In october 2011, the Congressional Budget Office published a report, “Trends in the Distribution of Household Income between 1979 and 2007,” showing that, during the period studied, aggregate income (as defined by the cbo) in the highest income quintiles grew more rapidly than income in the lower quintiles. This was particularly true for the top one percent of earners. This cbo study has been cited by the media and politicians as confirmation that income inequality has increased “substantially” during the period studied, and has been used to support President Obama’s claim that income inequality is a serious and growing problem in the United States that must be addressed by raising taxes on the highest income earners.

We will show that much of what has been reported about income inequality is misleading, factually incorrect, or of little or no consequence to our economic well-being. We will also show that middle-class incomes are not stagnating; in fact, middle-class incomes have risen significantly over the 29 years covered by the cbo study. Lastly, we will address assertions that the rich are not paying their “fair share” of taxes.

In our view, Americans should care about the well-being of the nation as a whole rather than whether some people earn more than others. To that end, the focus of public policy should not be on equality of income but on equality of economic opportunity. Policies designed to reduce income inequality inevitably involve redistribution of income through increases in transfer payments and marginal tax rates. But these policies discourage hiring and investment, which depresses economic growth and opportunity. In sharp contrast, policies designed to enhance equality of opportunity will increase economic well-being for all, most particularly those in lower income households.

Income inequality

Perhaps the most important question left out of almost every discussion about income inequality is, “Why should we care about it?”

Many of those who worry about high income inequality argue that it is an indicator of social injustice that must be remedied through redistribution of income (or wealth). Unfortunately, those who make this claim have not provided any generally accepted criteria for determining when an economic system is unjust. Nor have they provided a convincing argument that such injustice is widespread in the U.S. (In considering this issue, it is worth noting that Greece, Spain, and Italy all have substantially lower income inequality than the U.S. The same is true for Afghanistan, Pakistan, and Bangladesh.)

Measuring inequality using the Gini coefficient. There are at least five methodologies used to measure income inequality. The most commonly used is the Gini coefficient (also called the Gini index) developed by Italian statistician Corrado Gini. The Gini coefficient is a method of measuring the statistical dispersion of (among other things) income, consumption, and wealth. The figure of merit for the Gini coefficient for income inequality ranges from zero to 1.0, where zero represents total equality (all persons have identical incomes) and 1.0 represents total inequality (one person has all of the income). By this measure, the U.S. has substantially higher income inequality than almost all other industrialized nations. In 2010, the Census Bureau reported that the U.S. Gini coefficient was .469, while the average Gini coefficient for the 27 European Union nations was .31.1

The U.S. Gini coefficient cited here comes from an annual report of the Census Bureau, which uses what it calls “money income” in its measurement of income inequality.2 Money income, which is the definition of income typically used in public references to inequality, consists of cash income only, does not subtract taxes, and excludes the value of noncash transfer payments (such as nutritional assistance, Medicare, Medicaid, and public housing), as well as many other components of income. In addition to transfer payments, which are a substantial portion of income at the low end of the income scale, some of the other missing components of income are: employer-provided fringe benefits (primarily retirement benefits and health insurance, which can amount to as much as 30 percent of income 3), capital gains,  imputed rent from owner-occupied housing, and increases in the value of home equity. We believe excluding these items renders this measure of income inequality relatively meaningless. However, since this measure is so frequently cited, it is worth noting that during the 29-year period covered by the cbo study, inequality of money income (as measured by the Gini coefficient) grew only about 10 percent.4

Using a comprehensive definition of income provides a much more meaningful measure of income inequality. The Census Bureau’s 15th definition5 is superior because it includes most of the income items listed above and subtracts taxes. Based on this more relevant definition, income inequality declined 1.8 percent during the sixteen-year period between 1993 and 2009, when the Gini coefficient dropped from .395 to .388.

An important shortcoming in the October 2011 cbo report is its almost singular focus on income as a measure of economic well-being, when there is a clear consensus among economists that the best measure of living standards over the long term is not income, but consumption. Focusing on consumption rather than income provides a very different picture of inequality.

There is a body of research indicating that consumption inequality is not only substantially lower than income inequality, but has been declining in recent years. For example, a 2006 study, “Economic Inequality Through the Prisms of Income and Consumption,” conducted by the Bureau of Labor Statistics (bls) found that in 2001, the Gini coefficient for consumption was only .280 (almost 30 percent lower than the Gini for comprehensive income, and about 40 percent lower than the Gini for money income), indicating that inequality with respect to this most meaningful measure of living standards is relatively modest. Moreover, according to the bls, during the fifteen-year period between 1986 and 2001, consumption inequality went down slightly; from a Gini of .283 to a Gini of .280.

It is important to note that two working papers from the National Bureau of Economic Research, one by Orazio Attanasio, Erik Hurst, and Luigi Pistaferri and the other by Mark Aguiar and Mark Bils, suggest that the Consumer Expenditure Survey (ces), which is the source for much of the research on consumption inequality, may have some measurement error that biases the figure downward. By contrast, a study by Bruce D. Meyer and James X. Sullivan addresses some of the measurement issues regarding the ces and in the process shows that consumption inequality is declining.6 In their study, Meyer and Sullivan construct an alternative measure of inequality to the Gini coefficient using the widely accepted “90/10 ratio,” which compares the consumption of those in the 90th percentile to those in the bottom 10th percentile. They document that these measures were fairly stable in the 1990s, and then tended to decline after 2005. Specifically, the 90/10 ratio declines by twelve percent during the eighteen-year period between 1990 and 2008.

Current methodologies measure only market consumption rather than total consumption.

Finally, we note that consumption inequality may be even lower than reported by any of the studies cited above. This is because current methodologies measure only market consumption rather than total consumption, which is the sum of both market (purchased) and nonmarket (home-produced) goods. This is important because lower-income households consume a disproportionate amount of goods produced in the home (what economists call “home production”), including home-cooked meals, household-provided child care, and household home improvements and maintenance. Economists have estimated that home production is around one-third of gdp, yet this form of consumption is not counted in the total when measuring consumption inequality.

Our conclusion from this analysis is that consumption inequality is considerably lower than income inequality. This is because consumption expenditures are made with after-tax dollars and are influenced by many factors other than money income, including transfer payments, family savings, barter, imputed rent from owner-occupied housing, income from the underground economy, and assistance from family and friends.

Alternative method of measuring income inequality. Another way of measuring income inequality is to divide the population into quintiles (twenty percent of the population in each quintile) ranked according to the aggregate income in each segment. Then, by comparing the growth in real income in each quintile, we can determine whether income inequality has grown or declined. The October 2011 cbo report emphasized this methodology. The table below shows the inflation-adjusted increase in household income7 in each quintile (with a separate breakout for the top one percent; figures for the “fifth” quintile are actually for the 81st to the 99th percentiles, i.e., all but the top one percent):

growth of income; share of income

The headline news from the cbo report was that the income of the top one percent of earners grew 278 percent, and that income in the higher-income quintiles grew more than income in the lower quintiles. But the study also showed that during the period reviewed, aggregate income grew 62 percent; income in the middle three quintiles grew “just under” 40 percent; income in the middle quintile grew 35 percent; and even the lowest quintile grew 18 percent.

In our view, these growth rates are worst-case, inasmuch as we believe the cbo’s figures have understated real income growth during the period. Note that these data are inflation adjusted using the “Consumer Price Index-Urban-Research Series” (cpi-u-rs) methodology, which many economists believe overstates inflation. For two major reasons, we along with many economists consider the “Personal Consumption Expenditure Deflator” (pce) a more accurate measure of inflation.8 First, the cpi is based on the pricing over time of a fixed market basket of goods. It does not account for the fact that households routinely substitute out goods when their prices rise relative to other goods of comparable utility. For example, a consumer may substitute apples for bananas when the price of bananas rises relative to the price of apples. In this example, the cpi would overstate inflation because it would assume that consumers would still buy the same number of bananas at the higher price. In contrast, the pce takes these substitution effects into account. Second, the cpi measures only prices paid by urban consumers, while the pce measures the prices of all consumption goods, wherever they are purchased. Since lower-income households tend to live disproportionately in non-urban areas, the exclusion of their purchases by the cpi further biases the index toward overstating inflation and, in this case, understating income growth. Income growth using the pce is presented below along with the cbo data.

Growth in real after-tax income from 1979 to 2007

If we are right that the pce is the more accurate measure of inflation, then real income growth during the period was significantly higher than reported by the cbo. When inflation is adjusted using the pce, real income growth in the first quintile was 40 percent higher, growth in the middle quintile was 24 percent higher, and growth in the fifth quintile was 8.6 percent higher. Thus, the cpi has not only understated real income growth but has overstated the rise in the level of income inequality.

The impact of globalization on income inequality. It is noteworthy that the income growth in the U.S. during 1979 to 2007 was achieved in a time of rapid globalization and technological change in the world economy. Globalization, which effectively is a breaking down of trade barriers, has put upward pressure on income inequality in most of the industrialized nations as the production of goods and services has migrated to countries with lower labor costs. While this process has raised living standards in developing countries, it has reduced jobs or suppressed wages in many developed countries. As a result, many of these nations have experienced an increase in income inequality. A 2011 study from the Organisation for Economic Co-operation and Development reported that from the mid-1980s to the late 2000s, income inequality increased in seventeen of the 22 oecd countries for which long-term data series are available.9 Seven of the most advanced oecd economies — Canada, Finland, Germany, Israel, Luxembourg, New Zealand, and Sweden — experienced greater increases in inequality than the U.S.

Measuring economic well-being

We believe the focus on income inequality is misguided. The most important finding of the cbo report is not that income grew more in the higher quintiles than in the lower quintiles; it is that income in all quintiles grew. And, as measured by the pce, incomes grew significantly faster than reported.

America’s economy has outperformed all other industrialized nations. The vast majority of Americans have fared well over the period of the cbo study. In fact, the U.S. economy has been the best-performing large economy in the world as measured by per-capita gdp and median standard of living. According to the oecd, per-capita gdp in the U.S. in 2010 was $46,600, which is 47 percent higher than the $31,800 average per-capita gdp in the eu nations in that year.

In addition to substantially higher gdp per capita, the U.S. has a significantly higher standard of living than almost all of the most advanced economies. According to “The Luxembourg Wealth Study,” the data source used by the oecd for international comparisons, in 2002 (the latest year for which results were available), median disposable personal income in the U.S., adjusted to reflect purchasing power parity, was 19.3 percent higher than in Canada; 68 percent higher than in Finland; 45 percent higher than in Germany; 59 percent higher than in Italy; 31 percent higher than in Norway; 73 percent higher than in Sweden; and 31 percent higher than in the United Kingdom.10

The figures for gdp per capita and median income understate America’s economic performance advantage because the median age of the U.S. population (36.8 years) is about four years lower than the average median age in the European Union and almost eight years lower than in Japan. Age, as a proxy for experience, is a significant contributor to income until individual earnings peak sometime between age 50 and 55. In addition to higher median incomes, Americans also have higher median net worth, a further contributor to the difference in standards of living.

There is no question that until the recent recession, the U.S. economy performed well in the 25 years from 1983 to 2008 compared to other advanced economies. During this period, real compound annual gdp growth in the U.S. was 3.3 percent (slightly ahead of the long-term trend line), substantially greater than the growth of its g-7 counterparts, which on a weighted-average basis grew only 2.3 percent per year. Moreover, in the recent recession, the U.S. economy contracted less than the world’s other advanced economies. Specifically, U.S. gdp shrunk 3.5 percent in 2009, which was 25 percent less than the 4.7 percent contraction experienced by the non-U.S. g-7. And in 2010, the U.S. economy grew 3.0 percent, 42 percent more than the 2.1 percent growth for the non-U.S. g-7.11

While in the aggregate, the U.S. economy performed much better than the economies of its g-7 counterparts, this was in part because the U.S. had higher population growth. The most meaningful measure of economic performance is based on per-capita statistics, which removes the effect of population growth. On a per-capita basis, the U.S. once again outperformed, albeit by a smaller margin. During the comparable period, real compound annual gdp per-capita growth in the U.S. was 2.1 percent, higher than the 1.8 percent weighted average growth of other members of the g-7. Again using per-capita numbers, in 2009 the U.S. economy contracted about 4.3 percent, which was less than the 5.1 percent contraction recorded by the other g-7 members. And in 2010, the U.S. economy grew 2.1 percent, 31 percent higher than the 1.6 percent growth of non-U.S. g-7 countries.

Further evidence of the superior economic performance of the U.S. economy comes from a comparison of unemployment rates. The average unemployment rate in the United States from 1982 to 2007 was 6.0 percent, compared with 9.0 percent in France, 8.3 percent in Germany, and 7.7 percent in the United Kingdom.

The average unemployment rate in the U.S. from1982 to 2007 was 6.0 percent, compared with 8.3 percent in Germany.

The not-stagnant middle class. As noted earlier, the claims that incomes in the U.S. have been stagnant “for decades” are at odds even with the arguably understated income growth data from the cbo report, which show that income in the middle three quintiles grew “just under 40 percent.” And as we have seen, using the pce deflator, incomes in the middle three quintiles grew about 48 percent.

While these growth rates are somewhat below historical averages, they are impressive inasmuch as they occurred during a period of rapid globalization and technological change. In any event, it is clearly wrong to say that middle class income growth during the period was “stagnant.”

America’s poor: Putting poverty into perspective. Currently, about 46 million Americans live below the official federal poverty line. But the data suggest that by some measures America’s poor have a somewhat higher standard of living than is commonly believed.

Based on a standard established in 1965, a family of four reporting $22,300 or less in money income in 2010 was considered poor and thus eligible for government support. By this standard, about 15 percent of Americans are currently judged to be poor, roughly the same percentage as was reported in 1965, and up from 12.5 percent before the recession. However, in his 2008 book The Poverty of “The Poverty Rate,” American Enterprise Institute scholar Nicholas Eberstadt makes a compelling case that the government measure for the “official poverty rate” is seriously flawed. His assertion is based largely on the fact that the reported income of people defined as poor has increased only about 10 percent since the standard was set, while other measures of well-being have increased substantially more.

Most notably, consumption expenditures by the lowest-income Americans have consistently exceeded reported income, and, this difference “has widened tremendously over the decades since the official poverty rate made its debut.” Specifically, Eberstadt reports that according to the Department of Labor, in 1960–61 consumption expenditures in the lowest quartile were 112 percent of reported income, rising to 140 percent (in the lowest quintile) in 1972–73, and 198 percent (in the lowest quintile) in 2005. Thus, a family claiming $22,300 in income in 2005 would have reported about $44,000 in expenditures in that year. As noted earlier, the gap between reported income and consumption is filled by various categories of government transfer payments (including Medicaid, food stamps, subsidized housing, the Earned Income Tax Credit, Temporary Assistance for Needy Families, etc.), family savings, imputed income from owner-occupied housing, barter, support from family and friends, and income from the underground economy.

The average U.S. household lives in about 845 square feet per person, or 2.3 times the average European household.

Additional data on the discrepancy between reported income and the actual well-being of the poor comes from research by Robert Rector and Rachel Sheffield of the Heritage Foundation. In recent years, shortly after the release of the Census Bureau’s report “Income, Poverty and Health Insurance Coverage in the United States,” Rector has issued his own report on American poverty. In the September 13, 2011, report, “Understanding Poverty in the United States: Surprising Facts about America’s Poor,” the authors examine various data on the living conditions of poor households. For example, the data show that on average America’s poor live in housing that totals 515 square feet per person, about 40 percent more per person than the living quarters of the average European household. (The average American household lives in about 845 square feet per person, or 2.3 times the average European household.) They also present data that show “There is little or no evidence of poverty-induced malnutrition in the United States.” And that “nutrient density (amount of vitamins, minerals, and protein per kilocalorie of food) does not vary by income class.”

In addition to food, clothing, and shelter, some of the most meaningful indicators of well-being are the properties and amenities that make life more comfortable or enjoyable. Based on data from the 2009 “American Housing Survey,” Rector and Sheffield report that 42 percent of poor households own a home (median price: $100,000); 80 percent have air conditioning; 98 percent have a color tv (65 percent have two or more); 99.6 percent have a refrigerator; 98 percent have a stove and oven; 75 percent have a car or truck (31 percent have two or more); 81 percent have a microwave oven; 78 percent have a dvd or vcr; 64 percent have a satellite connection; and 25 percent have a dishwasher.

Our purpose is not to make light of the deprivations the poor suffer every day. There is no doubt that the poorest Americans struggle mightily, and that too many Americans are poor. But these data are useful in understanding the difficulties in defining poverty, and for constructing effective policies aimed at helping those in need.

Why U.S. income inequality is higher

There does not appear to be a clear consensus among economists as to why inequality in the U.S. is higher than in other industrialized nations.

There are many factors that contribute to income inequality, at least two of which are common to all countries and are unalterable. They are: differences in individual ability and preferences (defined as the capacity and desire to earn) and differences in age (this latter factor is currently in evidence in the U.S., as 80 million aging baby-boomers are passing through their peak earnings years). As discussed above, a third influence on inequality in almost all countries during the last 30 to 40 years has been globalization.

In addition to these common factors, we believe there are several factors specific to America that have put upward pressure on income inequality. Some have enabled certain segments of the population to earn extraordinary incomes, and some have caused certain segments to lag behind

Extraordinary incomes at the top. The most influential factors enabling the growth in incomes in the U.S. appear to be:

  • Greater economic freedom: Despite increasing infringements on U.S. economic freedom in recent years, America’s lower aggregate taxes, less stringent regulation, and more business-friendly economic environment have made the U.S. the most productive nation in the world and a place where more people can become rich, or even “super rich,” than in other countries.
  • A highly developed entrepreneurial culture: During the 17th, 18th, and 19th centuries, huge waves of self-selected individualists braved enormous perils to come to America for a better life. It can be argued that entrepreneurship, if not actually in the dna of most Americans, is at a minimum a strong component of the nation’s cultural heritage.
  • The “electronics revolution”: We believe that the invention of the integrated circuit by Robert Noyce and Jack Kilby in 1958 ignited a modern version of the 19th century Industrial Revolution, which has been unparalleled in history. This revolution, which has given us first the microprocessor and ultimately the Internet, along with thousands of other remarkable technological advances and offshoots, has raised living standards and quality of life for much of the world population. This has been especially true in the U.S.
  • A large, highly developed venture capital industry: America’s venture capital industry, which took off in the late 1960s and is unrivaled in size and sophistication by any other country’s, has combined with America’s entrepreneurial culture and the electronics revolution to produce more innovation and higher productivity than any other country in the world.

Yes, these factors have combined to produce massive wealth for a few — but have also contributed to raising incomes for most.

Suppression of incomes at the low end. In addition to globalization and technology, another important factor putting downward pressure on incomes in the U.S. has been the substantial influx of low-skilled, low-income immigrants into the U.S. workforce over the past 30 years.

But as noted at the beginning of this essay, we and others believe that an even more important cause of lagging incomes in America is inequality of opportunity. There is considerable debate over what impedes equality of opportunity. Many assert that institutional racism and sexism is a major factor; others argue that the political system is rigged in favor of corporations and the rich. These explanations seem to have as many detractors as they have advocates. But one cause of lagging incomes on which there is broad agreement is America’s substandard k-12 education system. We believe that a solution to this problem would do more to reduce income inequality and increase prosperity than any other public policy fix.

Have tax cuts increased inequality? A common claim is that the rate cuts for capital gains and dividends under President Clinton together with President George W. Bush’s cuts in marginal rates and further cuts in capital gains and dividend rates raised income inequality. But the evidence does not support that claim, inasmuch as the Gini coefficient for comprehensive income during the period 1993–2009 (the period in which almost all of the Bush and Clinton tax cuts took effect) did not change. A plausible explanation for this is that the Bush tax cuts reduced taxes on people with lower incomes more than it did on people with higher incomes. For example, under Bush, the lowest marginal rate, 15 percent, was lowered to 10 percent (a 33 percent reduction), while the highest marginal rate, 39.6 percent, was lowered to 35 percent (a 12 percent reduction). In addition, under Bush, the child credit doubled and the earned-income-tax credit increased significantly, further reducing the tax obligations of lower-income earners.12 This has almost certainly contributed to the increase in workers who pay no federal income tax, which now totals about 47 percent of tax filers.

Is the increase in the ranks of the “super rich” a problem? Much of the concern today about income inequality is about the income levels of the highest income earners in the country, such as the top one percent or the top 0.1 percent. Critics assert that the rising share of income held by this group is a problem that public policy should address through income redistribution. We disagree.

The period of rapid and substantial accumulation of wealth for some individuals has not just coincided with, but has contributed to, one of the best periods of economic growth in U.S. history. Much of the wealth created during this period was amassed by extraordinarily talented entrepreneurs who made their money by developing new technologies and products that have raised productivity and living standards for all. We can only imagine what the U.S. and the world economy would be like without: Bob Noyce, Gordon Moore, and Andy Grove (founders of Intel); Bill Gates (Microsoft); Steve Jobs (Apple); Larry Page and Sergei Brin (Google); Fred Smith (Federal Express); Larry Ellison (Oracle); Sam Walton (Wal-Mart); Bob Swanson (Genentech); George Rathman (Amgen); Howard Schultz (Starbucks); Jeff Bezos (Amazon); and the founders of hundreds of other remarkably successful companies established in the last 40 years. These multi-millionaires and billionaires started entire new industries or completely transformed old ones. In this process, they have been responsible for directly or indirectly creating tens of millions of jobs.

To be sure, not all of the newly rich are creators. For example, many come from the financial sector. But the financiers, investors, and bankers are an essential part of the economic ecosystem that has made the U.S. economy superior to any other. Moreover, it is not possible, and is almost certainly counterproductive, to enact policies based on value judgments about who should be allowed to get rich and who should not.

New York Times columnist Paul Krugman said in his October 20, 2002, article that “if the rich get more, that leaves less for everyone else.” Is that really true? Is the U.S. economy a zero-sum game? If the above list of billionaires had not been so successful, would the rest of us be better off today? Surely not.

So how should we think of these creative millionaires and billionaires: as contributors to income inequality, or as driving forces behind American prosperity? They are both. But why should we care how high their incomes are when the vast majority of us have prospered greatly from their success?

Do the rich pay their “fair share”?

The answer to this question should start with an agreement on an accepted definition of fair. But those who assert that the rich do not pay their fair share have not provided such a definition.13

Here is how the federal income tax burden was distributed in 2009 (the most recent year for which data are available):14

Distribution of federal income taxes (2009)

The U.S. income tax system is, by any measure, quite progressive. In fact, according to a study released in 2008 by the oecd, the U.S. federal income tax system is the most progressive of any of the 24 countries in the “oecd-24,” which includes Canada, Japan, Australia, and all of the richest European nations: Germany, France, the United Kingdom, Italy, the Netherlands, Norway, Switzerland, Luxembourg, and Sweden. In fact, the U.S. progressivity index is 22 percent higher than the average for the 24 countries.15

Are payroll taxes regressive? Many of those who assert that the rich don’t pay their fair share contend that both the Social Security and Medicare tax systems are regressive, and that the regressiveness of these systems substantially or wholly offsets the progressiveness of the federal income tax system. But studies show that both systems are themselves progressive.

Payroll taxes are collected for the express purpose of providing income supplements and medical care during retirement. In the case of Social Security, income is taxed proportionately up to a cap that is currently set at $110,100. Those who assert that the tax is regressive argue that the cap results in a decline in taxes paid as a percentage of income as income rises above the cap. But this argument omits two relevant facts: the amount of each beneficiary’s Social Security income at retirement is also capped; and higher-income workers get less back as a percentage of their contributions than lower-income workers do. Moreover, Social Security income is subject to the income tax, which is taxed progressively. Thus, overall, the system is progressive, not regressive. The Social Security Administration itself has found that the tax is indeed progressive, with retirement benefits about halfway between a pure defined-contribution program and a flat dollar benefit amount. It has also analyzed progressivity for future cohorts and predicts that progressivity may stabilize around its current level.

In the case of Medicare, the amount paid into the system is proportionate to income and has no cap. So a person with a lifetime income of $5 million will pay five times as much into the system as a person with a lifetime income of $1 million. Since the benefits provided by Medicare (paid health care) are on average essentially the same for each beneficiary, all other things being equal, the system is progressive. In support of this assertion, recent research by Darius N. Lakdawalla and Jay Bhattacharya shows that, based on the “Medicare Current Beneficiary Survey,” the poorest groups receive the most benefits at any given age, and that this advantage in benefits received significantly outweighs the effect of their higher death rates relative to wealthier recipients. The result is that Medicare is a highly progressive public program.16

Taxes on corporate dividends and capital gains are taxes on corporate income that has already been taxed once.

Do higher-income earners pay lower tax rates? The latest argument in favor of raising the taxes on higher-income earners is that the rich pay lower average rates than lower-income earners. This claim has been given currency by the famed investor Warren Buffett, who recently announced that he paid a lower rate of tax on his income than did his secretary. Since most of Buffett’s income comes from dividends and capital gains (which are taxed at the rate of fifteen percent), and assuming Buffett pays his secretary well, it is understandable that the rate shown on his tax returns would be less than the rate paid by his secretary. However, the relationship between Buffett’s low tax rate and that of his secretary is a statistical outlier. According to the cbo, the rich, on average, definitely pay higher income tax rates than lower-income taxpayers (see table below).

Moreover, as the Wall Street Journal among others has noted, Buffett has ignored the fact that the taxes on corporate dividends and capital gains are taxes on corporate income that has already been taxed once at rates as high as 35 percent, not including state taxes. (Thirty five percent is the statutory rate of tax on corporate income; the average rate is 25 percent.)

To be fair, not all of the corporate income tax is born by shareholders. Most economists agree that a significant part of the corporate tax may be borne by labor or passed through to consumers. While there is no consensus on the portion borne by shareholders, we believe a reasonable estimate is that roughly 50 percent of the corporate tax burden falls on shareholders, in which case, a typical shareholder would pay a combined tax rate of 27.5 percent (if the company paid the average rate) and as much as 32.5 percent (if the company paid the statutory rate). In any event, these rates are substantially higher than the average income tax rates paid by 99 percent of taxpayers.

But what if all federal taxes are considered? Do the rich enjoy lower tax rates? Again, according to the cbo, the answer is no. The combination of income taxes, payroll taxes, excise taxes, and shareholders’ share of the corporate income tax makes up about 94 percent of federal taxes on individuals. As shown in the table below, when all these taxes are included, in 2007 the top one percent paid an average tax rate of 29.5 percent, substantially higher than the rates paid by those below the 99th percentile.  

Average tax rates as a share of income (2007)

In its report, the cbo makes the standard assumption that the incidence of each federal tax is entirely borne by the individual or organization reporting the income. A conceptually superior approach would be to estimate how the incidence of these various federal taxes indirectly impacts different income earners beyond the direct effect reported by the cbo. For example, our view is that only about 50 percent of the corporate tax is borne by capital, and the rest is borne by labor, consumers, and other stakeholders through lower wages and cash flows. Since there are no generally accepted estimates of tax incidence across these taxes, this adjustment is beyond the scope of this essay. Suffice it to say, we believe that the cbo’s basic conclusion is correct: Higher-income earners pay higher rates of tax when all federal taxes are taken into account.

Taxing income from capital is both economically inefficient and inequitable. The current discussion about the “Buffett Rule” highlights the fact that a considerable fraction of the income of top income earners is capital income (e.g., dividends and capital gains), and that extracting more tax revenue from top earners will require significant increases in capital-income tax rates. But doing so is at variance with the standard view in the public finance literature on the perils to economic efficiency of taxing capital income. Many studies find that in the long run capital-income tax rates should be set near zero, even if society chooses to redistribute income. Taxing capital is particularly harmful because it is supplied very elastically. This means that even small increases in taxes on capital income lead to a lower capital stock, which, relative to other forms of taxation, depresses labor productivity, wages, and living standards.

In addition to economic efficiency considerations, we believe that taxing any income from savings and investment is inequitable. Here’s why: Assume two people, Angelina and Brad, have exactly the same lifetime earned income, but Angelina saves ten percent of her after-tax income and Brad saves nothing. In this hypothetical, if income from savings is taxed, Angelina will pay more lifetime tax than Brad, simply because Angelina saved. We believe this is clearly inequitable.

So what is a “fair share”? The U.S. tax system is more progressive than that of any other advanced economy. Higher-income workers already pay a substantially disproportionate amount of the income tax relative to their share of income. The top five percent pay 44 percent more in taxes than the bottom 95 percent, while 47 percent of tax filers pay no tax at all. The bottom 50 percent of filers pay only 2.3 percent of taxes, and the bottom quintile gets money back.

Based on these facts, how does one make a case that the rich are not paying their fair share?

Equality of opportunity

We are unaware of persuasive evidence that reducing income inequality will increase economic well-being for the majority of citizens; in fact, America’s superior standard of living and economic growth relative to other advanced economies is evidence to the contrary. For arguably the most commonly used measure of inequality and for the Census Bureau’s most comprehensive definition of income, inequality has not risen since 1993. Moreover, the rise in income inequality that occurred before that year appears to have been, at least in part, a byproduct of the remarkable success of a group of entrepreneurs who in the past few decades created countless jobs and contributed substantially to the higher living standards we all currently enjoy.

Increasing taxes on America’s most productive earners — those who create most of the jobs in our economy — will depress economic growth and reduce opportunities for the less fortunate. Rather than focusing on income inequality, policymakers should address the very real impediments to achieving equality of opportunity, particularly for the youngest and least-skilled workers among us. We believe such efforts should begin with fixing our k-12 education system, which is failing to train many young Americans to be competitive in today’s global labor market. If we can solve this problem, we will enable future generations of young people to climb the economic ladder and achieve the economic success that has long made the United States the world’s leading economy.

1. Gini coefficients cited here come from the Census Bureau report “Income, Poverty, and Health Insurance Coverage in the United States: 2010,” “The cia World Factbook 2010,” and Eurostat, the official statistical office of the European Union.

2. “Money income” is the first of fifteen definitions of income used by the Census Bureau.

3. “Employer Costs for Employee Compensation” (Bureau of Labor Statistics, December 2010).

4. The ten percent rise in income inequality is an estimate that excludes a 4.8 percent off-trend jump in the Gini coefficient between 1992 and 1993, caused primarily by a change in the way data were collected. This change in methodology biased the Gini calculation upward, leading the Census Bureau to warn that the periods before and after 1993 should not be compared.

5. For more, see and (these and subsequent weblinks accessed July 2, 2012).

6. Meyer and Sullivan’s paper is available at

7. The CBO defines income for the October 2011 report as “market income” (labor income, business income, capital gains and other capital income, and retirement income received from past service), plus transfer payments, less taxes.

8. The cpi-u-rs is constructed by the Bureau of Labor Statistics, while the pce is constructed by the Bureau of Economic Analysis and is used to measure gross domestic product. Both price indices are widely used to deflate consumption expenditures for the purpose of measuring real income (see We note that the Federal Reserve seems to have shifted focus from the cpi to the pce for some of the same reasons we identify (see

9. “Divided We Stand: Why Inequality Keeps Rising” (oecd, 2011), 23–24.

10. Emilia Niskanen, “The Luxembourg Wealth Study: Technical Report on lws Income Variables” (June 2007), available at “The Luxembourg Wealth Study” organization was founded in 1983 and is one of the few sources for international comparisons of income.

11. Data on U.S. gdp growth are from the Bureau of Economic Analysis. Data on g-7 (ex-U.S.) growth are from oecd.stat.

12. The eitc is now the largest cash transfer program for lower-income workers. See Hilary Hoynes, “The Earned Income Tax Credit, Welfare Reform, and the Employment of Low-Skilled Single Mothers” (University of California Davis, 2008).

13. See Kip Hagopian, “The Inequity of the Progressive Income Tax,” Policy Review 166 (April-May 2011). An unabridged version is available at

14. The data on shares of income and taxes are from “Summary of Latest Federal Individual Income Tax Data” (The Tax Foundation, October 2011).

15. “Growing Unequal? Income Distribution and Poverty in oecd Countries” (oecd, 2008), 106

16. Darius N. Lakdawalla and Jay Bhattacharya, “Does Medicare Benefit the Poor?” Journal of Public Economics 90 (2005), available at

overlay image