The Widening Income Gulf
by Isaac Shapiro and Robert Greenstein

Table of Contents

I.  The CBO Data and the Use of Projections for 1999

II. Recent Income Trends

III.  Changes in Income Shares

IV.  Other Data Produce Similar Findings

V.  Wealth Disparities are Much Larger than Income Disparities

VI.  Policy Implications

Appendix I

Appendix II

State and Local Taxes

View press release.

Congressional Budget Office data issued this summer indicate that after-tax income has increased dramatically since 1977 for the highest-income one percent of the population but risen only modestly for those in the middle of the income spectrum and declined for those in the bottom fifth. The CBO data, which start in 1977 and include projections for 1999, are widely regarded by analysts as the best data available on income and tax trends. These data include various forms of income that standard Census data miss, such as capital gains income and income from the Earned Income Tax Credit.

The CBO data indicate that after-tax income is more heavily concentrated among the richest one percent of the population — and also among the most affluent 20 percent of the population — than at any time from 1977 to 1995, the years for which historic CBO data are available. (CBO data are not available for years before 1977 or for 1996 through 1998. Census data on before-tax income are available going back to 1947; they show that before-tax income is at least as concentrated today at the top of the income scale as at any other time in the past half-century.)

An analysis of the CBO data finds:

Figure 1
Figure 1

Indeed, just the increase in the income of the top one percent of the population since 1977 is estimated to substantially exceed the total income of the bottom 20 percent of the population this year.

While most of these trends have been driven by underlying economic developments, a significant share of the increase in the after-tax income of the top one percent of the population is due to substantial net tax cuts that high-income households have received since the late 1970s. Even with the tax increases on high-income households that the 1990 and 1993 deficit reduction packages contained, these households pay a substantially smaller percentage of their income in federal taxes today than they did in 1977. If the richest one percent of households paid the same percentage of income in federal taxes in 1999 as this group paid in 1977, these households would pay an average of at least $40,000 more in taxes this year. In other words, these households have received an average tax cut of more than $40,000. This $40,000-plus average tax cut for the top one percent of households is greater than the entire average before-tax income of the middle fifth of households, which is projected to be $38,700 per household in 1999. (These and all other dollar figures in this analysis are in 1999 dollars.)

These exceptional income gains at the pinnacle of the income spectrum — along with the increasing disparities in income between the very wealthy and other Americans and the growing concentration of income among the most well-to-do — should be considered as the nation enters a debate over competing uses for projected budget surpluses, including a debate over whether to provide substantial tax cuts and to whom to direct any tax cuts that are provided. The tax bill Congress approved August 5 would direct the lion's share of its tax cuts to those with high incomes, which would further increase the concentration of after-tax income among those at the top of the income scale.

 

The CBO Data and the Use of Projections for 1999

The Congressional Budget Office has long compiled information on income and taxes based primarily on Census and IRS data. In July of this year, CBO released tables that provide a consistent measure of such income for selected years from 1977 to 1999. The CBO tables provide data for every other year from 1977 through 1995 and then provide 1999 figures reflecting CBO's projections for income and taxes. This analysis examines the projected 1999 data and compares them to data for previous years. The CBO tables released in July 1999 include data on both before-tax and after-tax income, as well as on federal tax rates, by income group. We calculated from these CBO data the shares of after-tax income that each income group would receive.

The CBO data are widely respected as the most complete measure available of disposable income because, in contrast to the standard data from the Census Bureau, the CBO data use an income measure that captures nearly all of the principal sources of income. For example, the CBO data include capital gains income while the official Census data do not. In addition, the CBO data allow for an assessment of income among the highest-income groups in the population and take into account the effect of federal taxes. The Census data do not show the full income of those at the very top of the income scale or show income after taxes.

Just the increase since 1977 in the income of the top one percent of the population will substantially exceed the total income this year of the bottom 20 percent of the population.

The CBO data also differ from Census data in that the CBO data reflect the reality that larger households have greater needs than smaller households. CBO assesses and categorizes the income of each household according to how much income it has relative to a standard of need for a household of its size. (See Appendix I for a more detailed comparison of the CBO data and the traditional income data from the Census Bureau.)

CBO's projections for 1999 were based on actual income for 1995 and CBO estimates made in January 1998 of how much various components of income, such as wage and salaries and capital gains income, would grow from 1995 to 1999.(2)

Why did we undertake an analysis of projected data for 1999 instead of relying on the actual data for 1995? First, the 1995 data do not account for the tax cuts enacted in 1997; the 1999 projected data do begin to show the effects of these cuts.(3) Second, the 1995 data do not reflect the considerable income growth that has occurred in the last four years while the 1999 data incorporate a reasonable estimate of this growth. For these two reasons, the 1999 data allow for the better assessment of current after-tax income distribution and trends.

Data from Different Sources Yield Similar Conclusions

The CBO data indicate that both the top one percent of the population and the top 20 percent will receive a larger share of the national after-tax income this year than at any time since at least 1977. Similarly, Census data available through 1997 on before-tax income show that the top five percent and top 20 percent of the income spectrum received as large or larger share of the before-tax income that year as in any year since Census began collecting these data in 1947.

CBO projects the top twenty percent of the population will receive 50.4 percent of the national after-tax income in 1999. The Joint Committee on Taxation’s projections for 1997, the last year for which the Committee has issued such data, show that the top 20 percent of households would receive effectively the same proportion of after-tax income, 50.5 percent.

Data compiled in substantial part from surveys the Federal Reserve Board has conducted show that wealth is more concentrated among those at the top than at any point since the Depression. Wealth is much more concentrated than income, which as this report indicates, is rather highly concentrated itself.

The use of projections naturally raises questions of reliability: when the actual figures for 1999 are available will they differ considerably from the projections used here? Will the key conclusions of this analysis be upheld?

The available evidence indicates that the actual CBO data for 1999 will be quite similar to the projected data(4) and that the central conclusions of this analysis will not change. The CBO projections are consistent with an array of data from other sources which indicate that income and wealth inequality are now at post-World War II highs and show a distribution of after-tax income that is similar to the distribution in the CBO projections.

As explained later in this paper, Census data through 1997 show that the share of before-tax income that the top 20 percent of the population receives grew as much just between 1995 and 1997 as the CBO projections indicate this group's share of before-tax income will grow between 1995 and 1999. The Census data also show that the concentration of before-tax income at the top of the income scale is at or tied for an all-time high for the post-World War II era. In addition, the most recent data from the Congressional Joint Committee on Taxation on the distribution of after-tax income show a distribution similar to CBO's projection for 1999. Finally, other data show that the richest households own a larger share of the nation's total wealth than at any time since the Depression.

While it is extremely likely that after-tax income will be more concentrated in 1999 than in the years from 1977 to 1995, it is more difficult to assess whether it will be significantly more concentrated in 1999 than in 1997 or 1998.(5) (CBO data for those two years are not yet available.) Significant changes in income concentration typically do not develop in a one- or two-year period, and economic data for the period from 1997 to the present do not provide a clear answer to this question. Wage gains since 1997 have been strong at the bottom of the wage distribution. On the other hand, these wage gains have been partially offset among some groups at the bottom of the income scale by larger-than-expected declines in the amount of welfare cash assistance received. At the top of the income scale, capital gains income appears to have remained at high levels and may have grown since 1997.

The all-time highs in the concentration of income that this analysis finds reflect a comparison of the projected data for 1999 to actual data for years from 1977 to 1995. It is possible that income concentration from 1997 to 1999 will be relatively constant, that the concentration in 1999 will modestly surpass the concentration for 1997 and 1998 or that income concentration will prove to have been slightly greater in 1997 or 1998 than in 1999. What is clear is that in the years that comprise the end of the 20th century, the income disparities between the highest-income Americans and the rest of the population will be greater than at any other point in at least the past quarter-century.

 

Recent Income Trends

Income disparities are greater in the years at the end of the 20th Century than at any time in at least the previous quarter century.

The CBO data show that when averaged across all households, after-tax income is expected to be 20 percent higher in 1999 than in 1977, after adjusting for inflation. But it is high-income households that pull up this average. The increase in average income is much greater than 20 percent for upper-income households and much lower for middle- and low-income households. (The CBO data do not track the incomes of particular households over time. Rather, the CBO data compare the incomes and taxes of households comprising certain parts of the income spectrum in one year to the households that comprise the same parts of the spectrum in another. For example, those households constituting the fifth of households with the highest incomes in 1999 are compared to the fifth of households with the highest incomes in 1977. Each year, somewhat different households make up the top fifth of households.)

Most of the growth in incomes at the top of the income scale occurred between 1977 and 1989. From 1977 to 1989, the average after-tax income of the top fifth of households climbed 33 percent. By contrast, between 1989 and 1999, average after-tax income of this income group is projected to rise only seven percent. Even so, growth in after-tax income has been somewhat uneven during the most recent decade. As compared to the seven percent average gain among the top fifth of households from 1989 to 1999, the middle fifth of households is projected to experience average income growth of only three percent, while the bottom fifth of households is expected not to gain at all.

When looking at the top one percent of the population, it is instructive to examine trends in both before-tax and after-tax income. The before-tax income of the top one percent of households is projected to grow 96 percent between 1977 and 1999. But the after-tax income of this group is projected to grow 115 percent, about one-fifth more. Similarly, the share of the national before-tax income that the richest one percent of the population receives grew less over this period than the group's share of the national after-tax income.

These trends reflect the fact that despite tax increases on high-income households enacted in 1990 and 1993, this group received large net tax cuts over the 1977-1999 period. CBO projects that the richest one percent of households will pay a significantly smaller share of its income in federal taxes in 1999 than in 1977.

The CBO data show, using conservative assumptions, that between 1977 and 1999 this group will receive a federal tax cut of 13 percent, worth an average of more than $40,000 in 1999 to households in the top one percent of the population. (According to CBO, the effective federal tax rate for the top one percent of households was 39.7 percent in 1977 and is expected to equal 34.4 percent in 1999. This represents a 13.4 percent reduction in the effective tax rate.)(6) Changes in federal tax policy over the period have exacerbated, rather than lessened, the large growth in income disparities that has occurred in the private economy.

The net effect of changes in tax policy since 1977 has been to provide the richest one percent of households an average tax cut worth more than $40,000 per household.

The trends are somewhat different for the most recent 10-year period, from 1989 to 1999, than for the 1977-to-1999 period as a whole. Although the before-tax income of the top one percent of households is expected to increase 12.2 percent during the most recent 10-year period, the after-tax income of this group is expected to rise at a much slower 3.2 percent rate. This reflects the increased proportion of income that the top one percent pays in federal taxes in 1999, compared to what it paid in 1989.(7) Nevertheless, the proportion of income that the richest one percent of households will pay in federal taxes in 1999 remains well below the proportion it paid in the late 1970s.

 

Changes in Income Shares

As a result of these wide divergences in income growth, the already-wide gaps between those with the highest incomes and other Americans have widened further.

Figure 2Figure 2

Figure 2 shows the shares of after-tax income for income groups in 1977 and 1999.(10) Some of the changes in the percentage of after-tax income that various income groups receive may seem relatively small. For example, in 1977, the middle fifth of the population received 16.4 percent of the national after-tax income; by 1999, this share is expected to slip to 14.7 percent. Yet seemingly small shifts in shares of national after-tax income reflect differences of tens of billions of dollars in the amounts of income that various groups receive.

Table 1

Shares of National After-Tax Income, 1997 vs. 1999

 

1977

1999 projected

Top One Percent of Population (2.7 million people in 1999)

7.3%

12.9%

Bottom 38 percent of Population (100 million people in 1999)

16.9%

12.9%

Source: CBPP analysis of CBO data

Table 2

Average After-Tax Income in 1999

 

 Projected

If group’s share of after-tax income had been the same as in 1977

The amount by which the group’s average income would be higher or lower

Lowest Fifth

$8,800

$12,100

$3,300

Second Fifth

20,000

23,700

3,700

Middle Fifth

31,400

34,900

3,500

Fourth Fifth

45,100

48,400

3,400

Highest Fifth

102,300

89,800

-12,500

Top One Percent

515,600

289,700

-225,900

Source:  CBPP analysis of CBO data; figures rounded to the nearest hundred dollars.

 

Other Data Produce Similar Findings

The distribution of after-tax income also can be assessed by examining data from Congress' Joint Committee on Taxation. This source does not provide data over a range of years but does provide recent data. Although CBO and the Joint Tax Committee use different methodologies, the results that their data yield concerning the share of after-tax income that households at the top of the income distribution receive are similar.

The latest such Joint Committee tables were released in 1997 and include projections for that year. These tables do not take into account the tax cuts enacted in 1997, which were of disproportionate benefit to high-income households. Those tax cuts are reflected in the CBO data.

Census data on before-tax income, which are available through 1997, also show the marked trend toward greater income disparities. The Census data show that in 1997, the proportion of the national before-tax income that the top 20 percent of households received — as well as the proportion the top five percent received — were at or tied for the highest proportions on record since the Census Bureau began collecting these data in 1967.(12) (Census does not publish data on the top one percent of households.) Meanwhile, the share of the national before-tax income received by the broad middle class — the middle 60 percent of the income distribution — was tied for the lowest on record. The share of before-tax income the bottom 20 percent of households received also was tied for the lowest level on record. (See Appendix II for a discussion of how the inclusion of in-kind benefits, which mitigate income disparities somewhat, and state and local taxes, which exacerbate income disparities somewhat, would affect these calculations. Census data on an alternative measure of income that includes in-kind benefits for years since 1979 show that when these benefits are counted, income disparities still were greater in 1997, the latest year for which these data are available, than in any other year in the period.)

In addition, the Census Bureau data show that the share of before-tax income that the top 20 percent of households receives increased by 0.7 percentage points just between 1995 and 1997. This is the same percentage-point rise in the share of before-tax income that the CBO figures project to occur for the top fifth of the population between 1995 and 1999.(13)

Figure 3
Figure 3

The Census Bureau also has compiled data on the shares of national before-tax income that families, as distinguished from households, receive. (The data on families do not include single individuals; such individuals are included in the household data used throughout this report). These data are of interest because it goes all the way back to 1947. The data show that the share of family income before taxes that those at the top of the income spectrum receive is as great or greater now than in any year on record, going back one-half century.

 

Wealth Disparities are Much Larger than Income Disparities

These findings underscore the extent to which income has become more concentrated in the United States. Federal Reserve Board and other studies show that the top fifth of households possesses an even larger share of the national wealth than of the national income.

Edward Wolff of New York University has complied data on the distribution of wealth in the United States over the past 75 years. These data indicate that wealth has become more concentrated in recent years than at any other time since the Great Depression.(14) Wolff's findings, based in substantial part on the Federal Reserve Board's Surveys of Consumer Finances, indicate that:

Wealth has become more concentrated in recent years than at any time since the Great Depression.

 

Policy Implications

Income and wealth trends since 1977 primarily reflect the effects of changes in the private economy. Earnings and capital gains trends, changes in technology and international trade, and the weakening of unions are among the factors that appear to be driving the rise in income disparities. Given these trends toward widening wage and income gaps, it would be reasonable to expect modifications in tax and other government policies to compensate at least modestly. At a minimum, government policies ought not to exacerbate these trends.

Certain governmental decisions, however, appear to be ignoring this backdrop. Most recently, the tax bill Congress approved in early August would grant large tax cuts to high-income households. When the tax cuts were fully in effect, the top one percent of households would receive an average tax cut of $31,800 a year, according to the Treasury Department, while the average tax cut for the bottom 60 percent of households would be $166 per year.

The top fifth of households would receive 78.5 percent of the legislation's tax-cut benefits when the bill was fully in effect, while the bottom three-fifths of households would receive 7.5 percent of the tax cuts. In addition, by repealing the estate tax, which affect the estates of only the wealthiest one percent to two percent of individuals who die, the bill would lead to still-larger concentrations of wealth. Also, by repealing what is known as the individual alternative minimum tax — which was designed to ensure that high-income people do not largely avoid taxation — the bill would enable some very high-income individuals to use a sufficient number of tax shelters and other tax devices so they no longer owe any significant amount of federal income tax.

This is a dubious set of priorities. As these data indicate, as a group, only high-income households are dramatically better off than in the past. Moreover, there is no economic need for a large tax cut for the wealthy; the economy is running at full tilt with current tax rates, and the stock market is booming at current capital gains rates.

Tax reductions such as those in the pending legislation seem unwise compared to other policy options. Sounder uses for these resources could include greater debt reduction; strengthening the long-term financial security of Medicare and Social Security; public investments that hold promise for improving long-term productivity growth (such as investments in education and training, infrastructure, research, and early intervention programs for children), measures to lower the U.S. poverty rate (including the still high child poverty rate); and tax cuts that accord a greater share of their tax reductions to the middle class and the working poor. Available resources could be used in sounder ways than to exacerbate the highly uneven and increasingly disparate distribution of income in the United States.


APPENDIX I
The Congressional Budget Office Data vs. Census Data

The CBO data reflect a fuller measure of income than the Census data do. The CBO data include capital gains income. They also take into account the effect of federal taxes, including the Earned Income Tax Credit. The standard measures of income that the Census Bureau publishes do not consider taxes and do not include capital gains income.

The failure to account for capital gains income in its standard series means that the Census Bureau information understates income at the top; another aspect of Census data has the same bias. Census Bureau data reflect a mechanism known as "top coding." Census data, for instance, recognize only the first $1,000,000 of earnings from a worker's primary source of employment. If an individual earns more than that the individual is recorded as having earnings of exactly $1,000,000.

CBO is able to estimate all earnings, including those above the Census cutoff. CBO also is able to account for capital gains income and taxes because of its use of Statistics of Income data from the Internal Revenue Services.

The CBO data differ in one other notable respect from the data the Census Bureau issues — CBO assigns households to income quintiles in a somewhat different manner than the Census Bureau does. The Census Bureau groups households strictly by the amount of income they have, so that the 20 percent of households with the lowest incomes become the bottom fifth, or quintile. CBO, on the other hand, accounts for the fact that larger households have greater income needs than smaller households. Accordingly, it adjusts each household's income according to the household's size, using economy-of-scale factors, before assigning the household to a quintile. This method accounts for the fact that a $20,000 income is far more ample for a single-person household than for, say, a household of five.

More specifically, CBO uses adjusted family income (AFI), which is family income divided by the poverty threshold, and CBO arranges households from low to high on the basis of this measure. So, for example, when the report says that according to the CBO data, the fifth of households in the middle of the income spectrum is projected to experience an increase in average after-tax income of eight percent from 1997 to 1999, that means that AFI (income adjusted for family size) is projected to increase by eight percent over that time period. Most economists consider AFI to be a superior measure of economic well-being.


APPENDIX II
Income Distribution, In-kind Benefits, and State and Local Taxes

The standard Census data on income, as well as the CBO data analyzed here, do not include the value of in-kind benefits. Since low- and moderate-income households receive more assistance from government in-kind benefits than high-income households do, these benefits moderate income disparities somewhat. But the effect is small. Even if such benefits are included, the share of national income these households receive remains quite modest. In addition, the inclusion of in-kind benefits has little effect on the trends in the distribution of income over time; income disparities have grown sharply over the past two decades whether or not in-kind benefits are considered.

This can be seen by examining various experimental measures of income that the Census Bureau employs. These measures, which currently provide information for years from 1979 to 1997, incorporate several types of income not included in the official Census income data — capital gains income, in-kind benefits, and the effects of federal income and payroll taxes, including the Earned Income Tax Credit.

Under all of these experimental income measures, income disparities have grown substantially since 1979. That is, growth in in-kind benefits has not been large enough to alter the trend of widening income inequality. Under the most comprehensive of these experimental measures, income disparities were wider in 1997 than at any other time since 1979. (Technically, disparities appear to have been wider under the Census experimental data in 1986 than in 1997. The 1986 data, however, are distorted and not a useful basis for comparison.)(15)

Moreover, the share of national resources that the bottom fifth of households receives does not increase that much when in-kind benefits are considered. Even when an expansive measurement of in-kind benefits that includes some valuation of Medicare and Medicaid benefits is used, the share of national income received by the bottom fifth of households rises by only 0.7 percentage points. When Medicare and Medicaid are excluded from measures of income as many experts believe they should be, but other in-kind benefits are counted, the share of national income the bottom fifth receives increases by 0.4 percentage points.

 

State and Local Taxes

While the inclusion of in-kind benefits would slightly moderate the extent of income disparities for any year, the inclusion of state and local taxes would operate in the opposite direction. The CBO data on after-tax income, as well as the Treasury Department and Joint Tax Committee data cited in the report, reflect income after federal taxes are subtracted; they do not subtract state and local taxes. Since state and local taxes are generally regressive — that is, low- and middle-income households pay a larger share of their income in these taxes than upper-income households do — the distribution of after-tax income would be shown to be more uneven if state and local taxes were taken into account.


Endnotes:

1. Here and elsewhere in the report, references to changes in after-tax income over time refer to changes in family income after an adjustment is made for family size (see Appendix I for a description of this adjustment). The average adjusted family income for a group in a certain part of the 1999 income spectrum, such as the 20 percent of households in the middle of the income spectrum, is compared to the average adjusted family income for the group in the same part of the income spectrum in 1977.

In this report, reference often is made to a group consisting of 20 percent of households or a fifth of households. The CBO data series actually divides the population into five groups or fifths with equal numbers of people, not equal numbers of households. For simplicity, we use the phrase "fifth (or 20 percent) of households" and "fifth of the population" interchangeably.

2. CBO Memorandum, Estimates of Federal Tax Liabilities for Individuals and Families by Income Category and Family Type for 1995 and 1999, May 1998, pp. 5 -6. The projections for 1999 used in this report were first made public in May 1998, but CBO released only in the past few weeks a revised time series that makes these data consistent back to 1977.

3. The tax cuts enacted in 1997 phase in over time. While many of these tax cuts are fully in effect as of 1999, some of the key tax reductions are not. For example, after 1999, taxpayers at somewhat higher income levels will become eligible to make deductible contributions to Individual Retirement Accounts.

4. Some information can be gained from examining how the CBO economic forecast for 1999, which was issued in January 1998 and relied upon for the CBO projections of after-tax income in 1999 used in this analysis (we will label this the "January 1998 forecast"), differs from CBO's most recent economic forecast for 1999 ("the July 1999 forecast").

The more recent forecast for capital gains income suggests that, if anything, this paper understates the likely concentration of after-tax income in 1999. The amount of capital gains income CBO forecasts for 1999 is considerably higher in the July 1999 forecast than in the January 1998 forecast on which this analysis relies. Since capital gains income is highly concentrated among those at the top of the income spectrum, this suggests that the actual amount of after-tax income the top one percent of the population receives will be higher in 1999 than this paper projects. CBO's forecast for the level of capital gains income in 1999 has risen from $331 billion in the January 1998 forecast to $395 billion in the January 1999 forecast.

The forecast for the overall size of the economy also has risen. The size of the economy (the Gross Domestic Product or GDP) is projected to be about three percent larger in 1999 under CBO's July 1999 forecast than under its January 1998 forecast. This suggests the overall increase in after-tax income from 1977 to 1999 might be larger — i.e., about three percent larger — than CBO originally projected and that all income groups might be a bit better off than under CBO's projection. Such a development would not affect the trends this paper discusses. For example, this paper reports that the after-tax income of the bottom 20 percent of the population is projected to be nine percent lower in 1999 than in 1977. If after-tax income turns out to be three percent higher for this group in 1999 than the CBO projections assumed, the decline in the average income of this group over this period would be six percent instead of nine percent.

5. It also is likely that after-tax income will be more concentrated in 1999 than in 1996, both because capital gains realizations have grown dramatically since 1996 and because Census data on before-tax income, which do not include capital gains income, show that before-tax income grew faster at the top of the income scale between 1996 and 1997 than in the rest of the income spectrum.

6. The figure that the top one percent of the population received an average tax cut of more than $40,000 was computed as follows: we took the difference between the percentage of income that the top one percent of households paid in federal taxes in 1977 and the percentage this group paid in 1999 and multiplied this difference by the group's projected average before-tax income for 1999. The CBO data show the top one percent of households paid an average of 39.7 percent of before-tax income in federal taxes in 1977 and will pay 34.4 percent in 1999. We multiplied the difference — 5.3 percent — by $786,000, the projected before-tax income of the top one percent of households in 1999. The result is an average tax cut of $41,700.

These calculations are rough benchmarks that are likely to understate the true tax cut for the top one percent of the population. This is so because these calculations do not adjust for the difference in the average effective federal tax rate that occurs when income rises. Since the federal tax system is progressive, a large rise in income leads to an increase in the percentage of income paid in taxes even if the tax code remains unchanged. Since the before-tax income of the top one percent of the population rose dramatically from 1977 to 1999, one would have expected the average effective tax rate of this group to rise if the tax system had remained the same. Instead, the average effective tax rate fell, underscoring how substantial the tax cuts have been.

Stated another way, comparing the average effective tax rates of the top one percent of households over the two periods, as we did, understates the size of the tax cut this group of households received. If the average income of the top one percent had been as high in 1977 as it is projected to be in 1999, the top one percent would have had a much larger amount of income taxed at the highest marginal tax rate in 1977, which was 70 percent, and thus would have had a higher average effective tax rate in 1977 than the 39.7 percent the CBO data show for that year. Hence, changes in the tax rate since 1977 actually reduced this group's taxes by more than 5.3 percent of its income.

7. The CBO data indicate that the top one percent of households paid 28.1 percent of income in federal taxes in 1989, a much lower percentage than in the 1970s. This figure rose to 36.5 percent in 1995 before dropping to 34.4 percent in 1999 (for an average tax cut of $16,500 between these two years). Under current law, it should decline further in coming years as more of the tax cuts enacted in 1997 take full effect.

8. As these figures indicate, the share of after-tax income received by the richest one percent of the population is expected to increase by 5.6 percentage points from 1977 to 1999. This would significantly exceed the 4.2 percent share of the national after-tax income that the bottom fifth of the population is expected to receive in 1999.

9. This gain in the share of after-tax income that the top fifth of the population receives is concentrated among the top 10 percent. The share of after-tax income that the next-to-the-top 10 percent of the population received is projected to decrease between 1977 and 1999.

10. Totaling the percentages in Figure 2 yields a figure of more than 100 percent. This is because CBO excludes households with negative incomes from the bottom quintile but includes such households when calculating total national income. If the households with negative income were included in the bottom quintile, the share of after-tax income this quintile receives would be lower.

11. Switching to before-tax income, the projected amount of average income in 1999 is $38,700 for the middle fifth of the population. For other groups, the projected amounts of average before-tax income are: bottom fifth, $9,200; second-poorest fifth, $23,200; next-to-the-top fifth, $57,900; top fifth, $144,300; and top one percent, $786,000.

12. The shares of before-tax income that the top five percent and top 20 percent of households received were both at their highest figures ever, but the differences between the shares for a few of the recent years are unlikely to be statistically significant.

13. The Census Bureau data show the share of before-tax income received by the top fifth of households rising from 48.7 percent in 1995 to 49.4 percent in 1997. The CBO data indicate that this share will rise from 53.2 percent in 1995 to a projected 53.9 percent in 1999. (Recall that the CBO data show the top fifth of households with more income than the Census data because the Census data do not capture a substantial amount of income that high-income households receive; see Appendix I.)

14. Edward N. Wolff, Top Heavy: A Study of the Increasing Inequality of Wealth in America, A Twentieth Century Fund Report, 1995, and Edward N. Wolff, "Recent Trends in Wealth Ownership," for Benefits and Mechanisms for Spreading Asset Ownership in the United States, forthcoming.

15. The 1986 data reflect a one-time jump in capital gains income that year, as high-income investors rushed to sell assets and "cash in" their capital gains before substantially higher capital gains tax rates took effect in 1987 as a result of the 1986 Tax Reform Act. The data for 1986 consequently reflect an artificially elevated level of income among those at the top of the income scale and provide a skewed picture of the share of the national income that high-income households were receiving in the mid-1980s. If the 1997 experimental data are compared to 1985 or 1987, income disparities are found to have been significantly wider in 1997 than in the mid-1980's.