Poorly–designed policies that may initially appear “family-friendly” can impede progress toward gender equality in two different ways—by making it costly for employers to hire or promote workers suspected of having costly family commitments (e.g. women of childbearing age) or by encouraging workers with such commitments to drop out of paid employment for so long that their prospects of advancement on the job are permanently damaged.
In other words, increases in women’s job tenure and/or reductions in work-family stress may be purchased at the expense of women’s earnings and career advancement.
Awareness of these possible tradeoffs—and efforts to eliminate them—have informed most recent efforts to reform family policy in the U.S. That’s why it was so disappointing to read Claire Cain Miller’s recent contribution to the New York Times Upshot feature emphasizing the “backfire” problem and claiming that “researchers say” there is no simple way to avoid it.
In fact, there are a number of simple ways, which, ironically, Miller herself mentioned almost immediately. Many of the over 600 comments posted by readers online reiterated the most obvious remedies: fund paid family leave and childcare through social insurance rather than requiring employers to pay; create gender-neutral policies aimed to help employees with care responsibilities, broadly defined (rather than focusing on maternal commitments); create and enforce laws against family responsibility discrimination.
Many of the family policy researchers I know were dismayed by the title and tone of Miller’s piece. Janet Gornick at the City University of New York’s Graduate Center circulated a call to respond.
My colleagues at the University of Massachusetts Amherst were quick off the mark:
Joya Misra, Michele Budig, and Irene Boeckmann from Sociology immediately wrote up a short summary of their detailed cross-national research on family policies showing how careful policy design, as well as changing cultural norms, contribute to positive outcomes for women’s employment and earnings. Marta Murray-Close from Economics pointed out that the relative size of effects matters: the benefits of a large long-term increase in women’s employment may well outweigh any short-term reductions in promotions for a handful of women at the upper reaches of corporate management.
Silke Staab, a research specialist at UN Women, offered some perspective on Miller’s discussion of Chilean policies, which led off her piece. Yet, requiring Chilean employers to provide child care for women have made it expensive for them to hire women. But that was partly the intent of the law, which was originally passed in 1917 when relatively few mothers were employed. Staab observes that today, in Chile, almost everyone agrees this policy is a really bad idea. (For more discussion see the new UN Women report on Progress of the World’s Women
I focused my own attention on an unpublished paper by economist Mallika Thomas, which Miller summarized, purporting to show that passage of the Family and Medical Leave Act (FMLA) of 1993 in the U.S. increased women’s employment but reduced their likelihood of promotion by an even larger percentage.
I found this claim startling, because the FMLA covers less than one half of U.S. workers, is not restricted to maternity leave, and, because it is unpaid, is not taken up by many of the workers who are covered. According to a Department of Labor report published in 2012, more than half of all leaves are taken for employees’ own illness. Yet Thomas describes her paper as a study of “mandated maternity benefits.”
Ploughing through an on-line version I found a really interesting—and, at first glance, sophisticated–model showing how the FMLA may have changed employers’ perceptions of women employees, making it harder for them to identify those with the strongest career commitments. This information problem, Thomas argues, led to more uncertain estimates of future labor supply, which in turn, made it riskier to promote any women at all.
She has a point here: employers have an incentive to discriminate against women in promotions if their future effort on the job is harder to predict than that of men. But Thomas doesn’t seem to include any explicit consideration of the effects of the FMLA on turnover, and this seems like a pretty serious flaw.
The FMLA, like many of the corporate policies explicitly adopted by many firms for their professional and managerial employees, reduces job turnover by making it more likely that mothers will return to the job after their leave is over, rather than quitting to stay home with their baby for a few months and then seeking another job.
In this sense it reduces uncertainty. Indeed, the loss to a firm when an employee leaves the firm is almost certainly greater than the loss that might occur because the same employee takes the maximum twelve-week FMLA leave even if she also lowers her future work intensity somewhat as a result of ongoing family responsibilities.
At the very least, turnover effects belong in any theoretical model of the effects of the FMLA. Reduced turnover is central to arguments that both unpaid and paid family leave can create benefits for both employers and employees (See, for instance, this recent Wall Street Journal article).
The empirical results that Thomas reports could result from the fact that the policy may increase heterogeneity among women employees, because more women stay on the job rather than quitting to take time out for family care. The average probability of promotion may be lower because the promotion pool now includes more women who, under the previous policy regime, would simply have dropped out.
I hope Thomas will respond to this point in her future work. I also hope that the New York Times will improve its reporting on these issues by consulting a broader spectrum of policy experts. David Leonhardt, managing editor of the Upshot, suggested back in 2010 that feminists had been lax on family policy. I challenged that view right away on the Economix blog.
Oops, maybe my efforts backfired. But somehow, I don’t think that’s the problem.