Consider the Evidence

“Far from leaning against economic inequality, U.S. schools make it worse.” This sentiment, from a recent Clive Crook op-ed, expresses a view that’s commonplace on both the left and the right, and among both proponents and opponents of school reform.

It’s wrong. Americans do leave the schooling system more unequal in cognitive and noncognitive skills than when they enter it. Yet that inequality is less — probably much less — than it would be in the absence of schools. Schools don’t increase inequality; they just don’t do enough to overcome the inequality produced throughout childhood by differences in families, neighborhoods, peers, and other influences.

How do we know that? First, children are vastly unequal in ability when they enter the school system at age five or six. This is due partly to genetics and partly to environmental differences.

Second, we have evidence from the natural experiment that is summer vacation. During those three months out of school, the cognitive skills of children in lower socioeconomic status (SES) households tend to stall or actually regress. Kids in high-SES households fare much better during the summer, as they’re more likely to spend it engaged in stimulating activities. In his book Intelligence and How to Get It, cognitive psychologist Richard Nisbett concludes that “much, if not most, of the gap in academic achievement between lower- and higher-SES children, in fact, is due to the greater summer slump for lower-SES children” (p. 40).

Without schools this pattern would be magnified, and the gap in cognitive and noncognitive abilities at age 18 almost certainly would be much greater than it now is.

This by no means implies our educational system is doing fine. It could and should do much better at helping children from disadvantaged environments. But saying it currently makes things worse suggests the situation is hopeless. Instead of promoting reform, that undercuts it.

When social scientists first began noticing and studying the rise in earnings and income inequality in the United States, much of the focus was on technological change. The idea is that in the past generation technology — especially computerization — has advanced more rapidly than skills, so employers have bid up pay for those able to use and improve new technology and reduced pay for (or gotten rid of) employees less adept at doing so.

Though this remains perhaps the single most popular explanation, many are skeptical. In their book The Race between Education and Technology, Claudia Goldin and Lawrence Katz offer an especially compelling critique. They suggest that the pace of skill-biased technological advance actually hasn’t changed much over the past century. What distinguishes recent decades, they contend, is that growth of educational attainment has slowed. Here’s their key picture (the vertical axis shows the share with a college degree):

Among Americans born between 1875 and 1950, the share getting a college degree rose more or less continuously. According to Goldin and Katz, as they became a sizeable portion of the labor force (assume a lag of about 30 years), inequality held steady or declined despite technological progress. But among those born between 1950 and 1965, who became an important part of the labor force beginning around 1980, college completion was pretty much flat, falling for males and increasing just slightly for females. This, say Goldin and Katz, is the key to the rise in earnings inequality that began at that time.

I think there’s something to this story, but I’m not sure it takes us very far in understanding the rise in U.S. earnings inequality or that it points us toward a solution.

For one thing, comparative evidence doesn’t seem especially supportive of the Goldin-Katz hypothesis. The following chart shows changes in earnings inequality from 1979 to 2004 (the most recent year of available data) by changes in average years of schooling completed over the same period. There is a negative association, as Goldin and Katz would predict, but it’s mainly a function of the United States; if we remove the U.S. the relationship largely disappears.

Second, the Goldin-Katz story says pay for college graduates has jumped because their supply stopped growing around 1980. But the key features of the rise in U.S. income inequality are soaring incomes among the top 1% of households (especially the top 0.1%) and slow income growth in the bottom half of the distribution.  A slowdown in the supply of college graduates is unlikely to be the key to either of these two developments. It might seem implicated in the latter until we recall that college graduates have never accounted for more than a third of working-age Americans.

Finally, the above chart from Goldin and Katz indicates that college completion began rising again for cohorts born in 1980 and after. Does this mean earnings inequality will soon level off or perhaps even decrease? Absent other changes, I wouldn’t count on it.

Education is important for individuals and for society, and I certainly favor efforts to improve both its quality and its quantity. But it doesn’t seem to me likely to get us very far in reversing the rise in American income inequality.