You are using an outdated browser.
Please upgrade your browser
and improve your visit to our site.
Skip Navigation

Inequality is a Symptom, and More Years in School May Not Be the Cure

Raghuram G. Rajan has an excellent piece up on TNR’s website (“Let them Eat Credit”). Without trying to do too much violence to his argument, I would summarize it as follows: Growing income inequality in the United States has done tremendous damage to our economy. The most important cause for this inequality (supported by well-known research by Goldin and Katz) is that, although technological progress requires the labor force to have ever greater skills, our educational system has not kept pace by providing the labor force with greater sufficiently improved human capital. Rather than address this reality, politicians over the past decades have used expansion of credit to allow people to increase consumption through borrowing. This had reduced the political pressure to solve the underlying problem, but the chickens are now coming home to roost. If the central problem is that too much of the U.S. workforce is unqualified for the good knowledge-related jobs that are will be created by its economy, then the solution is to upgrade skills and education.  

I’m supposed to criticize here, but let me just suggest a few friendly amendments.

First, while I certainly agree that excessive inequality will create destructive social friction, I believe that it is better considered a symptom and statistical artifact than as a cause. The striking thing about Rajan’s chart of real income change between 2002 and 2008 isn’t so much the difference between the percentiles, but that most jobs paid less in real terms in 2008 than in 2002. That is, I don’t think most people would consider it a sign of dysfunction that real wages rose 1.7 percent for jobs at the ninetieth percentile between 2002 and 2008—the issue is that jobs at the tenth, twenty-fifth, fiftieth, and seventy-fifth percentile all pay less in real terms than they did six years earlier. The root problem is middle class wage stagnation, exacerbated by income inequality and declining income mobility.

Second, what is the cause? I agree with Rajan that America’s inability to produce a workforce with sufficient human capital to reliably grow income quickly for the broad middle class is the core driver. But where we differ is on the cause of the cause: I think that he overly privileges technology as a cause and schools as a solution. This misconception proceeds directly from Golden and Katz, who famously described the U.S. workforce as being in “a race between education and technology.” 

Goldin and Katz’s book is excellent. I cited it approvingly and used data from it in a National Review article last year (“Factory Guy”). Their mapping of the wage structure of early twentieth-century manufacturing industries is worth the price of the book by itself. Further, I fully believe in their basic thesis, at least at a high level of abstraction, which I would state as roughly: Workers in our economy are in a race between development of as-yet-non-commoditized cognitive capabilities on one hand, and wage reductions as capabilities that are commoditized through technological advances on the other. But, in my view, they fail to make the case either that additional years of schooling consistently create the relevant capabilities, or that technology is the uniquely important cause of commoditization. In other words, I would say that we are in a race between capabilities and commoditization, rather than education and technology.

What are other potential causes beyond technology of this commoditization that in turn is causing middle class wage stagnation and growing income inequality? It turns out that the list tends to make both parties squirm.

At the top is globalization. The decreasing relevance of large-scale war under Pax Americana combined with the economic re-emergence of Western Europe and Japan by the 1970s and the Asian heartland more recently have created trans-national labor pools through a mix of outsourcing, immigration, and importing labor content via shipped manufactured goods. We move the stuff, the jobs or the people; but, in all cases, labor in Indiana increasingly competes with labor in India. Ceteris paribus, this creates upward pressure on wages for the most skilled, and downward pressure on wages for the less skilled.

Other contingent factors have mattered as well. For example, American domestic production of oil peaked in 1971; oil imports doubled between 1970 and 1975; and OPEC was able to drive large price increases. This oil shock was directly regressive, but it also tended to disproportionately harm those industries that were the source of high-wage union jobs. As another example, in what is probably inextricably both cause and effect, "non-­distributive services" (finance, professional services, health care, and so on) became in 1970 a larger part of the private economy than goods-­producing industries. This shift to services tended to enhance the prospects of the cognitive elite at the expense of traditional industrial workers.

Finally, there is the elephant in the room: the breakdown of the traditional social norms among the less educated and successful.

The United States went through a huge set of social changes that began to have effects on a mass scale in the 1970s. American society has, to some extent, re-normalized and incorporated these new ways of living. The new normal, however, is different from the old normal. To begin with, certain strands of the old bourgeois consensus have frayed, and others have simply disappeared, at least for some parts of the ­population. The wealthier and better-educated segments of our society, for ­example, have re-established the primacy of stable families and revived their intolerance of crime and public disorder. But they have combined this return to tradition with very non-traditional attitudes about sex, ­masculinity, and overt piety.

But, while affluent and educated Americans are returning to this (somewhat) traditional family model, the poor and less educated are not. The gap between rich and poor today is also a gap in cultural norms and mores to a degree unparalleled in our modern experience. The overall divorce rate exploded in the 1970s but has since returned to just about its 1960 level for those with a college education. For the less ­educated, however, the rate has continued to climb—and women without high-school diplomas are now about three times as likely to divorce within ten years of their first marriage as their college-educated counterparts.

Child-rearing has seen a similar split. In 1965, almost no mothers with any level of education reported that they had never been married. Today, this still holds true for mothers who have finished college: Only 3 percent have never been married. But that figure stands in stark contrast with the nearly 25 percent of mothers without high-school diplomas who say that they have never been married. In fact, last year, about 40 percent of all American births occurred out of wedlock. And about 70 percent of African American children—as well as a majority of Hispanic children—are born to unmarried ­mothers. But this situation obtains for low-wage, non-­college-­educated whites as well: It is estimated that about 70 percent of children born to non-Hispanic white women with no more than a high-school education and income below $20,000 per year were born out of wedlock.

The level of family disruption in America is enormous compared to almost every other country in the developed world. Of course, out-of-wedlock births are as common in many European countries as they are in the United States. But the estimated percentage of 15 year olds living with both of their biological parents is far lower in the United States than in Western Europe, because unmarried European parents are much more likely to raise children together. It is hard to exaggerate the chaotic conditions under which something like a third of American children are being raised—or to overstate the negative impact this disorder has on their academic achievement, social skills, and character formation. There are certainly heroic exceptions, but the sad fact is that most of these children could not possibly compete in the contemporary American economy.

But, even if globalization, bifurcation of social norms, and other factors have combined with relentless technological change to create a problem, that does not mean that the best way to address them is to spend money to keep more people in schools for more years. The the evidence that this works isn’t there.

Study after study has shown that, at a minimum, there is no clearly demonstrable educational benefit from more aggregate spending on schools. The classic study is Hanushek’s meta-analysis of more than 400 published, peer-reviewed analyses of spending increases versus outcome changes showing, in aggregate, what is essentially a distribution of measured effects scattered around zero. To be fair, Hanushek had an epic methodology debate with Greenwald et al., who claimed to find a small positive effect. Brookings did a book in which a number of scholars weighed in, including Hanushek and Greenwald, and I think it would be hard to conclude upon reading it that there is good evidence of a relationship between aggregate spending and outcomes. I think it’s a fair statement of analytical opinion that there is no widely accepted evidence that there is a material causal relationship between aggregate spending and student achievement. For what it’s worth, this is consistent with accepted opinion, to my knowledge, in other OECD countries.

Rajan points, however, not only to aggregate spending on education, but also to evidence for specific kinds of innovations, such as new technology that might enable to escape “Baumol’s Disease,” and new curricular and other interventions subjected to rigorous randomized trials. But the macro effects of these technologies or interventions remain almost entirely speculative. The specific link in Rajan’s piece is to a paper that describes the experimental protocol for a program test that is currently underway, rather than one that has been shown to work. This is telling. Randomized testing of educational interventions appears to have been shockingly rare until about ten years ago, and, consistent with the experience of introducing such measurement into therapeutics, criminological, social welfare, and other areas, it has in recent years mostly served to show how rare it is to find any scalable, replicable intervention that can materially improve academic performance.

I believe that improvements to our schools are possible, need to be part of how we approach the problem of wage stagnation, and should be pursued aggressively. These reforms are almost certainly going to include dragging the education industry through the kinds of changes—becoming more entrepreneurial, flexible, and market-driven—that have transformed large-scale institutions throughout our society. This is not a matter of simply “being more like Apple” or whatever, and we don’t yet know the unique manifestations of this evolution in education. Rajan is right that it will require that we run experiments and “learn quickly from them and scale up those that are most promising,” But it is also the case that we will never really do that without a huge political battle to wrest authority from providers and give it to consumers.

And, more important, improving schools is not going to be close to enough. I believe strongly, for example, that we need a new approach to immigration that reconceptualizes immigration as recruiting rather than law enforcement. Though it is unfashionable to say it now, we need to figure out how to extend the market revolution to sectors of the economy that remain protected from it by political power. We have to figure out some way to prevent the self-destructive spiral that limits the social productivity of a huge fraction of the American population. We need to understand the limits to American power and attempt to match our defense expenditures to finite, achievable goals, rather than imperial fantasies. These examples are obviously drawn from a right-of-center perspective, and the list could be extended, but I think it gives some idea of the scale of the challenge that we face.