College graduation rates have been steadily increasing among White, Black, and Hispanic populations for the last five decades, although both Black and Hispanic students lag far behind their White counterparts. Additionally, the graduation rates for men in all three groups have been on a much slower growth trajectory than for women. That began around 1970, effectively reversing the gender gap in favor of men that had existed for generations before that. In fact, Hispanic women in the 1990 birth cohort are poised to notch higher college graduation rates than Black males. What is behind these shifting patterns and what does the future hold? A new working paper aims to find out.
A trio of researchers extend their previous work, which analyzed the career and family decisions of women born between 1935 and 1975, by adding the education component, now also including males in their modeling, and stretching the timeline. Data in both studies come from the Current Population Survey (CPS) of the U.S. Census Bureau.
The new study starts by creating a model of individual and family decision-making for students born between 1960 and 1980 (as well as fitting a broad range of other demographic and labor market outcomes over the same period), with groups of students broken out in ten-year cohorts. The model assumes that White, Black, and Hispanic individuals have common preferences for school, work, marriage, and fertility/family size. These preferences are assumed to differ by gender, but are held fixed over cohorts. The three factors that do differ across cohorts (and by gender/race) are parental background, consisting of mothers’ education, parent’s marital status, and parent’s immigration status; labor market conditions, specifically, job offers and wage levels; and what are termed “marriage market conditions,” or the frequency and quality of marriage offers.[1] By adjusting the various components and running several robustness checks, they are confident that the model describes the college graduation patterns observed…and more labor, education, and family-related outcomes besides. They conclude that changes in parental background factors account for one quarter of the observed growth in women’s college graduation from the ’60 to ’80 cohort. Marriage market factors account for another one-fifth and labor market factors explain the rest.
The researchers then use their model to project outcomes forward for the student cohorts born in 1990, 2000, and 2010. They predict that White women’s graduation rates will plateau, while that of Hispanic and Black women will grow rapidly. But the aggregate graduation rate for women will grow very slowly due to the combination of these trends. For just a few specifics: Since the college graduation rate of White mothers increased from 38 percent in the 1990 cohort to 51 percent in the 2010 cohort, the graduation rates of their daughters are predicted to hit a peak and flatten out (at around 53 percent) because the remainder of the cohort consists of women whose mothers did not graduate. The college graduation rate of White men (married or not) is predicted to grow slightly faster than for White women, narrowing but not closing the gap between them. For Black women, the college graduation rate of mothers increased from 13 percent in the 1980 cohort to 19 percent in the 1990 cohort and 28 percent in 2010. This will result in a predicted boost of the graduation rate of their daughters from 24 percent (1980 cohort) to 28 percent (1990) to 33 percent (2010). For Black men, the graduation rate is predicted to rise modestly over time, increasing the gap between them and Black women. The pattern for Hispanic individuals—both men and women—is similar to that of Black individuals, for the same reasons.
We must consider that these projections are models only and are built on trend data from decades earlier. More on this in a moment. But it is interesting that, in their conclusion, the authors discuss some more prosaic reasons underlying both the observed and predicted gender gap outcomes. For example: Women’s rate of return on the investment required to earn a college degree has increased faster than men’s in recent years—in both the labor and marriage markets (corny and misogynistic as it may sound, they’re talking about meeting and marrying college-educated men as a measure of marriage prospect “quality”). And while the labor market returns for men are still higher, women’s returns have increased quickly as more and more enter and remain in the workforce.[2]
And in the end, that is the moral of the story: Despite being able to show mathematically—and empirically—that the college graduation rate’s gender gap has strongly reversed in favor of women in the last fifty years across the racial and ethnic groups studied, men still fare better in the labor market due to a combination of longstanding societal and structural factors. At the same time, modern-day and near-future externalities will likely shake the foundation of both the statistical model and the real world outcomes. To name just a few items: There is evidence that the long-rising “college wage premium” may have plateaued overall (and even begun shrinking) for graduates versus their non-degreed peers, there are fast-moving economic changes today that might impact college degree holders more strongly than others, and the marriage and fertility rate declines are real and likely accelerating. It all combines to feel like a perfect storm of headwinds threatening to cause women—along with all non-White college graduates—to slip further back down the economic ladder in the future despite their degrees.
SOURCE: Zvi Eckstein, Michael P. Keane, Osnat Lifshitz, “What Explains Growing Gender and Racial Education Gaps?” NBER Working Papers (June 2025).
[1] There’s a lengthy Appendix describing how all of these factors are defined and quantified; they are all too complex to be summarized here.
[2] A recent Pew Research report finds that, in 2023, 87 percent of college-educated women were in the labor force, the highest estimate on record.
