UC Berkeley

Access to higher education is a key determinant of lifetime earnings in the U.S. Since the 1960s, selective public universities have admitted students mostly on the basis of standardized test scores and other measures of academic preparation. In this dissertation, I employ quasi-experimental and structural research designs to investigate the efficiency and economic mobility ramifications of these 'meritocratic' admissions policies. I focus on the selective University of California (UC) system, with each chapter analyzing a newly-constructed longitudinal dataset that links all 1994-2018 UC applicants and most 1975-2018 UC enrollees to their national college enrollment, major choice, and degree attainment (whether at UC or elsewhere); their UC student transcripts (for UC enrollees); and their 2000-2019 California wages.

Chapter 2 studies race-based affirmative action, which broadened lower-'merit' underrepresented minority (URM) college applicants' access to UC campuses until the policy was banned by a ballot proposition in 1998. I employ a difference-in-difference research design to show that ending affirmative action caused underrepresented minority (URM) freshman applicants to cascade into lower-quality colleges. The well-known "Mismatch Hypothesis" implies that this cascade would provide net educational benefits to URM applicants, but URM applicants' degree attainment declined overall and in STEM fields, especially among less academically qualified applicants, and URM UC applicant’s average wages fell in turn. These declines are not explained by URM students’ performance or persistence in STEM course sequences, which were unchanged after Prop 209. Complementary regression discontinuity and institutional value-added analyses suggest that affirmative action's net wage benefits for URM applicants exceed its (potentially small) net costs for on-the-margin white and Asian applicants. These findings provide the first causal evidence that banning affirmative action exacerbates socioeconomic inequities and suggest that loosening meritocratic admissions policies may generate efficiency and economic mobility gains.

Chapter 3 further analyzes the efficacy of test-based meritocracy in college admissions by evaluating the impact of a grade-based "top percent" policy implemented by UC between 2001 and 2011. Eligibility in the Local Context (ELC) provided large admission advantages to the top four percent of graduates from each California high school. I first employ a regression discontinuity design to show that ELC led over 10 percent of barely-eligible applicants from low-opportunity high schools to enroll at selective UC campuses instead of less-selective public colleges and universities. Half of those participants were URM, and their average SAT scores were at the 12th percentile of their UC peers. Instrumental variable estimates show that ELC participants' more-selective university enrollment caused large increases in five-year degree attainment and annual early-career wages. I then analyze ELC's general equilibrium effects by estimating a structural model of university application, admission, and enrollment with an embedded top percent policy. I find that ELC and counterfactual expansions of ELC substantively increase disadvantaged students’ net enrollment at selective public universities. Reduced-form and structural estimates show that ELC participants derived similar or greater value from more-selective university enrollment than their higher-testing peers, providing further evidence that access-oriented admission policies at selective universities can promote economic mobility without efficiency losses.

In Chapters 4 and 5, both coauthored with Aashish Mehta, I turn from meritocratic college admissions policies to the meritocratic allocation of lucrative fields of study. We study a popular class of policies -- which we term 'major restriction' policies -- that prohibit students with poor introductory course grades from earning their preferred college major. Chapter 4 employs a difference-in-difference event study design around the implementation of 28 major restrictions at four UC campuses since the 1970s to show that the policies are binding and differentially impact URM students and students with absolute (not comparative) academic disadvantage, closely paralleling the function of meritocratic college admissions policies in decreasing educational access for disadvantaged lower-'merit' students. A student-level extension of the event study design shows that major restriction policies tend to lead female and URM students to relatively lower-average-wage majors, generating cross-major stratification.

Chapter 5 focuses on one specific major restriction policy -- which limited access to the UC Santa Cruz economics major between 2008 and 2012 -- and uses a regression discontinuity design to show that lower-GPA students prohibited from declaring the economics major earned $22,000 (46%) lower annual early-career wages than they would have as economics majors. A decomposition of this wage effect shows that the return to majoring in economics would likely have been above-average for the near-threshold students rejected from the economics major, once again suggesting the potential for efficiency and economic mobility gains in implementing a less 'merit'-oriented allocation policy.

In sum, this dissertation presents a collage of evidence from three educational allocation policies suggesting that the reallocation of selective higher education to disadvantaged students with relatively poorer measured academic preparation can promote both economic mobility and allocative efficiency, with those students' net education and wage gains exceeding their crowded-out peers' net losses. These efficiency findings undermine the primary justification for the 1960s implementation of meritocratic admissions policies at public institutions.