histories to estimate a dynamic model that captures the tradeoffs workers face in pursuing education later in life. A key contribution of the framework is to flexibly quantify worker skills using a data-driven approach. I find substantial heterogeneity in the earnings benefits of returning to school, but that these effects are limited or negative for those at the margin of enrollment. Chapter 2, co-authored with Justine S. Hastings, Christopher A. Neilson, and Seth D. Zimmerman, studies how the choice of college and major are related to earnings in a context where selection bias can potentially confound direct measurement. We leverage administrative student data and the centralized structure of the Chilean higher education system to estimate a generalized Roy model of selection into colleges and majors. We find an important role of match effects and sorting on causal gains in generating earnings dispersion, indicating a large role for sorting into fields of study based on comparative advantage. Chapter 3, co-authored with Sharada Dharmasankar, studies how economic shocks propagate through a local labor market by studying the effects of mass layoff events on an entire labor market. By leveraging matched employer-employee data, we can identify sharp layoff events that affect a specific labor market, cleanly identify which workers are directly and indirectly exposed to a layoff event. We find that mass layoff events permanently scar local labor markets with some evidence of spillover effects.