The Wage-Gap Debate Is Based On Junk Science

Nowadays, rarely a week goes by free from news about the so-called gender pay gap. There is even now a “wage-gap day” on November 10.  This is the day when women allegedly start to work for free for the rest of the year. The remaining 51 days are 14% of the year — a figure corresponding to the wage gap between women and men.

One can hardly find more mindless approach to this issue, however, as the idea behind the “working for free” narrative is an affront to any serious study of the of the wage-gap issue. 

It is usually asserted that the gap is a product of discrimination and sexism. But any scientific approach to the issue requires that if we’re going to make such an assertion, but we must take into account other variables that affect wages, such as the self-sorting by employees,total work experience (including working time), and education. Once we correct for such variables the adjusted wage gap is generated, which is significantly smaller. It is closer to the 2-5% range, depending on the study. Without such adjustments one is obviously comparing apples and oranges. We could as well say that a truck driver earning 12-times less than a banker is working for free since February, almost for the whole year.

I do not want to deal here, however, with the obvious distinction between the statistical wage gap and the adjusted wage gap — even though the popular press constantly and passionately refuses to learn anything from the existing research. I would like to focus here on what is left after the adjustment: that small, but still positive wage gap of 2-5%.

Let us pause for a while and analyze how the adjustment is made. There are various widely used statistical tools to do it, but their general methodological core is the same. We classify workers according to some objective and easily recognized features such as education, working time, sector and so forth. Furthermore, by using econometric analysis, we try to see the statistical connection of each change (increase or decrease) of a particular objective variable resulting in changes of the wage level. After the filtering is done we can recognize how much education, working time, and experience can contribute to higher income. Yet that does not fully bring the wage gap difference to zero.