This seems to assume that preferences don't vary with race and gender, which isn't true. This is extremely well-documented with respect to gender, and it's called "The paradox of gender equality"[0]. TL;DR: Men and women are different from each other in lots of ways, including what they prefer to work on. As you increase gender equality in a society (moving from, for instance, Saudi Arabia at one extreme to Norway at the other), you see a systematic /increase/ in the difference in what men and women end up working on. More gender equality --> fewer women engineers, more women doing child care.
Edit: oh,and I would encourage anyone looking in to the literature to view it in the correct light: the field is overtly hostile to this data, and that greatly skews what gets published. And here it is anyway.