Consider two young men. They’re similar in every way. One chooses an elite college, one doesn’t. Research has shown both end up earning the same later in life.
But in the case of two women, the one at the elite college ends up on average earning 13.9 percent more two decades later.
Why did choosing more-selective colleges change the trajectories of women but not men?
The short answer is marriage, according to a working paper by economists Suqin Ge of Virginia Tech, Elliott Isaac of Tulane University and Amalia Miller of the University of Virginia that built on an earlier analysis of women entering college in 1976. The new paper was circulated recently by the National Bureau of Economic Research.
When accounting for women who left the workforce after college, the new findings show that women who went to more-selective colleges earned significantly more — almost entirely because they became more likely to work and less likely to marry.
The experience of more-selective colleges appears to change women’s ambitions and expectations. Two women — both daughters of working mothers with similar qualifications — could make very different career decisions depending on where they went to college. The one who attended a more prestigious university would be more likely to pursue a career.
“They’re earning more just because they’re working more. If we limit the sample to people who are working, we don’t find an effect,” Miller said.
“Married women in our sample have substantially lower earnings and lower labor-force participation than other women,” she added.
A woman graduating from a more selective school was 3.9 percentage points less likely to end up married two decades later than a woman who was accepted at similar schools but chose to attend one where average SAT scores were 100 points lower.
That difference in marital status has a huge impact on lifetime earnings potential. The married women were about 18.6 percentage points less likely to work outside the home overall.
In addition, attending a more selective school boosts women’s odds of going on to an advanced degree, by 4.8 percentage points, which also increases earnings potential.
The results hold true even for women whose mothers were working, which tends to increase the likelihood they will pursue college and careers.
Ge, Isaac and Miller are revisiting an influential 2002 paper by Stacy Dale, now of Mathematica Policy Research, and Princeton University economist Alan Krueger. Dale and Krueger analyzed detailed testing, application and acceptance data for thousands of students who entered 30 colleges and universities in 1976, and followed up with them two decades later.
While students who attended more-selective universities earned more overall, Dale and Krueger found the effect evaporated when comparing students at elite colleges with students who were accepted at the same or similar institutions but chose not to attend.
The exception? Low-income students, whose earnings rose if they attended more-selective schools.
Dale’s work has been cited more than a thousand times by researchers and the media and read by countless aspiring economists. One of them was U-Va.’s Miller, who was then working toward a Ph.D. in economics at Stanford University.
She recalled being excited by the results and approach and said her first thought was that someone ought to investigate the marriage-market impact of attending an elite school.
“I married someone I met in college,” she said. “We were dating at the time. So maybe that was part of it.”
Many recent studies have gone beyond paychecks to measure other benefits of higher education, including health, status and social networks. This work might explain why top schools continue to attract record numbers of applicants, despite studies showing their surprisingly low impact on earnings. The marriage market could be one reason.
“For a long time parents wanted to send their kids to a very good school in hopes that they would find their spouse there,” Dale said.
UCLA economist Paola Giuliano, who has published several works on labor-force and family status, noted several recent studies had found that “there is some cost in the marriage market for women who do well.”
University of Chicago Booth School of Business economist Marianne Bertrand found that couples were less likely to match and marry if the woman’s income exceeds the man’s — and that they were more prone to break up. “Couples where the wife earns more than the husband are less satisfied with their marriage and are more likely to divorce,” Bertrand wrote in the Quarterly Journal of Economics in 2015 with Booth colleague Emir Kamenica and the National University of Singapore’s Jessica Pan.
Dale and Krueger’s original work focused only on respondents with full-time jobs. Because married women are less likely to work, they were disproportionately excluded from the study.
When Miller and her collaborators included all graduates in their analysis sample, they found the marriage-market effect she had wondered about years before. Women who attended a college with higher SAT scores were eight percentage points more likely to end up with a spouse with an advanced degree. There was no similar effect for men.
University of Colorado at Boulder economist Murat Iyigun’s 2009 American Economic Review analysis showed education has a much larger impact on women’s future earnings and marriage-market outcomes than it does for men. His work, conducted with Pierre-André Chiappori of Columbia University and Yoram Weiss of Tel Aviv University, didn’t address a school’s selectivity, but he said that otherwise Ge, Isaac and Miller’s findings were in line with their model.
Miller said the earlier study was in some ways a product of its era. The 1970s, when the women graduated, was “a particularly interesting time in history for college women,” she said. But even today, it’s relevant to consider how family concerns shape women’s lives.
“Even with a college degree and even with an advanced degree, women tend to be the ones who are facing more challenges in terms of having a successful career and family life,” Miller said.
Dale, who did the original study, said Miller and her colleagues may be onto something. But “there’s a lot of choices women make that are probably really hard to model,” she said.