Creative Destruction

April 18, 2006

Inside Higher Ed On The Gender Pay Gap

Filed under: Economics,Feminist Issues — Ampersand @ 1:08 pm

Here’s a recent article from Inside Higher Education about a new study examining the wage gap between female and male professors. The study itself sounds useful, but what interested me is all the dubious assumptions about the wage gap embedded in the article (and perhaps in the study itself).

Explaining the Gender Gap in Pay

Why do female professors earn less than male professors? Some charge that gender bias is at play, while others insist that once factors such as experience are accounted for, the gaps aren’t consequential.

There may be truth to both views, according to research findings presented this week at the annual meeting of the American Educational Research Association by Paul D. Umbach, an assistant professor of higher education at the University of Iowa.

An example of how the media misrepresents stories in order to seem “objective.” It’s not true that the study found “truth to both views.” The controversy is between those who say “human capital factors account for part of, but not the entire, pay gap” versus those who say “human capital factors account for all of the pay gap.” This study found that about two-thirds of the pay gap could be attributed to human capital factors, but almost a third could not be.

Far from finding “truth to both views,” as the article reported, this study supports the feminist view and refutes the “human capital accounts for everything” view. But saying that would have compromised the faux-objectivity news writers specialize in.

Umbach used a series of databases to calculate the gender gap in pay over all, and then to account for all kinds of factors other than gender bias that may contribute to the salary gap. In the end, he found that looking at those factors decreases the size of the gap, but that it remains meaningful.

Leaving all factors out, the mean salary for women in the professoriate was 21.8 percent less than that for men. Add all the possible explanations and their impact, and the gap shrinks to 6.8 percent.

Before anyone says “6.8%” isn’t much, imagine coming into work tomorrow and being told that they’ve decided to give you a 7% pay cut. And remember, that’s an average pay gap. But in practice, the pay gap tends to get larger over the course of a career (see the discussion of “cumulative causation” in this post); so what starts out as a small and relatively managable pay gap can grow very large by the end of a career.

For example, the mean differential favoring men was $12,649 in English literature, $24,845 in chemical engineering, and $23,294 in economics. But these comparisons included men and women at all stages in their careers — so the senior faculty members with higher salaries (and who are more likely to be men) tilt the sample significantly.

What’s not being counted here? Benefits. This arguably means that this study will underestimate any pay gap, because more seniority, and higher rank, is commonly linked with higher-value benefits.

So then Umbach ran a series of analyses designed to compensate for that and other factors. Years of seniority were factored in, as were books and articles written, career patents, whether the person was receiving outside support for research, professorial rank, and the general job market in the discipline (based on percentage of new Ph.D.’s who are employed), among other factors. When all of those factors were added, the gap still remained, at 6.8 percent.

There are not clear explanations for the gap, leaving open the possibility that bias is at play, Umbach said.

It’s true that bias is a possible explanation for part or all of the unexplained 6.8%. What bothers me is the implicit, unjustified assumption that the “explained” factors can’t themselves reflect bias. But if job discrimination against women exists in academia, is there any reason to assume that sexism has nothing at all to do with factors like who gets grants for outside support, and whose articles are published?

For instance, they list “rank” as one of the factors that explains pay. But if bias exists, one likely way for gender bias to be expressed is that men might be more easily promoted to full professor positions. By implicitly assuming that “rank” and other human capital factors are discrimination-free zones, this study’s design may overlook significant forms of gender bias.

Another example is the assumption that women get paid less because women spend less time working and accrue less experience. This is no doubt true, but causation also goes in the other direction: women work less because they get less reward for working. (This is called a “feedback effect.”) To some degree, then, women’s lesser experience is not only a cause but also a result of gender bias.

But he said that other parts of his study suggest that the bias may not be a simple preference for men, but may relate to biases based on disciplines and on how faculty members spend their time.

For instance, Umbach found that as the proportion of females in a discipline increases, the mean salaries drop — for men and women.

This is something feminists have long argued, and that many other studies support. Gender wage discrimination is not just (or even primarily) a matter of women being directly discriminated against, but instead a matter of work done primarily by women being undervalued. In this way, even men who work in underpaid female-dominated occupations could be said to be hurt by the gender wage gap.

Another factor that negatively correlates with salaries is the percentage of time spent teaching: The greater a discipline’s time spent on teaching, the lower its salaries — for men and women. The more outside research funding, the higher the salaries.

In one respect, Umbach said, those findings don’t suggest bias because male and female faculty members in the discipline are affected equally. But when these figures are coupled with other studies suggesting, for example, that female professors may spend more time on teaching, questions are raised about underlying bias.

“We know that women tend to be employed in disciplines with a lot of other women, in disciplines without as much funded research, in disciplines with more time teaching,” he said. “Is the reward structure more male? Are we creating structures that reward men?”

I’d say that worries about “structures that reward men” are legitmate, but have to be extended beyond what this article discusses. One major reason for women’s on average lower wages is that women who are mothers tend to spend less time in the workforce (both in terms of years in the workforce, and in terms of how many hours worked per year) while they take care of their children. As I wrote in an earlier post, many feminists believe that in a non-sexist society, fathers and mothers would share equally in childcare – or at least, that fathers would take on a larger degree of childcare than they do now. Therefore, any “parenting wage penalty” in a nonsexist society would be split more evenly among men and women. The fact that women are virtually the only ones hit by the parenting wage penalty doesn’t prove that sexism no longer exists; on the contrary, it shows that sexism still matters, and has a big negative impact on women’s wages. (It also has a negative impact on men’s contact with their families.)

But to take it a step further, arguably that there’s a “parenting wage penalty” at all is a sign of sexism. Why isn’t the workplace designed to accommodate parenthood? The American job market was designed for men – in particular, it was developed in a society in which workers were had a wife at home to take care of the kids. Society has changed, but our jobs haven’t, and that works to the disadvantage of all working mothers (and to mothers who would like to work, but can’t find a job that will give them the flexibility they need to combine work and motherhood). Isn’t it sexist to expect mothers to fit into a work system that was designed for a Father Knows Best family?



  1. […] (This post has been cross-posted at Creative Destruction. If you have trouble posting comments here, try the cross-posted version.) […]

    Pingback by Alas, a blog » Blog Archive » Inside Higher Ed on the Gender Pay Gap — April 18, 2006 @ 1:10 pm | Reply

  2. Who “designed” the work system?

    Comment by Robert — April 18, 2006 @ 1:48 pm | Reply

  3. Thousands of managers and executives, each contributing their little piece to the design. Or, metaphorically, society as a whole, acting through those individuals. I don’t see how the metaphorical use of “designed” is problematic here.

    Comment by pdf23ds — April 18, 2006 @ 2:57 pm | Reply

  4. It’s problematic because it leads to false conclusions. Why was the workplace not “designed to accomodate parenthood”? Because “design” isn’t what happened. If some person or small group had sat down to design a workplace system, they would have accommodated parenthood – it would be an obvious need, from a top-down perspective. But nobody sat down to design such a system; instead, millions of individual employers and hundreds of millions of employees mutually negotiated a vastly complex net of individual contracts, within a framework of incentives and regulations established by government. Within that vast system of contract, accommodations for parenting have been subordinated to other values – sometimes with the cooperation and choice of all parties, sometimes not.

    Saying “the design of the workplace” makes it seem as though some small group somewhere needs to change their priorities and change them now, bucko! When in reality, no such simple decision would have any effect.

    Instead, family-friendly workplaces must come from an explicit valuation of those accommodations in the contracting decisions made by millions of people.

    The real parenthood problem comes in the fact that kids are an externality to the workforce. It doesn’t do IBM much good to have parents as workers instead of single people, and the parents are going to incur higher costs for IBM. (“Sorry I missed the meeting, chief, Rebecca was throwing up everywhere and…”) Single people have their own issues but its parents who flake out the most, big time, in my own experience as an employer (and as a parent). Parents, male and female, probably need to be willing to recognize their reduced value to the employer, and accept a negative wage premium in exchange for the employer’s concessions to flexibility and flakiness.

    Comment by Robert — April 18, 2006 @ 3:07 pm | Reply

  5. I should note that I basically agree with Amp about the existence of a pay gap, which is largely predicated on differential treatment of parents, where parent generally equals mother.

    I disagree about whether the locus for changing this is in familial arrangements, or in the workplace. I don’t believe that parents should get any net benefit from their status in the workplace, and that probably means that there will be pay gaps between parents and non-parents; if this disproportionately affects women because women are disproportionately the primary childcare providers, I don’t see that as intrinsically problematic.

    It WOULD be problematic in scenarios where the mother is NOT the primary childcare provider, the father is, but the mother still is perceived as such and treated as such in the workplace. That would be gross discrimination and unacceptable.

    Comment by Robert — April 18, 2006 @ 3:50 pm | Reply

  6. Or, perhaps, when women’s caretaking and men’s is perceived differently.

    I have to disagree with the ‘parents flake more’ observation. In my experience, it’s the young singles who flake the most, but they tend to be excused for it because nobody has hangups about 20somethings in the workplace they way they do about mommies in the workplace.

    Kids are an externality in workplaces that focus on short-term goals (can we get somebody for the Tuesday shift?) and have regular turnover. For workplaces that want a stable, skilled, long-term workforce, you can’t beat families. It’s a lot harder to say “Bag this job” when your job feeds other people, not just an IKEA habit.

    Comment by mythago — April 18, 2006 @ 5:01 pm | Reply

  7. A female parent who takes off work to take care of her kids will often be seen as inherently flaky because she’s a parent.

    On the other hand, a male parent who takes time off to care for his kid will often be seen as *more* reliable – because he’s taking his responsibilities seriously.

    Comment by Mandolin — April 18, 2006 @ 6:31 pm | Reply

  8. Young singles as a class may flake more, but that’s selection bias. Among the group of people that are non-flaky by nature, it’s the parenting ones who end up flaking because of real problems. Yeah, the 20somethings didn’t show up for work because of the Phish concert – but I didn’t hire them anyway. I hired their more-responsible younger sister, and she shows up every day. (Or would show up, if I wasn’t running a virtual company. I don’t know where the hell she herself was; the work got delivered, so I’m happy.)

    Comment by Robert — April 18, 2006 @ 6:54 pm | Reply

  9. Young singles as a class may flake more, but that’s selection bias.

    Your whole argument is based on anecdote, so I don’t know why you’re throwing around ‘selection bias’. Non-inherently-flaky people without parents also have cars break down, parents get sick, pets develop sudden illnesses, and so on.

    Comment by mythago — April 18, 2006 @ 9:44 pm | Reply

  10. But all of those things are true for people with kids, too. I don’t get an increase in car reliability just because my wife popped out a baby; I do get a predictable uptick in mandatory doctor visits.

    You’re right that my viewpoint is purely anecdotal (I could claim its empirical, and thus vastly superior to mere theories!) but what I mean by selection bias is that in my hiring decision, I’ve deliberately weeded out people who are intrinsically or basically flaky. Among the survivors, everyone is basically responsible, but I notice that people with additional life burdens like kids are more likely to have those burdens interfere with their economic productivity. I’m only dealing with the responsible subset of workers; among THAT set, I’ve found more problems with parents.

    Hope that makes things clearer.

    Comment by bobhayes — April 18, 2006 @ 11:18 pm | Reply

  11. I do get a predictable uptick in mandatory doctor visits.

    Rather, your wife does, and that’s what your employer and probably your colleagues expect of both you and her. That’s what we were talking about originally, yes?

    You may well get an uptick in responsibility–now you have another little mouth to feed and cover with health insurance.

    The question is whether parents, and specifically mothers, overall, are such a drag on productivity that it makes sense for employers to heavily disfavor them as employees. I don’t think that’s been shown.

    Comment by mythago — April 19, 2006 @ 11:23 am | Reply

  12. Rather, your wife does…

    No, I’m the primary caregiver. (Sexist!)

    The question is whether parents, and specifically mothers, overall, are such a drag on productivity that it makes sense for employers to heavily disfavor them as employees. I don’t think that’s been shown.

    Heavily? Probably no. Somewhat? I think indubitably. I’ve just come off a three week cycle of being repeatedly short on a production quota because of a parent whose kids have had a run of bad health. If we hadn’t built up a track record of awesome previous performance, we would probably have lost the client – who contributes about 25% of our net.

    Comment by Robert — April 19, 2006 @ 11:51 am | Reply

  13. No, I’m the primary caregiver.

    As is my husband, but I think we can agree that we’re both weirdos and an employer’s first assumption is not going to be that you or Mr. Mythago are the ones rushing the kids to the pediatrician.

    Somewhat? I think indubitably.

    You mean, anecdotally.

    Comment by mythago — April 19, 2006 @ 2:57 pm | Reply

  14. All I have is anecdotally. I can’t use general statistics (which I don’t have, or even know if they exist) to make predictions about the small group of people I actually work with.

    Comment by Robert — April 19, 2006 @ 3:16 pm | Reply

  15. One would expect the pay gap to decrease in fields in which the employees aren’t tied as closely to deadlines or strict work schedules. Lawyering would be one of those. According to the anecdotes I’ve heard, male lawyers often have families, where female lawyers are much more often single. Perhaps the long hours expected in the profession interfere with drawing conclusions, though.

    Comment by pdf23ds — April 19, 2006 @ 4:40 pm | Reply

  16. “‘For instance, Umbach found that as the proportion of females in a discipline increases, the mean salaries drop — for men and women.’

    This is something feminists have long argued, and that many other studies support. Gender wage discrimination is not just (or even primarily) a matter of women being directly discriminated against, but instead a matter of work done primarily by women being undervalued. In this way, even men who work in underpaid female-dominated occupations could be said to be hurt by the gender wage gap.”

    I’m perfectly willing to believe that there might be a gender based pay gap. But this particular portion of the argument doesn’t require gender bias. It is perfectly normal when there is more competition for a job that either the employer becomes more choosy (qualification requirements go up) or the pay goes down. When the supply of something goes up (in this case number of people willing to do a particular job is augmented by women coming into the field) the price for it goes down. This is also one of the reasons why “women’s jobs” were often lower paying. When nurse, teacher and secretary were the only major professions available to women there was a larger pool of women competing for those jobs–allowing much lower prices than you would expect in a market where men and women competed for all jobs equally. Now I think there was a prestige (or lack thereof) issue with these jobs which is partially gendered, but the simple economics of the situation contributed greatly to the pay difference.

    Comment by Sebastian Holsclaw — April 19, 2006 @ 6:19 pm | Reply

  17. You’d also expect wage depression resulting from the “uota effect.” If a rational employer isn’t free to make rational economic decisions then they may compensate for having to maintain gender and racial balances in order to avoid threat of lawsuit by undervaluing employees by gender and race. You can see this play out in this study, The Impact of Gender on the Review of the
    Curricula Vitae of Job Applicants and Tenure
    Candidates: A National Empirical Study
    where those who are doing the hiring do not disciminate when they can base their decisions on objective data:

    In contrast, when men and women examined examined the highly competitive curriculum vitae of the real life scientist who had gotten early tenure, they were equally likely to tenure the male and female tenure candidates and there was no difference in their ratings of their teaching, research and service experience.

    However, when it came to examining the candidates who were fresh out of grad school, the study noted:

    In one such study, 238 academic psychologists, 118 male and 120 female, evaluated a résumé submitted in application for an assistant professorship that was randomly assigned a male or female name. Both male and female participants gave the male applicant better evaluations for teaching and research and were more likely to hire the male applicant

    The report also noted that there is a greater degree of skepticism regarding the qualifications of the female candidates:

    These cautionary comments include such comments as, “We would have to see her job talk,” “It is impossible to make such a judgement without teaching evaluations,” “I would need to see evidence that she had gotten these grants and publications on her own.” Such cautionary comments on the male tenure candidate’s vitae were quite rare.

    You see the same behavior exhibited towards minorities. The decisionmakers are all aware of the distortions caused by Affirmative Action policies where the less qualified are given a boost and sought out so as to meet management diversity targets. The tokenism distorts rational decisionmaking therefore there is more skepticism when analyzing unknown situations where the information can’t be trusted. It’s simply an expected value calculation where one of the variables has to be modified.

    The problem that I can see is that once the decisionmakers can trust the information generated from within their organization there is likely to be an inertia that has built up over the years regarding salary increments which make a “catch-up” or “equalization” difficult to implement within the organization. You see this happening in other circumstances as well, such as when newly hired candidates who are in demand are given starting salaries near or above those employees who were in the organization for a longer period. This disjoint between the markets – one is a pool of new candidates and the other is a pool of existing candidates – is often arbitraged by the existing candidate leaving the organization and finding a new employer who will recognize their value. The problem for the quota beneficiaries is that if they try this maneuver they present the same “information distortion” to the new employers – did this female candidate achieve her accomplishments as a result of her talent or was she a quota beneficiary? Then the discount is applied once again so that the new employer can make their own assessment of the candidate’s skills.

    Comment by TangoMan — April 19, 2006 @ 8:52 pm | Reply

RSS feed for comments on this post. TrackBack URI

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Create a free website or blog at

%d bloggers like this: