Everyone know the standard news report, at least once a year "Women earn 80% of what men do" or "Women make 76 cents for every dollar earned by men in comparable positions." It appears over and over, with slightly adjusted numbers. And, when the story itself isn't running, the newscasters report how those numbers compare with last year's numbers, and what that indicates.

The problem is, those numbers just don't mean anything.

In as series of comments I mentioned one of the more serious problems, the failure to control for a number of considerations. And, even when they do "control" for realistic differences (such a women taking maternity leave much more often than men take paternity leave), the adjustments are often little better than guesswork.

Let us start by looking at maternity leave. The studies generally try to calculate how much it costs a company to hire and train replacements during maternity leave, they then multiply this by the frequency of maternity leave, subtract it from the average male wage, compare to female wages and, voila!, they think they have accounted for everything.

Not even close. First, a replacement is rarely as productive as a regular employee. They need to include not only the training and hiring costs, but also figures for lost productivity. Second, on a related note, there is institutional memory, those things which are known by employees but not written down. If the woman out on maternity leave is privy to such information, and her absence makes such information unavailable, it may cause other losses as well. Third, while a woman is out on maternity leave, she is not practicing her skills. IN some professions this is not important, but in some professions (advertising, fashion, some technical fields) a six, or even three month absence can translate into a significant loss of skill and require some time to get back up to speed. Finally, though fewer women do it now than in the past, there are still women who decide after giving birth that they want to be stay at home mothers and resign. This imposes additional costs on the company, as they planned only for a temporary replacement, and now they have to start over again finding a full time replacement. It also means a skilled and experienced employee will now be replaced with a novice. All of which increases the potential costs of a female employee.

Of course, each of these only increases costs by a few cents, most likely. But they are just a few considerations. There are many others, which can also contribute a few cents. Women sue for sexual harassment or sexual discrimination at much higher rate than men, increasing employer legal costs. Studies have shown that, in general, women favor family and personal time over time dedicated to work. It is not an absolute, but on average, if women are less available for overtime or travel, they will be promoted less quickly and will be given less favorable reviews, reducing their pay relative to similarly trained men. Also, in the case of those women who do leave their careers for a family, when they return to the workforce, their training is not as current as newcomers who have been more recently educated, making their skills less up to date and depressing their wages. And the list goes on and on. All, admittedly, very small differences, but, when put together, enough to account for a pretty large disparity, on average.*

But, let us ignore that and look at some other weaknesses of these studies.

Another major problem is what the studies compare. Some, fairly, compare the same position at the same company, which is probably the only fair test. Others compare jobs with similar descriptions at different companies, which is less fair. As anyone can tell the researchers, different companies pay differing amounts, and if, by chance, one company hires more women, that normal variation in pay can skew the numbers to show a disparity which is likely not sex related.

The worst studies do something even less honest. They argue that some fields have been traditionally male and higher paid and some female dominated and lower paid, so they try to compare "comparable jobs", arguing a truck driver is the same as a waitress or a nurse administrator requires the same education as an engineer. The problem with this should be obvious. Jobs do not pay based solely on education, and five years of education in engineering is a very different thing than five years of training in painting still life or analyzing iambic pentameter. All may be valuable skills, but the difficulty of one means that there will be fewer who can successfully complete the program, and thus they will be in short supply, elevating wages.

But there is an even bigger problem with the "comparable job" theory, the idea that wages are the outcome of some deterministic function based on education and experience. Nothing could be farther from the truth. Wages are based solely upon the increase in value the employer anticipates from the employee. Education and experience may play into that, but they hardly determine it.

Think about this: The educational requirements for a star baseball player, a Hollywood superstar, a prostitute, and a McDonald's cashier are identical (that is, none required). In fact, the McDonald's cashier actually needs some minimal education and receives training, so should be the highest paid of the lot by the theory these studies often use. (Then again, the prostitute is self-employed, and thus free of discrimination, so she should really be the best paid if we accept that discrimination keeps women from earning as much as men.) So, if wages are a function of education alone, why are starlets in Hollywood earning so much more than cashiers?

But, really, none of this matters. The truth is wages are not "set" by anyone. An employer pays what he thinks he must to hire and retain an employee so long as he expects the employee will return more than he costs. Of course the employer will try to pay as little as possible, but the fear of an employee being hired away by a competitor tends to keep pay close to the break even point. Sex, race, and any other non-economic factor cannot really play a role, unless the employer is willing to accept a loss, as, refusing to pay blacks or women what they are worth will result in those employees being wooed by a competitor.

And that is the single best reason I see to doubt these studies, human greed.

If women are uniformly underpaid, some clever businessman would think "Women earn only 80% of what men do? Well, I can pay then 85%, get only the best women, and still save 15% on labor costs while getting an elite labor force." And, as the returns would be so great, I don't think there is so much sexism that a business man would turn down a 15% cut in labor costs.

But that is where the problem arises for the studies. Because if we agree that the first situation would happen, wouldn't another clever businessman pay 90%, still cut costs by 10%, and still make a killing? And then wouldn't another think a 5% cut in labor costs is still pretty good and pay 95%? And eventually, wouldn't women be earning pretty close to what men do?

So, if women really are being so dramatically underpaid, it means that every business on earth, even those owned and run by women, are so incurably sexist that they would rather keep women oppressed than earn a huge profit. I just do not believe it.

Far more likely the studies are flawed than such a massive conspiracy of sexism exists.
 
But I am sure many will disagree, and quite vocally.

---------------------------------------------------------

* Any of these individual items can be argued, and perhaps one or more of my theories is wrong, or perhaps I have overlooked some significant factor. My point is not to list all the factors, just to point out that the studies trumpeting wage disparities commonly fail to account for all the many things which can influence an employer's decision.