# What Is the Average Federal Individual Income Tax Rate on the Wealthiest Americans?

By Greg Leiserson, Senior Economist (CEA); and Danny Yagan, Chief Economist (OMB)

Abstract: We estimate the average Federal individual income tax rate paid by America’s 400 wealthiest families, using a relatively comprehensive measure of their income that includes income from unsold stock. We do so using publicly available statistics from the IRS Statistics of Income Division, the Survey of Consumer Finances, and Forbes magazine. In our primary analysis, we estimate an average Federal individual income tax rate of 8.2 percent for the period 2010-2018. We also present sensitivity analyses that yield estimates in the 6-12 percent range. The President’s proposals mitigate two key contributors to the low estimated rate: preferential tax rates on capital gains and dividend income, and wealthy families’ ability to avoid paying income tax on capital gains through a provision known as stepped-up basis.

When an American earns a dollar of wages, that dollar is taxed immediately at ordinary income tax rates.[1] But when they gain a dollar because their stocks increase in value, that dollar is taxed at a low preferred rate, or never at all.[2] Investment gains are a primary source of income for the wealthy, making this preferential treatment of investment gains a valuable benefit for the wealthiest Americans. Yet the most common estimates of tax rates do not fully capture the value of this tax benefit because they use an incomplete measure of income. This analysis asks: what was the average Federal individual income tax rate paid by the 400 wealthiest American families’ in recent years, determined using a more comprehensive measure of income?

*How the wealthy enjoy low income tax: preferred rates on an incomplete measure of income*

The wealthy pay low income tax rates, year after year, for two primary reasons. First, much of their income is taxed at preferred rates. In particular, income from dividends and from stock sales is taxed at a maximum of 20 percent (23.8 percent including the net investment income tax), which is much lower than the maximum 37 percent (40.8 percent) ordinary rate that applies to other income.

Second, the wealthy can choose when their capital gains income appears on their income tax returns and even prevent it from ever appearing. If a wealthy investor never sells stock that has increased in value, those investment gains are wiped out for income tax purposes when those assets are passed on to their heirs under a provision known as stepped-up basis.

*Analyzing a more comprehensive measure of income*

Preferred tax rates on income from stock sales (“realized capital gains income”) and from dividends feature prominently in commonly cited tax rates as well as in our analysis.

An important feature of our analysis that is less common in existing estimates of tax rates is that we include untaxed (“unrealized”) capital gains income in our more comprehensive income measure as they accrue.[3]

Measuring income in this more comprehensive manner matters relatively little for estimating most families’ tax rates, as most families have few investment assets.[4] However, it matters greatly for the wealthiest families for whom such unrealized and thus untaxed gains are a large share of their income. Like all other forms of income, unrealized capital gains income can be tapped to finance consumption and can improve financial wellbeing.

A common reference point for defining income in economics is known as Haig-Simons income.[5] Pre-tax Haig-Simons income equals families’ change in wealth, plus taxes and consumption. We define our income measure as families’ change in wealth plus easily estimable taxes. Our definition of income is more limited than the Haig-Simons definition because it excludes consumption and other taxes, but it is a simpler way to include a substantial share of capital gains in income and can be implemented with publicly available data.[6]

*Primary estimate and sensitivity*

In our primary analysis, we estimate that the 400 wealthiest families paid an average Federal individual income tax rate of 8.2 percent on $1.8 trillion of income over the period 2010–2018, the years from the last decade for which the necessary data are available. Two factors that contribute to this low estimated tax rate include low tax rates on the capital gains and dividends that are taxed, and wealthy families’ ability to permanently avoid paying tax on investment gains that are excluded from taxable income. The 2022 President’s Budget proposes to raise the capital gains and dividend tax rate and to virtually end stepped-up basis for the highest-income Americans, thereby ensuring these investment gains are subject to income tax.

The true tax rate of interest may differ from our primary estimate. First, excluding consumption and some taxes from our measure of income suggests that the 8.2 percent estimate is actually higher than the tax rate measured relative to a truly comprehensive measure of income. Second, the data and methods on which we rely are imperfect. Different estimation choices illustrate the estimate’s sensitivity. For example, varying the analytic assumptions underlying one of our key methodological choices (discussed in greater detail in the technical appendix below) causes the estimate to vary from 6 to 12 percent.

The tax rate we estimate is, of course, sensitive to changes in policy. The top capital gains tax rate was 15 percent between May 2003 and 2012 and has been 20 percent (23.8 percent including the net investment income tax) since 2013. In addition, the Federal individual income tax is only one tax. Focusing on the individual income tax sheds light on the structural limitations of that tax and the scope for reforms, such as curtailing the ability of the wealthy to avoid paying tax on their investment gains through stepped-up basis. However, alternative tax rates could also be estimated that account for other taxes, such as the payroll tax, estate and gift tax, corporate income tax, and taxes paid to foreign governments. Moreover, one could use alternative definitions of income or adopt various approaches to the treatment of certain subsidized activities such as charitable giving. Finally, we focus on an extended time period (2010–2018), which helps to ensure that our analysis reflects the long-run reality of positive asset incomes despite short-run fluctuations. However, we also present alternative estimates using different start and end years, as well as an estimate that replaces the Forbes 400 wealth in 2009 with the average for the period 2008–2010.

Our primary estimate of 8.2 percent is much lower than commonly cited estimates of top Federal individual income tax rates. For example, the Joint Committee on Taxation (2021) estimates that the 2021 Federal individual income tax rate on the top 0.4 percent of families ranked by income (i.e., the 715,000 families with income over $1 million) will be 26 percent. Our analysis differs by (a) analyzing a smaller group of families (the top 0.0002 percent) ranked by wealth, and (b) including unrealized capital gains income in the income measure. See the end of the technical appendix for additional discussion of how our analysis compares to commonly cited estimates.

*Method*

We emphasize that any estimate of tax rates on the wealthiest is uncertain and open to refinement, due to current data limitations.

We detail our method in the technical appendix below. In a nutshell, our method is as follows: we take the IRS Statistics of Income (SOI) Division’s Federal individual income taxes paid by the top-400-by-*income* families and multiply it by an adjustment factor constructed using the Survey of Consumer Finances (0.63) to convert it to an estimate of taxes paid by the top-400-by-*wealth*. We then estimate a more comprehensive measure of income as the change in Forbes 400 wealth, plus our estimates of the top-400-by-wealth’s Federal individual income taxes paid and State-and-local tax deductions (estimated similarly based on Statistics of Income data). The ratio of the two yields our estimated tax rate.

*Conclusion*

In this analysis, we used publicly available data to estimate the average Federal individual income tax rate paid by America’s wealthiest 400 families, using a relatively comprehensive measure of their income that includes income from unsold stock. In our primary analysis, we estimated an average tax rate of 8.2 percent for the period 2010–2018. We also present sensitivity analyses that yield estimates in the 6–12 percent range.

Preferential capital gains rates and stepped-up basis—a provision of tax law that allows wealthy taxpayers to wipe out unrealized capital gains for income tax purposes when they pass assets to their heirs—contribute to this low tax rate. The 2022 President’s Budget would increase capital gains rates and virtually end stepped-up basis for the highest-income Americans, thereby ensuring their investment gains are subject to income tax.

__Technical Appendix__

This technical appendix documents how we combine publicly available data from the IRS Statistics of Income (SOI) Division, the Federal Reserve Board’s Survey of Consumer Finances (SCF), and Forbes magazine’s estimates of the wealthiest 400 Americans to estimate the Federal individual income tax rate paid by the wealthiest 400 families. For reference, Forbes estimates that the 400 wealthiest Americans in 2018 had wealth ranging from $2.1 billion to $160 billion. Our tax rate estimate is dollar-weighted: it is an estimate of total Federal individual income taxes paid by the wealthiest families, divided by an estimate of those families’ income. We focus on the period 2010–2018 and report estimates for alternative periods as well.

We first describe the basic idea of the estimation procedure and then go through the details. We divide an estimate of the Federal individual income taxes paid by the 400 wealthiest families by a relatively comprehensive estimate of their income. For the numerator, we start by estimating the taxes paid by the families with the highest reported income on tax returns. Then we estimate how the income of the highest-*wealth* families compares to the income of the highest-*reported-income* families and use that as an adjustment factor to estimate the taxes paid by the highest-wealth families. For the denominator, we use changes in the reported wealth of the Forbes 400 to estimate the income of the 400 wealthiest families.

*Numerator: estimated Federal individual income taxes paid by the wealthiest 400 families*

The numerator of our tax rate estimate equals estimated 2010–2018 taxes paid by the wealthiest 400 families. We construct the numerator by estimating the 2010–2018 taxes paid by the highest-reported-income families, then multiply by an adjustment factor based on the Survey of Consumer Finances (SCF) to account for the fact that highest-reported-income families are not the same as the highest-wealth families.

SOI published estimates of the taxes paid by the 400 highest-reported-income families annually from 1992 through 2014. In a first and straightforward step, we extend this series through 2018. To do this, we rely on estimates of the total Federal individual income tax paid by the top 0.001 percent, available from SOI annually from 2001 through 2018. For the years 2001 through 2014, when both estimates are available, the ratio of taxes paid by the top 400 to taxes paid by the top 0.001 percent varies in only a small window around 0.59. We therefore estimate the taxes paid by the 400 highest-income families for the period 2015 through 2018 by assuming that it is 0.59 times the taxes paid by the top 0.001 percent for this period. Our SOI-based estimate of 2010–2018 taxes paid by the 400 highest-reported-income families equals actual SOI top-400 taxes for years 2010-2014, plus our estimates for years 2015–2018.

Our SOI-based estimate of 2010–2018 taxes paid by the 400 highest-reported-income families surely exceeds the 2010–2018 taxes paid by the 400 highest-wealth families: some of the wealthiest families have lower reported income and pay less tax. For example, Warren Buffett was a member of the 2015 Forbes 400, but his voluntarily-released 2015 tax return information indicates 2015 adjusted gross income of $11.6 million (Cohen 2016). The thresholds for top percentile groups in 2015 in the SOI estimates show that $11.9 million was required to be in the top 0.01 percent (about 14,000 families). Thus, Buffett was not even in the top 14,000 tax units ranked by reported income, let alone the top 400 ranked by reported income. Moreover, he paid $1.8 million in Federal individual income tax in 2015, far less than the $36 million average for the top 0.001 percent or the $9 million average for the top 0.01 percent. As a result, the 2015 taxes paid by the top 400 families ranked by reported income would overstate the 2015 taxes paid by the top 400 families ranked by wealth.

Hence, we must convert our SOI-based estimate of taxes paid by the highest-reported-income families into an estimate of taxes paid by the highest-wealth families. We do so by multiplying the SOI-based estimate by an adjustment factor of 0.63, constructed as follows from the Survey of Consumer Finances which contains information both on approximate reported income and on wealth.

A formula helps to clarify the method. Our goal for the numerator of the tax rate is to estimate taxes paid by the families with wealth rank 1 through 400, which we write as: TAX^{W}_{1,400}. The SOI data give us an estimate of the taxes paid by the families with reported-income rank 1 through 400: TAX^{I}_{1,400}. Ideally, we would multiply TAX^{I}_{1,400} by an adjustment factor (in blue) equal to the ratio of the tax paid by the 400 highest-wealth families to the tax paid by the 400 highest-reported-income families:

However, the ideal adjustment factor cannot be directly measured in publicly available data. The best available data source—the SCF—lacks information on taxes paid and excludes the Forbes 400 wealthiest from the survey sample by construction. We make two assumptions that allow one to approximate the ideal adjustment factor using *reported incomes* among families *ranked 401 through 1400* (i.e., the “next-1,000” groups ranked either by reported income or by wealth), which is approximately the rest of the top 0.001 percent. First, we assume that the highest-reported-income and highest-wealth groups pay the same average tax rate.[7] Second, we assume that the ratio of the reported incomes for the next-1,000 groups is the same as the ratio of reported incomes for the top-400 groups. Under those assumptions, one can replace the ideal adjustment factor with the following alternative adjustment factor that uses only next-1000 information:

Estimating the reported income of the next 1,000 by wealth *I ^{W}*

_{401,1400}is relatively straightforward: the SCF excludes the top 400 by wealth, so we simply use the reported income of the wealthiest families in the SCF.[8]

Estimating the income of the next 1000 by income *I ^{I}*

_{401,1400}is more challenging, as it depends on how much overlap there is between the Forbes 400 and the top 400 by reported income. If there is full overlap, then none of the top 400 by reported income should be in the SCF. We could then estimate

*I*

^{I}_{401,1400}using the SCF observations with the highest reported incomes; doing so would yield an adjustment factor of 0.44, similar to Saez and Zucman (2019).[9] At the other extreme, if none of the Forbes 400 is in the top 1400 by income, then the appropriate SCF observations to use would be those with reported income ranks 401 through 1400. Doing so would exclude many high-reported-income families from the calculation and thereby yield a higher adjustment factor of 0.66.[10] A higher adjustment factor leads to a higher resulting tax rate estimate.

We lean toward the conservative side of the spectrum: we assume an overlap of 100 and estimate *I ^{I}*

_{401,1400}using the reported income of the SCF observations that represent families ranked 301 through 1300. Doing so, we obtain an adjustment factor of 0.63. Thus, our estimate of taxes paid by the wealthiest 400 families equals the SOI-based taxes paid by the 400 highest-reported-income families multiplied by our 0.63 adjustment factor.

*Denominator: estimated income*

We divide our estimate of taxes paid by the wealthiest 400 by our more comprehensive estimate of the wealthiest 400’s income: their estimated change in wealth, plus easily estimable taxes. This income measure excludes consumption and other taxes, which would cause us to understate Haig-Simons income and therefore overstate the Federal individual income tax rate on Haig-Simons income.

To estimate comprehensive income for the 2010–2018 period, we begin by subtracting the total wealth (net worth) of the Forbes 400 in 2009 from the total wealth of the Forbes 400 in 2018.[11] We then add two additional components of Haig-Simons income that are available in tax data: Federal individual income taxes paid (estimated above) and State and local individual tax deductions (estimated from the same SOI data using the same 0.63 adjustment factor).[12]

*Estimate and sensitivity*

Using our estimated numerator and denominator, our primary estimate for the 2010–2018 Federal individual income tax rate for the wealthiest 400 is 8.2 percent. For the numerator, we estimate that the wealthiest 400 families paid $149 billion in Federal individual income taxes, equal to the $237 billion paid by the highest-income families in the SOI data multiplied by 0.63. For the denominator, Forbes estimates suggest that the wealthiest 400 experienced a change in wealth for the period 2010–2018 of $1.62 trillion. Adding the $149 billion of estimated Federal individual income taxes and an analogously estimated $46 billion in State and local taxes, we estimate the wealthiest 400’s income for the period 2010–2018 to be $1.82 trillion. Dividing $149 billion by $1.82 trillion, we obtain 8.2 percent.

Appendix Table 1 presents a sensitivity analysis for different periods. Column 1 repeats our main analysis for time periods that begin in years other than 2010 (all ending in 2018). Our analysis for 2018 alone yields 8.5 percent, for the most recent five years yields 9.8 percent, and for the most recent 20 years yields 10.2 percent. Column 2 repeats the exercise for time periods that end in 2014, which is the last year that does not rely on extrapolated top-400 Federal individual income tax data (though it also includes fewer years with the higher post-2012 capital gains tax rates). Our analysis for years 2010–2014 yields 6.2 percent and for 2014 alone yields 6.3 percent. In addition, replacing 2009 Forbes 400 wealth with its 2008–2010 average yields an estimate of 8.6 percent.

Our 0.63 adjustment factor is estimated with error. One of the most substantial risks to the accuracy of the estimate is if the income of the top 400 by wealth relative to the top 400 by income differs consistently from the corresponding ratio for the next 1000. Other uncertainties include sampling or non-sampling error in the SCF data on which we rely. Under the extreme assumption that the highest income families are the highest wealth families, the adjustment factor would equal 1, and the average tax rate for 2010-2018 would be 12.3 percent. If, on the other hand, the highest wealth families have only 43 percent of the income of the highest-income families following Saez and Zucman’s (2019) analysis, the average tax rate would be 5.8 percent. If the highest wealth families systematically differ from the next 1000 for whom information is available in the SCF, the ratio could, in principle, be even lower.

We define our more comprehensive measure of income such that it is systematically lower than (pre-tax) Haig-Simons income, which includes all taxes and consumption. The SOI data contain information on one additional category of expenditure that could be included: deductible contributions to nonprofit organizations. When including these deductible contributions (estimated in the same way as we estimate State and local taxes) in comprehensive income, we obtain an estimate of 7.9 percent.

Forbes 400 wealth is surely measured with error. An active literature studies and assesses wealth measurement at the very top of the wealth distribution (e.g., Kennickell 2009; Johnson, Raub, and Newcomb 2013; Piketty 2014; Kopczuk 2015). Saez and Zucman (2016) use capitalized income tax returns to, over some periods, estimate faster growth in top wealth than does the SCF while mostly taking Forbes as given. In ongoing work, Smith, Zidar, and Zwick (2020) do not publish top 400 estimates, but generally estimate slower growth in top wealth, which could be consistent with Forbes being misled, unable to value nontraded assets, or unable to observe gifts or debt. Higher growth in top wealth would lead to lower tax rates while lower growth in top wealth would lead to higher tax rates. For example, if the Forbes 400 overstates top wealth growth by one-third, our estimate would be 11.7 percent.

Combining data across three cross-sectional data sources yields some inconsistency in time periods studied. Our target population is the wealthiest in each year of the period examined based on end-of-year wealth. The Forbes 400 data are released each fall. The 0.63 adjustment factor is based on families ranked by income in year *t*, compared to families ranked by wealth when surveyed at some point in year *t+1* (though the reported wealth may or may not be current as of the time they were surveyed). Since the wealthiest families change over time, subtracting Forbes 400 totals across years understates the income of the wealthiest at the end of each year, which leads to overestimated tax rates.[13]

Our tax rate estimates are substantially lower than commonly cited top Federal individual income tax rates produced by the Congressional Budget Office, Joint Committee on Taxation, the Department of the Treasury, and the Tax Policy Center.[14] These estimates differ from ours in three key respects. First, and most fundamentally, the Congressional Budget Office, Joint Committee on Taxation, the Treasury, and the Tax Policy Center estimate tax rates relative to income measures that largely exclude unrealized capital gains. The analyses thus find substantially higher tax rates than we do because they, to varying degrees, exclude the untaxed income that motivates this analysis in favor of more accurately estimated cash income flows. Second, tax-preferred realized capital gains are a larger share of income for the top 400 than they are for the larger top groups for which these other estimates are produced. Third, we examine income tax rates by wealth rather than by income, and unrealized capital gains may be even more concentrated among high-wealth families than high-income families.

We conclude this technical appendix by emphasizing the fundamental uncertainty in our estimates. We hope that our analysis stimulates further estimation and direct measurement of income tax rates inclusive of unrealized capital gains income and by wealth group.

__References__

Bricker, J., P. Hansen, and A.H. Volz. 2019. “Wealth concentration in the U.S. after augmenting the upper tail of the survey of consumer finances.” *Economics Letters *184. (Link)

Cohen, Patricia. 2016. “Buffett Calls Trump’s Bluff and Releases His Tax Data.” *New York Times*, October 10. (Link)

Congressional Budget Office. 2021. “The Distribution of Household Income, 2018.” (Link)

Johnson, B., B. Raub, and J. Newcomb. “A New Look at the Income-Wealth Connection for America’s Wealthiest Decedents.” SOI Working Paper. Washington: Internal Revenue Service Statistics of Income. (Link)

Joint Committee on Taxation. 2012. “Overview Of The Definition Of Income Used By The Staff Of The Joint Committee on Taxation In Distribution Analyses.” JCX-15-12. (Link)

Joint Committee on Taxation. 2021. “Overview Of The Federal Tax System As In Effect For 2021.” JCX-18-21. (Link)

Kennickell, A.B. 2009. “Ponds and Streams: Wealth and Income in the U.S., 1989 to 2007.” Finance and Economics Discussion Series 2009-13. Washington: Board of Governors of the Federal Reserve System. (Link)

Kopczuk, W. 2015. “What Do We Know about the Evolution of Top Wealth Shares in the United States?” *Journal of Economic Perspectives *29, no. 1: 47-66. (Link)

Piketty, T. 2014. *Capital in the Twenty-First Century. *Cambridge: Harvard University Press.

Saez. E. and G. Zucman. 2016. “Wealth Inequality in the United States since 1913: Evidence from Capitalized Income Tax Data.” *The Quarterly Journal of Economics *131, no. 2: 519-578. (Link)

Saez, E. and G. Zucman. 2019. “Progressive Wealth Taxation.” Brookings Papers on Economic Activity Conference Draft. (Link)

Smith, M., O. Zidar, and E. Zwick. 2020. “Top Wealth in America: New Estimates and Implications for Taxing the Rich.” Working paper, Princeton Economics. (Link)

Tax Policy Center. 2021. “T21-0134 – Average Effective Federal Tax Rates – All Tax Unites, By Expanded Cash Income Income Percentile, 2021.” (Link)

U.S. Department of the Treasury Office of Tax Analysis. 2020. “Distribution Table: 2021 001; Distribution of Families, Cash Income, and Federal Taxes under 2021 Current Law.” (Link)

[1] A wage earner may defer taxation—subject to statutory limits—by contributing to a retirement savings account. Other generally applicable tax benefits may also reduce a worker’s tax rate.

[2] The minority of capital gains, realized within one year of acquiring the underlying asset, is taxed at ordinary rates.

[3] Unrealized capital gains are the increase in the value of assets even before the assets are sold. A wealthy individual who purchases corporate stock worth $100 million that subsequently increases in value to $200 million over the next ten years has accrued $100 million of unrealized capital gains income over that period. These unrealized capital gains are a major source of income for the wealthiest Americans.

[4] For example, the Federal Reserve’s Distributional Financial Accounts estimate that, as of the first quarter of 2021, the top 1 percent of families by wealth held 54 percent of the value of corporate stocks and mutual funds, compared to 11 percent for 50^{th}–90^{th} percentiles, and less than 1 percent for the bottom 50 percent.

[5] For example, the Joint Committee on Taxation (2012) states, “Economists generally agree that, in theory, a Haig-Simons measure of income is the best measure of economic well-being.”

[6] This approach would be less informative for middle-class families because they consume a much larger share of their income.

[7] We lack direct evidence on the average tax rate paid by the highest-wealth families. In principle, the average tax rate could differ in either direction. The highest-income families could pay a lower average tax rate because they are high-income due to large single-year capital gains realizations that are taxed at low rates. Alternatively, the highest-wealth families could pay a lower share of their tax-return income in taxes due to large charitable deductions.

[8] The Survey of Consumer Finances intentionally excludes from its sample anybody included in the Forbes 400 due to privacy concerns. However, some Forbes 400 wealth may be represented by families included in the Survey of Consumer Finances sample, and some additional observations are also excluded from the SCF sample. Bricker, Hansen, and Volz (2019) propose a method for augmenting the SCF with the Forbes 400 data without double counting. We simplify by assuming that there is a sharp cutoff between the two and do not rely on the Survey of Consumer Finances to compute any aggregates.

[9] Note, however, that the assumption of full overlap would imply that the desired ratio for the top 400 would be one. In effect, under this set of assumptions, the value for the next 1000 would be a poor guide to the value for the top 400. In implementing the procedures described in this section, we use the 2001-2019 SCFs and average the resulting annual adjustment factors across years in order to increase our effective sample size.

[10] Technically, because the SCF is a survey with sample weights, we mean the observations that, when weighted, represent these ranks. We split SCF observations that cross relevant rank boundaries, allocating a proportionate share of the observation’s weight to each side of the boundary.

[11] The date for which the Forbes 400 estimates wealth has varied over time. In 2020, Forbes used market prices near the end of July. In 2019, Forbes used market prices in September. For simplicity, we treat the Forbes 400 as end-of-year wealth estimates. The Forbes list is a mix of person-level information and immediate-family information. By treating it as family-level (more precisely, tax-unit-level) information, our estimate of income could potentially be somewhat conservative, though we anticipate this effect is small.

[12] We impute the State and local tax deductions for the top 400 for the period 2015–2018 as the top 400’s share of the top 0.001 percent’s State and local tax deductions in 2014 multiplied by the total deductions of the top 0.001 percent for 2015–2018. The 2017 tax law limited the State and local tax deduction. For 2018 only, we impute total State and local tax deductions of the top 0.001% tax units by multiplying 2018 top 0.001% Federal taxes by the ratio of total 2014-2017 State and local tax deductions to total 2014-2017 top 0.001% Federal taxes.

[13] A further timing issue that could arise is if the wealthy systematically realize capital gains only when they are not in the top 400 by wealth. In this case, the tax rate of the top 400 by wealth in each year could understate a life-cycle estimate of the tax rate of the extremely wealthy. However, it is not clear that—if this is a concern—our estimate of the tax rate of the extremely wealthy is affected by it. If it were the case that the extremely wealthy systematically do not realize their income when they are in the top 400 by wealth, our adjustment factor may be an overestimate of their taxable income.

[14] The Congressional Budget Office (2021) recently estimated that the average Federal individual income tax rate of the highest-income 1 percent of households was about 24 percent for the period 2014 through 2018. Other analysts focus primarily on forward-looking estimates. The Treasury Department (2020) estimated that the average Federal individual income tax rate of the highest-income 0.1 percent of families in 2021 would be 23 percent, and the Tax Policy Center (2021) estimated that it would be 25 percent. The Joint Committee on Taxation (2021) estimated that the tax rate for families with incomes of at least $1 million would be 26 percent. The 2017 Tax Act reduced individual income tax rates in 2018. However, this effect is small relative to the difference between the average Federal individual income tax rate on the wealthiest that we estimate and the estimates cited here.