Richard Freeman accurately describes the recent rise in wage and income inequality in United States. He speculates about its consequences, and proposes five policies to alleviate it.

Wages were no more unequally distributed in 1990 than in 1940. They were even more unequally distributed during the Depression and in earlier periods of economic stress. Our democratic society was preserved then, albeit with strain. Democracy and the social order can survive at higher levels of earnings inequality than those experienced in the 1960s.

What is different now is the trend. After 1940, inequality was substantially reduced. The reduction came because the real wages of the least skilled increased more rapidly than the real wages of the skilled, although both grew substantially in the post-war period. A whole generation became accustomed to this trend which started early in this century. One could make a good living even if one dropped out of school.

What fuels the post-1980 rise in wage inequality is the absolute decline in real wages for the unskilled caused by a shift in demand away from them. Until recently, such a decline only occurred during depressions or severe crises when wages of all workers decreased. Now, however, the wages of the unskilled are declining when the wages of the skilled are rising. Expectations about success in the labor market have dramatically changed for a whole generation of unskilled workers. Standard measures of earnings inequality understate the severity of the problem because they exclude nonearners. Among the least skilled, nonemployment and lack of any earnings has substantially increased.

When Freeman and others write about the problem of inequality, they are concerned about this phenomenon, not inequality per se, although some of Freeman’s proposals seem directed toward solving the problem of inequality and not the problem of declining real wages. The principal issue is to raise living standards at the bottom. Redistributive policies aimed at eliminating inequality miss the point. Our current problem is not that some are doing better than others, but rather the despair among the least-skilled, who have become detached from the modern economy. If the trend continues, it will promote participation in the social pathologies of crime, welfare, and illegitimacy.

In addressing the problem of the unskilled, it is important to take a long view while at the same time recognizing the problems of social dislocation caused by the new labor market. In general terms, Freeman is right when he speculates that the most effective long-run policy is to target interventions early in life. More specifically, all the available evidence points to the great long-run value of raising the skill levels and motivation of the very young. Research in psychology and economics indicates that skill begets skill; early learning promotes later learning. Investment in the education and training of the very young earns a far higher return than investment placed in a teenager or middle age adult. John Donohue of Stanford Law School and Peter Siegelman of Yale Law School have estimated that the benefits of reduced crime alone more than pay for the costs of an enriched Head Start program targeted toward disadvantaged youth.1

With older workers, matters are more complicated. Policies that improve skills can help, but my research and that of others demonstrates that such policies are very costly, even if applied to unskilled and low-ability youth in their late teens, and certainly if applied to older workers. The economic return to education and training for these people is so low that a more useful policy would be to subsidize their employment. Job subsidies offer an attractive alternative to welfare: they promote employment, integrate the unskilled into the economy, and provide them, their communities, and their children with the dignity and social benefit of work. I have developed this argument elsewhere.2

Subsidies can take many forms. The Earned Income Tax Credit is one program that creates financial incentives for low-income persons to work. Other subsidy programs would induce employers to hire low-skill workers at high wages by compensating them for the gap between wages and productivity of their unskilled workers. Targeting job subsidies too narrowly and at too low a benefit level can backfire, however, and it is important to avoid stigmatization. If subsidies are given only to very low-skill workers with severe motivation problems, and the subsidy is not sufficiently large, research by Gary Burtless of Brookings reveals that employers view the subsidy as a warning label and do not hire subsidized workers.

In contrast, policies that reduce the demand for the unskilled, like minimum wage laws or increases in union wage scales, are generally bad ideas. They increase the wages of those who remain employed but at the same time have the perverse effect of increasing inequality among the least skilled, and reducing their employment.