Saturday, January 03, 2004

Money for Your Lingerie? (Or Retirement)

It is ever so Out these days to raise a peep about discrimination against women. If you are so unfashionable as to complain, you are promptly labeled a whiner. Hip young post-feminists will advise you to get over your "victim psychology," Dr. Laura will tell you to pull up your socks, and the women's magazines will advise you to try sexy lingerie for your problems.

This is how Molly Ivins begins her column on the gender wage gap, and she's got a point. It IS unfashionable to complain, and it IS fashionable to ignore the wage gap. That women make, on average, only 76 cents to every dollar men make, well, who cares? It's much more fun to learn pole-dancing and to discuss what is sexxee. Besides, lots of people will start screaming at you with red faces if you as much as peep about it.

On top of that, understanding the reasons for the gender wage gap is really hard. It's something economists spend years learning to do, so thinking about the wage gap can be daunting to even the most intellectually alert individual. (If you wish to have a very good introduction to the wage gap, read Ampersand's series at the Alas, A Blog. You won't get anything of same calibre here.)

Maybe that's the reason for the current meme that it's ok to treat the wage gap as something completely debatable, from whether it even exists to what causes its current size. I don't buy that personally, but I'll be glad to show you how it's done here.

Sadly, a little bit of economics is needed before the fun can begin: Basically, there are three possible reasons for the gender wage gap (and yes, all economists out there, I know that I am oversimplifying here):

1. Women can't / Men can (performance differences)

2. Women don't want to /Men want to (preference differences, or as some prefer, choice)

3. Women are not allowed to /Men are allowed to (discrimination)

In reality, the question how much each of these contributes to the wage gap is an empirical one: it should be answered by using proper data and careful methods. But in the medialand the game is played differently: simply pick one or two of the three based on which you like the best. Then write up something emotional to support your choice.

Even Molly Ivins falls for this. Her choice is number 3: discrimination:

Of course, more women could afford sexy lingerie if we weren't still the victims of wage discrimination. Equal Pay for Equal Work is the oldest demand in the feminist repertoire, and everyone gives it lip service; even the anti-feminists assure us that they certainly believe in equal pay for equal work.

Molly, the Equal Pay Act of 1963 applies only to men and women doing exactly the same job. Unfortunately, the vast majority of men and women don't work in the same jobs. Even if every single firm obeyed the Equal Pay Act (which not every single firm does), women would still earn less, because so many work in female-dominated low-pay occupations. If discrimination is to be the culprit, you need to look much deeper into the structures of labor markets.

The spunky ladies of the right-wing Independent Women's Forum choose number 2., choice, combined with the auxiliary argument that all nondiscriminatory factors should be taken into account before examining any residual discrimination:

According to the "wage gap" claim, women earn only 76 cents on the male dollar. But, this is a deliberately misleading claim that fails to account for a number of commonsense facts about women's workplace experiences. Women actually earn 98 cents on the dollar when factors such as age, education, and experience are taken into account. It is critical to compare apples to apples. For example, women often leave the work force to raise children and later return.

The claim about women earning 76 cents on the male dollar is not deliberately misleading, ladies. It's pretty much correct, subtract or add a few cents, depending on the year. What you are trying to say is that not all of the 24 cent difference can be attributed to discrimination without first controlling for variables such as education and work experience. This is true. (But age shouldn't be one of these variables to control for. It's sometimes used as a measure for work experience when the latter is not available in the data, and if you insist to include it even with work experience you are assuming that ageism is an ok thing to affect wages.)

What isn't true is all that stuff about women earning 98 cents on the dollar. This was the result from one study which compared highly-educated single, childless men and women in their first jobs. Lo and behold, the study found that the entry-level wages were almost the same for both sexes. This is a) not proof of equal wages more generally, b) not proof of lack of discrimination and c) not even new.

Economists have known for a long time that all sorts of people earn about the same starting salaries for a given set of education. This would be the case even if they were all hired by the most rabid bigot you can imagine, simply because it is against the law to discriminate in hiring (Title VII of the Civil Rights Act) and because it is against the law to pay different wages for the same job (Equal Pay Act of 1963). The rabid bigot (in my discussion a woman-hater) would have to wait a little longer before she or he could start throwing obstacles in front of the female workers. This could happen through not promoting women, through not giving them opportunities for on-the-job training , or through various forms of sexual harassment. (Though all of these are illegal, too, they are much more difficult to prove than obvious wage differences at entry level.)

I'm not saying that any of these alternatives would necessarily happen. Many women might simply opt out of promotions because of their childcare duties, as the IWF argues. But we don't actually know how much choice and discrimination affect women's earnings development over time. To imply that it's all choice ("For example, women often leave the work force to raise children and later return.") is part of the opinion-game.

My final example of the game is a real peach. It's from an article in the Mensnewsdaily:
The fact is, the "wage gap" disappears when you take into account such factors as training, years in the workforce, travel requirements, degree of physical labor, and risk to life and limb .
And truth be told, men essentially have no choice -- they are expected to be the primary breadwinner in order to support their wives and children . So they accept the high-paying, dangerous jobs that women are unwilling to accept.
In contrast, women have a broad range of options: Be a full-time mom, take on a part-time job, or do volunteer work.
So the so-called "wage gap" is really a "choices gap." And the feminist campaign to level wages really amounts to equal pay for unequal work.

Here the gentleman in question (one Carey Roberts) tells us that there is no discrimination, at least in his opinion. The wage gap totally disappears when proper attention is paid to the requirements of the job. Unfortunately, no evidence is provided for this opinion, probably because it doesn't exist. The gentleman appears to see the reasons in the wage gap mainly in number 2., choice, though including the degree of physical labor as one of the important explanations for higher wages and referring to equal pay for unequal work also smacks of an attempt to bring in number 1, performance differences.

The concept of choice is used in a very weird manner: women have all the choices but men have none. It seems more likely that if men feel constrained in their work choices, so do women. In particular, the society in the U.S. expects women to be responsible for child-care, and most economists agree that women's duties at home are one of the main determinants of the gender gap.

Some argue that these responsibilities make women 'choose' occupations which allow for flexibility in work hours and leave taking, even when this flexibility is at the cost of lost earnings and pensions. Many traditionally female occupations (the so-called pink collar jobs) do have these characteristics, though it should be noted that many high-paying jobs also have flexibility.

Do women make these sorts of choices? Many do, probably, but they are not free choices in the way a choice between chocolate and vanilla ice-cream might be. The work women do at home is valuable for the wider society; it produces the future tax-payers and soldiers and politicians, and the quality of these future citizens is of interest to everyone. At the same time, many observers would like the costs of these choices to fall totally on the women and their families. Also, even women who would not voluntarily choose to stay at home with their children may be pressured into this through societal disapproval of working mothers and life-long absorption of messages promoting a certain mothering style.

So it's not all that clear that the choices that women (or men) make are in some sense free and purely private. But it's also very difficult to prove to what extent women's earnings are affected by discrimination in the labor market.

Though discrimination clearly exists (as evidenced by court cases won by the plaintiffs and by audit studies in general), the current level of economic studies doesn't allow us to conclusively separate choice-based, performance-based and discriminatory reasons for earnings differences. What they can tell us is that once we have taken into account all the performance and choice-related aspects that we have data on, the remaining unexplained difference in men's and women's earnings could be due to discrimination. It could also be due to something else that we have inadvertently omitted.

Not a comfortable state of knowledge. But it's at least much more honest than the opinion games being played across the political spectrum. A recent, relatively well-done congressional study* shows that this unexplained difference amounts to roughly 20 cents on the dollar. (It's important to keep in mind that this study looks at men and women who work either full-time or part-time, while the gender gap in earnings as commonly discussed applies only to full-time workers. So the figures from this study are not directly comparable with the 76 cent figure in the preceding discussion.)

The raw data in this study show women earning, on average, only 44 cents for each dollar men earn, but factors such as working hours (especially in the form of women's part-time work), education, work experience and the number of small children the worker has at home explain a large portion of this difference. Still, these can only bring the women's adjusted earnings to 80 cents per each men's dollar. What causes the remaining difference is unclear. It could be caused by discrimination or possibly by women seeking jobs which allow greater flexibility. But note that some of the earlier factors could be discriminatory, too. For example, women might be working part-time due to pressures from societal gender roles rather than through free choice.

Why talk about something this boring? Well, it affects how much lingerie you can have. It also affects your old age, because retirement incomes go hand in hand with earnings. Women over 65 have an average annual income of $14,200 compared to men's $39,000.

It also affects families, as some observers have started noting. Whatever Mensnewsdaily might say, the fact is that many women and men live together in families, and the children and male partners of women also suffer when the women suffer.
*The link to this study gives you an Adobe pdf file. If you want to download it, go here, but ignore the title of the article: it's incorrect.

Postscript: The gender wage gap I discuss here is based on comparing men and women who work full-time, though the specific study I discuss also includes part-time workers. I wish to confess that I haven't bothered checking if this year's figure is exactly 76 cents or not. But it's in the ballpark!