it's all about the edges of distribution though. If equality of the means mattered, countries across the world would be very similar economically, which unfortunately is not the case.
It claims that while averages for men and women are the same, you get more men at the high IQ end of the spectrum and also more men at the lower end of the IQ spectrum (which balance out the average).
The whole argument is predicated on what happens at the edges of the normal distribution, not what happens at the center.
We don't imprison random people from the population, we imprison the most aggressive (in case of violent crime at least).
We hire people with top IQ for engineering and science related endevours.
We promote people who work the hardest ("industriousness")
The most ambitious people get the top jobs.
It doesn't matter if on average two groups are the same for those qualifiers, if only the top 1% are selected, then you would expect which ever group has the flattest bell-curve distribution (and thus more "on the edge") to be over-represented (whether it is gender, "race", age). If it isn't, then racism/sexism/whateverism is at work.
Caveat: with top IQ in engineering, an important factor that only people who "like things more than people" apply, other way around for loads of other fields.
/apologies to Garrison Keillor
edit:
https://en.wikipedia.org/wiki/Lake_Wobegon#The_Lake_Wobegon_...
(and yes Keillor's "Lake Wobegon" stories [ibid.] are hilarious)
Therefore, if on some metric, you have two groups who are average but one with greater deviation, and then you select among the top 5% of the combined group, you will select more of the group with greater deviation even though the groups are average. You can test this with simulations if you want.
This isn't much of a problem until you apply it to everyday demographics, at which point it can result in uncomfortable conclusions.