Research

Social scientists have long been concerned with the slowing convergence of the gender wage gap in the United States since the 1990s. While research has proposed various explanations for the stalled convergence, the contributions of persistent gender segregation in college majors and rising earnings inequality between majors remain underexplored. This article examines how trends in the gender wage gap among college-educated workers correspond to changes in the composition of majors among men and women, as well as to changes in gender-specific wage returns to different majors. Using a detailed decomposition method and data from the National Survey of College Graduates, I find that 30 percent of the observed gender wage convergence from 1989 to 2023 can be explained by a greater decline in the share of education degrees among women, a historically lower-paying field. Yet, two countervailing forces hindered further convergence. First, men increasingly earned degrees in electrical engineering and computer science at a higher rate, a field associated with a large wage premium. Second, declining wage returns to education degrees disproportionately undermined women’s wages, given women’s continued overrepresentation in the field. Together, these two forces offset 53 percent of the gender wage convergence, which can help explain the stalled progress toward gender pay equality among college graduates.

How do schools affect socioeconomic inequality in academic achievement? Most prior studies isolate the contribution of a single mechanism rather than analyze how multiple potentially competing mechanisms jointly generate inequality. This article develops a unified theoretical and empirical framework that integrates three mechanisms: (1) unequal access to high-quality schools, (2) heterogeneous returns to high-quality schools, and (3) differential selection into high-quality schools by students’ family socioeconomic status (SES). Analyzing data from the 2001–2007 Taiwan Education Panel Survey (TEPS) with a novel nonparametric causal decomposition framework, we find nearly offsetting mechanisms: unequal access benefits higher-SES students and increases the SES achievement gap by 8 percent; heterogeneous returns benefit lower-SES students and decrease the SES achievement gap by 10 percent; and differential selection into high-quality schools may benefit higher-SES students and explain about 3 percent of the gap. Therefore, high-quality schools in this system contribute nearly nothing on net to the SES achievement gap. However, policies that expand access to high-quality schools for lower-SES students could narrow the SES achievement gap.

Sociologists have long debated whether college is the “great equalizer,” yet little is known about how the equalizing power of a college degree varies across fields of study. Drawing on the recent literature on school-to-work linkages, I theorize that fields with higher linkage strength—the degree to which a program structurally channels graduates into a particular set of occupations—reduce the influence of family origin and therefore promote greater mobility. To test this hypothesis, I calculate field-level linkage strength using the 2010–2012 American Community Survey and analyze data from the 2010–2019 National Survey of College Graduates. Using a residual balancing technique to address selection, I estimate origin–destination rank–rank slopes by field of study and examine the effect of linkage strength on field-level slopes. Results show that linkage strength significantly attenuates the origin–destination slope, even after controlling for institutional selectivity, graduate-school sorting, and STEM curriculum. Supplementary analyses using the National Longitudinal Survey of Youth 1997 with richer controls yield similar results and provide no evidence that the findings are driven by selection on pre-college ability. These findings suggest that the organizational structure linking fields of study to the labor market independently shapes the equalizing potential of a college degree, with important policy implications for weakening the intergenerational persistence of inequality.