Putting the Gini Back in the Bottle

For some decades after the end of World War II, income inequality was not an important political issue in the United States. Voters, their own fortunes improving steadily and their children entering university, were not particularly sensitive to the subject. The Democratic Party claimed to speak for the working man and, after 1960, adopted the Black man; but income disparity remained in the subtext of the party’s discourse. Adlai Stevenson’s 1956 acceptance speech for the Democratic presidential nomination devoted just six sentences to income disparity; three of them referred to the family farm.

This state of affairs began to change sometime around 1990. The nation’s rate of economic growth inflected downward. Middle-class incomes didn’t stagnate but lost the sense of inevitable improvement upon which the country had come to depend. With the Iron Curtain raised, the world opened up; the faces of newly minted billionaires became more common in the daily press. By the end of the 1990s, a subject that once was the domain of professional statisticians could arouse passion and be exploited for partisan advantage.

One problem in forming an intelligent opinion on income inequality is its confounding complexity. This is a subject of numbers, each one of which has multiple moving parts and is subject to interpretation. Consider two commonly used measures of income disparity, the Gini coefficient and the share of national income earned by the top 1 percent. The Gini coefficient is the measure of the area beneath a Lorenz curve. It has been adopted by economists as a comprehensive index of income distribution: A Gini of 0 means everyone earns the same amount. A Gini of 1 means one person earns everything. Published Gini coefficients for the United States run from .39 to .51; the share of income going to the top 1 percent is said to be between 8.8 percent and 22 percent. The differences depend upon how one defines family income and, indeed, upon how one defines a family. Is income measured before or after taxes? Does it include government transfers—and, if so, which ones? Cash transfers like Social Security and unemployment insurance, or the really big ones that come in kind, like Medicaid? Should income include capital gains when earned or when realized? The entrepreneur who takes a firm public enjoys a big payday when selling some of the stock. Millions of middle-class families with IRAs and defined contribution pensions earn billions of dollars in capital gains almost every year—but realize them only over the course of their retirement, a great difference for the Gini coefficient.

Numbers can be and are manipulated. Even careful readers of the national press can believe that the problem of inequality is either a national crisis or a statistical illusion. Let us try to separate the science from the polemics. A warning: There are no simple answers.

In 2018 the U.S. Gini coefficient—using calculations by the Congressional Budget Office (CBO) based on income data reported to the IRS, ignoring tax payments, and including capital gains as realized and federal cash transfers but not entitlements—was .49. That is 10 percent higher than in 1988, when the Gini was just below .44, and some 15 percent higher than in 1970, considered by some as the last year of the Golden Age of postwar equality. The year 1988 was a watershed. The changes to the federal tax codes implemented since 1980 had wiped out most of the tax shelter industry, and the rich reported income much as they do today. In that year, a family in the 90^th percentile of income earned just over six times as much as a family in the 10^th percentile. By 2018 the same ratio was about ten; and for a family in the 95^th percentile, it was more like twenty.

The CBO provides a second measure that takes account of all positive and negative tax payments and also puts a cash value on federal and state entitlements by imputing to families the income they would need to buy the implied insurance. These changes lower the measured coefficient by just under 10 percent; and, since entitlements have risen so much faster than overall income, they eliminate much if not all of the increase in inequality since 1988. The economist Richard Burkhauser and his colleagues, writing in Papers and Proceedings of the American Economics Association, report about the same result for a slightly different period. As these data suggest, though market forces over the past thirty-some years have generated a substantial rise in income disparity, government intervention through graduated taxation and safety net programs has mitigated almost all of it.

This conclusion, while seemingly dramatic, is not too surprising in light of the size of entitlement programs. Medicare and Medicaid alone consume some 7 percent of U.S. GDP.

The left wing of the Democratic Party, trying to force through the largest peacetime spending bill in the nation’s history, hates both the CBO results and the inference drawn from them. The last thing they want to hear is that government programs have gotten much of the job done. Their strongest criticism of the CBO is that the big entitlements don’t provide actual income—i.e., cash. There is something to this point, but not a lot. True, recipients of entitlements, given the opportunity, might well sell their entitlement rights for a cash sum smaller than the government’s cost of providing them. We don’t know how much smaller; but treating the entitlements as dollar-for-dollar equivalents of cash income overstates their value. Nonetheless, only the ideologically blind would deny that these benefits have done much to elevate living standards and have ironed out a good bit, if not all, of the rise in inequality over the past three decades.

It is all but impossible to compare inequality across countries. Differences in definitions and reporting are just too great. The CBO relies upon income data from the IRS. What would their counterpart rely upon in Italy—or France, or any country where the wealthiest families own large private enterprises from which income reporting is often a matter of discretion?

These impediments, however, have not stopped the World Bank from publishing Gini coefficients for just about every country on Earth. For what they are worth, they show the United States at the head of the inequality league table, with a Gini of .41 compared to coefficients between .32 and .36 for the other large, high-income countries. The Scandinavian countries report even lower figures, but what does one make of a country with a population smaller and less diverse than Los Angeles County?

All of the measurement problems that beset U.S. data apply at least equally to other countries, including the treatment of capital gains. The relatively few households outside the United States that hold common stock do so through retirement funds, which smooth out gains, pay annuities, and dampen measured income disparities. The wealthiest families never report capital gains.

Despite all the foregoing caveats, almost no one denies that market forces have opened income gaps over the past thirty years and that without government intervention, inequality would be a lot worse than it is. Why is this the case? To understand the forces at work, we should concentrate on the work of economists who have done serious research in the field, some of whom come to it after long involvement with labor markets and income determination. Some of the plausible explanations are not economic but sociological. Indeed, some writers at mostly conservative publications write off the entire phenomenon of income inequality as due to either sociology or demographics. Among them, the two favorite explanations are “assortative mating” and changes in the patterns of family formation.

Assortative mating, the tendency for equally educated couples to marry, has at least a superficial appeal. Over the past fifty years, the fraction of married couples each of whom holds at least a B.A. has risen almost without interruption, from 11 percent in 1980 to over 25 percent by 2015. At the same time, the modal occupation of the female partner has gone from elementary school teacher to something like banker or lawyer. Hasn’t this increased income inequality? Unfortunately, no. Most of the rise in highly educated couples reflects the general rise in higher education. Assortative mating affects income distribution only if it exceeds the coincidence one would expect from random chance. In this sense it has actually declined over the past fifty years. Lasse Eika and his colleagues at the Federal Reserve Bank of New York have convincingly shown that assortative mating raises the level of measured inequality by about 5 percent, but it has done so for a long time and cannot account for the recent rise in measured Gini coefficients.

This fact, however, does not mean demographic factors have played no role in the trend toward inequality among families. Thanks to the rising divorce rate, especially at lower income levels, and the increase in out-of-wedlock motherhood, the number of single-parent households has exploded in the past four decades. If the number of single-adult households had simply risen along with the general population since 1980, there would be ten million fewer of them today. The calculations by Eika and his colleagues suggest this proportional rise has been responsible for perhaps as much as one-third of the rise in income disparities. A number of writers have dismissed this trend as itself a consequence, not a cause, of inequality. There could be something in that argument, but, at best, the causation goes two ways: Someone who wants to become poorer need only get a divorce.

More, according to the 2020 census data, half of all married college graduates have spouses with a high school diploma or less. That is, there are as many couples of mixed educational attainment as there are those who are uniformly educated. One doubts that this is true of graduates of the fifty or so most competitive schools in the country. Instead, there is clearly a second level of assortative mating going on: It is the graduates of the better schools, who then go on to higher earnings, who have a high propensity to marry one another. Researchers have so far ignored this part of the mating dance because the data are not easy to come by; but it could go some way to explaining the discrepancy between the common perception of assortative mating and Eika’s dismissal of it.

One characteristic of wage inequality that has captured the attention of researchers is its ubiquity. Kevin Murphy and Robert Topel of the University of Chicago report that the 99^th percentile of earners has pulled away from the 95^th, the 59^th from the 55^th, and the 19^th from the 15^th. Similar things have happened in Europe and in Asia, with a general worldwide rise in the premium paid to skill. Indeed, some writers maintain that all the rise in income disparity over the past forty years is due to the increased premium paid to higher education at all levels. Accordingly, economists have looked to broad trends in technological change for explanations.

David Autor of MIT and his co-authors, for example, have developed models emphasizing the steady rise in technological work requirements in the 20^th century. Events of the past few decades, they maintain, have simply continued this trend albeit perhaps at an accelerated pace. The great expansion of state universities in the mid-20^th century produced a steadily increasing stream of graduates, which kept wage premiums in check; but in recent decades, college graduation rates have crested. Murphy and Topel also note the stagnation in the supply of educated workers. The portion of male high school graduates who finish college by age twenty-three has been stuck at 15 percent since 1980. College graduation rates have continued to rise among men, but at ages between twenty-five and forty. These and other authors see the operation of demand and supply: The relentless evolution of technology has raised the demand for educated workers. The education system kept pace for a while, but in recent decades has either flagged in its efforts or approached the limits of its capacity.

Technology’s power to generate income disparity has been intensified by globalization, the flowering of international commerce that followed the end of the Cold War. The thought that international trade can depress the wage of low-skilled labor has a long provenance going back at least to the work of Paul Samuelson and Wolfgang Stolper in 1945. Samuelson and Stolper showed that the expansion of trade between a capital-rich country like the United States and a labor-rich country like China would lead to a—somewhat counterintuitive—redistribution of income toward capital within the United States. Defining capital broadly, to include human capital, produces something that sounds a lot like what this country has experienced.

Imports from China have grown by a factor of ten since 1990, with almost no proportional growth in exports. Even writers like Paul Krugman who were initially skeptical now concede that many Chinese imports are really at the low-skill end of the high-tech manufacturing process and are probably depressing manufacturing wages in the United States. Autor and his colleagues, studying manufacturing at the local level, estimate that 21 percent of the decline in employment in U.S. manufacturing between 1990 and 2007 can be attributed to imports from China. This decline displaced some 1.5 million workers who had to find work in other industries, further depressing wages.

As Thomas Piketty has emphasized, the move toward capital has been even broader and more powerful than Samuelson and Stolper envisioned, and it has produced winners as well as losers. Aside from the rise in the rewards to higher education, new ideas of all kinds are now worth many multiples of what they once were. Between 1980 and 1989, in constant dollars, the total market value of all firms at the end of their first day’s trading after their IPOs was $354 billion; for 2010 to 2019, the total was just under $2 trillion. Part of this change simply reflects the rise in market valuations; but it also reflects the fact that since the fall of the Soviet Union, a new firm can look forward to not just the United States but the entire world as its market. For much the same reason, bestselling novels now run to tens of millions of copies; paintings by living artists command millions of dollars before the paint even dries.

The share of corporate profits earned by U.S. companies from operations outside the United States has gone from 15 percent in 1990 to nearly 50 percent today. Many firms, like Nike, that once manufactured abroad exclusively for distribution in the United States now sell abroad as well. Just about the only U.S. contribution is the management. When the process began, the United States had a near-monopoly on the world’s better schools of business administration. Other countries have been slow on the uptake. MBA programs are now sprouting all over the world, but in the meantime it is another case of demand for skilled labor outpacing supply.

It is no surprise that a problem about which there is so little agreement has produced policy prescriptions bearing no resemblance to one another. Data from the Congressional Budget Office suggest two things: Existing entitlement programs have done a great deal to raise living standards at lower income levels, and the Earned Income Tax Credit is a powerful device for assuring the material sufficiency of low-skilled workers. This latter program has a long history, going back to the Nixon administration in the early 1970s, and could probably be expanded with a lot less political friction than the proposals generated by the current administration. Still, it would probably not be enough.

In a world divided between participants and spectators, those left out, regardless of their material conditions, inevitably suffer a sense of self-abasement. People need careers that will give them some self-respect. Several writers have proposed a variety of solutions, some of which are quite radical, like dismantling the entire meritocracy.

Ostensibly liberal writers inevitably turn to the nation’s public schools as the way to prepare more young people to live productive lives. Very few, however, admit the size of the restructuring needed to make the system effective. Even Krugman, whom one would think would know better, seems to believe it is merely a matter of money.

Gary Becker, who, before he died, was arguably the world expert on these matters, advocated a greater investment in education—but warned of disaster if the obsession with inequality were so great that Congress acted to lower the returns to the already educated rather than raise the level of those currently left behind. Even Becker, however, was not optimistic about the country’s capacity to expand its base of educated citizens without restructuring the way public schooling is delivered and without devoting considerably more resources to the task.

Paul DeRosa has taught economics at Columbia, ran a hedge fund for some thirty years, served on the research staff of the Federal Reserve Bank of New York, and has written extensively on topics related to economics.