An Analysis of New IRS Income Data
by Isaac Shapiro, Robert Greenstein, and Wendell Primus

This report in PDF
Press Release:  HTML or PDF
Appendix:  HTML or PDF

To access the PDF files, right-click on the underlined text, click "Save Link As," download to your directory, and open document in Adobe Acrobat Reader.

Recent data from the Internal Revenue Service indicate that disparities in income between those with very high incomes and other Americans have widened sharply in recent years, with the average after-tax income of the top one percent of taxpayers rising $121,000 — or 31 percent — between 1995 and 1997, while the average after-tax income of the bottom 90 percent of taxpayers increased three percent. These IRS data reflect actual information from tax returns.

While assessments based on Census data have suggested income disparities might have remained flat since about 1993, the new IRS data show this is not the case. As many researchers in the field long have noted, Census data alone cannot answer questions related to income trends, because they miss very large amounts of income received by people at the top of the income scale. For instance, the official income data the Census Bureau publishes each year do not include capital gains income, which totaled $356 billion in 1997 according to the IRS information and doubled in just two years from 1995 to 1997. Reflecting the incompleteness of its data for those at the top of the income scale, the Census Bureau does not publish income information for the top one percent of the population. The IRS data are the sole reliable source of data on the incomes of these individuals.(1)

The IRS data reflect adjusted gross income, as reported on income tax returns, as well as the federal income taxes paid on that income. The IRS data do not include forms of income that are not included in adjusted gross income, such as certain government cash benefits. They include data on income taxes but not payroll taxes or other taxes. We focus here on adjusted gross income after federal income taxes, which we refer to as "after-tax income." All income figures are adjusted for inflation and expressed in 1997 dollars.

This analysis focuses on changes between 1995 and 1997. The Congressional Budget Office has published extensive data on income through 1995, relying heavily on the IRS data;(2) what is new here are data covering the years from 1995 to 1997. This analysis also provides some data for the 1993-1997 period, since people looking only at Census data sometimes think that income disparities stopped widening around 1993.

The principal findings of this analysis include:

There are some limits to the IRS data. The IRS data are best for measuring income and income trends at the top of the income distribution. They are more limited in measuring incomes at the very bottom of the distribution. They do not include households that have incomes so low that they do not file a tax return. In addition, the adjusted gross income measure does not include sources of income that are especially important to low-income households, such as certain forms of government cash assistance and in-kind benefits. (See the Appendix for a full discussion of the strengths and weakness of the IRS data in comparison to other data sources.)

The IRS data also do not break out information in a manner that would permit examination of income trends just among those in the middle of the income spectrum, such as the middle fifth or middle three-fifths of taxpayers. The closest approximation of this group that can be derived from the IRS data consists of the next-to-the-top quarter of tax filers (i.e., those whose income places them between the 50th and 75th percentiles of income for all tax filers). Among filers in this group, average after-tax income rose 3.4 percent between 1995 and 1997, the same amount as the rise in income for the bottom 90 percent of filers as a whole.

As a result of these sharp differences in income growth, income disparities between the highest-income individuals and the rest of the population grew substantially between 1995 and 1997, reaching their widest point over the period the IRS has collected these data. The data go back to 1986.

Table 1
After-tax Income Trends for the Top 1% and Bottom 90% of Tax Filers
IRS Data, 1986-1997
(income figures adjusted for inflation and expressed in 1997 dollars
income is adjusted gross income, minus federal income taxes)
Average After-tax Income Top 1% Bottom 90%
1986 $273,562 $23,451
1987     339,490 23,834
1988 448,223 23,750
1989     417,239 23,527
1990 403,753 23,145
1991     357,587 22,721
1992 398,508 22,982
1993     366,150 22,757
1994 369,535 22,939
1995     396,545 23,041
1996 448,704 23,168
1997     517,713 23,815
Change 1993-97 41.4% 4.6%
1995-97 30.6% 3.4%

Note on these data: The after-tax income of these tax filers (particularly in the bottom 90 percent) is lower than would be expected for two reasons; 1) families can have more than one tax filer and their income would not be combined into one family; and 2) not all sources of income, most notably government cash assistance, are counted in adjusted gross income. Working in the opposite direction, the data subtract only individual income taxes. Other taxes such as payroll taxes or the corporate income taxes are not considered.

Source: CBPP compilation of data from David Campbell and Michael Parisi, "Individual Income Tax Rates and Shares, 1997," SOI Bulletin, Internal Revenue Service, Spring 2000, Tables 5 and 6.

Also of note, the average income tax rate paid by the top one percent of filers declined between 1995 and 1997, from 28.7 percent to 27.9 percent. This appears to reflect two developments. Capital gains income, which is taxed at a lower rate than other income for those in the top income bracket, constituted a larger share of the income of these filers in 1997 than in 1995. In addition, legislation enacted in 1997 cut the capital gains tax rate, starting in that year.

These data on income trends are among a number of factors that merit consideration as policymakers deliberate over what uses to make of projected budget surpluses. Information on the effect of various tax and spending proposals on different income groups should inform these deliberations. For example, a recent analysis of the tax cuts that the House, the Senate, or both chambers approved before the August recess found that when these tax measures (which carry a substantial cost) would be fully in effect, half of the tax-cut benefits they provide would accrue to the five percent of taxpayers with the highest incomes.(4)

The remainder of this analysis examines these findings in more detail.

 

The Internal Revenue Service Data

The Internal Revenue Service regularly issues data showing annual trends in adjusted gross income(5) and income taxes for various groups of tax filers. The IRS recently issued such data for 1997.(6) These data are particularly useful in that they include comprehensive income information for people at the top of the income spectrum.

The IRS data are significantly more useful in this respect than the more widely cited data from the Census Bureau. The standard data on income that the Census Bureau publishes do not include capital gains income. The IRS data show that income from capital gains totaled $356 billion in 1997, with two-thirds of this income going to the 1.5 percent of tax filers with the highest incomes, those with adjusted gross incomes exceeding $200,000. The Census also place an upper limit on the amount of certain types of income counted for any individual, disregarding income above these amounts; this is done for confidentiality and other reasons.(7) The Census also relies on a survey method that may lead to significant under-reporting of income by those at the top of the income scale. As a combination of these and other factors, the Census data significantly understate the income of those at the top of the income spectrum, as well as their recent gains in income (as the next section of this paper documents).

A third source of data on income and tax trends is a data series that the Congressional Budget Office maintains. The CBO data series is based heavily on the IRS data for its information about income at the top of the spectrum, while making several improvements in the IRS data. The IRS data do not cover the entire population; they do not cover households that lack any adjusted gross income, have negative income, or have incomes so low they do not file tax returns. CBO supplements the IRS data with Census data so it can cover the entire population. In addition, CBO data take into account the effect of all federal taxes, while the IRS data reflect federal income taxes only.

The latest CBO data on actual income and tax trends cover years through 1995. CBO also has issued income and tax projections for 1999.(8) Last year, the Center on Budget and Policy Priorities issued an analysis of the CBO data through 1995 and the CBO projections for 1999.(9) The next set of CBO data on actual income and tax trends will reflect the new IRS data for 1997.

 

Recent Changes in Income

The IRS data indicate that both between 1995 and 1996 and between 1996 and 1997, income growth was striking among the one percent of filers with the highest adjusted gross incomes. In the first of those two years, average after-tax income grew 13 percent among this group, after adjusting for inflation. It climbed 15 percent in the second year. Over the two-year period, the average after-tax income of this group rose 31 percent.

While this paper focuses on after-tax income because it is a better measure of the amount of income a family has at its disposal, it bears noting that the income trends described here have been driven primarily by changes in before-tax income and not by changes in the tax system. Between 1995 and 1997, the average before-tax income of the top one percent rose by 29 percent, compared to an increase of 31 percent for after-tax income. Among the bottom 90 percent of filers, average before-tax income increased 3.6 percent, while after-tax income rose 3.4 percent.

Despite the differences in methodology between the Census Bureau and IRS measures, both show similar rates of income growth for the bottom 90 percent of the population between 1995 and 1997. The Census information on before-tax income shows an average increase in income over these two years of 3.6 percent for the bottom 80 percent of the population, the same rate of income growth as the IRS data show for the bottom 90 percent of filers (see Table 2).

The two data sources diverge only for those at the very top of the income spectrum, with the Census data failing to capture the rapid income growth for this group because those data miss capital gains and certain other income these people receive. For example, the Census data show a rate of income growth for the top five percent of households between 1995 and 1997 that is less than half of the income growth the IRS data show to have occurred among the top five percent of filers.(10) (This comparison uses the top five percent because Census data are not broken out for the top one percent.)

Table 2.
IRS and Census Data Diverge at the Top,
But Follow Similar Pattern for the Rest of the Population

Real change in income,
adjusted for inflation

1995-1997

1993-1997

At the top Top 1% Top 5% Top 1% Top 5%
IRS, after-tax income 30.6% 19.2% 41.4% 26.5%
IRS, before-tax income 29.1 19.7 41.2 28.3
Census, before-tax income na 8.3 na 11.6
For the bulk of the population* Change
1995-1997
Change
1993-1997
IRS, after-tax income, bottom 90% 3.4% 4.6%
IRS, before-tax income, bottom 90% 3.6 5.1
Census, before-tax income, bottom 80% 3.6 6.8
*Note: The IRS breaks its data out for the bottom 90 percent of tax filers but not the bottom 80 percent. The Census data are broken out for the bottom 80 percent but not the bottom 90 percent.

 

Significant Gains at the Top Likely to Have Occurred Again in 1998

Since 1997, there have been gains across the income spectrum. Census data indicate that average before-tax income for the bottom 80 percent of households grew by 3.2 percent from 1997 to 1998, after adjusting for inflation. Other economic data, including ongoing wage growth, suggest the broad population has continued to experience increases in income since then.

While the data the IRS has released to date do not break out overall changes in income for different income categories between 1997 and 1998, these data do indicate, however, that income gains continued to be significantly larger for the very top of the income spectrum than for the rest of the population and that income disparities consequently continued to widen. The preliminary data for 1998 that the IRS has issued indicate that between 1997 and 1998, the total amount of capital gains income rose about 20 percent, after adjusting for inflation.(11) This is a very substantial increase (although it represents roughly half the rate of increase in capital gains income that occurred in 1996 and 1997).(12)

Capital gains income is heavily concentrated. The preliminary data for 1998 indicate that 72 percent of the capital gains income received that year went to filers with adjusted gross incomes of $200,000 or more. These data suggest that in 1998, the incomes of the top one percent of tax filers are likely once again to have risen robustly — and considerably faster than the income growth that average Americans secured.

 

Income Disparities

The disparate patterns in after-tax income from 1995 to 1997 between those at the top of the income spectrum and the rest of tax filers resulted in a significant widening of income disparities. The IRS data show that the percentage of the national after-tax income going to the top one percent of tax filers increased significantly in these years, while the share going to the bottom 90 percent of tax filers declined.(13)

The IRS data go back to 1986. The share of after-tax income that the top one percent of filers received in 1997 is the highest recorded over this period. The share of after-tax income the bottom 90 percent of tax filers received is the lowest.


End Notes:

1. The CBO data described later in the report provide an accurate snapshot of trends at the top because CBO relies on the IRS information for this income group.

2. The CBO data also included income projections for 1999.

3. The top one percent of tax filers used here are those whose adjusted gross income, before taxes, places them in the top one percent of filers. In 1997, these were filers with adjusted gross incomes of more than $250,736.

4. See Robert Greenstein, James Horney, and Iris Lav, "Tax Cuts That the House or Senate Has Approved Would Consume All of the Realistically Available Surpluses," Center on Budget and Policy Priorities, July 28, 2000 -- /cms/index.cfm?fa=view&id=495. This report examined data on the distribution of Congressional tax cuts from the Citizens for Tax Justice.

5. The IRS study states, "Adjusted gross income is 'total income,' as defined by the tax code, less 'statutory adjustments' (primarily business, investment, or certain other deductions, such as payments to a Keogh self-employed retirement plan, certain deductible contributions to an Individual Retirement Arrangement (IRA), and self-employed health insurance deductions). Total income includes, for example, salaries and wages, taxable interest, dividends, alimony, and net amounts from such sources as business income, rents and royalties, and sales of capital assets."

6. David Campbell and Michael Parisi, "Individual Income Tax Rates and Shares, 1997," SOI Bulletin, Internal Revenue Service, Spring 2000. These IRS data cover tax filers with positive adjusted gross income. The vast majority of the population falls into this category. Those not included are tax filers who have negative AGI or incomes so low they do not file tax returns.

7. For example, the highest salary that can be recorded is $999,999. That is, anyone with a salary above this amount is considered to have a salary of $999,999.

8. The CBO projections for 1999, which were first released more than two years ago, indicated that income disparities would continue to widen between 1995 and 1999. The new IRS information indicates the disparities are indeed growing and, in fact, are growing considerably more rapidly than CBO estimated they would. This is occurring because income is growing faster at the top of the income scale than CBO projected it would, due in no small part to the surge in capital gains. In making its projections for 1999 (based on actual data for 1995 and broad economic trends through 1997), CBO assumed that capital gains income would total $331 billion in 1999. The new IRS data show that capital gains income already had reached $356 billion by 1997. Preliminary IRS data suggest that capital gains income will surpass $420 billion in 1998.

9. Isaac Shapiro and Robert Greenstein, The Widening Income Gulf, Center on Budget and Policy Priorities, September 4, 1999 -- /cms/index.cfm?fa=view&id=705.

10. The difference between the growth rates among the top five percent in the Census and IRS distributions largely reflect the Census Bureau's failure to capture growth among the top one percent. The IRS data show a similar rate of average before-tax income growth for those in the 95th through the 99th percentiles of income (9.1 percent) as the Census Bureau data show for the top five percent (8.3 percent).

11. Preliminary IRS data indicate that capital gains income equaled $420 billion in 1998 (adjusting for inflation and expressed in 1997 dollars), an 18 percent increase from the final 1997 level of $356 billion. In recent years, the preliminary data on capital gains income have understated the final figure by three percent to four percent. Assuming the understatement for 1998 follows the same pattern as the understatement for 1997, the final capital gains income figure for 1998 would be $434 billion, a 22 percent increase from 1997.

12. From 1995 to 1996, capital gains income rose by 44 percent, after adjusting for inflation. From 1996 to 1997, it rose 38 percent. The consistent pattern of significant growth in capital gains throughout the 1995 to 1998 period indicates that the sharp overall income growth between 1995 and 1997 at the top of the income scale was not a reflection of a one-time jump in capital gains due to the capital gains tax cut that began to take effect in 1997.

13. The share of after-income tax income the top one percent of filers received climbed from 12.1 percent in 1995 to 14.7 percent in 1997, an increase of more than one-fifth. This is an enormous increase for a two-year period. The share of income that the bottom 90 percent of tax filers receive declined from 63.2 percent in 1995 to 60.7 percent in 1997.