The Truth About Income Inequality

John H. Hinderaker & Scott W. Johnson
Center of the American Experiment Minneapolis, Minnesota
December 1995

Executive Summary

Dramatic stories portraying a widening inequality of income and a disappearing middle class in America have permeated the media in recent years. These stories have been based, almost universally, on biased analyses which manipulate and distort the underlying economic facts.

Data generated by the Census Bureau, the Bureau of Labor Statistics, the Federal Reserve and other nonpartisan sources refute claims commonly made by the left. For example, data from such agencies show that:

  • Differences in family income largely reflect differences in how many members of a family actually work — and how hard they work.
  • Income is about as widely distributed today as at any time in American history, and income is nowhere nearly as concentrated among top earners as it was in the period before World War II.
  • Americans in all income groups have prospered, or have failed to prosper, together. Gains by upper-income Americans have not come at the expense of middle- or lower-income Americans. Nor has anyone else gained in those periods when higher-income families have lost ground.
  • The best era in recent history for middle-income and lower-income American families was the Reagan boom, which lasted from 1982 to 1989. During that period, middle-class families saw their real incomes grow by an average of 12.6 percent, while lowest-quintile families saw an average increase in real earnings of 12.9 percent.
  • There is a remarkable degree of income mobility in the American economy. Today’s “poor” are tomorrow’s “rich,” and vice versa. To a considerable extent, income distribution reflects obvious demographic factors. For example, workers in their prime earning years generally make much more money than those who are starting out in life or are retired.

Basic economic data do not support the gloomy picture promoted by the left. Moreover, the data show that the last 25 years do not represent an undifferentiated upward or downward trend for middle- and lower-income Americans. The Carter years, for example, were an economic disaster; the Reagan years were a bonanza; and the years since have been characterized by stagnation and downward drift.

The data certainly do not suggest that higher taxes, increased government spending, or hyper-regulation of the economy will benefit middle-income or lower-income citizens.

In recent months, the debate has been expanded to include the distribution of wealth (as opposed to just income). Liberal claims here, too, are based on manipulated data and faulty reasoning. More fundamentally, the wealth inequality debate should focus on what public policies will aid the accumulation of wealth by more, not fewer, American families.


To borrow a fistful of boxing metaphors, pound for pound and jab for jab, this paper on income inequality pummels as many myths into submission as has any American Experiment publication over the last half-dozen years. In doing so, John H. Hinderaker and Scott W. Johnson cite numerous examples of statistical mistakes and misrepresentations when it comes to how income and wealth have come to be distributed (or more accurately, earned) in the United States. Just as importantly, they show how such fictions have already seriously contorted national debate and policymaking.

In simplest form, “The Truth About Income Inequality” argues that our nation is not growing economically skewed in unprecedented and undemocratic ways, contrary to rising assertions in various quarters about the supposed unfairness of the “Reagan” and “Republican” revolutions. Sure, loads of numbers have been thrown about since the 1980s “proving” that America is growing less fair. But as Messrs. Hinderaker and Johnson persuasively demonstrate in detail, those numbers and interpretations are flat-out wrong. Here’s an example:

The left routinely issues warnings about the galloping death of the middle class, arbitrarily defined here as families with incomes between $15,000 and $50,000 annually, in constant 1990 dollars. As Messrs. Hinderaker and Johnson write, there is “a kernel of truth in this contention,” but not as commonly understood. Yes, they acknowledge, a smaller percentage of families were in this income category in 1990 compared to 1970. But all of the net statistical decline, they show, was caused by members moving into a higher income category, above $50,000.

Having said and relayed this, you should know that I’m generally one of the first in my circle of associates and friends to recognize that making a living — supporting a family — can be increasingly tough for young and poorly trained men and women. Certainly it was more possible a generation and two generations ago for ill-equipped breadwinners to actually win, or at least scrape by, than is frequently the case now. This is but another way of saying that the economic premium on a decent education has unsurprisingly and rightly grown over time.

Likewise, given a decade and more of corporate downsizing and “reinventing,” I would like to think myself quick to recognize that many workers, perhaps especially older ones, feel less secure about their jobs and livelihoods than they once did. Once more, as with the nervousness of many younger workers above, these worries can be well-grounded.

But as the following pages meticulously show, in no way are such understandable concerns synonymous with a radical skewing or fraying of America’s economic compact. As the authors make convincingly and quantitatively clear, the United States remains a nation of extraordinary opportunity and mobility. Or, from another angle, the evidence is bountiful that the “rich” are not getting richer, as is routinely claimed, at the expense of a newly trapped caste of left-behind Americans.

This is the second time that John Hinderaker and Scott Johnson have teamed up on a Center paper, as they wrote the very good “Lani Guinier Reconsidered” in August 1993, right after her nomination as U.S. Assistant Attorney General for Civil Rights appropriately collapsed. Mr. Hinderaker received his law degree from Harvard University and his undergraduate degree from Dartmouth College. Mr. Johnson did his law work at the University of Minnesota and also attended Dartmouth as an undergraduate. They both practice law in Minneapolis and write frequently as a team for local and national publications.

“The Truth About Income Inequality” is not only a superb piece of argumentation, it is also a first-rate piece of journalism, as in addition to mastering arcane documents, Messrs. Hinderaker and Johnson interviewed key researchers and players in and out of government in this exceedingly technical, not always high-minded controversy. The essay in hand is a major breakthrough and contribution to straight talk and intellectual honesty, and my colleagues and I in American Experiment are terrifically proud to publish it. My great thanks to them.

My large thanks as well to Mr. and Mrs. Fred T. Lanners, Jr., for their generosity in underwriting this project and in assuring its wide dissemination. They are valued, long-time supporters of the Center.

American Experiment members receive free copies of almost all Center publications, including “The Truth About Income Inequality.” Additional copies are $4 for members and $5 for nonmembers. Bulk discounts are available for schools, civic groups and other organizations. Please note our phone and address on the previous page for membership and other information.

Thanks very much, and as always, I welcome your comments.

Mitchell B. Pearlstein
December 1995


The proposition that income inequality is a problem, and a growing one, has become a staple of the left. Assertions of a “widening income gap between rich and poor Americans” and a “disappearing middle class” appear daily in the nation’s newspapers as the refrain of liberal politicians and sympathetic journalists.1 By dint of repetition these assertions have attained the status of conventional wisdom; most commentators no longer consider it necessary to cite evidence to support them. Examples of this phenomenon could be multiplied endlessly, but a typical instance is the claim by Michael Lind, senior editor of the New Republic writing in the New York Times Book Review, that “income inequality in the United States has reached proportions not seen since the Great Depression.”2 No authority was deemed necessary to support this rather startling claim — which, as the data set forth below show, is entirely false.

Not surprisingly, politicians — usually Democrats, but also including such Republicans as Pat Buchanan — have seized on this conventional wisdom and sought to turn it to partisan advantage. In their hands, assertions of unprecedented income inequality invariably accompany pleas for higher taxes, more government spending on social welfare programs, or expanded government regulation of the economy. Some commentators and politicians call explicitly for redistribution of wealth from the rich to the poor.3

Before adopting measures that seem calculated to hamper economic growth and retard the creation of wealth generally, the relevant data should be scrupulously examined. The purpose of this paper is to review the basic data on income distribution in America, both contemporary and historical, in order to evaluate the truth of the proposition that the distribution of income is unfairly and increasingly skewed toward the rich, to the detriment of the poor and middle class. Equally important, we will try to put the data on income distribution into the larger context of work, achievement and opportunity in a mostly-free economy.

For it goes without saying that in a free economy, some people will always earn much more money than others, and a given person will, during some stages of his life, earn much more than in other stages. It is therefore far from clear that there is such a thing as a “right” distribution of income, and to our knowledge, no critic of current patterns of income distribution has set forth any meaningful criteria by which the “right” income distribution can be determined. Instead of articulating such criteria, critics have generally been content to argue that income inequality is “increasing” or “widening” and have assumed that it goes without saying that such a trend, if true, would be bad. Both the belief that income inequality is increasing, and the assumption that any such increase must necessarily be bad, need to be critically examined.

Inequality of income and inequality of work

Discussions of income distribution are inherently misleading in one fundamental respect: Income is not distributed, it is earned. This fact, while obvious, is often overlooked. Upper-income Americans are routinely derided as “the rich,” as though they typically achieved their status by clipping coupons. Both politicians and journalists often seem to think that “the rich” are at best lucky, and at worst criminal. The implicit message is that upper-income citizens should be grateful that the government allows them to keep some of their money. Richard Gephardt’s reference to people who earn good incomes as “winners” in “the lottery of life” is one infamous example of this attitude.

But upper-income Americans did not win a lottery. To a remarkable extent, differences in income relate directly to differences in work. For the most part, upper-income American families do better than lower-income families because they work more.

In 1990, of the families that comprised the lowest 20 percent in household income — the bottom quintile — 55 percent had no income earners at all. And even among the bottom-quintile family members who were employed, only 24 percent worked full-time. On the other hand, in the highest quintile of income (i.e., the top 20 percent), 83 percent of families had two or more income earners. And, amazingly, 11 percent of families that ranked in the top income quintile included four or more income earners.4

In other words, most families with high incomes contain no corporate chief executives or professional baseball players, much less lottery winners. The “fortunate fifth,” in Robert Reich’s endlessly repeated phrase, isn’t fortunate — it is hard-working. Most upper-income American families simply work more jobs and longer hours than their neighbors. It is fair to question whether the fact that harder work is typically rewarded with higher incomes really constitutes “inequality.”

The supposedly disappearing middle class

Since the late 1980s, liberals have argued that the middle class is disappearing. They have portrayed America as a country increasingly divided between rich and poor, with fewer and fewer families in between. Statistics have been manipulated shamelessly to make this claim appear plausible. In particular, many left-of-center commentators have argued that, since 1970, the percentage of American families in the “middle class” — defined as those with incomes between $15,000 and $50,000 — has declined.

There is a kernel of truth in this contention, as the chart below shows. The percentage of families in the arbitrarily defined “middle class” did decline between 1970 and 1990, from about 63 percent of all families to about 55 percent. But this statistical “decline” of the middle class did not occur because its members were losing ground. On the contrary, all of the net statistical “decline” of the middle class was caused by its members moving into the higher income category — families with incomes in excess of $50,000 in constant 1990 dollars. That category grew from 24 percent of the population in 1970 to 32 percent in 1990.

Distribution of Family Income

In other words, the middle class hasn’t declined, it has become more prosperous. While some politicians may be nostalgic for the days when a large majority of Americans were locked into an income level arbitrarily defined as “middle class,” it is hard to see any reason to complain about the fact that many people have moved into a higher income category.

The table below gives a more detailed picture of the fortunes of middle-income Americans. The table divides the population into five income quintiles, and charts the average income of each quintile from 1973 to 1991. All figures are in constant 1991 dollars, so all gains are real and not the result of inflation.

Real Average Family Income Since 1973 (in constant 1991 dollars)

Year Lowest Fifth Second Fifth Middle Fifth Fourth Fifth Highest Fifth
1973 10,746 23,451 34,457 47,090 80,794
1974 10,584 23,104 33,811 46,321 79,216
1975 10,205 22,226 33,119 45,351 77,481
1976 10,444 22,780 34,068 46,549 79,567
1977 10,282 22,865 34,483 47,588 81,584
1978 10,599 23,588 35,499 48,991 84,099
1979 10,765 23,750 35,870 49,395 85,589
1980 10,199 22,904 34,695 48,140 82,433
1981 9,782 22,126 33,958 47,682 81,741
1982 9,256 21,785 33,370 47,332 83,371
1983 9,236 21,823 33,648 47,964 84,381
1984 9,547 22,413 33,658 49,563 87,341
1985 9,675 22,711 35,132 50,356 90,627
1986 9,990 23,501 36,471 52,115 94,926
1987 10,157 23,872 37,069 53,053 96,956
1988 10,197 23,848 37,111 53,298 97,792
1989 10,359 24,184 37,571 54,055 101,780
1990 10,247 23,900 36,808 52,935 98,377
1991 9,734 23,105 35,851 51,997 95,530
Source: Census Bureau

A number of conclusions can be drawn from this table. Perhaps the most obvious is that all income groups rise and fall together. When upper-income Americans prosper, so do middle-income and lower-income Americans. Conversely, when high earners are stagnating, so are income-earners in all other categories. Critics of income inequality often state or imply that the prosperity of upper-income Americans comes at the expense of other groups. They believe that economics is a zero-sum game, so that, in George Gilder’s classic formulation, wealth causes poverty. But history provides no support for that claim.

Focusing specifically on middle-income Americans (i.e., the third quintile), the table shows that the average income of the middle quintile declined in real terms from 1973 to 1975. It then increased from 1975 to 1979. There was a substantial decline from 1979 to 1982, reflecting the “stagflation” of that era.

Then, from 1982 to 1989, the economic well-being of middle-income Americans steadily increased. Average income grew every year during that period, adding up to a 12.6 percent real growth in middle-quintile family income. In 1989, middle-class family income reached its highest level in American history. While average middle-quintile incomes have fallen off since then, the data yield no support for the idea that the middle class is undergoing a historic decline in its fortunes.

Moreover, the data in the above table, which are based on family income, unfairly understate the growth of real income. Since the 1970s numerous social trends have reduced the number of adults in the average family. Some of these trends are benign; e.g., an increasing average age of marriage. But by far the most important trend reducing the size of families, and thereby depressing median and average family income, has been the high divorce rate, combined with increasing numbers of people who have children but who never marry. The impact of these trends on family income is obvious: If two married people work, and if they each earn $50,000, they have one household with a $100,000 income. If they divorce (or, for that matter, if they never marry), they are two households, each having a family income of $50,000 — representing, in the economic statistics, a 50 percent decline, even though the total amount earned is the same.5

Thus, the fragmentation of the American family has significantly depressed statistics based on family incomes. A truer measure of economic trends, as opposed to social trends, is individual income. As Thomas Sowell and others have pointed out, real individual income — i.e., income per person in constant dollars — grew by about 40 percent from 1969 to 1992, far outpacing average growth in family incomes.6

The chart below shows real per capita income in constant 1992 dollars over the period 1970-1992.

Per Capita Money Income

Year Constant Dollars (1992)
1970 $10,793
1975 12,028
1980 13,275
1985 14,360
1989 15,904
1990 15,444
1992 15,033
Source: Census Bureau

These Census Bureau data show substantial and reasonably steady gains in real personal income over recent decades. Individual income figures, like the family income data reviewed above, demonstrate that far from losing ground, the American middle class has significantly improved its economic well-being since 1970.

Concern over middle-class incomes is often linked to the fear that high-paying manufacturing jobs are being lost to foreign competition, and are being replaced by low-paying or “dead end” jobs in the service sector. While it is undeniable that the American economy in recent years has created an enormous number of new jobs, news reports often deride these jobs as mere “hamburger flipping.” If it were true that the modern economy were creating only or predominantly low-paying jobs, there would be legitimate grounds for concern about the future of middle-income Americans. But again the facts fail to support doomsayers’ claims.

While the data are not all in on the current expansionary phase of the economy, the job creation that occurred during the expansion of the 1980s has been thoroughly analyzed. It is well known that between 1982 and 1989, the economy created more than 19 million new jobs. What is less well known is that a large majority of these jobs were in categories normally regarded as high-quality. The chart below shows job creation in the 1980s, broken down into the classifications used by the Bureau of Labor Statistics.

Jobs Created January 1982-December 1989

Job Category Number (Millions) Percentage Increase % 1989 Median Earnings
Managerial/Professional 7.600 33.1 $32,873
Production 2.194 19.0 25,831
Technical 6.630 21.8 20,905
Operators 1.374 8.2 19,886
Services 2.120 16.8 14,858
Farming -0.116 -3.7 13,539
Total/Average 19.892 20.3 $23,333
Source: Bureau of Labor Statistics (employment); Bureau of the Census (earnings).

Fully 89 percent of the jobs created during the expansion of the 1980s were classified by the Bureau of Labor Statistics as Managerial/Professional, Production and Technical — the three highest-paying categories. The largest growth was in the Managerial/Professional category, and the growth in production employees almost exactly equaled the growth in service workers. Once again, the data provide a far more positive picture of middle-class prosperity than the conventional wisdom purveyed in the American press.

None of this is to suggest that economic conditions are perfect for middle-income Americans. In particular, rising tax rates have squeezed the after-tax incomes of most Americans. Middle-income Americans have indeed lost purchasing power since the historic peak of 1989, the end of the Reagan era. And the data accumulated so far seem to indicate that the current expansion is not producing the same proportion of high-paying jobs as that of the 1980s. But on balance it is simply wrong to suggest that the American middle class is shrinking or is in the throes of a long-term decline.

What about the poor?

One constant theme of the liberal critique of the economy is that the rich are getting richer while the poor are getting poorer. Superficially, Table 2 above might seem to support that thesis, since, while the fortunes of upper-income and lower-income families have risen and fallen together, the net effect of 18 years of ups and downs has been an 18 percent growth of real incomes in the highest income quintile, and a 9 percent decline for families in the bottom quintile.

This comparison, however, is badly distorted by the family breakup that has decimated poor families in the last two decades. More than any other group, low-income Americans’ average family incomes have been depressed by family fragmentation.

This point is illustrated by a related set of data, covering a slightly different time period and focusing on Black Americans. Between 1967 and 1988, real income per Black household rose only 7 percent, suggesting stagnation in Black incomes. But this suggestion is misleading. Over the same time period, real income per Black person rose by 81 percent!7 The difference largely reflects the effect of family fragmentation on average family income.

Moreover, as the statistics cited above show, the fundamental reason why the bottom 20 percent of families don’t make much money is that they are not working much — as of 1991, 55 percent of those families had no member who earned any income at all. Given that the poorest segment of the population, however defined, consists of people who earn little or no money, and given that technological progress will continue indefinitely to increase both the total amount of wealth created and the amount that can be earned by the most successful and hard-working individuals, it is simply a fact of life — or a fact of statistics — that the “gap between rich and poor” is destined to grow. This fact in itself is not noteworthy. The real issue is whether those at or near the bottom of the economic heap still have the opportunity to move up.

Opportunity and income mobility

Conventional discussions of income distribution proceed by dividing the population into income groups — generally quintiles — and following the progress of those quintiles over time. This is what Table 2, for example, does. Such analysis obviously can be useful. But it also tends to be misleading, because it is hard to discuss the gains and setbacks of various income groups without implicitly assuming that the same individuals persist in those groups over time. In other words, when we say that Table 2 shows bottom-quintile families losing ground to the stagflation of the late 1970s, then making steady gains during the Reagan boom, and then losing ground again in the 1990s, it is natural to visualize a single family and to conclude that such a family wound up, when all was said and done, about where it started out. This assumption could lead to the pessimistic conclusion that America is, after all, a stratified society where the poor are doomed to remain poor, and to drop ever farther behind the growing prosperity of those at the top of the ladder.

Such a conclusion, however, would be utterly false. The fact is that America, to an extent that may be unprecedented in its or any other country’s history, is a land of opportunity and upward mobility. When we follow individuals through time, as opposed to following statistically defined income groups, we find a picture of growth and opportunity.

Several recent studies have shed light on this critical issue. In 1992, the Treasury Department analyzed tax returns filed by 14,000 taxpayers for all of the years from 1979 to 1988. For each year, they determined the distribution of income among all 14,000 filers and assigned each tax filer to an income quintile for that year. They then compared the economic status of each tax filer in 1979 with the status achieved by 1988.

The results were startling. The Treasury Department study found that very few of those who started out poor (i.e., in the bottom 20 percent of income) in 1979 remained poor. On the contrary, 86 percent of those who were in the bottom quintile in that year had moved to a higher quintile by 1988. And the Treasury Department found that a person who was “poor” in 1979 was more likely to have become “rich” by 1988 than to have remained “poor.” Only 14.2 percent of those in the bottom 20 percent in 1979 were still in that category in 1988, while 14.7 percent had not only improved their lot, but had moved all the way to the top income category.8

This study tells us something critically important about America’s “rich” and “poor.” They are, in large part, the same people at different times in their lives. “Poor” people certainly include some who are demoralized, disabled or dissolute. But to a greater extent, “poor” family units consist of students, young people who are just starting out in life, recent immigrants, and retired people. The authors of this article, and many of those who will read it, were once classified as “poor” in government statistics. This year’s bottom 20 percent are not the same as last year’s, or those of the year before. Over time, the vast majority of low-income people find jobs, work their way up, and spend their peak earning years in far more prosperous circumstances. They are replaced at the bottom by young people, immigrants and others who then proceed to work themselves into a higher category.

Income mobility, of course, is a two-way street. The “rich,” like the “poor,” are an ever-changing group. Contrary to popular conviction that the rich only get richer, the Treasury Department study found that of those taxpayers who had ranked in the top 1 percent in income in 1979, 58 percent had dropped out of that category by 1988.

Further confirmation of the remarkable degree of mobility in the American economy comes from a second study, conducted for the Urban Institute by Isabell Sawhill and Mark Condon. Sawhill and Condon followed a procedure somewhat like the Treasury Department study. They identified a group of families and recorded their incomes in 1977, arranging them into quintiles of family income for that year. Then, in 1986, they recorded the family incomes for the same people. Rather than arranging the 1986 incomes into quintiles, they calculated the average family income, as of 1986, for each of the five groups that had been identified in 1977. The results of the study are shown below.

American Family Income

Quintile 1977 Quintile Members
in 1977
1977 Quintile Members
in 1986
Percentage Change
Lowest $15,853 $27,998 77
Second 31,340 43,041 37
Middle 43,297 51,796 20
Fourth 57,486 63,314 10
Highest 92,531 97,140 5
All 48,101 56,658 18

Sawhill and Condon found that the group that experienced the largest percentage increase in income, on average, were those who started out in the bottom quintile. Those families saw their real incomes increase a remarkable 77 percent, on average, by 1986. The second-poorest group experienced the second-largest increase — an average of 37 percent. The smallest average increase between 1977 and 1986 was that of the highest income group. Families in the top 20 percent in family income saw their average real incomes increase by only 5 percent in nine years. On the basis of these data, Sawhill and Condon characterize the period from 1977 to 1986 as an era when the poor were getting wealthier fast, and the rich were getting wealthier slowly.9

These income mobility data are of tremendous significance to the income inequality debate. They show that much of what passes for inequality between rich and poor is really inequality between generations.10 America is not a country in the process of stratifying into permanently separate and unequal camps, one rich and one poor. Rather, it is a country that offers its young people and immigrants unprecedented opportunity to achieve financial success.

If the rich person and the poor person are, in large part, the same person at different stages of life, it is reasonable to ask whether there is anything objectionable about the “inequality” that results from the fact that a mature, experienced, established worker can earn more money than one who, in the early stages of his career, lacks those advantages. It appears clear that far from being objectionable, this type of “inequality” is an inherent and desirable characteristic of an advanced society.

This point can be made by a simple illustration. Imagine two societies. The first is a primitive peasant society in which the typical employment is near-subsistence farming. The average “career path” is flat. A person can readily learn the simple skills necessary to earn his livelihood and, having learned them, his productivity and consequently his earnings will be essentially constant, until failing strength and health lead to a declining income in his later years. In such an economy, a typical earnings history might look like this:

Next imagine a more specialized, more modern economy, where technical skills, experience, a network of relationships and mature judgment are all conducive to advancement. A typical earnings history in such an economy might look like this:

In the first economy, there will be far less income inequality than in the second. Demographic factors — the fact that experience and the qualities that go with experience can significantly increase earning power — will make the second, more advanced economy far more unequal in its distribution of income. At any given time, experienced workers will be earning much more than less experienced workers. Yet virtually anyone would consider this a good thing. In the second society, opportunity is far greater and the typical person will attain significantly greater material well-being over his or her lifetime.

Therefore, it is not at all clear that “increasing income inequality,” which is universally decried by pundits, is necessarily bad. Widely unequal earned incomes are the desirable concomitant of an advanced economy. Until left-wing critics put forth a coherent theory as to what constitutes the “right” income distribution — a theory that must, among other things, take into account the demographic factors reviewed above — there is no clear ground on which to attack the pattern of income distribution now prevailing in the United States.

What about the rich?

Most of the data discussed above are organized in terms of income quintiles. While this approach is imperfect for the reasons already discussed, it is a widely used and practical way of tracing the fortunes of upper-, lower- and middle-income Americans. But of course, nowhere near 20 percent of Americans can be considered “rich,” and much of the publicity that has surrounded the income inequality issue has been directed at the rich, often embodied as corporate executives and professional athletes. Some readers will no doubt want to know whether the 1990s are really a golden era for the genuinely rich, as critics claim.

Since any era is a good time to be rich, this question calls for historical perspective. And in historical context, it is simply not true that top income earners today are faring unusually well. On the contrary, it appears clear that until quite recently, income was much more unequally distributed than it is today. The chart below shows the percentage of the total aggregate personal income that was earned by the top 5 percent of income earners from 1913 to 1991.

Percent of aggregate income earned by top 5 percent of income earners

Source: 1913-1946, Eugene Smolensky and Robert Plotnick, Inequality and Poverty in the United States: 1900 to 1990; 1947-1990, U.S. Bureau of the Census, Current Population Reports, Series P-60.

In 1913, the richest 5 percent of the population earned about 30 percent of aggregate income (excluding capital gains). That proportion stayed about the same until 1933, when it slowly began to decline. There was a more rapid decline during World War II, and by 1947 the share of the top 5 percent of income earners had fallen to 17.5 percent of total income. It then fell to about 15.5 percent, where it stayed through most of the 1960s and 1970s, and inched back up to around 17.5 percent in the 1980s, exactly where it was in 1947. Thus, by historical standards, the share of income being earned by the top income group is relatively low, and has remained essentially unchanged in recent years.

The other significant point that emerges from the above chart is that, within a context of long-term decline in the proportion of income earned by the top group, income inequality tends to widen in times of prosperity, and narrow in times of recession. Thus, short-term drops in the income share of the top 5 percent occurred in the recession years of 1949, 1957-58, 1973-74, 1980-81, and 1990.

Upon reflection, this should not be surprising. In times of prosperity, more people are earning more money, and it takes a higher income to qualify for the top 1 percent, 5 percent, or any other statistical yardstick. Nearer the bottom of the income spectrum, incomes tend to be more fixed. So it should be expected that in boom periods, income inequality may increase somewhat, while in times of recession, income inequality will tend to narrow. And, in fact, the table above shows that this is exactly what has happened throughout the post-war era. “Increasing income inequality” is a statistical construct that, in practice, is likely to be synonymous with prosperity.

Putting the statistics to one side, there is no question that at the extremes of talent, industry or luck, a person in today’s economy can earn an extraordinary amount of money. Amounts earned by the most successful individuals do seem stratospheric. Why is it that Madonna earns more money than did Elvis? Why does Kirby Puckett earn more than Ted Williams did? And why does Steven Spielberg earn more than Howard Hawks did? Superior talent is probably not the answer. The modern world has created opportunities for the talented and ambitious to leverage their abilities beyond what was possible in past eras. Billions of people around the world buy compact discs, attend movies, and watch television. Worldwide communication and global markets have opened unprecedented opportunities not only in the entertainment industry, but in many other businesses as well. Those same billions of people wear jeans, eat hamburgers and drive cars. Small wonder that talented inventors, entrepreneurs and businessmen can earn a great deal of money.

The more fundamental question is, why should anyone consider it a bad thing for someone else to succeed? There is, in some quarters, a conviction amounting almost to superstition that money earned by one person must somehow be at the expense of another. But all of the economic data reviewed above refute that crude notion of how the world works. At some point, one must ask whether, at bottom, the whole contemporary campaign against inequality rests on any foundation more solid or more praiseworthy than naked greed and envy.

The role of the Congressional Budget Office in generating misleading data

The Census Bureau data reviewed above will be new, and perhaps surprising, to many readers. This is not because those data are hard to find. They aren’t. Nor is there any particular doubt about their accuracy. But the facts about income distribution or, more accurately, income dispersion, as measured by the Census Bureau, have gone virtually unreported in the press. In fact, the media have imposed a near blackout on the relevant Census Bureau figures. Instead, reporters, editors and the media generally have collaborated with liberal politicians in publicizing alternative “data” that purport to show an alarming decline in the fortunes of middle-income Americans, and skyrocketing income growth in the top income group, particularly during the 1980s.

Virtually all of the “data” purporting to document these trends originated in the Congressional Budget Office. Throughout most of the 1980s and continuing until 1994, the CBO operated as an arm of the Democratic Party. One of its functions was the production of alarmist “data” to support the political agenda of the party that then controlled the CBO. The CBO carried out this function only too well.

Beginning in 1987, at the request of Democratic Sen. George Mitchell, the CBO produced “data” purporting to show two things: First, that the income of average families was actually declining; and second, that a majority of the economic gains during the Reagan boom went to the top 1 percent and other categories of high-income earners. These propositions were music to the ears of liberal politicians, who immediately began using them for partisan purposes. Bill Clinton, for one, prominently featured both of the CBO’s themes in his 1992 presidential campaign.11

How could the CBO produce data so grotesquely at odds with the Census Bureau’s numbers and other reliable sources of information? A full account of the statistical chicanery underlying the CBO’s “data” is beyond the scope of this paper. Indeed, the full story behind the CBO’s manipulation of data has never been told.12 But the following examples will serve to show how far the CBO was willing to compromise itself for the political purposes of the then-Democratic congressional majority, and how misleading its statistics are.

The CBO’s data were created by mixing IRS tax return data with Census Bureau data to derive figures purportedly measuring average income levels in various percentiles of income earners, including the top 1 percent.13 The Census Bureau does not release data on the top 1 percent of income earners because it considers such data too unreliable; the CBO was therefore responsible for the creation of data regarding the top 1 percent of income earners in the context of its calculation of income growth for all income groups generally.

In order to produce an apparent decline in average middle-class family income, the CBO introduced a systematic bias into the data by redefining and reshuffling “families” from Census Bureau data. The Census Bureau keeps statistics separately for “families” and “unrelated individuals.” (The Census Bureau also keeps statistics combining these two categories under the heading of “households.”) Census Bureau figures show that between 1980 and 1989, real income for the middle quintile of families increased by 8.3 percent, while real income for the middle quintile of unrelated individuals increased by 16.3 percent. The CBO manipulated these Census Bureau data by combining “families” and “unrelated individuals” into the single category of “families.” In other words, single persons were defined as “families” of one. Since demographic trends produced more rapid growth in the number of unrelated individuals in the 1980s, and since families on average have far higher incomes than unrelated individuals, combining these groups into a single category of “families” greatly depressed average “family” incomes. 14

Thus, the CBO claimed that even though middle-quintile families’ incomes increased by 8.3 percent and middle-quintile individuals’ incomes increased by 16.3 percent, the CBO’s new category of middle-quintile “families” saw their incomes decline by 0.8 percent over the same period. The CBO thus performed the remarkable trick of combining two positives — strong growth in the incomes of both families and unrelated individuals — and getting a negative!

The CBO’s data purporting to show an overwhelming share of the income growth that occurred during the 1980s as going to the top 1 percent of income earners are also the product of statistical manipulation. In addition to the CBO’s artificially created “family boom,” the CBO used other methods to distort the income data further and exaggerate the income growth of the top income groups. First, the CBO, unlike the Census Bureau, included capital gains along with earned income. Second, all income figures were indexed for inflation, but capital gains, which often accrue over many years, were not adjusted for inflation. This trick skewed the apparent distribution of income in the direction of taxpayers who reported capital gains. Third, the CBO included all capital gains in its figures, but further exaggerated their effect by limiting net capital losses to $3,000. Fourth, the CBO included partnership income but excluded partnership losses as well as rental losses.15 Finally, the CBO completely ignored income mobility in its analysis. These blatant methodological distortions caused Robert Bartley to describe the CBO income data as “worthless, or more bluntly a lie.”16

The CBO published its “data” without fully disclosing its methods. It took two years of heroic detective work by Christopher Frenze, then a minority staff member of the Joint Economic Committee of Congress, to uncover the technical errors and deliberate distortions that underlay the CBO numbers. Frenze, who is now a senior economist on the staff of the Joint Economic Committee, published a series of papers that exposed the CBO’s manipulations.17

When the CBO’s methods were revealed, it was unable to defend them.18 First the CBO distanced itself from the estimates of enormous income gains by the top 1 percent. Later it abandoned its most popular and defective family income data, including measurements used to portray family incomes in a manner inconsistent with Census Bureau data.

But the CBO’s express and tacit admissions that its statistics were misleading has not deterred the left from continuing to rely on its faulty data, virtually to the exclusion of other, more reliable sources of information. On the basis of CBO income data, for example, influential MIT economist Paul Krugman has asserted that the Reagan years make “a picture of simultaneous growth in wealth and poverty unprecedented in the twentieth century” and that the “growing gap between rich and poor was arguably the central fact about economic life in the 1980s.”19 Astonishingly, in his book devoted to this and other related topics, Krugman saw no need to acknowledge, disclose or account for the Census Bureau data summarized above. Krugman served as a Clinton campaign adviser in 1992.

Liberal politicians have similarly based their complaints about income inequality almost exclusively on the CBO’s discredited reports. Martin Sabo, currently the ranking minority member of the House Budget Committee, is just one example of a liberal politician who has relied on the CBO’s discredited numbers to support explicitly statist legislative proposals.20

It is unfortunate that some politicians will seize on any purported evidence, no matter how false or misleading, that will buttress their position. But what is even more appalling is the role played by the American press. Newspapers and other media commentators have, nearly without exception, accepted the jury-rigged CBO data at face value. Even after the CBO itself backed off its calculations and abandoned the most indefensible of its practices, the old, discredited data continue to circulate without criticism in the press. To our knowledge, not a single American newspaper has undertaken a critical analysis of the CBO’s data, nor, to our knowledge, has a single newspaper covered the scandal of how the CBO manipulated data, and how its manipulations were exposed by Christopher Frenze. The willingness of our press to abandon its usual attitude of critical skepticism in order to promote the personal political agendas of editors and reporters is one of the most troubling aspects of the entire income inequality debate.

Inequality of wealth

When the focus shifts from inequality of income to inequality of wealth, animus against “the rich” emerges even more clearly. While most people recognize that income must be earned, and therefore requires work, references to “wealth” conjure up images of privilege and easy money — even though, for most of us, our “wealth” consists of our house, our car, and our pension plan.

In recent months, there has been a great deal of interest in the purportedly increasing inequality of wealth distribution, triggered largely by the publication of a book by economist Edward Wolff.21 Wolff’s book, published by the liberal Twentieth Century Fund, analyzes household wealth over the period from 1983 to 1989 and concludes that, while the total amount of household wealth increased dramatically during that time, the increase was unequally distributed, with the largest percentage of the gain going to the wealthiest families. Wolff argues that this represents an acceleration of a trend that began in the late 1960s. Based on these data, he urges that the federal government adopt a tax on wealth for the purpose of redistributing it more broadly through federal spending programs.

Not surprisingly, Wolff’s book has generally been received in the press with uncritical adulation. Nevertheless, although Wolff argues his case ably, it is not at all clear that the data he relies on support his conclusions.

In the first place, Wolff’s exclusive concern with the distribution of wealth obscures the more fundamental fact that the period from 1983 to 1989 was a time of unparalleled creation of wealth. Wolff acknowledges that during those years, the heart of the Reagan boom, American families added $4 trillion to their total net worth, so that average net worth grew by 20 percent per family and median net worth grew by 8 percent. The period from 1983 to 1989 saw a rapid acceleration in the pace at which wealth was created; mean marketable wealth grew at almost twice the rate at which it grew from 1962 to 1983.22 In view of these incontrovertible facts, the exclusive focus of Wolff and his followers on the distribution of gains in wealth seems one-sided at best, and perverse at worst.

Moreover, Wolff’s own data show that the alleged increase in wealth inequality from 1983 to 1989 was hardly unique. In fact, his data indicate that from 1973 to 1979, inequality increased more sharply than during the expansion of the 1980s.23 In other words, according to Wolff’s own data, the “increasing inequality” from 1983 to 1989 represented the moderation of a previously existing trend.

The foregoing comments take Wolff’s figures at face value. But in fact there is great doubt about the accuracy of Wolff’s numbers. Economist John Weicher of the Hudson Institute has analyzed the same Federal Reserve survey data relied on by Wolff, but has come to substantially different conclusions.24 Like Wolff, Weicher analyzes the Federal Reserve survey data in detail and reviews the technical issues involved in defining wealth and measuring its distribution.25 The technical issues that divide Weicher’s analysis from Wolff’s are beyond the scope of this paper, but the following main points can be distilled from the debate.

First, Wolff defines “wealth” narrowly, and in such a way as to maximize the concentration of “wealth” at the top of the economic scale. For example, he does not include the value of automobiles in his definition of “wealth.” Since most people own cars, the inclusion of their value in the definition of wealth yields a broader pattern of wealth distribution. More important, Wolff does not include either private pensions or the present value of Social Security benefits in his definition of “wealth.” Wolff has acknowledged the extreme importance of this factor; by his own calculation, including the value of Social Security and pension benefits reduces the percentage of total wealth owned by the wealthiest 1 percent of all families, as of 1989, from 39 percent to 21 percent.26

Second, even using the same definitions of “wealth,” Weicher’s analysis of the Federal Reserve survey data reaches conclusions quite different from Wolff’s. Weicher finds that wealth may or may not have become less equally distributed between 1983 and 1989, depending on how various technical questions are resolved. On some assumptions, the distribution of wealth became more equal during that time; on other assumptions, it became less equal. At most, if any increase in inequality occurred, Weicher finds that it was so slight as to be “more or less on the margin of [statistical] significance.”27

Third, Weicher, unlike Wolff, gives considerable emphasis to the enormous creation of wealth during the 1980s, and the benefits that wealth creation conferred on American families. He points out that between 1983 and 1989 the net assets of the average household increased from $151,000 to $181,000.28 Average wealth grew by about 24 percent for white households and 35 percent for Black households. And, whereas in 1983 the richest 1 percent of U.S. households owned about 39 percent of the “wealth” in the country, in 1989, they owned about 36 percent.29 These figures do not include such assets as private pensions and the present value of Social Security and Medicare benefits.

Finally, Weicher, unlike Wolff, has tried to address the question of economic mobility. We have already seen that comparing income groups over time is misleading because the composition of those groups is always changing. Weicher tried to analyze whether the same is true with respect to wealth. In other words, were the top 1 percent of “rich” families in 1983, who are said to have gotten a disproportionate share of the gains in wealth between that year and 1989, in fact the same people who were in the top 1 percent in the latter year? This question could not be answered definitively because “longitudinal” data from the Federal Reserve survey have not been made public. So Weicher tried to address the issue by referring to the self-descriptions given by the 1983 and 1989 groups on such matters as asset composition. Comparing the self-descriptions of the top 1983 group to the same information for the top 1989 group, Weicher found that there were very substantial differences. He concluded that “‘the richest 1 percent’ don’t seem to be the same people at the end of the boom as at the beginning, by and large.”30

Thus, despite the media attention that has been devoted to claims of increasing inequality in the distribution of wealth — virtually all of it uncritical — it is far from clear that such inequality is in fact increasing.

But there are broader issues embedded in the economists’ technical debates. To say that a relatively small number of families own a substantial percentage of the country’s wealth is another way of saying that, for whatever reason, most people are not saving much money. And there is some truth to this claim. Economists have for years been decrying Americans’ relatively low rate of savings. But why is it that so many Americans do not save, and therefore do not accumulate much wealth? Part of the answer may be sheer imprudence. But we would propose two additional explanations.

First, taxes are too high. The Tax Foundation has calculated that the average American family now spends more on taxes than on food, clothing and shelter combined. Small wonder that for most families, there isn’t much left to save. If our government seriously wants to expand the number of people who are able to accumulate wealth, it should cut their taxes. To the extent that more people are able to save money rather than pay it to the government, wealth will be more broadly distributed.

It is on this point, rather than on the technical economic issues, that we part company most decisively with Wolff and the media pundits who follow his lead. Wolff’s proposed solution to the “problem” of unequal distribution of wealth is — another tax! Wolff proposes to add to the existing tax burden a new tax on wealth. This will not increase the number of Americans who can accumulate wealth; it will decrease their number. This, we suspect, is not what Americans want. The goal of economic policy should not be to prevent Americans from accumulating wealth. The goal should be to facilitate the creation and accumulation of wealth by as many people as possible. Toward this end, the best thing the government can do is reduce its spending and cut taxes.

We suspect that a second factor may also be at work, closely related to the first. A primary motive for saving has always been a desire to provide for one’s old age. Until relatively recently, it was considered mandatory to save money in order to be assured of a stream of income during one’s retirement years. This retirement income was needed, among other things, to pay the medical bills that are likely to be incurred as a person ages. The need for retirement income has always been a powerful incentive to save and invest.

With the advent and steady expansion of Social Security and Medicare, that calculus has changed. The perceived need to save for old age has declined drastically, and at the same time, Social Security and Medicare taxes have made saving more difficult for everyone, and nearly impossible for some.31 The result is that many people save less and thereby accumulate less “wealth” than they would in the absence of these programs. In exchange, however, they have a right to receive Social Security and Medicare benefits, which properly should be counted as part of their “wealth.” As noted above, this is a factor of tremendous quantitative significance. Simply adding the value of Social Security benefits and private pensions to the definition of a household’s wealth reduces the share of “wealth” owned by the top 1 percent of families by nearly one-half. If the value of Medicare benefits were included as well, the apparent concentration of assets in the top 1 percent of families would probably be cut by more than half. Conversely, privatizing these programs and allowing taxpayers to retain the income they now pay in taxes to support them would go far toward broadening, and thereby equalizing, the distribution of wealth. Whether on balance this would be desirable can of course be debated. But it hardly makes sense to raise taxes to a level which makes the accumulation of wealth by most families difficult if not impossible; then complain that most wealth is concentrated in a relatively small number of families; and finally propose to raise taxes again to “solve” the problem of unequal distribution of wealth.


We may reasonably ask: Why now? Why, when by historical standards both income and wealth are broadly distributed; when the distribution of income mostly reflects the distribution of work; and when economic mobility has never been greater, has income inequality assumed a central role in our public discourse? The answer, we believe, has more to do with politics than economics.

In its current form, the debate over income inequality began in the late 1980s, and arose out of the then-heated, and still-polarizing, debate about Ronald Reagan and his administration. The Reagan Administration never lacked critics, but by the late 1980s those critics faced an intractable problem: By any conventional yardstick, the American economy had performed extraordinarily well during Reagan’s term in office. Gross domestic product increased by one-third, inflation was brought under control, and unemployment fell. The Reagan Administration oversaw the longest peacetime expansion in American history. The extent to which Reagan’s policies deserve credit for these trends is of course a legitimate subject of debate, but sustained economic growth, declining unemployment, and a stunning victory over rampant inflation gave his critics very little to work with.

Hence income inequality. In the late 1980s, a virtual cottage industry sprang up among liberal economists and journalists. This cottage industry was the production of statistics designed to make the economic growth of the 1980s look bad. These statistics were based not on the familiar concepts of median income, unemployment and inflation, but focused rather on such previously obscure categories as income quintiles and the Gini inequality index.

The 1980s are now receding into history, but the income inequality debate rages on. It has, in fact, assumed a broader and even more significant role in our political discourse. Left-of-center critics have gone beyond using income inequality to discredit the growth of the 1980s, and with it, the Reagan Administration. Indeed, in a spirit of apparent bipartisanship, they now generally assert that the purported trend toward growing inequality has been going on for some 25 years. But this apparent bipartisanship is, in reality, nothing less than an attempt to sweep under the rug the lessons of recent economic history.

For middle-income and lower-income Americans, the last 25 years have not represented an undifferentiated upward or downward trend. On the contrary, there have been sharp ups and downs in the economic well-being of American families. The Carter era was an economic disaster. Real income of middle-quintile American families fell by 5 percent between 1979 and 1981. Conversely, the Reagan years saw a great increase in the well-being of middle-income families: a 12.6 percent gain in real income from 1982 to 1989. The years from 1990 to 1995 have been characterized by stagnation and ever-growing tax and regulatory burdens.

These basic facts of recent economic history suggest that all Americans, especially middle- and lower-income Americans, are best served by policies of sound money, low marginal tax rates, reduced government regulation and restrained growth in government spending — the precise opposite of the policies prescribed by liberal critics of income inequality. Of course, many factors other than government policy are at work in the economy, and it is fair for economists and politicians to debate the extent to which credit or blame for economic growth or decline should be attributed to government action. But it is absurd to try to paper over the very different results that have accompanied different government policies with vague references to a “25-year trend toward increasing income inequality.”

The reason for the persistence of income inequality as a political issue, notwithstanding its lack of empirical foundation, appears clear. It provides an all-purpose justification for the redistributionist impulse that is the core of contemporary liberalism. Every increase in spending on social programs, and every concomitant tax increase, can be explained and justified by the claim that the economy is tending toward ever-widening disparities in economic well-being, which in turn must be counteracted by ever-expanding government action. If government distributes more and more resources to the poor, it is only fair, since the private sector is “distributing” more and more resources to the rich.

In truth, the liberal appetite for “inequalities” to remedy has become insatiable. Redistribution of income, which was not recognized by the founders of this country as a legitimate object of government, has now become its chief purpose. The traditional functions of government have dwindled into relative insignificance. National defense now accounts for only about 17 percent of total federal spending. And an American Bar Association study concluded that the entire apparatus of the civil and criminal justice system consumes less than 3 percent of public spending. These traditional functions of government are dwarfed by its principal contemporary purpose: Taking money from one person in the form of taxes and giving it to someone else. To sustain this vast structure of redistribution, liberals must always claim, with increasing hysteria, that inequality is growing and “income gaps” are expanding. If the public were ever to realize fully the hollowness of these claims, the entire redistributionist enterprise would collapse.

The “equality” espoused by the critics of income inequality is equality of outcome or of result. This notion of equality appropriates the language of America’s founders, but it nevertheless strikes at the heart of the founders’ understanding of equality, which was based on equality of rights. For the founders understood that equality of outcome is impossible and undesirable, given the different abilities with which each person is born. Thus James Madison wrote in the most celebrated of the Federalist papers that the “first object of government” is protecting the “different and unequal faculties of acquiring property” which necessarily results in the “possession of different degrees and kinds of property,” or inequality of outcome.32 The liberal critics of income inequality have an argument not just with the facts or with their current political adversaries, but with the authors of the American Constitution itself.

1For a convenient summary of leading articles, see Christopher Caldwell, “Inequality: Fact and Myth,” Forbes MediaCritic, Fall 1995, pp. 45-55.
2Michael Lind, “Why the Rich Get Richer,” New York Times Book Review, Sept. 24, 1995, p. 15.
3Rep. Martin Olav Sabo, D-MN, for example, has introduced an “Income Equity Act” (H.R. 620) in which he proposes to limit the deductibility of executive compensation to 25 times that of the lowest paid worker in the same firm.
4Richard B. McKenzie, What Went Right in the 1980s (San Francisco: Pacific Research Institute, 1994), p. 107. McKenzie notes the contention of Lawrence Mead that a major share of household income differences can be explained by the amount of work, with the rich working far more than those in lower income groups. See Lawrence M. Mead, The New Politics of Poverty: The Nonworking Poor in America (New York: Basic Books, 1992).
5The Census Bureau itself has speculated on the possible impact of demographic trends on family and household income statistics. Among the trends noted by the Census Bureau are a growing elderly population, the growing number of persons living in nonfamily situations, and the growing percentage of families with a female householder, no husband present. Bureau of the Census, Money Income and Poverty Status in the United States, 1989, Current Population Reports, Series P-60, No. 168, Sept. 1990. This issue is discussed informatively by Robert L. Bartley, The Seven Fat Years (New York: Free Press, 1992), pp. 274-277.
6Thomas Sowell, “Observations,” Forbes, April 24, 1995, p. 86.
7Sowell, op. cit.
8U.S. Treasury, Office of Tax Analysis, “Household Income Mobility During the 1980’s: A Statistical Assessment Based on Tax Return Data” (Washington, DC: U.S. Government Printing Office, June 1, 1992).
9Isabell Sawhill and Mark Condon, “Is U.S. Income Inequality Really Growing? Sorting Out the Fairness Question,” Policy Bites, no. 13 (Washington, DC: Urban Institute, 1992)
101991, fewer than 9 percent of the families in the bottom quintile included anyone in the prime earning ages of 45-54.
11For a convenient example, see the 1992 Clinton campaign manifesto Putting People First (New York: Random House, 1992).
12For a brief account of the political origin of the CBO data, see Bartley, op. cit., pp. 271-272, 278-280.
13Authors’ interviews with Christopher Frenze, November 13, December 5 and December 8, 1995. Frenze is the senior economist on the staff of the Joint Economic Committee of the United States Congress. Frenze’s several papers exposing the CBO’s errors and distortions are summarized in the Republican Views section of the Joint Economic Committee’s 1994 Joint Annual Report, pp. 157-167.
14See Christopher Frenze, “Cooking the Books with CBO Family Income Data–Part II,” Joint Economic Committee, October 5, 1992.
16Bartley, op. cit., p. 280.