April has come and gone, and, with it, the highly publicized Equal Pay Day.  It’s the day of the year on which women are said to have finally reached pay parity with men from the previous year; women working full-time, it turns out, only earn 77%, 78%, or 79% of what full-time male workers earn.
But there’s a problem with those figures, no matter which you choose: they account for exactly one confounding factor.
One.
In economics, there’s this great, smart-sounding, Latin phrase, ceteris paribus, which means “with other conditions remaining the same.” It’s a quick way of communicating the idea that we’re trying to figure out exactly how much impact one factor has in determining an outcome.
(It’s also fun because injecting Latin words into an article like this really makes the author look like he knows some stuff; but I digress.)
Back to our numbers: we’re trying to discover just how much influence sexist discrimination has on women’s earnings relative to men’s, we’ve controlled for full-time status, and now we’re saying that–ceteris paribus–sexism alone accounts for a gap of 21% to 23% between men and women?
Come on.
Granted, controlling for full-time status is important, but surely we can do better than one measly factor. I mean, there are a ton of other things that play into comparing the earnings of men and women, right?
For example, shouldn’t we try to compare women and men who studied the same things in college and selected careers in the same fields? Seems relatively important–and when we do that, the American Association of University Women finds that women actually make 93% of their male coworkers.
What about comparing work experience?  Lengthy career interruptions? And, heck, what about overtime?  Controlling for those things, a report prepared for the US Department of Labor found that women actually make 95% as much as their male coworkers.
I’d say a rise in women’s pay from 77% to 95% is quite an improvement, wouldn’t you agree?
But wait, there’s more!
Starting a family makes up a large (and growing) proportion of the total wage gap, and women tend to accept “family-friendly fringe benefits” (flexible hours, child-care, and parental leave, for example) in lieu of higher wages. When we account for these kinds of non-wage compensation, the gap in total compensation falls to 3.6%.
Neat, right?
Clearly, controlling for confounding factors is important; but as encouraging as all of these numbers are, we should take a step back and analyze some assumptions behind these comparisons.
For one thing, we seem content to uncritically accept the notion that, at some point, when we’ve controlled for everything we can possibly think of, the remaining disparity will necessarily indicate discrimination; but it’s not obvious that this is the case.  Stanford economist Thomas Sowell writes, “Where there are very significant differences in known factors between one group and another, it would be reckless to assume that all remaining unknown factors are the same.”
Another questionable assumption is that people–not just women–are single-mindedly concerned with maximizing their income, but that, too, is far from obvious.  In fact, individual success is subjective and multidimensional.  Using relative income as a means to compare men and women doesn’t tell us as much as we might wish it would because it ignores the complexity of success.
But even if we accept that comparing wages is a useful means of measuring sexual discrimination, we’re dealing with averages here, and while it’s important to understand what these numbers are telling us, it’s just as important for us to understand what they’re not telling us.
Take this study published in the American Economic Review, which finds that, when controlling for a host of relevant factors, women earn 97.5% of what men earn.  Are we supposed to believe that that number is telling us that every female worker makes precisely 97.5% of what her male co-workers earn?  
Not at all–but it does give us insight into the fact that there’s considerable wage variation between individuals.  
Let’s say that the average male income is $50,000 per year, meaning that an average adjusted female income of 97.5% is $48,750. Sure, half of these women make less than $48,750, but the other half makes more than that. It’s the same thing with men:  while half make more than $50,000 per year, the other half makes less.
This means that among the half of women who make more than $48,750 and the half of men who make less than $50,000, there’s considerable overlap, with a not inconsequential number of women out-earning men.
Given this overlap, sexual discrimination as an explanation for disparities in income loses some of its intellectual appeal–after all, if Jack makes more than Jill, what accounts for Jane, who makes more than Jack?  What about John, who makes less than Jill?
A more comprehensive explanation of the data might be that men and women make different choices on the basis of differences in values–and, empirically, there’s actually very good reason to believe that this is the case. Research shows that men and women differ widely in choice of major while attending college. As a result, fewer women than men enter certain fields–STEM, for instance–while the converse is also true–nursing is just under 9% male.
Furthermore, a recent study found that women choose to enter STEM fields less often in countries where they enjoy a higher degree of gender equality.  Coupled with research that finds that men and women in countries with more gender equality diverge more widely in personality–even as they converge in valuing self-actualization–these facts suggest that empowered women tend to prefer pursuing careers they enjoy over jobs that merely pay them well.  
University of Chicago economist Steven D. Levitt puts it this way: “Rather than interpreting women’s lower wages as a failure, perhaps it should be seen as a sign that a higher wage simply isn’t as meaningful an incentive for women as it is for men.”
Don’t get me wrong, here: none of this should be taken to mean that sexism doesn’t exist–it most certainly does–but the data, and especially this analysis, should be encouraging to those of us who are concerned about the economic well-being of women in 2018.