Federal Tax Burden, 1979--2001: CBO

Introduction

This page attempts to answer two questions:

• How high is the federal tax burden?
• How does the burden vary with income?

The size of the tax burden, and how it varies with household income, is portrayed in Graph 1. The distribution of pretax household income itself is portrayed in Graph 2 and Graph 3.

The data and discussion are based on three publications from the Congressional Budget Office (see references [1], [2], and [3]). Because any distributional analysis of a tax burden must make certain assumptions, the figures arrived at by the CBO most likely differ from other such tables, though it is likely the overall patterns and trends would be similar.

It should be emphasized that these data are for federal taxes only. Because total state and local taxes are generally far less progressive than total federal taxes, the overall tax burden in the United States is less progressive than that illustrated here.

What Is Being Measured?

The CBO analysis calculates effective tax rates faced by American taxpayers. An effective rate is the total tax paid divided by the taxpayer's income. This raises three questions:

• What taxes are being considered?
• What income is being considered?
• Should taxes and income be considered by individual, family, or household?

Taxes

Taxes considered

The CBO analysis considers these taxes:

• The individual income tax;
• the corporate income tax;
• payroll taxes; and,
• excise taxes.

Behind all business taxes are people bearing those taxes---be they shareholders, bondholders, proprietors, employees, or consumers. Hence taxes on firms (businesses) are distributed to people (actually households; see below).

Tax incidence assumptions

The incidence of a particular tax refers to the ultimate bearer of the burden of the tax. This may not be identical to the person or business who actually remits the tax to the government because of the phenomenon of tax shifting. Tax shifting, while an important object of study, is often hard to verify and account for.

Consider, for example, payroll taxes, some of which are remitted by the employee (meaning the taxes are taken out of the paycheck), and some by the employer. Almost all economists agree that the employee share of payroll taxes is born by the employee. However, economists also almost uniformly agree that the employer share of payroll taxes falls not on the employing business, but rather on the employee: the employer shifts the burden to the employee by lowering her paycheck a corresponding amount. (This shifting may not be deliberate; rather, it reflects supply and demand in the labor market.)

The assignment of tax incidence can be a controversial subject.

CBO makes the following assumptions on tax incidence:

• Households bear burden of individual income and payroll taxes, which they pay directly.
• The employer share of payroll taxes falls entirely on the employee.
• Corporate income taxes fall on owners of capital in proportion to income from interest, dividends, rents, capital gains.^<2>
• Excise taxes fall on households.^<3>

It is important to note that if part of the burden of a tax is shifted to a party, then the income of that party is taken to have increased by the same amount. (This is why business taxes are considered to be ultimately part of household income; see below.)

Income

The following outlines the components of income included in the CBO's analysis:

• Cash income, taxable and tax exempt, including wages, salaries, self-employment income, rents, taxable and nontaxable interest, dividends, realized capital gains, cash transfer payments, and retirement benefits
• Business taxes, including corporate income taxes, the employer's share of Social Security, Medicare, and federal unemployment insurance payroll taxes (imputed to households, as per the assumptions on tax incidence above)
• Employees contributions to 401(k) retirement plans
• All in-kind benefits (Medicare, Medicaid, employer-paid health insurance premiums, food stamps, school lunches and breakfasts, housing assistance, and energy assistance)

Note that CBO:

• uses the Census Bureau's fungible value measure for government in-kind transfers;
• does not adjust capital gains for inflation, and does not include unrealized capital gains or imputed rents on owner-occupied housing (see [1], pp. 23--24); and,
• double counts retirement income (see [1], p. 21).

Households and taxpayer rankings

Because some people (such as underage children) are financially dependent on others, and because people (typically families) consume and save together, individuals are often not the unit over which income and taxes are measured in tax analysis. CBO uses the households---people sharing a single housing unit, regardless of the relationships between them---not the family (see [1], pp. 19--20). The effective tax rate faced by a particular household in a particular year is the ratio of taxes paid by the household to income earned by the household (where taxes and income are imputed as above).

To describe how effective tax rates vary with income, households are ranked by income. Because larger households have greater needs than smaller households (all other things being equal), household incomes are first adjusted by dividing by the square root of the size of the household ([1], p. 24). Note that

• this adjustment is used only to rank the households; it plays no role in computing the effective tax rate; and,
• using the square root allows for the fact that there are typically economies of scale in consumption, particularly housing.^<4>

Once households are ranked by income (adjusted for household size), they are split into five quintiles. The adjusted incomes of households in the second quintile are higher than those in the first; those of the third are higher than those in the second; and so on. It is important to note that the quintiles are chosen so as to have equal numbers of people, and hence do not have equal numbers of households.

Households with negative income are not included in the first quintile. (They are, however, included in any totals appearing in the CBO analysis.)

Finally, the CBO tables often include numbers broken out for the top 10, 5 and 1 percentiles, in addition to the five quintiles.

Some caveats

The following caveats must be considered. For further discussion of these and other concerns, see references [2] and [1].

• The effective tax rates amount to average, not marginal, rates.
• The distortionary effect of taxes on the economy are not measured.
• Government benefits, whether or not directly tied to taxes, are not considered.
• The data are not longitudinal. Thus, a given quintile does not contain the same households from year to year.
• Taxes and income are considered on an annual basis only. Hence income cannot be smoothed over many years, and lifetime models of income and consumption cannot be entertained.
• Effective tax rates and their distribution among different income classes change for reasons other than changes in tax law. These include
  • Changes in the national composition of income (e.g., wage and salary income versus corporate income)
  • Changes in the distribution of income between different income classes
  • Demographic shifts
• Careful consideration to business cycle effects must be given when considering trends in incomes and taxation over time

Federal tax burden

As discussed above, for each year, CBO computes the income and taxes for each household; ranks the households; and divides the households into income classes. The effective tax rate for each income class in a given year is defined to be the ratio of the total taxes imputed to households in the class to the total income. (The ratio is then expressed as a percentage.) The resulting trends in effective tax rates are plotted below.

Graph 1: Federal tax burden, 1979--2001

Data source

The data for Graph 1 were taken from the portion of [5] labeled "Total Effective Federal Tax Rate."

Pretax Income

The next two graphs show how much income each household grouping received from 1979--2001.

Graph 2: Average pretax income, 1979--2001

This graph portrays the trend in pretax income in constant 2001 dollars for the various household income groups. As outlined in the discussion above, the definition of income used is broader than just cash income, so it may be difficult for an actual household to compare its income to the numbers portrayed here.

Note that income is portrayed on a logarithmic scale.

Data source

The data for Graph 2 were taken from the portion of [6] labeled "Average Income (2001 dollars)...Pretax Income."

Graph 3: Shares of average pretax income, 1979--2001

This graph portrays the fraction of national income (as imputed to households in the procedure outlined above) for each household grouping.

Note that households belonging to the top 1% also belong to the top 5%; households in the top 5% belong to the top 10%; and the top 10% belongs to the highest quintile.

Data source

The data for Graph 3 were taken from the portion of [6] labeled "Share of Income (Percent)...Pretax Income."

Footnotes

References

[1]	Congressional Budget Office, "Effective Federal Tax Rates, 1979--1997" (October 2001)
[2]	Congressional Budget Office, "Effective Federal Tax Rates: 1979--2001" (April 2004)
[3]	Congressional Budget Office, Effective Federal Tax Rates, 1997 to 2000 (August 2003)
[4]	Joel Slemrod and Jon Bakija, Taxing Ourselves: A Citizen's Guide to the Great Debate over Tax Reform (2000), Cambridge
[5]	"Table 1A: effective federal tax rates for all households, by comprehensive household income quintile, 1979--2001," in Congressional Budget Office, "Effective Federal Tax Rates: 1979--2001" (April 2004)
[6]	"Table 1C: number of households, average income and income shares, and income category minimums for all households, by household income category, 1979--2001," in Congressional Budget Office, "Effective Federal Tax Rates: 1979--2001" (April 2004)