[*PG993]THE MATTHEW EFFECT AND FEDERAL TAXATION

Martin J. McMahon, Jr.*

For whosoever hath, to him shall be given, and he shall have more abundance; But whosoever hath not, from him shall be taken away even that he hath.

—Matthew 25:29

Abstract:The “Matthew Effect” is a synonym for the well-known colloquialism, “the rich get richer and the poor get poorer.” This Article is about the Matthew Effect in the distribution of incomes in the United States and the failure of the federal tax system to address the problem. There has been a strong Matthew Effect in incomes in the United States over the past few decades, with an increasing concentration of income and wealth in the top one percent. Nevertheless, there has been a continuing trend of enacting disproportionately large tax cuts for those at the top of the income pyramid. Neither economic theory nor empirical evidence supports the argument that these tax cuts increase incentives to save, invest, and work. A growing body of economic literature supports the thesis that economic inequality impedes economic growth instead of fostering it. Furthermore, in a modern industrialized democracy, most of what everyone earns is attributable to infrastructure created by society acting through government. Paradoxically, public concern with increasing economic inequality is not matched by opposition to tax legislation that delivers vastly disproportionate benefits to the super-rich. This Article suggests that future tax legislation ought to mitigate the Matthew Effect rather than enhance it.

Introduction

The term, the “Matthew Effect,” was coined by sociologist Robert K. Merton in 1968 based on the passage from the Gospel of Matthew [*PG994]in the epigram.1 “Put in less stately language, the Matthew Effect consists in the accruing of greater increments of recognition for particular scientific contributions to scientists of considerable repute and the withholding of such recognition from scientists who have not yet made their mark.”2 The Matthew Effect is not limited to the context in which Robert Merton first coined it. More generally, it is a synonym for the well-known colloquial aphorism, “The rich get richer and the poor get poorer.” This Article is about the Matthew Effect in the distribution of incomes in the United States and the failure of the federal tax system to address the Matthew Effect.

Over twenty years ago, economist Paul Samuelson observed, “If we made an income pyramid out of a child’s blocks, with each layer portraying $1000 of income, the peak would be far higher than the Eiffel Tower, but most of us would be within a yard of the ground.”3 Things have changed a lot since then, and things have changed little since then. The peak is higher, but most people are still in essentially the same place. During the last two decades of the twentieth century, the distribution of incomes and wealth in the United States reached levels of inequality that have not been seen since the Roaring Twenties. Although the “Roaring Nineties,” as the decade was labeled by Joseph E. Stiglitz, might have been “the world’s most prosperous decade,”4 the prosperity was not spread around. The data indicate that a very small number of people garnered an overwhelming amount of the increase in incomes and wealth in that decade, as well as in the prior decade.

Between 1947 and 1997, median family income (in constant dollars) grew by 122%.5 Ninety-one percent of this growth, however, occurred before 1970. Between 1979 and 1997, average household before-tax income grew by nearly one-third in real terms, but that growth was shared unevenly across the income distribution. Average income for households in the top quintile rose by more than one-half, while average income for the middle quintile increased by only 10% [*PG995]and average income for the lowest quintile decreased slightly.6 Income growth at the very top of the distribution was even greater. Average before-tax income in 1997 dollars for the top 1% of households more than doubled, rising from $420,000 in 1979 to more than $1 million in 1997. Inequality continued to increase in the late-1990s.7

During the 1950s and 1960s, family income inequality decreased, but the tide changed after 1969, and through the last three decades of the twentieth century income inequality increased.8 Nevertheless, the federal tax system did little to ameliorate the increasing economic inequality. Prior to 1982, high marginal rates at the top had some redistributive effects. Redistributive effects were reduced as a result of the rate reductions in the Economic Recovery Act of 1981 (the “1981 Act”), and after the Tax Reform Act of 1986 (the “1986 Act”), the redistributive effect of the income tax was relatively low.9 Adoption of the 39.6% bracket in 1993 increased the redistributive effects of the income tax, but redistribution decreased again as a result of the reduction in capital gains rates in 1997. As of 2000, the redistributive effect of the income tax was somewhat less than it was in the early 1980s, although it was somewhat greater than it was in the early 1990s.10 As we move into the new millennium, however, recent changes in the federal tax system presage a decreasing role not only in redistribution, but in mitigation of vast disparities in income and wealth. Since the inauguration of the Bush Administration11 in 2000, there have been three major tax acts, which have reduced significantly the tax burden of the super-rich, while handing out small change to everyone else.

Part I of this Article examines in detail the increasing concentration of income and wealth in the top 1%, and particularly within much narrower cohorts near the top of the top 1%, that has occurred [*PG996]over the past twenty-five years.12 This Part demonstrates the strong Matthew Effect in incomes in the United States over that period. The super-rich are pulling away from everyone by so much and at a rate so fast that the fact that incomes of many households at the bottom and in the middle have stagnated, or even fallen in constant dollars, has been obscured by ever increasing per capita income—a false talisman of progress because it obscures distributional issues.

Part II discusses the distribution of after-tax income and wealth in the United States in recent years.13 Wealth and income levels are highly correlated. This Part describes the increasing disparity in after-tax incomes, particularly the rate at which the amount and share of total after-tax income of the top 0.5%, and even of smaller cohorts further toward the top of the income pyramid, are growing relative to everyone else. Moreover, the share of wealth owned by the super-rich is growing even faster than its share of income. This Part demonstrates that the federal tax system has failed to respond adequately to take into account ever increasing income inequality.

Part III examines changing effective federal tax rates over the last two decades of the twentieth century, examining with more precision the aspects of the federal tax system that have failed to respond adequately to ever increasing income inequality.14 After first discussing the various major legislative changes in this period, the Part then examines the shifting burdens, measured by effective tax rates on different income cohorts, of the various federal taxes individually and collectively. Part IV then reviews the economic literature on the effect of these changes on the progressivity of the tax system. It concludes that by the close of the twentieth century the tax system was not raising revenue as fairly and was doing less to mitigate inequality than it had in the middle of that century.

Part V describes the Republican tax policy agenda for the new millennium, as embodied in tax legislation enacted in 2001 through 2003 and discusses the projected distribution of the benefits of the massive tax cuts enacted in that brief period.15 The projections show that the tax cuts disproportionately favor those at the top of the income pyramid with very small tax cuts going to everyone else, even the upper middle class and the merely rich, in contrast to the super-rich.

[*PG997] Part VI deals with economic issues.16 It starts by demonstrating that theory does not support the argument that the tax cuts were necessary to spur incentives to save, invest, and work, and that the empirical evidence of the effect of tax cuts on savings and investment clearly contradicts the claims made by supporters of the tax cuts. Next, this Part examines the rapidly growing body of economic literature supporting the thesis that economic inequality impedes, rather than fosters, economic growth. Thus, not only do the tax cuts not spur economic growth, but because they increase inequality, they probably impede economic growth. This Part then examines empirical data that debunk the notion that “a rising tide lifts all boats,” by demonstrating that increasing incomes for the few and decreasing incomes for the many can occur even though the economy is “growing,” that is, the national GDP is increasing. Distribution thus counts. Finally, the Part briefly notes the disastrous long-term economic effects of the massive federal budget deficits largely attributable to the Bush tax cuts.

Part VII discusses the philosophical basis for a highly redistributive tax system, arguing that in a modern industrialized democracy, most of what everyone earns is attributable to infrastructure created by society acting as a whole, principally through government.17 It rejects the notion that individuals have the first claim to everything that they earn, and although it does not label it as such, adopts a more communitarian approach.18 This Part then briefly discusses the deleterious effect of increasing concentrations of wealth on the future health of democratic institutions.

Part VIII examines the paradox of public concern with increasing economic inequality, thinking it undesirable, while simultaneously supporting tax cut legislation that in fact delivers vastly disproportionate benefits to the very wealthy—the super-rich.19

The Conclusion suggests that it is time for the tax system to address these problems by substantially increasing progressivity at the top of the income pyramid. Marginal tax rates should be increased for incomes in excess of $500,000, and as incomes increase to progres[*PG998]sively higher levels, additional rate brackets should be added to impose substantially higher marginal rates on incomes in excess of $1 million and particularly on incomes that exceed $5 million. Future tax legislation ought to mitigate the Matthew Effect rather than enhance it.

I.  The Distribution of Before-Tax Income

F. Scott Fitzgerald was right when he had a character quip, “Let me tell you about the very rich. They are different from you and me.”20 Both income and wealth in the United States are highly concentrated in a very small percentage of the population, and wealth is somewhat more concentrated than income. Although the data from various sources often use slightly different measures of the relevant unit and the precise measurement of income or wealth, the pattern is consistent. One percent or less of the population is remarkably different than everyone else. Compared to everyone below them, the top 1% are in a class by themselves. But that is not all. Although the data are not as complete for subgroups within the top 1%, there are enough data to indicate that even the top 1% is not a homogenous group. The cr�me de la cr�me—the top 0.01%, or even smaller subgroups—have so much income and wealth that these groups merit separate consideration in any analysis.

There are several methods for comparing income distributions, including actual money incomes for different income classes, incomes for different groups with reference to an index number, for example, as a multiple of the poverty rate, household income ratios, the percentage of national income received by different income classes, and the Gini index.21 These are all valid measures of income inequality [*PG999]and changes in income inequality.22 A wealth of data is available on income distributions, and it reveals very unequal distributions no matter which method is chosen.

A.  The Rich Are Getting Richer

1.  The “Merely Rich” Are Running Away from the Pack

Recent data from the Congressional Budget Office (the “CBO”) provide the best perspective on the phenomenal growth of the income of those households in the highest income cohorts and the ever increasing income inequality over the past two decades. Other data are available from the Census Bureau and the Internal Revenue Service (the “IRS”), but the CBO data provide the more comprehensive perspective.23 Although the Census Bureau data are available for a number of definitions of income, they do not generally include capital gains and do not adequately break out those households within cohorts smaller than the top 5%.24 IRS data, which are based primarily on adjusted gross income (the “AGI”) shown on tax returns, do not adequately reflect economic income, do not consistently identify the top 1% and top 5% cohorts, and are available with respect to taxpayers rather than households. One important piece of information from the Census Bureau data, however, is that between 1979 and 2001, the income ratios—the multiple of the average income of the lower percentile that is the average income of the higher percentile—for every income percentile above the fiftieth percentile increased relative to lower income percentiles, whereas the income ratio of the tenth to [*PG1000]fiftieth percentile was the same in 2001 as it was in 1980.25 The data thus demonstrate that although those in the middle of the income distribution did not gain relative to the poor, the households in the top half pulled away from households below them. Furthermore, within that top half, households relatively higher in the income distribution pulled away from households relatively lower in the income distribution at an increasing rate.

The most recent CBO data show not only that the income inequality inexorably increased throughout the last two decades of the twentieth century, but that income inequality—particularly with respect to the rate at which those at the very top of the income pyramid pulled away from everyone else—increased in the 1990s more than in the 1980s.26 In 2000, before-tax income was more concentrated in the top 1% than at any time since 1929.27 The increasing income disparities between the top 40% and the bottom 60% between 1979 and 1993 was attributable to the combination of a decline in real income of the bottom 40% and stagnation of the income of the middle quintile, coupled with modest income growth for the fourth quintile and significant income growth for the top quintile, particularly for the higher cohorts within the top quintile. From 1993 to 2000, the three lowest quintiles experienced a not insignificant increase in real incomes. Nevertheless, due to dramatic increases in their incomes, the upper income quintiles—particularly the top 10%—actually pulled away from the lower income quintiles at a much greater rate in the mid-to-late 1990s than they did in the period from 1979 to 1993, as is demonstrated in the following table.

This table vividly demonstrates that the real incomes of the top cohorts—the top 5% and the top 1%—grew dramatically more than did the incomes of all of the other cohorts.30 The top 5% saw its average income increase at nearly nine times the rate of increase for the mid[*PG1002]dle quintile and at nearly five times the rate of increase for the fourth quintile. The top 1% saw its average income increase at nearly fifteen times the rate of increase for the middle quintile and at nearly eight times the rate of increase for the fourth quintile. The bottom 60% saw two decades of nearly stagnant or very modest real income growth.31

Nevertheless, the manner in which the CBO data are presented masks the real disparity between the top 1% and the remainder of the top 5% and top 10%. Because the average income grows at an increasing rate, the averages for large cohorts, including the top 1%, are distorted. The data more accurately present the true picture if they are recalculated separately to state the average pre-tax income of the 81st to 90th percentiles, the 91st to 99th percentiles, and the top 1%.

The data in this table clearly illustrate that the top 20% is not a group that can be lumped together meaningfully when discussing income distribution and the role of taxes in effecting redistribution. The average income of the 81st through 90th percentiles is closer to the average income of the fourth quintile than it is to the average income of the 91st through 95th percentiles. Even the average income of the 91st through 95th percentiles is closer to the average income of the fourth quintile than it is to the average income of the 96th through 99th percentiles. And the average income of the 96th through 99th percentiles, although not even twice the average income of the 91st through 95th percentiles is dwarfed by the average income of the top [*PG1003]1%. More significantly, the top 1% saw its average income increase at six times the rate of increase for the 81st through 90th percentiles and at more than three times the rate of increase for the 96th through 99th percentiles. The top 1% is leaving everyone else in the dust.

2.  The “Super-Rich” Are Soaring Above the Merely Rich

Just as the aggregate data for the top quintile hide the extraordinary differences between the top 1% and the remainder of the cohort, the aggregate data for the top 1% hide extraordinary differences within that select group. Data for smaller cohorts within the top 1% are difficult to obtain, and when they are available they often are based on different income measures, different statistical descriptions, and a different base year for measuring changes in constant dollars. The preceding data—mostly CBO data—were based on an expanded income concept and described mean income for the respective cohorts. Another measure of differences between income classes is the threshold income for each income cohort. The greater the difference between the threshold income necessary to be included in the cohort and the average income of the cohort, the greater the income inequality within the cohort itself. The threshold incomes necessary to enter each quintile, and smaller cohorts within the top quintile in 1979 and 2000, using a comprehensive money income calculation starting from AGI (including capital gains, but excluding in-kind receipts), measured in constant 1982 through 1984 dollars, were as follows:33

According to these data, between 1979 and 2000, the threshold to climb out of the bottom 60% fell in real dollars, meaning that this [*PG1004]group would have fallen in relative terms even if the top of the income pyramid had no increase in real income. But the top 20% realized increasing real incomes. Even the threshold for the top 20% increased by only 8.2% in real terms. At the same time, the thresholds for the top 1% and top 0.5% roughly doubled, which is significantly more than the increase in thresholds for the top 10% and top 5%. The threshold for the top 0.25% increased by 153%, and the threshold for the top 0.1% roughly tripled, indicating that the super-rich are pulling away from the nearly super-rich at an astonishing rate.34

The increasingly elite status of the super-rich can be put in the perspective of the earlier CBO data by examining the thresholds for entry into the top cohorts in current 2000 dollars, which facilitates a comparison with the CBO data on average incomes. According to the CBO data, the average income of the top 1% in 2000 was $1,290,800. The threshold for entering the top 1%—albeit using a different, less comprehensive, income definition based on IRS data—was $354,035.35 The threshold for the top 0.5% was $554,335; for the top 0.25%, it was $902,317, and for the elite top 0.10%, it was $1,696,322.36 Within the top 1%, so much of the income was concentrated in the top 0.10% that even those at the threshold for the top 0.25% had an income of roughly only two-thirds of the average for the top 1%.37

[*PG1005] The story of super-elites does not stop with the top 0.10%. Within the top 0.10% is another ultra-elite, dubbed the “Fortunate 400.”38 Based on IRS data, in 2000, the threshold income for joining this group, constituting the top 0.00031% of tax returns, was an AGI of $86.8 million, and the average AGI was $173.9 million.39 To qualify for this group requires an income more than fifty times the threshold income for joining the top 0.10%. This jump within a cohort of less than 0.10% exceeds the gap between the credentials for entering the top one-tenth of 1% and entering the top 40%—the $1,696,322 annual income necessary to join the top 0.10% was forty-seven times more than the $36,014 annual income necessary to join the top 40% in 2000.

The Fortunate 400, however, is a fluid group that changes significantly from year to year—over the nine years from 1992 through 2000, a total of 3600 returns were identified as belonging to this group, with fewer than 25% of taxpayers within this group appearing twice and fewer than 13% appearing more than twice. Furthermore, incomes of the members of this club consist largely of capital gains—over 70% of the group’s total AGI in each of 1998, 1999, and 2000 was net capital gains.40 Nevertheless, it is worth noting that if the lowest income member of this elite group, with an income of $86.8 million, realized the group’s average percentage of AGI as capital gains in 2000, that taxpayer’s capital gains would have been nearly $62 million. If that were averaged over even a thirty-year holding period, annual income from capital gains alone would have exceeded $2 million in 2000 constant dollars, which would have put the taxpayer substantially above the threshold for the top 0.10% in every year.41 Thus, even the lumpiness of capital gains realizations does not affect the status of members of this [*PG1006]ultra-elite group as members of the super-rich class. Those who have made the Fortunate 400 even once are in an elite class.

Although the key to joining the Fortunate 400 may be capital gains, this is not true with respect to the remainder of the top 1%. The percentage of income realized as capital gains increases as the percentile of the income cohort increases, but recent observations indicate that wage and entrepreneurial income is the dominant form of income for all income cohorts. Only for the top 0.5% do capital gains approach or exceed 20% of the income.

The percentage of the income of the top 5% realized in the form of wages (including stock options), in contrast to capital gains and periodic income from capital, has increased steadily over the last half of the twentieth century, and the percentage of income realized as wages has grown dramatically for the smaller cohorts at the very top.

This change in income composition of the top 1% is not attributable, however, to changes in the pattern of realization of capital gains and periodic income from capital; both remain highly concentrated in the highest-income cohorts.43 Rather, the change in income composition is attributable to dramatic increases in the wage level of top earners relative to everyone else—the phenomenon of the winner-take-all market economy of the United States at the turn of the millennium.44 More than half of the “very top” taxpayers derived the major part of their income in the form of wages and salaries.45 The “working rich” dominate the smallest measured percentile cohorts, if [*PG1007]not the Fortunate 400. Part of this might be attributed to athletes and other entertainers, many of whom earn astronomical salaries,46 although there also is evidence to the contrary.47 But most top 1% income earners—admittedly in this context a group that include the “wannabes” as well as the truly “income-rich” are engaged in business or professions.48 The dramatic increases in the compensation of the chief executive officers (and certain other officers) of publicly held corporations also play a part. CEO pay has risen astronomically in the past forty years. Whereas the average CEO made forty-one times as much as the average worker in 1960, by 2001 the average CEO made 411 times as much as the average worker, and that was a decrease— possibly temporary due to a decline in the stock market—from the levels in the immediately preceding years, in which average CEO pay was as much as 531 times the level of the average worker.

What kind of pay are we talking about here in dollars? According to BusinessWeek’s “52nd Annual Executive Pay Scoreboard,” the average CEO’s pay in 2001 was $11 million—many times the multiple necessary for entry into the top 0.10%—and that was a 16% decrease from the 2000 average.50 As do all averages, however, the 2001 average presents a somewhat distorted picture. Lawrence J. Ellison, the CEO of Oracle, earned so much due to pocketing $706 million from exercising stock options, that the rest of the CEOs averaged only $9.1 mil[*PG1008]lion, a low not seen since 1997.51 Many, of course, made far less than even that average. To break into the top 40 required annual compensation of over $40 million.52 But one does not even have to be a CEO to garner such munificent compensation. At least ten executives below the CEO level made between $58 and $128 million in 2001,53 and annual pay of more than $1 million is common for the second banana in publicly held corporations.54

3.  The Super-Rich Are Taking a Substantially Bigger Slice of the Pie

So far, we have been examining relative average incomes of various percentile cohorts of the population. Another perspective on the distribution of incomes is to examine the percentage of total personal income realized by the various income cohorts. Over the final two decades of the twentieth century, the top 5% increased its share of national before-tax personal income at the expense of virtually every other group, including much of the top quintile. This small subset of the top quintile increased its share of incomes dramatically, and even most of that increase accrued to the top of the top of the pyramid. The data as presented by the CBO illustrate that the higher income cohorts gained at everyone else’s expense.

[*PG1009]The picture of income distribution in the last years of the twentieth century is stunning. In both 1997 and 2000 the top 20% of the households had more income than everyone else combined, as well as in all of the intervening years.58 The top 5% had more income than the bottom 60%. And the top 1% had nearly one-half as much as the bottom 40%. Expressed slightly differently, in 2000, the 1.1 million richest households measured by income had nearly one and one-half times more money than the 44.2 million poorest households measured by income.59

Again, the CBO presentation of the top quintile masks the difference between the top 1% and everyone else in the top quintile. Breaking down the top quintile into cohorts that exclude higher level cohorts reveals that the bottom half of the top quintile joined the bottom 80% in transferring a slice of the pie to those who were better off, and that within the top 10%, only the top 5% made any significant gains from 1997 through 2000, with most of those gains going to the top 1%.60

Significantly, in the late 1990s, the pattern of increasing income shares differed from earlier years. Unlike earlier periods, from 1997 to 2000 it was not the top 20%, or 10%, or even the top 5% that was gaining income share at everyone else’s expense. From 1997 to 2000, [*PG1010]the only income cohort that increased its share of total income was the top 1%.

Although the CBO data do not address the share of economic income realized by smaller cohorts within the top 1%, some such data are available, and they parallel the data regarding increased dollar incomes. As any income cohort is broken down into increasingly smaller cohorts, the share of incomes realized by each successive higher level income cohort increases at an increasing rate. A somewhat more precise view of the pattern within the top 1% is revealed by calculations by Thomas B. Petska, Michael I. Strudler, and Ryan Petska from IRS Statistics of Income data based on AGI—a very different base than that used by the CBO. Although the percentages differ somewhat, the pattern is consistent. Their data show that almost one-half of the share of income realized by the top 1% is realized by the top 0.10%.62 Thus, the top 0.10%—roughly 110,000 households out of nearly 110 million households in the Untied States at that time—had roughly the same income as the 1.9 million households immediately below them near the top of the pyramid. Their analysis of the data also shows that this elite 110,000 households realized a greater share of income than the 44.2 million households in the bottom 40%.63

B.  Income Mobility and the Fallacy of the Horatio Alger Myth

Americans believe in the Horatio Alger myth. They love, and believe in, rags-to-riches stories.64 Opponents of progressive taxation use such anecdotal stories of income mobility to fight progressive taxation [*PG1011]on the grounds that income inequality merely reflects life cycle differentials. But the data tell a different story.

Examination of available data leads to the inescapable conclusion that the Horatio Alger myth is exactly that, a myth. Although some Americans experience “significant” income fluctuations from year to year,65 the data do not support the conclusion that many households frequently move between broadly defined income classes.66 An Urban Institute study found that in both the 1970s and 1980s, about half of the people in either the lowest or highest quintile at the beginning of the period were in the same quintile ten years later.67 Another study found that about half of the young adults (ages twenty-two to thirty-nine) who were in the bottom quintile of the income distribution in 1968 still were in that quintile twenty-three years later, in 1991.68 More significantly, three-quarters of those who were in the bottom quintile in 1968 were in the bottom 40% in 1991.69 According to another study, only 13.8% of those who are in the bottom 30% for any given year have lifetime income in the top 30%, and only 2.6% of those who are in the top 30% for any particular year have lifetime income in the bottom 30%.70 Both top to bottom mobility and rags-to-riches mobility are thus quite rare.

Focusing on the top of the income pyramid, 90% of those in the top decile for their age cohort at age forty-nine were in the top two deciles at age seventy-nine, and only 2% of individuals in the top decile for their age cohort at age forty-nine had fallen below the top three deciles by age seventy-nine.71 At the top, then, almost all of the [*PG1012]mobility is up, not down.72 This finding is confirmed by other studies which show income mobility within one or two deciles, but not much income mobility across more dispersed deciles, within any particular age cohort.73 Furthermore, because the percentage of people changing income category from one year to the next declined somewhat between the late 1960s and the early 1990s, income mobility diminished. Thus, the sharp increases in income disparities reflect true growth in disparities and not merely a reshuffling of the income distribution.74

All of this implies that increasing “income inequality within a single year is mirrored by a similar increase in inequality over Americans’ lifetimes.”75 The data on income mobility support, rather than impugn, the case for graduated progressive taxation, both on the grounds of fairness and to effect redistribution.

II.  The Distribution of After-Tax Income and Wealth

A.  After-Tax Dollar Incomes

When all is said and done, what is most important is the distribution of after-tax income.76 By this measure, despite its progressivity, the federal tax system really has done little to ameliorate the increasing disparities in income over the last two decades of the twentieth century. The CBO data show increasing after-tax income disparities.

Mirroring the changes in before-tax income,79 calculations based on the CBO data show each successive income class climbing up the income distribution pyramid realized a greater percentage increase in after-tax income than the income group below it. The second and third quintiles pulled significantly ahead of the lowest quintile, while the fourth quintile pulled ahead of the middle quintile by more than the middle quintile pulled ahead of the lowest quintile. And the top quintile appears to be in a class by itself. The percentage increase in its after-tax income outpaced the fourth quintile by four times as much as the increase for the fourth quintile exceeded that of the middle quintile.

[*PG1014] As with before-tax incomes, the CBO’s presentation, which does not adequately break out the smaller income cohorts within the top quintile, masks the extent to which increases in averages for the top quintile, top 10%, and top 5% actually are attributable largely to enormous increases in income of the top 1%.

The rate of after-tax income growth of all of the top quintile, except the top 1%, more nearly resembled the rate of income growth of the third and fourth quintiles than it did the top 1%. The after-tax income of the top 1% increased by nearly 150 percentage points more than the percentage by which the after-tax income of the 96th through 99th percentile increased, whereas the after-tax income of the 96th through 99th percentile increased by only 38 percentage points more than the percentage by which the after-tax income of the middle quintile increased.

The differences between the increases in before-tax income and the increases in after-tax income from 1979 to 2000 illustrate the special status of the super-rich. On the one hand, because tax rates generally fell during the period from 1979 to 2000, the first four quintiles saw their after-tax income increase at a higher percentage than the percentage at which their before-tax income increased.81 On the other hand, the fifth quintile saw its after-tax income increase by a slightly lower percentage than the percentage at which its before-tax income increased. But this did not necessarily reflect increased progressivity. The tax cuts in the various tax acts in that time period were not distributed evenly across or within quintiles. Many of the tax cuts were effected through increases in the earned income tax credit, af[*PG1015]fecting the first and, to a lesser extent because of its phase-out rules, the second quintile.82 Other tax reductions were effected through items such as the child and education credits, which due to phase-out rules affected primarily the second through fourth quintiles, and the bottom of the fifth quintile.83

Breaking the top quintile down in to smaller income cohorts reveals that the 81st through 90th percentiles saw after-tax income increase at a higher percentage than the percentage at which its before-tax income increased. But the 91st through 99th percentile after-tax income increased by a lesser percentage than the percentage by which its before-tax income increased—a small difference for the 91st through the 95th percentile and a significant amount for the 96th through 99th percentiles. When we get to the top 1%, we discover the big winner. After-tax income grew by 17 percentage points more than before-tax income increased. No other income class saw after-tax income increase by more than 2.5 percentage points more than before-tax income increased.

B.  Shares of Total After-Tax Income

As illustrated in Part II.A, the top of the economic pyramid realizes an extraordinarily disproportionate share of before-tax income.84 Policy makers have not responded to this ever-increasing growth in the disparity of incomes with any changes to the tax system that would reallocate [*PG1016]the tax burden to reflect these significant changes in the relative ability to pay taxes. Instead of increasing the progressivity of the tax system to meaningfully mitigate the rate at which the gulf between the rich and poor is widening, the policymakers have allowed the after-tax gulf to widen dramatically. The CBO data on the shares of after-tax incomes realized by each income class confirm this conclusion.

On an after-tax basis the top 5% has gained a share of income at everyone else’s expense. Even the 91st through 95th percentile has lost income share to the top 5%.86 And within the top 5%, the top 1% has grabbed the biggest share of the bigger slice of the pie, leaving the 96th through the 99th percentile only an additional sliver—at least compared to the top 1%’s extraordinarily increased share of pie.

[*PG1017]C.  Shares of Total Wealth

Wealth is strongly correlated to income.87 Precise measurement of the distribution of wealth is difficult, because the data are difficult to collect.88 Nevertheless, the various sources reveal consistent patterns, although details of the data may differ. Wealth in the United States is even more highly concentrated than income.89 Like incomes, wealth has been becoming increasingly more concentrated,90 but not at the same rate as the rate of growth of the concentration of incomes.91 Wealth, however, is more concentrated at the top of the pyramid than income. By some estimates, for over a decade, the top 1% has held nearly 40% of the total value of net wealth, while the top 5% has held nearly 60% of net wealth.92

[*PG1019]Other estimates show less concentration in the top 1%, offset by a larger share held by the 90th to 99th percentile, but the measure of the difference is not so significant as to change the import of the data.94 In any event, the top 1% alone holds more wealth than the bottom 80% or 90%.95

Although consistent data for smaller cohorts within the top 1% are difficult to obtain, what data there are demonstrate that the same pattern that occurs with respect to income distributions occurs with respect to wealth. The top of the top is different from the bottom of the top. In 1989, the top 1%, by wealth, owned 33.5% of all assets. Of this, the 99th to 99.4th percentile held 7.4%, and the 99.5th to 100th percentile—the top 0.5%—held 26.1%.96 Net worth was even slightly more concentrated.97 More recent data from the Internal Revenue [*PG1020]Service Statistics of Income Division (albeit using a different baseline measure of wealth) indicate that in 1998 the top 0.5% held over three-quarters of the wealth held by the top 1%.98 This is consistent with data from the Federal Reserve Board’s Triennial Survey of Consumer Finances for 1998.99 The data consistently show the top 0.5% holding more than 25% of personal net wealth in the United States.

Furthermore, in examining the distribution of wealth, there is an analogue to the IRS’s Fortunate 400 highest income earners. For wealth, the group is the Forbes 400—a list of the 400 wealthiest Americans published annually by Forbes magazine. From 1989 to 1999 the threshold for joining this elite group grew by 74%, to $609 million, measured in constant 1998 dollars.100 The average wealth of the top ten individuals grew by 611% to nearly $27.1 billion. Some estimates conclude that from 1989 to 1999, the percentage of total wealth held by the Forbes 400 grew from 1.5% to 2.6%. (With the stock market decline in the early 2000s, the Forbes 400’s share of total wealth is estimated to have fallen to 2.2% in 2001.101) Other estimates conclude that this group’s share of total wealth grew from about 1% in the early 1980s to closer to 3% in 2002.102 This growth in wealth was not even, and it was not evenly distributed. Between 1989 and 1995, most measures of the wealth of the wealthiest people grew fairly modestly in real terms, but from 1996 through 1999, there were dramatic increases. But the data indicate that the most significant increases in wealth were at the very top, and they tapered off at lower levels.103

It is true that much wealth in the United States today is newly created. On the one hand, Forbes magazine’s list of the ten wealthiest persons in the country in 2003 includes self-made billionaires Bill Gates, Warren Buffett, Paul Allen, Lawrence Ellison, and Michael Dell. On the other hand, many fortunes are inherited; five of Sam Walton’s heirs—individually, not collectively—also made the top [*PG1021]ten.104 The Walton heirs’ status is just one illustration that much of the wealth in the United States is dynastic.105 Many other members of the club that constitute the Forbes 400 acquired their wealth by inheritance. In 1999, half of the Forbes 400’s fortunes originated with inherited wealth.106 A number of studies indicate that approximately 50% of the wealth in the United States is inherited.107 Thus, although there is churning of identities within this elite group, there is still a high degree of stability of high wealth status.108

Regardless of whether we are considering inherited wealth or self-created wealth, one thing is clear. Because high incomes and high wealth are highly correlated, if the progressivity of the tax system at the high end continues to erode, wealth will become even more concentrated in the future than it is now.

III.  Effective Tax Rates

A.  Individual Taxes

1.  The Individual Income Tax Rate Schedule

Progressivity has always been an essential element of the income tax in the United States, even though progressive income tax rates have always been controversial.109 The 1913 income tax had a low, relatively flat-rate structure with generous exemptions. Although high marginal rates—going to over 90%—were enacted to fund World War [*PG1022]I, those high rates were rolled back in the 1920s. Steeply graduated progressive rates, rising to higher than 90% at the top of the scale, reappeared in the 1940s in response to the need for revenues during World War II. High marginal rates for taxpayers in the top decile continued to characterize the income tax rate schedules until 1981.110 For the most part, however, until the inflation-driven bracket creep of the mid-1960s, the income tax system was largely flat-rate or mildly progressive for the masses, with steeply progressive surtaxes on a relatively small percentage of the population.111

Progressivity went into decline in 1981, partially rebounded in the 1990s, but has never recovered to its pre-1981 level. Between 1981 and 1985, the largest percentage cuts in individual income tax rates went to the highest income groups.112 The decline of progressivity began with the 1981 Act, which eliminated all marginal brackets above 50%.113 The eliminated brackets applied almost exclusively to current yield from capital, but the changes had the important ancillary effect of reducing the maximum rate on long-term capital gains from 28% to 20%. In general, income from capital got a big break. In addition, through adjustments in the remaining rate brackets, taxpayers in al[*PG1023]most every rate bracket received approximately a 10% rate reduction. Every prong of the 1981 Act reduced progressivity. The effect of elimination of the brackets above 50% is obvious, but it was even more anti-progressive than it appears on the surface because it was, to a large extent, tax relief for dividends, the receipt of which is highly concentrated in the highest-income classes.114 That the other two changes reduced progressivity is not quite so facially obvious, but it is equally certain. The benefits of reduction of capital gains rates inure disproportionately to high-income taxpayers because capital gains realizations are highly concentrated in high-income taxpayers.115 Even the 10% across-the-board reduction in rates reduced progressivity. Across-the-board percentage cuts increase inequality in private before-tax income because they reduce taxes of higher-income taxpayers proportionately more than they reduce taxes of lower-income taxpayers.116

Five years later the rate structure was radically changed by the 1986 Act, which reduced the rate brackets to 15% and 28%.117 Because of significant base broadening, the elimination of the preferential treatment of capital gains, and an increase in effective corporate tax rates, the 1986 Act resulted in decreased tax effective rates for all income classes below the top quintile.118 The lowest income classes benefited significantly from the 1986 Act, primarily due to expansion of the earned income credit, but across most of the spectrum, the 1986 Act was to a large extent distributionally neutral. Some analysts [*PG1024]find the changes to have been mildly progressive,119 although others find the changes to have mildly reduced progressivity.120 Nevertheless, because the net effect of most of the changes in the 1986 Act was to lock in the effect of changes in the tax acts in 1981, 1982, and 1983, from the perspective of the top 1% vis-�-vis everyone else, effective tax rates after the 1986 Act were less progressive than they were immediately before the 1981 Act.121

As subsequently analyzed by the House Committee on Ways and Means, the changes in the individual income tax burdens in the 1980s lopsidedly favored those at the top of the very top of the economic pyramid.

These data show the top 1% getting twice as much tax relief as the middle class, and an even higher multiplier of the tax relief than the nearly rich in the 96th through 99th percentiles.

The disproportionate tax cuts accorded to the very highest income class in the 1980s set the stage for the introduction in 1991 of the 31% bracket and in 1993 of the 36% and 39.6% brackets, the lat[*PG1025]ter applying to taxable incomes over $250,000 (indexed for inflation).123 The higher rates enacted in 1993 were intended solely to apply to those who benefited the most from the tax cuts of the 1980s. The new higher brackets initially affected less than 4% of taxpayers— those at the very top of the income distribution.124 Nevertheless, the increased progressivity fostered by those rate increases was somewhat ameliorated four years later when, in 1997, the rates on most long-term capital gains were significantly reduced, from 28% to 20%.125

2.  Payroll Taxes

Important changes in the 1970s and 1980s that dramatically affected the distribution of overall federal tax burdens had nothing to do with the income tax. In addition to income taxes, the federal government levies payroll taxes on wages and self-employment income. Payroll taxes were first introduced in 1935 to fund the Social Security system, but they now also fund Medicare (starting in 1965) and federal unemployment compensation. Payroll taxes are imposed on wages and self-employment income, starting with the first dollar, with no exceptions or exclusions,126 but they are dramatically reduced after wages or self-employment income exceeds an applicable ceiling for the year. Although the payroll tax rate is proportional with respect to its base, the burden of the tax is regressive. Regressivity results from the combined effects of the absence of a floor exempting some income and the imposition of a ceiling on wages subject to the largest portion of the tax.127 Statutory payroll tax rates have risen from 8.8% in 1967 to 15.3% currently.

The payroll tax (excluding unemployment compensation) currently consists of two components. The first component is the 6.2% [*PG1026]Social Security tax on wages below a specific ceiling that increases each year as wages in the economy generally increase. For 2003 the Social Security component of the payroll tax is levied on the first $87,000 of wages, without any exemptions, for a maximum of $5394. The second component of payroll taxes is the 1.45% Medicare tax on all wages, again without any exemptions, but without a ceiling. The payroll tax is collected from both the employer and the employee, so the Social Security component actually is 12.4% (a maximum of $10,788) and the Medicare component is 2.9%, for a total of 15.3%. Self-employed individuals pay these percentages on self-employment income, subject to the same ceiling on the Social Security component. From the employee’s side, payroll taxes are neither deductible nor creditable in computing income taxes, but self-employed individuals may deduct one-half of self-employment taxes in computing income taxes.128 Employers deduct their share of payroll taxes if the wages are paid in a profit-seeking activity.

Although income tax rates—particularly the rates imposed on the top of the income pyramid—have fallen dramatically over the last forty years, both the payroll tax rate and base have increased markedly during that period, dramatically increasing payroll tax receipts. Through 1949, the combined employer and employee Social Security tax was 2% (1% each), which was imposed on wages up to $3000.129 Over the years, as necessary to keep the system solvent, Congress increased both the rate and the ceiling. Starting in 1975, the ceiling was increased for inflation (except from 1979 through 1981 when ceilings were increased ad hoc by Congress).130 Generally speaking, the goal was to collect, in any given year, payroll taxes somewhat more than enough to pay current benefits, but not to fully fund accrued benefits.131 From 1974 through 1982, the combined (employer and employee) payroll tax rate (including Medicare, added in 1965) rose from 11.7% to 13.3%.132 In 1983, spurred largely by the urging of Alan Greenspan to better fund [*PG1027]future accrued benefits,133 Congress revised the Social Security system. The goal was to increase the excess of taxes collected over benefits paid out in order to further increase the balance in the Social Security trust fund.134 Under the Social Security Amendments of 1983,135 the combined Social Security payroll tax rate rose to 12.4% and the combined Medicare payroll tax rose to 2.9%.136 The 1983 legislation also provided a formula to increase the rate at which the ceiling on Social Security taxes was raised. As a result, the payroll tax burden has increased substantially, with higher rates and a ceiling on the Social Security portion of the tax that increases annually for inflation. From the mid-1960s through 2003, payroll tax receipts increased from approximately 20% to about 40% of federal revenues.137 Over the last twenty years, however, the growth of payroll taxes has been attributable to increases in the wage ceiling rather than rate increases. The rates have not been increased since 1990.

Payroll taxes in excess of those necessary to fund the Social Security system and Medicare on a pay-as-you-go basis are “invested” in the “social security trust fund,” which consists solely of a special issue of Treasury bonds.138 The proceeds of the sale of those Treasury bonds to the Social Security trust fund are then used for general governmental expenditures, for example, military, farm subsidies, interest on the national debt, federal payroll, and so forth.139 In other words, increased payroll taxes fund expenditures that generally are thought by most taxpayers to be funded by primarily the income tax. Over [*PG1028]$150 billion was added to the “trust fund” in 2002.140 At the end of 2002, “the trust fund” held approximately $1.4 trillion,141 meaning that cumulatively nearly $1.4 trillion dollars collected by the payroll tax ostensibly to fund Social Security and Medicare had been spent on general government operations. It is estimated that another $1 trillion will be added to the trust funds—that is, spent on general government operations—by the year 2007.142

B.  Effective Tax Rates

One of the most frequently referenced norms for determining progressivity is what is known as “effective tax rates.” An “effective tax rate” is determined by dividing tax liabilities (total or with reference to the allocable burden of a specific tax) by total income.143 This method of analysis generally defines income in a normative manner, including in income many items that are exempt from taxation.144 Tax liabilities taken into account generally include actual taxes paid. Therefore, the method does not employ tax expenditure analysis, under which normative taxes equal the normal statutory rate applied to a normative base, and the difference between the normative taxes so calculated and actual tax liability is treated as an offsetting government subsidy to the taxpayer.145 Because of their differing methodologies, effective rate analysis using actual tax liabilities and tax expenditure analysis are mutually inconsistent and cannot be applied simultaneously.146 Most, if [*PG1029]not all, of the available data relating to tax burdens reflect actual tax liabilities rather than normative tax liabilities, and the data regarding the distribution of tax expenditure benefits are far less refined, even when they are available. Moreover, in the political arena, tax expenditures generally are viewed as “tax relief.” Thus, this Part focuses on effective rate analysis exclusive of tax expenditure analysis.

The CBO has published two major studies on effective tax rates in the past few years, one covering changes in the period 1979 through 1997,147 and a second covering changes in the period 1997 through 2000.148 The CBO studies provide data not only with respect to the overall effective federal tax rates, but with respect to the impact of income taxes, payroll taxes, and corporate taxes on each income quintile, as well as the top 10%, top 5%, and top 1%.

1.  Income Tax

According to the CBO data, from 1979 to 1997, the effective income tax rate fell for the first four quintiles, but increased slightly for the top quintile.149 The CBO also shows that the smaller cohorts within the top quintile, the top 10%, top 5%, and top 1%, saw slight increases in their effective federal income tax rates from 1979 to 1997, after dipping substantially after the 1986 Act and before the institution of the 36% and 39.6% marginal brackets in 1993.150 From 1997 to 2000, the effective individual income tax rates for the lowest income quintile rose (but remained negative due to the refundability of the earned income credit). Effective income tax rates fell for those in the second and middle quintiles, remained constant for households in the fourth quintile, and rose for those in the highest quintile (including all smaller cohorts within the highest quintile).151 These changes [*PG1030]in effective individual tax rates, however, were not due to statutory changes. According to the CBO, the increased individual income tax rates were attributable to bracket creep—inflation-adjusted income growth pushed more households into higher tax brackets—and disproportionately high income growth realized at the very top of the income distribution, which combined to make a larger share of income subject to the highest tax rate.152

How can the CBO data show an increase in effective income tax rates when rates have been cut? Part of the answer lies in changes in the base, because effective rate analysis uses an expanded definition of income, not AGI or taxable income.153 But more importantly, even the CBO studies themselves caution that its methodology and data can “mask or even misrepresent information about subgroups or specific taxes.” The CBO points out that total effective tax rates can rise between any two years, even if effective rates for households in every income quintile fall.154 Likewise, effective tax rates for a quintile can rise even though statutory tax rates for a subgroup remain constant or fall. As shares of income shift upward—that is, the real income of a higher income cohort increases disproportionately to a lower income cohort, a higher percentage of income is taxed at higher rates.155 Finally, because of tax preferences, such as exclusions from taxable income for pension plan contributions, which are counted in total income in the CBO data, and the preferential rate for capital gains, shifting composition of income within an income class can affect the effective tax rate even if statutory rates remain unchanged. This last factor is very important. The data show that in the 1990s a smaller percentage of the income of top income earners was in the form of capital gains than it was in the 1980s and a larger percentage of the income of those earners was wage and other income taxed without any preference.156 In 1998 more than half of the “very top [*PG1031]taxpayers derive[d] the major part of their income in the form of wages and salaries. . . . [T]he ‘working rich’ celebrated by Forbes magazine have overtaken the coupon-clipping rentiers.”157

A study by Thomas B. Petska, Michael I. Strudler, and Ryan Petska has reached a different conclusion than the CBO regarding effective income tax rates.158 Petska, Strudler, and Petska employed a “retrospective income concept,” which uses the income and deduction items available in the 1979 to 1986 period as the base, and found that all income classes, except the lowest class and the top 10%, realized a substantial decrease in average tax rates from 1979 to 2000. In contrast to the CBO, however, they concluded that average tax rates for the top 1% of the income distribution decreased substantially from 1979 to 2000, with the top 0.10% of taxpayers having a 15.7% decrease, from 31.41% to 26.48%, and the remainder of the top 1% seeing a 14.07% decrease, from 27.43% to 23.57%.159 In contrast, the remainder of the top 10% of taxpayers saw less than a 5% decrease in average tax rates.

2.  Other Federal Taxes

Individuals’ overall tax burdens reflect not only the income tax, but also payroll taxes, corporate income taxes, excise taxes, and wealth transfer (estate, gift, and generation-skipping) taxes. Individual income tax receipts constitute slightly less than 50% of total federal tax receipts.160 In the late 1970s, corporate income taxes represented about 15% of federal tax receipts, but in recent years corporate income taxes have dropped to less than 10% of total federal taxes. In addition, various excise taxes collect slightly less than 4% of total taxes.

Payroll taxes are particularly important. In recent years, payroll taxes have risen to an amount equal to nearly 40% of federal tax receipts. Payroll taxes nominally are imposed to finance Social Security and Medicare specifically.161 If earmarking of these receipts for these specific transfer programs is accepted at face value, it might be difficult [*PG1032]to evaluate these taxes without considering the distribution of the benefits they provide. Nevertheless, because federal expenditures apart from the transfer programs for which payroll taxes nominally are earmarked far exceed taxes other than payroll taxes (primarily, individual and corporate income taxes), and payroll taxes far exceed current Social Security and Medicare expenditures,162 payroll taxes are to a large extent financing current general purpose government expenditures.163 Thus, in analyzing tax burdens, payroll taxes should not be considered to be any different than the individual or corporate income taxes.

All of these other taxes, which ultimately are borne by individuals, should be taken into account along with the income tax in determining the progressivity of the federal tax system. The CBO studies have done so by determining the effective rate for each of these taxes for the various income classes.164 In computing the effective rates, the CBO assumed, as do most economists,165 that the employer’s share of payroll taxes is borne by the employees. Thus, the amount of those taxes was included in employees’ income, and the taxes were treated as part of employees’ tax burden. The CBO treated corporate taxes as borne by owners of capital and allocated corporate taxes to households in proportion to their income from interest, dividends, rents, and capital gains.166 Finally, the CBO assumed that excise taxes are borne by [*PG1033]households according to their consumption of taxed goods (tobacco and alcohol) or—in the case of excise taxes that affect intermediate goods—in proportion to overall consumption. Under this analysis, corporate taxes fall more heavily on taxpayers in the higher-income classes; social insurance tax rates are higher for the middle-income classes; and excise taxes fall disproportionately on low-income households.167

a.  Payroll Taxes

Most households pay a larger amount in payroll taxes than in income taxes. As previously explained, economists generally agree that even though the employer nominally pays one-half of total payroll taxes, the entire burden is borne by employees.168 Taking into account both the employers’ and employees’ shares of payroll taxes, 70% or more of households have paid more in payroll taxes than in income taxes, and that has been true for every income category below the top quintile since 1988.169 Payroll taxes are regressive because they are based on a flat rate and for the most part are subject to a ceiling. According to the CBO data, effective wage tax rates are lower for the top quintile than for any other income class.

The overall effective payroll tax rate increased fairly steadily from 1979 to 1994, as Congress increased the levies to deal with financing Social Security and Medicare. Since 1994, however, the overall effective payroll tax rate has been falling, as an increasing percentage of [*PG1034]total income, mostly realized by the highest income households, is not subject to payroll taxes or is above the ceiling for the Social Security portion of the payroll taxes. Effective payroll tax rates for 1979, 1994 (the peak overall payroll tax rate), and 2000 were as shown in the following Table.

Effective wage tax rates have been higher than effective income tax rates for households in the first four quintiles for every year since 1984, and for households in the first three quintiles in all years of the CBO studies, and have continuously risen.171 Because of the inherent structure of the payroll taxes, the highest income cohorts have experienced the lowest percentage point increases in effective rates.

In analyzing overall tax burdens, payroll taxes properly are taken into account for two reasons. First, the relationship between payroll taxes paid and Social Security and Medicare benefits received is very tenuous.172 Second, as already noted, since 1983 payroll taxes have been set at a level that is significantly more than adequate to fund the Social Security system and Medicare on a pay-as-you-go basis, and the excess revenue from payroll taxes funds general expenditures. In this regard, it is important to note that, these “excess” payroll tax receipts made the most significant contribution to the transitory surplus of the late 1990s that was “returned to the taxpayers” by the Bush tax cuts of 2001 and 2003. The “refund,” however, did not go to the taxpayers who paid the payroll taxes; most of the “refund” went to the top 1% of the income pyramid.

[*PG1035]b.  Corporate Taxes

Corporate taxes are important to progressivity because corporate taxes are borne disproportionately by high-income taxpayers. As noted previously, the CBO studies treat corporate taxes as borne by owners of capital and allocate corporate taxes to households in proportion to their income from interest, dividends, rents, and capital gains.173 Under this assumption, high-income households bear a disproportionately large share of the burden of corporate taxes, and the tax is highly progressive.174 A decline in the effective corporate tax rate benefits high-income households more than other households.175

The percentage of total federal taxes represented by corporate income tax collections has fallen dramatically in the last forty years. Prior to 1968, corporate tax receipts consistently represented more than 20% of total federal taxes. Corporate income taxes fell below 10% of total federal tax receipts for the first time after the 1981 Act, which significantly reduced the statutory rates and provided much more generous cost recovery allowances (depreciation) than had previously been allowed. Through much of the 1990s, corporate tax receipts hovered around 11.5% of total federal taxes, before dropping back to about 10% in the last years of the twentieth century. In the first years of the twenty-first century, corporate income taxes plunged to 8% or less of total federal taxes.176

[*PG1036] In part, the decline in the relative importance of the corporate income tax to total tax receipts has been attributable to the increasing relative importance of payroll taxes. But the corporate tax itself has become less burdensome through both rate reductions and erosion of the base, the latter primarily through increasingly generous depreciation deductions.177 Some analysts also have attributed its decline to the rise of corporate tax shelters in the 1990s,178 but this proposition remains controversial.179 Nevertheless, it is undeniable that effective corporate tax rates (measured at the corporate level) have fallen significantly in recent years, even apart from any significant statutory changes.180 Consequently, effective corporate tax rates measured at the individual level also have fallen.

From 1979 to 2000, the overall effective corporate income tax rate for individuals fell from 3.4% to 2.5%, although the overall rate [*PG1037]fell to as low as 1.4% in 1982 and rebounded to as high as 2.9% in 1997. Comparison of 1979 and 1982 data for the “all quintiles” category helps to identify the 1981 Act as the key point in the decline of effective corporate tax rates.

As should be expected, most of the benefit of the decline of the corporate income tax inured to the highest-income cohorts—not the top quintile, not even the top 10%, but to the top 5%, and within that small group, mostly to the top 1%, whose effective corporate tax rate was halved. The magnitude of the decline depends on the share of the cohort’s income derived from capital,182 and the top 5%, and particularly the top 1%, realize a significantly greater proportion of their income as income from capital than do classes lower in the income distribution.183 The greater reduction in the impact of corporate taxes for the highest-income classes has reduced progressivity.

c.  Excise Taxes

Finally, the federal government imposes a variety of excise taxes, for example, gasoline, cigarette, and liquor taxes. Excise taxes claimed a fairly constant share of overall income—at or just under 1%—between 1979 and 2000 despite increases in statutory rates. But that consistent overall rate obscures significantly different effects within different income categories. Members of the lowest quintile first saw excise taxes increase from 1.6% of their income in 1979 to 2.6% in 1994, before dropping back to 2.2% in 2000. In 2000, the second quintile’s effective excise tax rate was 1.4%, while the third [*PG1038]quintile’s was 1.2%, both only 0.1 percentage points higher than in 1979. In contrast, the top quintile saw its excise tax rate drop from 0.7% in 1979 to 0.6% in 2000, while the top 1% excise tax rate dropped from 0.05% to only 0.3% over the same period. In short, the continuing overall effect was to make a regressive tax even more regressive.184 Excise taxes claimed more than five times the share of income from the lowest-income households than they claimed from the highest-income households.185

3.  Total Effective Tax Rate

In the end, what is important from the broadest tax policy perspective is not the progressivity of any one tax, but the progressivity of the tax system. One tax might be changed so as to enhance progressivity, whereas another tax is changed to lessen its progressivity. Neither of the changes standing alone provides an adequate viewpoint for public policy analysis.186

In its 1997 study, the CBO concluded that total federal taxes had become more progressive from 1979 to 1997. By this the CBO meant that the federal tax system had served to narrow the gap between taxpayers at