I am in broad agreement with Becker's excellent analysis.
As discrimination declines, replaced by affirmative action, explanations for lagging achievement that are based on discrimination lose their plausibility. They were never entrely plausible, given Jewish achievement in the face of fierce discrimination, though it is argued by Stephen Pinker in a recent issue of the New Republic that discrimination against Jews in the Middle Ages, by forcing them into middleman occupations where intelligence is a more valued asset than in farming or soldiering, resulted in the more intelligent Jews having a higher birth rate (because they were better off) than the less intelligent Jews and so, through the operation of natural selection,discrimination can be "credited" with some of the responsibility for the high average IQ of Jews today--even its genetic component. (Hitler may have had something to do with this as well, as it is plausible that the most intelligent European Jews saw the handwriting on the wall earliest and left Europe in the 1930s before it was too late.)
As Becker points out, the mean performance of women in college and university is superior to that of the men, but the variance of male performance is greater and as a result there are more male geniuses. There is no reason why the difference in variance should result in higher average male earnings; that higher average is probably the result of women's spending less time in the work force because of pregnancy and child care. Women's greater proclivity for child care may well have a biological basis, as may the difference in variance that I mentioned. In the "ancestral environment"--the term that anthropologists use to describe the prehistoric period in which human beings reached approximately their current biological state--women who were "steady" would have tended to have the maximum number of children, while natural selection might favor variance in male abilities because variance would produce some outstanding men who would tend to reproduce more than other men (including the "steadies") in the polygamous conditions of prehistoric society.
If the explanation based on evolutionary biology is correct, women will continue to be "underrepresented" in high-achievement positions in many fields; why anyone should care is beyond me. But it doesn't follow that their average earnings will continue to be significantly lower than those of men. Women's lesser commitment to the labormarket may be balanced by their greater ability than men to perform most jobs, assuming academic performance is a good proxy for aptitude for today's desirable jobs. With the decline in the importance of physical strength and stamina as a job qualification, women may be able to perform most jobs better than men on average, though men may continue to dominate the top--but also the bottom--tier of the labor market.
The achievement lag in black males is troublesome from a social standpoint, as it seems correlated with definite social pathologies, such as enormous overrepresentation in criminal activities. Moreover, it is a matter of a lower mean rather than less variance. If and to the extent that that lower mean is a result of lower IQ, not much can be done because IQ has a strong genetic component--and what is not innate may still be innate rather than cultural (a product of conditions in the womb, for example). The genetic and environmental influences on abilities interact, as Becker says, but in addition the genetic can influence the environmental: many low-IQ mothers may be un able to take care of themselves adequately in pregnancy, contributing to their children's having innate intellectual deficiencies due to poor material nutrition or health care.
Differences in the mean achievements of racial or gender groups must be kept in perspective. General intelligence (IQ) follows a bell-shaped distribution, and two bell-shaped distributions that have different means will still overlap to a great extent unless the means are very far apart. The differences will be greatest in the tails of the distributions.
The achievement lag of Hispanic males may be a transitional phenomenon; they may still be adjusting to an American male culture that is quite different from the "macho" culture of Latin America, which is not conducive to vocational achievement under modern American conditions.
Like Becker, I view affirmative action as a matter of choice for colleges and universities, at least when the institutions are private rather than public. Higher education is highly competitive, and I am reluctant to have the government tell its institutions what policies are best. Academic freedom implies a high degree of academic autonomy, including autonomy in the administration of the institutions of higher education. Personally, however, I would like to see a few of the top colleges abolish all preferences unrelated to academic merit--no athletic scholarships, no affirmative action, no favoritism for the children of professors or of major donors, and no legacy admissions. That would be a useful experiment in the benefits and perhaps costs of meritocracy. It would have the incidental effect of giving us a better idea of the extent of real differences across race and gender in academic capability.