The Major Divide: Student Sorting, Meritocratic Admissions & Upward Mobility

Abstract

Large earnings gaps by socioeconomic status (SES) persist among college applicants even where tuition is free. Using the universe of Danish applications linked to administrative earnings, I ask whether these gaps reflect low returns to selective majors, admission rules, or students' own choices. Leveraging GPA admission cutoffs from a centralized matching algorithm, I apply a regression discontinuity design (RDD) to examine whether low-SES students benefit from selective college majors. Scoring just above the cutoff for admission to a more selective major raises low-SES students' income by 9% at age 30, partially closing the class gap in earnings. For high-SES students I find no detectable gain. To disentangle the contribution of major-specific returns from that of students' choices as sources of inequality, I then estimate a joint model of college major choice and earnings in which returns depend on peer composition, using students' rank-ordered application lists to correct for selection on unobservables. Low-SES returns exceed high-SES returns, relative to their outside options, across most of the major selectivity distribution. Current GPA-based meritocratic admissions are inefficient in terms of maximizing earnings: implementing need-affirmative admissions, holding capacity fixed, would reduce the class gap by 15% through upward mobility of disadvantaged students. Notably, the class gap in earnings could largely be eliminated if students selected college majors based on returns, though this requires expanding capacity. Thus, disparities persist not because disadvantaged students experience low returns from selective majors, but because of how seats are allocated and which majors students choose.