While a growing body of evidence documents that gender differences in major choice contribute to the gender earnings gap among college graduates, less is known about what drives these decisions. In Chile, 82% of the $4,758-per-year earnings gap between male and female graduates is explained by differences in representation across fields. Why do men and women sort into such different fields? In this paper, the authors examine the drivers of these choices and what they imply for policies aimed at closing the gender earnings gap.
The key to the authors’ analysis is administrative data from Chile’s centralized college admission system, in which applying students submit a ranked list of their preferred majors. This allows the authors to compare later-life outcomes among students who preferred the same majors and had similar qualifications but, due to the admission system’s assignment process, graduated from different fields. The authors combine these admissions data with records on graduation, earnings, and fertility to estimate the causal effect of each field of study, on earnings and fertility, measured thirteen years after college application, separately for men and women. They find the following:
- Fields of study differ sharply in their economic payoffs. All fields except Humanities yield positive earnings returns, ranging from 17 to 80 log points, with Medicine, Business, and Engineering delivering the largest returns—between 55 and 80 log points. Within a given major, however, men and women experience similar economic payoffs.
- Fields also differ in their fertility consequences. College graduation reduces the likelihood of having a child by one’s early thirties—by an average of 6.6 percentage points for women and 13 percentage points for men—but the impact varies widely by major: the reduction is smallest among graduates in Teaching and Health and largest among those in Law and Humanities. Fertility is higher in fields that steer graduates toward careers with greater job security, more progressive gender norms, and smaller child penalties.
- Gender differences in major choice are largely explained by how students weigh these payoffs. Earnings returns predict demand for fields among both men and women, but for women, monetary considerations are less predictive and fertility impacts more so—suggesting that women place greater weight on balancing career and family when choosing a major.
- These differing priorities are a major driver of the earnings gap. In a counterfactual exercise, if men and women placed the same weight on fertility payoffs, the gender earnings gap among college graduates would narrow by roughly 30%.
Recognizing that shifting preferences would require changing long-standing norms, the authors then take students’ preferences as given and ask what admission policies could do in the short run. Leveraging the structure of the centralized assignment system and their causal payoff estimates, they simulate how expanding Engineering seats by 10%, 30%, or 50%, or reserving those shares of Engineering seats for women, would reshape representation, the gender earnings gap, total earnings, and fertility. They find the following:
- Broadly expanding seats in high-return fields widens the gap. Because men demand Engineering seats more strongly, they secure most of the newly available places: the representation gap grows by 7–20% relative to baseline, and the overall gender earnings gap increases.
- Reserving seats for women narrows it, at minimal cost. Gender quotas considerably reduce the representation gap and reduce the gender earnings gap by 3–12%, depending on the quota’s size—with less than a 1% change in aggregate earnings and less than a 0.5% decline in the share of applicants having children.
The findings carry two lessons for policymakers and educators. First, major choice is not only a matter of academic interest but also a decision about the future balance of career and family—and admission rules can move equity outcomes quickly when preferences are slow to change. Second, over the longer run, alleviating the career-family trade-off itself—through workplace flexibility or more egalitarian norms—may offer a path to convergence that does not rely on quotas.