What do gender pay gaps tell us? Here are some of my thoughts.

My table shows the data, taken from the latest ASHE survey. This shows that the median woman working full-time earns 9.7% less than the median full-time male employee. This is the same as the median pay gap reported recently by large firms.

However, the pay gap is much smaller for younger workers than older ones. This might be because there's a pay penalty for being a mum. If you take time off or go part-time after the birth of a child you lose pay progression which you never recoup. The fact that lesbians earn significantly more (pdf) than straight women is consistent with this.

Another explanation, though, is that older women suffered sexism in the 70s and 80s, which retarded their pay throughout their lives.

I think Kate Andrews has a point when she says this gap could be closed if firms simply hired more men into low-paid jobs and denied jobs to less skilled women. That, though, isn't just a point about the pay gap but a point about the tyranny of metrics generally: any simple measure can be gamed. The point about simple measures of the pay gap isn't that they should be closed anyhow but rather, as Ben says, that they should help raise questions.

The median gender pay gap, though, is not all the story. The pay gap for higher-paid women is much bigger. A woman earning more than 90% of other women earns 18.2% less than a man who earns more than 90% of men.

This gap in part reflects high inequality among men. If well-paid men earned less, the pay gap between male and female high earners would shrink.

I suspect there's a reason for this big gap. In top jobs pay is set not so much by a thick market as by bargaining over a joint surplus. And women are in many (pdf) circumstances less good at bargaining (pdf) than men. As Claer Barrett advises, women should be pushier.

Julian Jessop has a point when he says these gaps might not reflect overt discrimination by employees. They might instead be due to women sorting into lower-wage jobs. Airlines, for example, have big gender pay gaps because women tend to be low-paid cabin crew whilst men are higher-paid pilots.

Such gender-based preferences, however, are NOT the end of the story. For one thing, as Sarah O'Connor says, employers might perhaps do more to accommodate such preferences for example by not penalizing women who want to work part-time so much.

But there's another point here that is grossly under-appreciated. It's that preferences are not natural and given. As Simone de Beauvoir said, "one is not born, but rather becomes, a woman." Women's relative lack of pushiness or ambition, or their preference for less well-paid work, might be due instead to the way they are socialized.

To cite just two examples from conventional social science, Alison Booth has shown that girls from single-sex schools are as competitive as boys. This, she says, "suggests that observed gender differences might reflect social learning rather than inherent gender traits." And Marta Favara has shown that girls who go to single-sex schools are more likely to study maths and sciences than others. In both cases, all-girl schools (at the margin) discourage girls from acting girly and so steer them away from stereotyping.

In this sense, two things might both be true. It might be that the gender pay gap is not due to employers discriminating against women. But nevertheless it might also be true that women are the victims of sexism because of (among other things) how they are socialized.

There's an analogy here with another issue: Oxbridge's lack of admissions of poor students. Maybe there is little outright discrimination against them. But many poor students nevertheless suffer inequality because of wider social forces: poor schooling and mentoring, a learned lack of ambition and so on.

The point here is a simple one. Inequalities are not necessarily due to bad people deliberately doing bad things. They can instead be the result of impersonal mechanisms.