VoxEU Column Labour Markets

Why do highly educated mothers opt out of the labour force?

Highly educated women tend to opt out of the labour force at motherhood. This column explores why some professions (doctors) opt out less than others (MBAs). One crucial finding is that women who worked in a family-friendly environment are 10% more likely to remain working, suggesting a role for improved work-family policies.

Led by cover stories in the New York Times Magazine and Time Magazine (Belkin, 2003, and Wallis, 2004), the popular press has highlighted the propensity of highly educated women to opt out of the labour force at motherhood. Given the limited supply of places at top educational institutions, many of these reports include an implicit and at times explicit contention that these women are squandering society’s investment in their education. Others have countered that these women are being “pushed out” of workplaces that remain structured for men with stay-at-home wives.

The proportion of highly educated women that remains in the labour force varies widely by their field of training, according to the 2000 Census. For example, 91% of doctors continue working, compared to 80% of lawyers and 78% of managers.1 (The same pattern exists among mothers who graduated from Harvard, according to Goldin and Katz 2008, and Swiss and Walker 1993.)

These findings raise the question of whether the medical profession has something to teach other high-education professions about how to structure the workplace to allow women to maintain career momentum after having children.

Identifying differences in motherhood across professions

The problem is that there may be a selection bias: a woman who chooses to be a doctor may be very different from a woman who chooses to be a manager, and these differences may be systematically related to factors that influence the labour supply decision at motherhood. For instance, a married woman’s decision will depend in part on her husband’s earnings, and since many women meet their spouses in graduate school, they are likely to work in the same field and have spouses with similar earnings. We would also expect women to vary in terms of underlying (and unobservable) factors, such as the value they receive from their professional identity and their taste for spending time at home with their kids. We might not expect this taste for time at home to vary systematically with a woman’s taste for being a doctor versus a lawyer. Yet because of the much greater cost of medical training – which can only be recouped by years spent working rather than at home – we would expect a systematic pattern in who ultimately chooses to attend medical school versus law school.

The question becomes how to disentangle these two elements – do the differences in labour force participation across high-powered careers reflect the types of women who choose these jobs or the workplace environment these women experience? If it is the former, the patterns observed in the Census data may reflect systematic differences in women’s ex ante propensity to opt out at motherhood; if the latter, the data may instead reflect differing work environments that push women in some professions out the door at motherhood.

To try to disentangle these effects, we study the labour supply choices of approximately 1000 mothers who graduated from Harvard College between 1988 and 1991 (Herr and Wolfram, 2009). A benefit of focusing on this population is the enormous richness of information available for these women, including information on each woman’s college and high school experience, her post-graduate education, her occupation and firm if working, marital information including her spouse’s education and occupation, and her family structure. Given this rich professional data, we also had a career consultant provide salary estimates for each woman and her spouse based on reported education, occupation, firm, and geographic location.

Doctors, businesswomen, and children

Looking at these women 15 years after graduation, when they are approximately 37 years old, we find the same pattern seen elsewhere; among mothers with a graduate degree, MDs are the most likely to be working (94%), and MBAs are the least likely (72%).2 Among PhDs, 86% are working, among lawyers (JDs), 79%, among women with non-MBA masters, 74%, and among those with no graduate degree, 69%. We then consider whether these differences can be explained by variation in women’s observable characteristics, such as the number and age of her children, her husband’s salary, and her own potential earnings. Although we find that these factors enter into a woman’s propensity to work in the ways we would expect, they cannot explain the pattern across graduate degrees.

Similarly, when we control for factors that might proxy for a woman’s underlying taste for time at home with her children (e.g., her age at first birth), or the value she places on her professional identity (e.g., whether her graduate degree lines up closely with her undergraduate major), we again find that the level differences across graduate degrees remain large and strongly statistically significant. Even after controlling for a very rich set of factors, the pattern remains unchanged – MDs are still much more likely to work than any other group, MBAs are still less likely, with PhDs and JDs in between.

Family-friendly jobs

This suggests that some other factor, potentially the “family friendliness” of a given field, instead drives this pattern. To examine this hypothesis, we look at women’s work environments both 10 and 15 years after graduation, focusing on those women who had their first child during these five years. Ideally we would like to capture several dimensions of a woman’s job, including both variation in work-family policies (e.g., the ability to work part-time without facing a career penalty), and the structure of the job itself (e.g., the flexibility with which one person can be substituted for another). Although we cannot directly observe these characteristics, we build an approximate measure of the family-friendliness of a woman’s work environment based on where she works (the non-profit or government sectors, or a firm labelled ‘family friendly’ by Working Mother Magazine or the like), or for MDs, based on the average hours in a woman’s medical specialty.

As a start, we consider whether controlling for a woman’s work-environment 10 years after college can help predict if she will be working 5 years later, after the birth of her first child. We find that it does, even after controlling for the rich data discussed above, and even though we must rely on this very noisy measure of family-friendliness. Women who worked in a family-friendly environment are 10% more likely to remain working.3

Alternatively, the key may be the opportunity to switch to a job that allows a woman to combine work with family after she becomes a mother. Looking at the women who were working in a ‘normal’ (non-family-friendly) environment before kids, we find an interesting pattern – whereas JDs are equally likely to shift to a family-friendly job as to leave the labour force, MBAs are instead twice as likely to quit. These results suggest that women who select a JD face a larger set of family-friendly job alternatives, or that, for MBAs, the characteristics of the family-friendly options are especially unattractive.4


Overall, these results suggest that systematic variation in policies and professional norms across high-education careers may be an important factor in explaining female labour force participation rates. In other words, a woman’s work environment can play a role in pushing her out of the labour force at motherhood. In particular, women with MDs, and to a lesser extent JDs and PhDs, may face more appealing work choices at motherhood than MBAs. Put more strongly, our results suggest that improved work-family policies or changes to social norms could drive labour force participation rates of highly educated women closer to parity with men.


1 2000 Census 5% sample for women, ages 32 to 40, with at least one child in the house age 5 or under.

2 Because we only observe these women through their late 30s, we cannot consider re-entry patterns. It may be that the high exit rate among MBAs reflects a lower cost of time off, although other sources suggest that this is not the case (e.g., Swiss and Walker, 1993, Hewlett et al. 2005, and Bertrand, Goldin, and Katz, 2009).

3 One worry is that our measure of family-friendliness may indirectly pick up an element of the identity value of the job. For instance, women who select a nonprofit over a corporate career may do so in part because of the identity value it provides. Although we find a positive correlation between our measure and factors that may speak to the identity value provided by work, the relationship is fairly weak and likely does not drive this strong result.

4 Evidence from the sociology literature (e.g., Blair-Loy and Wharton, 2002, and Eaton, 2003), as well as surveys of women working in these careers (e.g., Hewlett et al., 2005), show that lawyers report equal or greater dissatisfaction with their careers as women in business, suggesting a relatively low value placed on their professional identity. And as reported by Bertrand et al. (2008), they face smaller penalties for time off than MBAs. These studies also show that lawyers work in a more family-friendly environment, with lower barriers to the use of flexible work schedules. This suggests that JDs’ higher propensity than MBAs to stay in the workforce arises from the difference in availability of family-friendly alternatives, rather than from variation in the satisfaction provided by their careers.


Belkin, Lisa. "The Opt-Out Revolution", New York Times, October 26, 2003.

Blair-Loy, Mary and Amy S. Wharton, "Employee's Use of Work-Family Policies and the Workplace Social Context”, Social Forces, 80(3): 813-845, 2002.

Bertrand, Marianne, Claudia Goldin, and Lawrence F. Katz, "Dynamics of the Gender Gap for Young Professionals in the Financial and Corporate Sectors", NBER Working Paper 14681, January, 2009.

Eaton, Susan C., "If You Can Use Them: Flexibility Policies, Organziational Commitment, and Perceived Performance", Industrial Relations, 42(2): 145-167, 2003.

Goldin, Claudia and Lawrence F. Katz, "Transitions: Career and Family Life Cycles of the Educational Elite", American Economic Review: Papers & Proceedings, 98(2): 363-369, 2006.

Herr, Jane Leber and Catherine Wolfram, "'Opt-Out' Rates at Motherhood Across High-Education Career Paths: Selection Versus Work Environment", NBER Working Paper 14717, February, 2009.

Hewlett, Sylvia Ann, Carolyn Buck Luce, Peggy Shiller, and Sandra Southwell, "The Hidden Brain Drain: Off-Ramps and On-Ramps in Women’s Careers." Harvard Business Review Research Report, March, 2005.

Swiss, Deborah J., and Judith P. Walker, Women and the work/family dilemma: how today's professional women are finding solutions, New York: J. Wiley, 2003.

Wallis, Claudia, "The Case for Staying Home", Time Magazine, March 22, 2004.

1,575 Reads