Comments

Frank Levy[1]
(Ph.D, 1969)

April 17, 1999

Like some others here, I came to Yale through the advice of a faculty member. I was an undergraduate at MIT. In the spring of 1963, I had been accepted to graduate school by both Yale and Stanford. I still was making up my mind when I happened to bump into Ed Kuh whose econometrics class I had taken. Ed asked me about my decision and I said I was leaning toward Stanford. "I don't know," he said. "Jim Tobin is a lot to give up." I thought about that and eventually decided to come to Yale. On the principal that sometimes you catch a break, about a month after my decision I heard that Herb Scarf had decided to leave Stanford and come to Yale as well. Each man was an important influence on my graduate education. I owe both men a great deal.

Several people, including Janet Yellen, have talked about the Yale tradition and the place of social concern in that tradition. I want speak in that tradition today and talk about income inequality in the economy -- both its extent and its prospects for the future.

At different times, people have advanced a variety of reasons why the subject of inequality should not be on the table. In the early 1980s, the argument was that income inequality was not really growing in the first place -- the problem lay in faulty statistics. That argument is now discredited. In the early 1990s, the argument was that inequality largely reflected the economy's valuation of genetic intelligence and so there was nothing to be done about it. In the mid-1990s, the argument was that inequality was not a cause for concern because earnings mobility within the income distribution was so strong (IQ not withstanding) that cross-sectional earnings inequality measures meant nothing. Most recently, we have the argument that inequality is caused by Pareto-superior changes that leave some people richer and everybody else no worse off.

As I read the data, none of these arguments withstand close scrutiny.

Where, then, are we? According to Census statistics, income inequality today is higher than it was in 1947 when Current Population Survey data was first collected by the Census. As Dan Weinberg can attest, these Census numbers have significant problems.[2] Nonetheless, I believe the 1947-1997 comparison is roughly correct and that should give us some pause.

In 1947, the bottom one-fifth of families was dominated by two groups. One was farm and other rural families with very low cash incomes. The other was the elderly who still worked in large numbers since Social Security and the private pension system were still in their infancy. As the number of farm families shrunk and Social Security and the private pension system matured, ceteris paribus, inequality should have declined.

But, as we have seen, inequality has not declined and I count at least four reasons for this.

One is the growing diversity of family structures: more families headed by single women on the one hand, more families headed by two earner couples on the other.

The second is the rising rate of return to education to which others today have referred. This rise is driven by a mix of technology and by international trade. There is, as you know, an extensive literature on just what this phenomenon means: the extent to it represents a payoff to time spent in the classroom versus a payoff to the characteristics of children who spend time in the classroom - their social skills, their genetic intelligence, and so on. My own sense is that it is a mix of these things including, as Jim Heckman emphasizes, social attitudes and skills that, while not genetic, are formed early in life.

The third factor is the way in which transportation and communication have combined local markets into national markets. Consider, for example, Yale College. In 1969, there were about 15 million persons between the ages of 16 and 19. (I use this readily available statistic as a crude proxy for the number of young people who might be interested in college.) According to Caroline Hoxby of Harvard, who was kind enough to help me with these figures, Yale College in that year had approximately 6,800 applications for about 1,000 places. In 1999, there were again roughly 15 million persons between the ages 16 and 19. This year, there were about 14,000 applicants for 1,250 places.

Many things have changed since 1969 beginning, as my wife reminds me, with Yale's full admission of women. But another thing that has changed is the increasingly national nature of labor markets in which a good brand name on a diploma is a real advantage. Thus a good student from Iowa now thinks about Yale where his or her 1969 counterpart might have only thought about the University of Iowa or Grinell.

There are, finally as Dan Hammermesh has noted, issues of attitudes and institutions -- the appropriate role of unions, the appropriate level of executive compensation, and similar subjects which come into relief when we compare the U.S. wage patterns to wage patterns in other countries. A case in point has arisen in the recent Daimler-Chrysler merger because Daimler executives -- executives of the acquiring company -- earn much less than Chrysler executives do.

Together, these factors help to explain the growth general inequality in the middle of the income distribution and the growth of very high incomes at the top of the income distribution. For example the U.S. Treasury's Statistics of Income reports that tax filers with AGI over $200,000 in adjusted gross income (AGI) -- about 1.3 percent of all tax returns -- account for 19.7 percent of all Adjusted Gross Income tabulated by the Treasury and account and pay something over one-third of all federal income taxes.[3]

All of this suggests that today's inequality is not only somewhat larger than in 1947 but different in its substance. When inequality reflected geography more than education, migration was a ready remedy. A 20-year-old man in a depressed area could take stock, see his prospects were limited and migrate to an urban area where, in many cases, he had the skills to get a factory job.

Today, a 20 year-old with limited prospects is likely one with a poor education. There are important steps he can take - junior college is a prime example -- but, as Heckman notes, by age 20, much of the die has been cast.

I do not believe the skill-biased nature of technology and trade will change any time soon. So it is incumbent on us to find institutional arrangements to help offset what the market is doing. There are no magic bullets here but the profession needs to be investigating the kinds of policies, including early interventions, that can make a difference.

In my student years, Yale had a reputation as a liberal, activist department but this is not just a liberal enterprise. Supporters of markets should favor it as well.

The reason is straight forward and is illustrated by a New York Times article today reporting on an Allan Greenspan speech in Texas. Greenspan argued that just because free trade hurts some workers, we cannot block continued expansion of free trade. The answer, he seemed to say, was working out methods of compensation.

Greenspan is exactly right. Ultimately, the rationale for economic growth is that when the pie expands, the winners have enough income to compensate the losers and still be better themselves. When the economy is enjoying Janet Yellen's 4.3 percent unemployment rate, such compensation may not be an issue. But when the next recession comes, inequality has the potential to create a situation where free markets are seen as just one more special interest enterprise. Various "European" market restrictions - trade barriers, mandates on job security, etc. -- will gain in appeal and the market economy will suffer. It is a result we should all try to avoid.

[1] Daniel Rose Professor of Urban Economics, Department of Urban Studies and Planning, MIT.

[2] Dan Weinberg, Yale Economics Ph.d. is the Chief of the Income Division of the U.S. Census and also attended the reunion.

[3] To keep this figure in perspective, it is important to note that many families with incomes below $25,000 have no federal income tax liability (though they do pay payroll taxes).