Wednesday, January 25, 2006

The Gender Gap. Part III: Addressing the Wingnuts

You should read the first two parts of this series first, to fully appreciate what I write here. The theory post is here and the empirical evidence post is here.

To begin with, an admission: The gender wage gap has indeed been misused by some feminists and some on the political left, to imply that the total difference in average earnings between men and women is evidence of direct wage discrimination of the type where women are not paid the same for exactly the same work as the one men do. I'm sure that some of the misuse is because this is not an easy topic to understand. Many right-wing and anti-feminist commenters in the field make the exact opposite mistake by arguing that NONE of the observed gap is due to discrimination of any kind.

The facts are in the middle, as I have mentioned in the second part of this series. But discrimination does exist, and the only relevant type of discrimination is not the one where women are paid unequally for the same work. Even this type of discrimination occurs, and not all of it gets fixed by the court system because we tend not to know what others earn at work.

The anti-feminist wingnut framing choices in discussing the gender gap in wages are interesting. The basic emotional trick is to turn the discussion to two issues: That believing in the existence of discrimination makes you into a whiny victim, and that such a focus demeans the great achievements of women: that they have narrowed the gap since 1950. Here are examples of both emotional strategies:

"These dependency divas are once again using misleading statistics to convince women they are victims in order to advance their big government agenda," says Carrie Lukas, director of policy at IWF.

The gender wage gap is a manufactured crisis borne of interest group politics, not a reflection of the reality of women's lives. The arguments advanced by feminist groups on Equal Pay Day are deliberately misleading. Worse, their underlying assumption is that women are incapable of making free and informed decisions about their lives-or for standing up for their right to fair compensation. I for one have more faith in the female sex. I know that we will continue to succeed-but on our own terms.


Are women victims? Or can they hold their own in the workplace? Women's inequality in society is a standard refrain in the popular media and thus in the conventional wisdom. The purveyors of that point of view assume that women are systematically discriminated against on all levels of society, especially in the workplace, and that this discrimination has kept them from reaching equality with men. But the facts show otherwise.

These are all selected from the writings of the wonderfully cheerful gals who belong to the Independent Women's Forum (IWF), a right-wing site that fights feminism. It is funded by Richard Schaife, by the way, which makes the "independent" part of the group's name a little hilarious. These ladies do most of the heavy lifting in anything political that touches women's rights, and thus most of my examples in this post comes from them. They love to use Warren Farrell as an expert! Farrell's opinions on issues concerning women include the argument that the contraceptive pill destroyed men's lives completely as in the past the pregnant women needed men to do stuff and now they don't. Farrell also believes that women lead privileged lives compared to men but he doesn't want to change anything. Mindboggling.

The emotional bag of trickery the IWF uses is not limited to victim-blaming and the exhortation for women to pull themselves up by their brastraps. It also includes a reverse type of victim blaming: Women are blamed for the wage gap because the IWF sees everything as a consequence of women's free choices. Even some women perhaps not in the IWF do this. Here is a happy-fuzzy example:

Women may never achieve parity with men in the workplace, but that is not bad news for women. Some will choose not to work, while others will set their sights to lead the top corporations in America. The majority of women will fall somewhere in between

Replace "women" with "blacks" in that sentence and see how it sounds then.

So much for the emotional approach the wingnuts and their henchwomen take. What they fear is government intervention in the markets, of course, and as most of them view markets almost as highly as they view the Christian god this fear galvanizes them to lie, to distort and to misinform. The common ways to misinform are the following:

1. Omission of any evidence that shows discrimination exists.

2. Emphasis on women's free "choice" as the "real" explanation for the gender gap.

3. Emphasis on "free" markets as incompatible with sex discrimination.

4. Biased interpretations of studies which look at a narrow section of workers.

The most common of these is the first one, omitting any information on labor market discrimination, or even pretending that nondiscriminatory reasons have accounted for ALL of the gender gap in earnings. Here are some examples with Echidne's commentary:

The reality is that, when considering men and women with similar fields of study, educational attainment, and continuous time spent in the workforce, the wage gap disappears. This is true for some women in high-paying "male" fields such as engineering, chemistry, and computer science.

The reality is no such thing. When we standardize for all known nondiscriminatory reasons why women might earn less, on average, we still find an unexplained remaining gender gap in wages, in study after study. To argue otherwise is untrue. There are fields which women's earnings are a much higher proportion of men's earnings than in other fields, true, and there are probably also fields where discrimination is nonexistent. But such cases are exceptions to the general rule. Consider the study I discussed in Part II of this series, one which used a very large sample of women and men from all walks of life, one that was commissioned by the Bush administration, one which standardized for marital status, for the number of children, for the age of the youngest child, for being a part-time worker and for the occupations people choose, presumably at least partly for their flexibility or safety or other desirable characteristics, we still could only account for a little more than one half of the initial difference in average earnings!

Note also that if we standardize for everything the ladies of the IWF want to see standardized, including extremely detailed occupational choice, the evaluations from bosses, (perhaps even the color of the hairbands women use) and so on, we are also going to hold discrimination constant, if it works through what the boss writes about female and male workers or if women are steered into certain jobs by the school system, parents and the firms themselves.

Here is another common message from the IWF:

Feminists have ignored how women's lives and goals differ from men's. In doing so they have overlooked the fact that it women's life choices -- not sex discrimination -- are responsible for the infamous wage gap.

The idea of the wage gap between men and women being the result of free "choices" by women (and men?) is the most common of all right-wing arguments. As I wrote earlier in this series, it is very difficult to measure "choice" in empirical research. But even if it was choice that caused women to focus more on household responsibilities the studies that do control for the variables that might reflect such choice still find a large chunk of the gender gap remaining. To imply that the "choice" has somehow made all further discussion unnecessary is just blatant lying. Or at least reflects some severe confusion of the concepts of "opinion" and "fact".

The anti-feminist or wingnut writings never mention the studies (such as the audit studies I mentioned in Part II of this series) which demonstrate that discrimination still exists. Some quickly glance through the issue by quickly typing "of course discrimination exists but", while some others argue that courts are there to take care of any discrimination that still lingers. Alito might remove that remedy, of course, because the federal laws depend on an interpretation of the Commerce Clause which Alito might not like. Also, remember the difficulty of finding out whether you actually make the same as your coworkers in the same job.

A slightly more sophisticated wingnut strategy is to drag out the idea of markets as a savior for women who might be discriminated against. Here is a good example of this one:

I was -- I hate to admit -- blinded by ideology. The market is a consummately rational institution. The logic of rational self-interest precludes the sort of bias belief in which fuels feminism. If you hire men over more-qualified women, simply because you think more highly of men or want to keep women down, you won't last long in commerce. You'll be put out of business -- the market's distinctive form of punishment -- by those who hire on the basis of ability. A rational firm ignores irrelevant considerations in its decisions, including its hiring decisions.

It's a recovering male feminist who makes the confession here. His reference is to the theory I discussed in Part I of this series, the one about employer discrimination against women. In that model nobody else discriminates against women, not coworkers, not consumers, only the owners or managers of firms. In that world everybody can spot, instantly, who is a productive worker. There are no information problems at all, no prejudice, no using the average characteristics of all women in place of the unknown true characteristics of the individual woman you consider hiring. And existing bigot firms don't have any way of stopping the entry of new brave nondiscriminators who would snap up all those underpaid women.

As I also discussed in Part I, it is even possible that the markets might punish a firm which does not discriminate. It won't matter very much if your wage costs are lowered by hiring all women to be your house painters, if prejudiced customers refuse your bids just because you are employing women. Or, as I mentioned in the first part of the series, think about what might happen to a firm which employs women in a Taliban-type economy.

In the search for links I came across the Wikipedia entry on the gender gap in wages. A wingnut slipped in and wrote the whole thing. It's actually a masterpiece of obfuscation, and you should read it for that reason. Here are some excerpts:

However, legislation has meant that women's wages hold up quite well to men's wages when comparing specific job categories. Among adults working between one and 34 hours a week, women's earnings are 115 percent of men's. Among part-time workers who have never married, and who thus confront fewer outside factors likely to affect earnings, women earn slightly more than men. These statistics suggest that skill level, tenure and working hours are influential factors in how gender determines wages.

"Among adults working between one and 34 hours a week, women's earnings are 115 percent of men's." Do you think that this fascinating finding about part-timers, the vast majority of whom are women, somehow cancels out the gender gap in earnings? Especially as the post gives us no way of knowing if all the nondiscriminatory factors have been controlled for here? For example, what if women who work part-time have more education and experience than men who work part-time? I don't know if this is the case, but the Wikipedia post doesn't tell us anything about this, either. Do you think that this particular snippet is so important that it should be included in such a short Wikipedia post? Perhaps.

Then the conclusions:

Women's work-life patterns and their occupational preferences are significant factors in determining wages. Rather than being "funneled" into low-wage, low-prestige and part-time positions, women often choose these occupations because of the flexibility they offer. After adjusting for these factors, scholars find that the difference between men's and women's earnings is very narrow.

"Rather than being "funneled" into low-wage etc. positions, women often choose these". How does the writer know this? When there is no way of really finding out? And note the nice way of concluding that the residual unexplained wage difference is "very narrow". Say, twenty percent?

The most sophisticated of the wingnut/anti-feminists strategies is to cite those studies which find a very small unexplained wage residual after all the nondiscriminatory factors are added. A favorite economist to use for these purposes is June O'Neill. I googled her today quite extensively, and find her writings have appeared in places such as the Manhattan Institute. This suggests to me that her political alliance is with the right. Nothing wrong with that, of course, but it does make me look for the devices of the right in the way she interprets her research.

O'Neill's recent approach to analyzing the gender wage gap has not been so much in taking a random sample of men and women who work and then evaluating the results by using statistical methods to control for nondiscriminatory variables (though she does this, too), but to focus on one value of the standardized variables and then to look at men and women who have that one value. She does this with age.

The idea is to compare young workers, or workers who have only recently entered the labor market. The benefit of doing this is that these workers are unlikely to have family obligations yet, and many of them are not even married. The disadvantage has to do with them being young workers, and I will talk about that later a little more. But here is a summary of O'Neill's research into this field:

When women behave in the workplace as men do, the wage gap between them is small. June O'Neill, former director of the Congressional Budget Office, found that among people ages 27 to 33 who have never had a child, women's earnings approach 98 percent of men's. Women who hold positions and have skills and experience similar to those of men face wage disparities of less than 10 percent, and many are within a couple of points. Claims of unequal pay almost always involve comparing apples and oranges.

As you will know by now (if you have read the first two parts of this series), we are not comparing apples and oranges when we hold the values of nondiscriminatory determinants of wages constant. We may have the problem of omitted variables, but O'Neill's approach has a different problem: she is comparing very young apples to very young apples.

Why would this be a problem? The reason is this: First, think of an employer who wants to discriminate against women, but needs to hire workers for a firm. Women apply for the jobs and they are qualified for the jobs. If this employer refuses to hire any of them, he or she might soon find a law suit or a nasty audit study done on the firm, because Title VII of the Civil Rights Act bans hiring discrimination on the basis of sex. So the bigoted employer must hire some of these women. Now, that is terrible, you might think. Couldn't this poor soul at least pay the women less, to sweeten the bitter pill? Ah, but here he or she will run into another problem: The Equal Pay Act, which made it illegal to pay women and men different amounts for the same job. The only recourse the poor bigot has is to try to steer women into those jobs that pay less or that keep women further away from the bigot physically. These would usually be the lower-ranked jobs in the organization. O'Neill's approach to studying the gender gap would assume (or so it seems to me) that the women have "chosen" the occupations they have. If she controls for the job description in her analyses she will find no discrimination, or hardly any. Unless the bigot goes ahead and breaks the law anyway.

It is only with time that a discriminating employer could affect women's earnings negatively, by promoting selectively or by refusing women the on-the-job training that is needed for promotions or by simply instituting a hostile work environment which encourages women to leave. I'm not saying that this sort of thing happens widely. I don't know if it does, but the point of focusing on only workers at the beginning stages of their careers is that we wouldn't catch most avenues of discrimination.

Second, focusing only on young workers may overlook discrimination for a different reason. Some bigots might like to have young women around and don't mind seeing the opening slots filled with a fair mixture of both women and men. What the bigots might have difficulty with is women in power, or sexually no-longer-attractive women in the place of work. Both of these reasons would only be picked up by data if it included older workers, too.

I'm not arguing that the gender gap between men and women might not widen over time because of the "choices' women make or because some other nondiscriminatory reasons. But focusing on only young workers does not prove that discrimination is no longer a problem at all.

This is how the wingnuts do the talking. I have not addressed the use of anecdotal evidence, of quoting a single woman who, say, threw away the chances to run the world in exchange for more time to learn to ride horses. That women, on average, earn less does not mean that no woman anywhere isn't earning lots. That discrimination against women exists doesn't mean that no man is ever treated unfairly. This is why anecdotal evidence doesn't prove anything about wider trends. In order to look at the problem we need to focus on the large outlines and on statistical studies.

The end of my gender gap series. I hope that you enjoyed reading it.