I'm not entirely sure what day it is, but my computer is telling me it's Saturday. This post originally appeared on June 26, 2009, under the title "Equal Pay for Equal Work". A number of very good comments are now, unfortunately, buried in the OS archives.

Has it been cloudy this past month? Some people might answer, "It depends on where you are." Others might say, "One or more regions reported clouds, so... yes." Or "Looking at data from all domestic weather stations and averaging over the past month, we get this measure for cloudiness..." or "Probably--but what should we expect for cloudiness?" In general, most of the answers are going to be variations on Yes, but we need to keep in mind that even when people agree, there can be differences in what they mean.

Next question: Is there a gender wage gap? The answers are mostly Yes here, too, but understanding the assumptions behind the variations is trickier--social issues are even harder to untangle than weather patterns. Not to mention that, unlike weather, a gender wage gap is something we can do more than complain about; we can fix it. This is my understanding, in part, of the issues.

When people talk about a gender wage gap, here's what they generally seem to mean: If a woman gets paid less than a man for doing the same job, that difference is the gender wage gap. We've probably all seen census reports showing that, on average in the U.S., women earn about three-quarters of what men do. Isn't that a decisive answer to the question of whether there's a gender gap? Not if we're thinking about comparisons between the "same" jobs.

For example, take a surgical unit in a hospital. If we were to break down earnings by gender, we'd find that the men make a great deal more money than the women. The reason why, though, is pretty obvious, and it doesn't have to do with unfairness in wages, per se: surgeons earn huge amounts of money and are more likely to be men; operating room nurses don't earn nearly as much and are more likely to be women. That is, it can happen that there's no gender wage gap among surgeons, and no gender wage gap among OR nurses, but when you put them together, the differences in income and gender proportions between the groups produce the illusion of a problem. That is, it's not a problem if we focus on the idea of same job, same pay.

So, to talk about the "same" job, we need to establish categories, such as "doctors" and "nurses" and "school teachers" and so forth. Is that enough, though? We can imagine a Dr. Marcus Welby who's nearing retirement and pulling down a huge income, even compared with a Dr. Quinn, who has just acquired her medicine woman license (sorry about the stereotypes--there aren't that many iconic women doctors on TV).  Is it fair to say that the two have the same job and to call the earnings difference a gender wage gap? Perhaps not.

Labor economists generally agree that a fair comparison involves taking into account these factors [PDF]:

• Education (level)
• Education (major field of study)
• Experience
• Experience2
• Job tenure
• Marital status
• Presence of children
• Part-time status
• Union status
• Firm size
• Region
• Urban size
• Occupation
• Job responsibilities
• Industry

If these don't entirely account for variation in earnings, then one of the remaining explanations is gender. And we're far from being done. This just touches the surface of a complex and extensive area of work (Google Scholar turns up over 100,000 hits for the terms wage + gap + gender), and there's still a good deal of disagreement between experts about how to interpret the data. Marie Drolet, who wrote the article I've linked above, says, "The issue of male-female wage inequality is complex and requires analysis from a number of different perspectives. The goal of this article is not to provide a single, definitive estimate of the wage gap..."

That is, even professonal labor analysts and economists shy away from saying, "This is the gender gap," which naturally makes it difficult to come up with broad generalizations about its magnitude. Not everyone is shy, of course (the non-shy ones tend to get quoted in the newspapers). You and I, though, we'll probably want to be careful.

But wait. Let's take Drolet's advice to heart. In particular, let's consider a perspective on the gender wage gap that I dismissed too easily above. Imagine that women have been shut out of entire categories of professions, either by law or by custom. Instead, they are tracked into other categories of professions that tend to pay less. By ignoring the earnings differences between not-the-same-jobs in this hypothetical situation, as I did above, I might be ignoring evidence for an unjustifiable gender wage gap.

Of course, this situation isn't quite hypothetical; legal barriers are now gone, but I've described the state of much of the labor market for most of the history of this country. How much influence do social or cultural factors have today? Arguably, quite a lot. Consider the occupation of teaching. Historically, school teachers have not been held in high social regard, and today they are widely (though not universally) regarded as being underpaid, given the importance of their work and job requirements often imposed on them. It turns out that 98% of preschool and kindergarten teachers in the U.S. are women; only 2% are men. Now, there are good explanations for this, but here's a startling statistic for contrast: In families in which mothers work, 25% of the time preschoolers are cared for by their fathers. That's a big difference. It suggests that there might be an untapped pool of men interested in working in this occupation but not doing it, for one reason or another. Social factors? It would be hard to dismiss the idea completely. I've used this example because the low end of the earnings spectrum tends to get less attention than the high end--but the issues there are comparable. For example, it happens that only 2% of Fortune 1000 companies have women as CEOs. Social factors? Again, hard to dismiss completely. And if the occupations dominated by men tend to be higher paying than those dominated by women, well...

The analysis that would be needed to say that women tend to be pushed in the direction of lower-paying jobs, or that that jobs mainly done by women are lower paying exactly because they're done by women, is messy. (I'd be surprised if there isn't work on this issue, but I don't know the relevant literature well enough to say for sure.) One reason for the messiness is that we're no longer just comparing numbers--we have to think about how and why people make decisions related to their employment. For example, you may have chosen to go to a specific college (or not), to work in a specific industry, to get married, to have children, to join a union, to live in some region of the country, to take on new job responsibilities. All of these might have affected your earnings--they might continue to do that today. Were any of these decisions influenced by people's expectations of you, expectations formed in part because you're a man or a woman? Would it have been easy to overcome those influences? Did you even recognize them at the time? It might be easy for a single individual to answer these questions, but generalizing over a population would be hard. 

And we haven't even reached the point of discussing what to do. (Of course, we should fight for the notion of equal pay for equal work--it's the law, after all. We should encourage girls in middle school to continue taking math and science classes, so that they might pursue engineering careers later in life. But how much should we worry about the persistence of strongly gender-influenced occupations like auto repair or dressmaking? Wouldn't it be interesting, though, if we didn't automatically assume that secretaries, teachers, nurses, and so forth were women, while soldiers, police officers, fire fighters, and engineers were men?  I know, I know--that's crazy. But it is the 21st century...)

I could say more, but I'm getting a bit out of my depth; I'm not an expert on this topic. As with all complex issues, I think it's interesting to look at it from different angles. If you have something to add to the picture, please do.