David Autor and Daron Acemoglu have a long review article. Much of it is focused on shifting from a model in which there are only two skill levels (high and low) to a more nuanced model, with at least three skill levels.

Most notably, while both the 90th and 10th percentiles of the weekly and hourly wage distributions diverged rapidly from the median from the early 1980s to the early-1990s— consistent with a rising return to skill in the canonical model— the gap between the 10th percentile and the median substantially contracted over the next 15 years even as the 90/50 gap continued to rapidly expand. This suggests that, at the very least, it may be useful to distinguish between high, middle and low skill workers rather than just between high and low skill workers

…Between 1979 and 1989, occupational employment growth was nearly monotone in occupational skill; occupations below the median declined as a share of employment and occupations above the median increased. Between 1990 and 2007, relative employment growth was again most rapid at high percentiles but it was also strongly positive at lower percentiles, with the growth of occupational employment shares almost as rapid at the 10th percentile as at the 80th percentile.

…in ‘middle-skill occupations,’ comprised of four categories: office and administrative support occupations; sales occupations; production, craft and repair occupations; and operator, fabricator and laborer occupations. ..between 1979-2007, their growth rate lagged the economy-wide average and, moreover, generally slowed across decades. These occupations were hit particularly hard after 2007, with absolute declines in employment ranging from 7 to 17 percent.

…All three subcategories of service occupations– protective services, food preparation and cleaning services, and personal care— expanded by double digits in…1999 through 2007…Notably, even after 2007, employment growth in service occupations was modestly positive despite the deep recession—

In 1979, the four middle skill occupations— sales, office and administrative workers, production workers, and operatives— accounted for 58.9 percent of employment. In 2007, this number was 47.5 percent, and in 2010, it was 44.5 percent.

Note that Autor’s work was a major underpinning for my essay What if Middle-Class Jobs Disappear? Autor and Acemoglu continue

a task-replacing technological change squeezes out the type of worker previously performing these tasks, thereby creating ‘excess supply.’Consequently, these workers are reallocated to tasks for which they have lower comparative advantage, which pushes
their wages down.

They next turn to the issue of educational attainment.

The stagnation of U.S. K-12 graduation rates did not slowly take root over decades, as one would expect from an educational system in secular decline. Rather, 75 years of improving U.S. secondary school graduation rates came to a sudden and nearly complete halt in the mid-1960s and remained more or less fi…xed at that plateau for the next three decades. It seems implausible that this sudden stop in U.S. educational progress is attributable exclusively to a discontinuous, permanent deterioration in the quality of the U.S. secondary school system in the mid-1960s. Other factors are thus likely to be at play.

I do not suppose that they care to invoke Charles Murray to explain the “sudden stop.”

Let me invoke one of W. Edwards Deming’s insights about quality control. The later you intervene in a process, the more you have to spend to achieve a given level of improvement. Thus, if you wait until someone is 20 years old to try to improve their human capital, that is likely to be less effective than helping them at age 10. That in turn is likely to be less effective than helping them at age 3. (I read James Heckman as saying that the returns on trying to help before age 3 are much higher than those afterward.) And I read the segregation-seeking behavior of the affluent as suggesting that their internal model predicts that if you want someone to be highly educated as an adult, the only reliable strategy is to make sure that they are the offspring of highly-educated parents, who grow up in a milieu of other affluent children.