Dealing with the other side on the gender wage gap

Though there is a very large literature suggesting that the gender wage gap is not down to discrimination, this is not a universal finding, even in new papers. Three recent studies, for example, allege that their evidence supports the gender discrimination model of the labour market. However, their methodologies cannot well account for alternate hypotheses (e.g. gender difference) and we would do well to look primarily at the work which does try and factor these possibilities in. The first, "Estimating gender differences in access to jobs" (2012 pdf, 2015 gated), from authors Laurent Gobillon, Dominique Meurs, and Sébastien Roux, finds that:

females have a lower access to jobs at all ranks in the wage distribution of job positions and that the access function is decreasing with the rank. At the lowest ranks, the probability of females getting a given job is 9% lower than the probability of males. The difference between these probabilities is far larger at the highest ranks and climbs to 50%.

But wait! Their data allows for three explanations!

First, females may apply less often for high-paid jobs because working hours are less compatible with family constraints. Second, there can be taste discrimination against females which increases with the rank. Third, there can be statistical discrimination such that the skill distribution is the same for males and females, but skills are observed with more uncertainty for females than for males by managers.

Turns out there are lots of existing papers suggesting statistical discrimination (i.e. not sexism/racism) explains a big fraction of differential labour market outcomes between groups. And we have lots of evidence that men and women have different preferences about work hours. Let's not point to taste-based (i.e. sexist) discrimination before we've considered more well-supported alternative hypotheses.

"Gender and Dynamic Agency: Theory and Evidence on the Compensation of Top Executives" (2015 pdf) by Stefania Albanesi, Claudia Olivetti and María José Prados is even stranger. They:

document three new facts about gender differences in executive compensation. First, female executives receive lower share of incentive pay in total compensation relative to males. This difference accounts for 93% of the gender gap in total pay. Second, the compensation of female executives displays lower pay-performance sensitivity. A $1 million dollar increase in firm value generates a $17,150 increase in firm specific wealth for male executives and a $1,670 increase for females. Third, female executives are more exposed to bad firm performance and less exposed to good firm performance relative to male executives.

But their data shows that this is more or less entirely explained by male executives being older and more experienced. They don't have the data to control for age and experience so they just don't!

Sure, this isn't quite how they tell it in their abstract and conclusion but what else does this mean?

The managerial power/skimming view of executive compensation can rationalize these differences based on the notion that female top executives are less entrenched than male top executives, due to their younger age and their relative difficulties in accessing informal networks.

Whereas we know that for otherwise similar male and female execs, women get promoted more aggressively and earn more.

"The gender wage gap among PhDs in the UK" (2015 pdf) by Ute Schulze finds a similar sort of thing. There is a gender wage gap among even highly talented and motivated people—those who manage to earn a PhD. But is this down to discrimination? Schulze thinks it is: even within fields and within academia the gap ranges from an average of £559 to £10,902.

But does Schulze control for the positions these people end up reaching—no. She is right that men get higher returns on their observable characteristics, but she hasn't observed enough characteristics to justify her conclusion. There is quite a lot of good evidence that academia isn't significantly discriminatory towards women, and this simple regression based study is not enough to turn that over.

It seems more plausible that the large differences in preferences observed even between highly talented men and women explain the gap here—with men taking on more competitive, harder and just more work and hence ending up with dissimilar labour market outcomes.

It's true that this could come from social/cultural pressure, but at the same time it could be primarily biological. What's more, it doesn't seem to lead to women rating their lives as worse, in fact quite the opposite. Raising children and doing work in the home tends to be related to women reporting higher happiness, well-being and life satisfaction.

So it's not clear to me that this new hat-trick of papers adds anything to what we already know about sex/gender discrimination in the workplace—it still seems like there just isn't that much of it.