Essays on economics of higher education

Abstract

This dissertation consists of three chapters. In Chapter 1, We use rich administrative microdata from Missouri to examine the potential to expand and diversify the production of STEM degrees at universities by tapping into the population of community college students. We find that the scope for expansion is modest, even at an upper bound, because most community college students have academic qualifications that make them unlikely to succeed in a STEM field at a university. We also find there is almost no scope for community college students to improve the racial/ethnic diversity of four-year STEM degree recipients. We conclude that it will be challenging to expand and diversify STEM degree production at universities with interventions targeted toward community college students. In Chapter 2, I extend the analysis of Qian and Koedel (2020) using the Educational Longitudinal Survey of 2002 from the National Center of Education Statistics. Using these data, I examine the potential to expand and diversify the production of STEM degrees at universities by tapping into the population of two-year college students. Consistent with Qian and Koedel (2020), I find that although the number of two-year college students is large, only a small fraction of these students possess academic qualifications that suggest they would succeed in a STEM field at a four-year university. Therefore, policies targeting two-year college students can only increase STEM-degree production by a small amount. I also find no evidence that such policies can improve female or minority representation in STEM fields, with the possible exception of Hispanic students, for whom my estimates cannot rule out a modest positive effect on representation. In Chapter 3, I use a recent data panel spanning the years 2001-2017 to study the effect of local-area unemployment on postsecondary enrollment and degree completion. My analysis extends the literature in several ways, most notably by (a) incorporating data well into the recent economic recovery from the Great Recession, (b) using improved (more accurate) measures of postsecondary enrollment, and (c) accounting for the attenuating effect of measurement error in calculated unemployment rates. Like in previous research, I find that postsecondary enrollment is countercyclical. I further show that the countercyclical enrollment pattern is concentrated among students in two-year and sub-two-year degree programs. There is suggestive evidence that men are more elastic than women in their enrollment response to unemployment, and unemployment rates have effect on degree completion, but my estimates are too imprecise to draw strong inference.