How much would you earn if you were the opposite gender?
The short answer is: It’s complicated. Policymakers should pay heed to the gender gap’s complexity and the interplay of factors that impact how much men and women are paid before deciding how best to proceed.
Measuring the gender pay gap is difficult — economists have been studying the gap for decades — precisely because women and men are not different only because of their gender. Women and men also have different work histories on average and different career aspirations in the long run.
These differences become exaggerated around the time of childbirth, but they are present very early in a professional’s career as well: in the choice of what to study, which jobs to apply for and which job to take.
Perhaps all of these differences between men and women reflect their response to discrimination or to social norms. But men and women also differ biologically. It would be surprising if none of this hardwiring manifest itself in differences in preferences concerning work.
Making the task of sorting out the gender pay gap even more complicated, it’s not as if preferences, norms and discrimination are independent of one another.
Facing such a complex web of interactions and factors, it can be difficult to understand the impact of policy designed to shrink the gender pay gap, and such policy can easily backfire.
The backbone of the equal pay movement — equal pay for equal work — assumes that men and women are equally productive in the same job. But the evidence does not support this claim.
Using Danish data, I studied the relative output of firms employing more men compared to firms employing more women and found that productivity differences explain about two-thirds of the gender pay gap. On average, women are 8-percent less productive than men, so we would expect them to be paid 8-percent less.
Why study firm output, rather than look at pay? Precisely because men and women differ on many dimensions — they have different occupations, different job titles, etc. Whatever difficult-to-observe differences there are between men and women, these differences affect both how much they are paid and also their productivity.
A large gap between pay and productivity would suggest that women are being paid less for the same work. But I do not find a large gap between the pay of women and their productivity.
Put simply, firms don’t seem to be making tons of money by hiring women to do as much work as men but then underpaying them. This means that if equal pay legislation really forced firms to change the pay of women, we may expect many firms to stop hiring women or to change the way they classify the jobs women do.
For example, maternity leave policy, though great in many ways, may negatively impact the way all women are paid. If leave is costly to employers — and my research studying parental leave expansions in Denmark finds some suggestive evidence that this is the case — then implementing mandatory paid leave can have negative consequences on the pay of all women.
This issue is completely understudied in the case of U.S. policy but is crucial to understanding the impact of paid leave reform. In addition to leave-taking, there are many (more important) differences between mothers and other employees.
I find that motherhood alone drives the gender productivity gap: Mothers are substantially less productive than men, and their lower productivity completely explains their pay gap.
Where does this lower productivity come from? I could tell you a few personal stories (and had to keep myself from sharing these with potential employers when they asked the same question), but they all come back to hours worked.
Full-time, non-hourly wage workers are often lumped together as working 40 hours a week, but there may be large differences between men and women in reality. In U.S. time-use data, full-time mothers work about five hours less per week than full-time fathers.
Research shows that even high-earning women spend substantially more time than men on household activities, and it is not surprising that this takes away from time at work. For mothers with very young babies, sleep deprivation is also a problem.
In contrast to mothers, women without children are as productive as men, but they are also paid less than men. For these women, the gap between pay and productivity is concentrated in their prime childbearing years.
Why are these women underpaid? A natural explanation is that employers pay young, childless women less than young, childless men because they expect these women will soon have children and reduce their hours.
To avoid reducing a female employee’s pay as soon as she has children (which would be an unattractive HR decision for a variety of reasons) firms may pay women less in anticipation of childbirth.
Consistent with this hypothesis, the gap between pay and productivity is larger for women who will have children within the next three years compared to other women.
So, where does this leave us? How we design policy to combat the gender pay gap will look very different if the source of the pay gap is employer discrimination, workers’ own choices or some combination of these.
The fact that pay seems mostly explained by true differences in output between mothers and other workers makes the task of “solving” the gender pay gap more difficult.
Yana Gallen is an assistant professor at the University of Chicago Harris School of Public Policy. She received a Ph.D. in economics from Northwestern University in 2016. Her research focuses on understanding the sources of the gender pay gap — preferences, discrimination and/or productivity.