Survey of Current Business: Federal personal income tax liabilities and payments: preliminary estima

THE estimates of Federal personal income tax liabilities and estimates of Federal personal income tax payments for 1959-2001 now reflect the incorporation of the results of the 2003 comprehensive revision of the national income and product accounts (NIPAs) (table 1). (1) They also reflect the incorporation of data for 1999-2000 from the revised Internal Revenue Service (IRS) estimates and newly available IRS tax return data for 2001 and revised estimates of personal income tax payments for 1993-2000. (2(

For 2001, personal income taxes on a liabilities basis were $142.2 billion (14.2 percent) less than personal income taxes on a payments basis. The relatively large difference reflected the slowdown in the economy in 2001, a reduction in capital gains realizations, which are volatile, and the first year of the tax reductions that were enacted in the Economic Growth and Tax Relief Act of 2001.

The estimates of payments and liabilities and their source data are defined below, and then the differences between liabilities and payments for selected years are examined. A brief discussion of the revisions to the estimates for 1959-2000 is included. In conclusion, a technical note presents the details about the estimates.

Payments and liabilities

In the NIPAs, Federal personal current tax receipts (on a payments basis) are calculated as the sum of withheld income taxes and net declarations and settlements (which consists of gross declarations and settlements less advance refunds), less refunds. (3)

Withheld income taxes are withheld at the source of the income, mainly for wages and salaries, or "income subject to withholding." Declarations are estimated tax payments for income that is usually not subject to withholding, such as self-employment income and capital gains. Settlements are additional taxes that are paid when tax returns are filed and as a result of audits; these payments are treated as payments on income that is not subject to withholding. Advance refunds are authorized by tax legislation that is usually designed to stimulate economic activity or to eliminate budget surpluses. (4) Refunds are recorded as negative payments when the refunds are made.

For certain analyses, payments may not be the most appropriate measure of taxes. For example, households may base their spending decisions, especially for major purchases, on their disposable income that is net of expected tax liabilities rather than of actual tax payments. As a result, taxes on a liabilities basis may be a more appropriate basis for analyzing the effects of taxes on personal spending.

BEA's estimates of Federal personal income taxes on a liabilities basis are derived primarily from estimates of "total income tax" from the Statistics of Income (SOI) Bulletin. (5) SOI estimates of total income taxes are the sum of income taxes after credits, including a portion of the earned income tax credit, and the alternative minimum tax. The SOI estimates are adjusted so that the coverage of the liabilities series is comparable with that of the NIPA payments series. (6) The data for these adjustments are from the SOI Bulletin and related publications and from the Data Book 2003. (7)

The major sources of the differences between liabilities and payments can be accounted for by taxes on income that is subject to withholding and by taxes on income that is not subject to withholding. For details about withholding and nonwithholding, see the "Technical Note."

Differences between liabilities and payments

BEA's estimates of Federal personal income tax liabilities partly differ from the NIPA estimates of Federal personal income tax payments because they are derived from different source data. The estimates of liabilities are derived primarily from the SOI estimates of total income taxes, and the estimates of payments are derived primarily from the Monthly Treasury Statement of Receipts and Outlays of the United States Government. (8)

The most notable differences between liabilities and payments usually reflect changes in tax laws. (9) For 1991-97, the differences between liabilities and payments were relatively small despite numerous changes in the tax law. (10)

For 1998, tax payments exceeded tax liabilities by $25.3 billion, primarily because of the implementation of substantial tax cuts under the Taxpayer Relief Act of 1997 that reduced liabilities. The most important provisions that reduced liabilities were the child tax credit and credits for higher deductions for tuition, both of which became effective on January 1, 1998. In addition, tax refunds were substantial because the withholding tables for tax year 1998 did not fully reflect the reduction in liabilities. For 1999-2000, no major changes in tax law affected tax payments and tax liabilities.

For 2001, tax payments exceeded tax liabilities by $142.4 billion. This substantial difference was partly accounted for by a slowdown in economic activity and by some of the changes in the 2001 tax laws that reduced liabilities. In 2001, gross domestic product slowed to a 0.5-percent increase from a 3.7-percent increase in 2000. The strong growth in taxable income in 2000 produced substantial increases in declaration and settlement payments in 2001, and the deceleration in taxable income reduced liabilities. Capital gains realizations fell 50.3 percent to $324.6 billion in 2001 from $645.4 billion in 2000; this fall also reduced liabilities. (11)

The difference in tax payments and liabilities was also partly accounted for by the Economic Growth and Tax Relief Act of 2001, which will be phased in over the next decade. For 2001, liabilities were reduced by a reduction in tax rates that also resulted in advance refunds, by a 0.5-percentage point reduction in all tax brackets, by the creation of a new 10-percent tax bracket for the lowest income level, and by the expansion of certain credits. (12)

Tax liabilities decreased $136.7 billion in 2001. The decrease reflected decreases of $92.8 billion in total income taxes, of $0.7 billion in assessments, of $6.9 billion in fiduciary income taxes, of $0.4 billion in "other miscellaneous items," and of $35.9 billion in advance refunds for 2001. (13)

Sources of revisions for 1991-2000

The revisions to tax payments for 1993-2000 reflected the incorporation of revised data for 1993-2000 from the Social Security Administration and small revisions to tax payments from taxpayers living overseas (table 2). (14)

The revisions to tax liabilities reflected the incorporation of updated SOI estimates for 1999 and 2000 from the IRS.

Technical Note

Income subject to withholding. Some of the differences between liabilities and payments may be due to the overwithholding that reflects the type of withholding. The IRS defines three types of withholding: Mandatory, optional, and backup.

Mandatory withholding applies to most wages and salaries, military retirement pay, supplemental wages, and certain other incomes, such as gambling winnings. For wages and salaries, liabilities differ from payments for several reasons, but the most important reason is overwithholding, which increases whenever major tax law changes are implemented that result in changes to withholding tables and whenever the Nation's economic growth changes.

The withholding tables that are issued by the IRS and used by employers to calculate the amounts to be withheld from their employees' wages and salaries are based on three simplifying assumptions.

The first assumption is that taxpayers use the standard deduction in calculating their income tax liabilities. However, taxpayers who itemize their deductions may be overwithheld if they underestimate the number of additional withholding allowances that are necessary to offset the excess of their estimated itemized deductions over the standard deduction. The second assumption is that each taxpayer's wages are constant, and the third assumption is that U.S. economic activity is constant throughout the year. If wages vary during the year and are therefore subject to varying withholding rates, overwithholding may result. In addition, overwithholding may result from the use of withholding for "forced savings" or from the failure to estimate growth in itemized deductions.

Overwithholding may also result when withholding for certain payments is based on flat rates rather than on the rates in the withholding tables.

When tax laws change, the changes to withholding tables may differ from the changes to liabilities either by timing or by amounts. Tax law provisions are usually effective on January 1, but the withholding tables are sometimes updated later. For example, the Economic Growth and Tax Relief Reconciliation Act of 2001 reduced all tax rates over 15 percent by a half of a percentage point, and the act applied retroactively to the entire calendar year 2001; however, new withholding tables were not issued until June 2001.

Continued from page 1.

Optional withholding permits the taxpayer to choose the rate of withholding for pensions and annuities, for unemployment compensation, for certain Federal Government payments, such as social security and tier 1 railroad retirement benefits, and for sick pay from sources other than employers.

Backup withholding applies to all types of nonwage income that is subject to information reporting, such as interest income, dividends, rental income, and gains from the sale of assets; it was initiated in 1984 as a compliance measure. For example, backup withholding is required if the recipient fails to furnish an accurate taxpayer identification number to the payer or if the recipient lacks certification that the income is not subject to backup withholding.

Income not subject to withholding. Income that is not subject to withholding includes self-employment income, capital gains, taxable social security benefits, and most interest, dividends, and pensions and annuities. For this income, liabilities differ from payments for two reasons. First, the proportion of the liabilities for the current year that must be paid in estimated taxes in order to avoid a penalty is less than 100 percent.

Second, liabilities differ from payments because of differences in timing. Settlements and the last installment of quarterly estimated taxes are for liabilities that were incurred in a year but that are paid in the next year, and refunds are made in the year after the liabilities were incurred. As a result, net payments of nonwithheld taxes during a year may not reflect the income for the year, and nonwithheld tax payments (declarations and settlements) tend to be less than liabilities.

BEA's estimate of tax declarations in 2001 includes tax payments on declared income from the 2000 tax year that were paid in the 2001 calendar year and tax settlements resulting from IRS assessments or audits for preceding tax years.

Overwithholding on wages and salaries tends to offset most of this shortfall from the liabilities on nonwithheld taxable income, and the net difference between total payments and total liabilities is smaller than the difference that would be expected by an examination of either withheld income taxes or nonwithheld income taxes.

In addition to the timing differences, there are measurement errors that cannot be isolated from the timing-basis differences. These errors include sampling and nonsampling errors with the sample data from the Statistics of Income (SOl) Division of the IRS, reporting and processing errors with the financial statements for the Federal Government and with employment tax return tabulations from the Social Security Administration, estimating errors in the NIPA payments series, and in the coverage adjustments made to the SOI data to derive the liabilities series.

Table 1. Tax Liabilities and Payments, 1959-2001

[Billions of Dollars]

                         Federal personal income taxes

        Liabilities basis (1)    Payments basis (2)    Difference

1959                     39.0                  38.5           0.5
1960                     39.9                  41.8          -1.9
1961                     42.7                  42.7           0.0
1962                     45.4                  46.5          -1.1
1963                     48.8                  49.1          -0.3
1964                     47.8                  46.0           1.8
1965                     50.2                  51.1          -0.9
1966                     56.8                  58.6          -1.8
1967                     63.7                  64.4          -0.7
1968                     77.5                  76.4           1.1
1969                     87.4                  91.7          -4.3
1970                     84.5                  88.9          -4.4
1971                     86.1                  85.8           0.3
1972                     94.3                 102.8          -8.5
1973                    108.9                 109.6          -0.7
1974                    124.4                 126.5          -2.1
1975                    117.3                 120.7          -0.4
1976                    142.8                 141.2           1.6
1977                    161.0                 162.2          -1.2
1978                    189.6                 188.9           0.7
1979                    216.1                 224.6          -0.5
1980                    252.3                 250.0           2.3
1981                    286.7                 290.6          -3.9
1982                    280.2                 295.0         -14.8
1983                    277.8                 286.2          -8.4
1984                    306.7                 301.4           5.3
1985                    331.5                 336.0          -4.5
1986                    374.9                 350.1          24.8
1987                    378.7                 392.5         -13.8
1988                    422.0                 402.9          19.1
1989                    440.1                 451.5         -11.4
1990                    453.4                 470.2         -16.8
1991                    455.4                 461.3          -5.9
1992                    483.1                 475.3           7.8
1993                    508.5                 505.5           3.0
1994                    540.3                 542.7          -2.4
1995                    592.9                 586.0           6.9
1996                    664.5                 663.4           1.1
1997                    742.5                 744.3          -1.8
1998                    800.5                 825.8         -25.3
1999                    892.4                 893.0          -0.6
2000                    994.2                 999.1          -4.9
2001                    857.6                1000.0        -142.4

(1.) The data sources are explained in footnote 13 of the text.

(2.) This series is presented in NIPA table 3.2.

Federal Personal Current Tax Receipts, 1999-2001

[Billions of Dollars]

                                         1999    2000     2001

Federal personal current taxes           893.0   999.1   1,000.0
  Withheld income taxes                  717.8   782.8     807.3
  Net declarations and settlements       300.4   350.3     341.6
    Gross declarations and settlements   300.4   350.3     377.5
    Less: Advance refunds                  0.0     0.0      35.9
  Less: Refunds                          125.2   134.1     148.8

Table 2. Revision to Federal Personal Income Tax
Liabilities and Payments, 1991-2000

[Billions of dollars]

                Liabilities basis

        Previously
        published     Revised    Revision

1991         455.4      455.4         0.0
1992         483.1      483.1         0.0
1993         508.5      508.5         0.0
1994         540.3      540.3         0.0
1995         592.9      592.9         0.0
1996         664.5      664.5         0.0
1997         742.5      742.5         0.0
1998         800.5      800.5         0.0
1999         891.5      892.4         0.9
2000         989.6      994.2         4.6

                   Payments basis

        Previously
        published     Revised    Revision

1991         461.3      461.3         0.0
1992         475.3      475.3         0.0
1993         505.4      505.5         0.1
1994         542.5      542.7         0.2
1995         585.6      586.0         0.4
1996         662.9      663.4         0.5
1997         743.9      744.3         0.4
1998         826.4      825.8        -0.6
1999         894.0      893.0        -1.0
2000         999.0      999.1         0.1

                     Difference

        Previously
        published     Revised    Revision

1991          -5.9       -5.9         0.0
1992           7.8        7.8         0.0
1993           3.1        3.0        -0.1
1994          -2.2       -2.4        -0.2
1995           7.3        6.9        -0.4
1996           1.6        1.1        -0.5
1997          -1.4       -1.8        -0.4
1998         -25.9      -25.3         0.6
1999          -2.5       -0.6         1.9
2000          -9.4       -4.9         4.5

NOTE. The estimates for 1959-90 were not revised.

(1.) The quarterly series for tax liabilities for 1959-2001 is available upon request; e-mail mark.ledbetter@bea.gov.

(2.) For more information, see Eugene P. Seskin and Daniel Larkins, "Improved Estimates of the National Income and Product Accounts for 1929-2002: Results of the Comprehensive Revision," SURVEY OF CURRENT BUSINESS 84 (February 2004): 7-29.

(3.) The estimates of these components are published annually in NIPA table 3.4; quarterly estimates of Federal personal current taxes are published monthly in NIPA table 3.2. These estimates are also available on BEA's Web site at <www.bea.doc.gov>.

(4.) The Economic Growth and Tax Relief Reconciliation Act of 2001 created a new 10-percent tax bracket, which is responsible for the advance refunds of $300 for most single tax filers or married couples filing separately and the refunds of $600 for most married couples filing joint returns. In the NIPAs, advance refunds are treated as a "special factor"; they are considered early settlements for overwithheld income tax payments that result from retroactive tax law changes.

(5.) Internal Revenue Service, "Individual Income Tax Returns, 2001," Statistics of Income Bulletin (Washington, DC: U.S. Government Printing Office, Fall 2003).

(6.) The SOI estimates are raised by adding the following: Recapture taxes from new computations of investment levels, low-income housing, and other tax credits for the preceding year; assessment from audits, net of refunds on amended returns (Form 1040X); and fiduciary income taxes. The SOI estimates are reduced by the exclusion of the portion of the earned income credit that is used to offset social security and penalty taxes and by the exclusion of income taxes paid by U.S. citizens living abroad for a year or more. In the NIPAs, the income taxes paid by these citizens are also excluded from estimates of personal current tax receipts.

(7.) Internal Revenue Service, Data Book 2003, publication 55B (Washington, DC: U.S. Government Printing Office, 2004).

(8.) The NIPA estimates of personal income taxes are derived primarily from Monthly Treasury Statements (MTS). First, using the combined MTS data for income taxes and social security taxes, BEA makes a timing adjustment to convert withheld taxes to an accrual basis. Second, the social security tax portion is subtracted on the basis of a tabulation of employment tax returns from the Social Security Administration. Other adjustments, such as the reclassification of interest charges on late taxes and of taxes paid by residents of foreign countries, are also made.

(9.) For a discussion of the tax law changes and its effects on liabilities and payments for 1972-1990, see Thae S. Park, "Federal Personal Income Tax Liabilities and Payments, 1959-1997" SURVEY 80 (March 2000): 13-15.

(10.) The following major tax law changes affected liabilities in 1991-97: The decrease in the maximum tax rate on net long-term capital gains from 28 percent for 1991 to 20 percent for 1997; the increase in the maximum tax rate on other types of income from 31 percent for 1991 to 39.6 percent for 1997; the increase in the maximum earned income tax credit from $1,192 for one qualifying child for 1991 to $2,210 for 1997; the phaseout of certain itemized deductions, beginning with 1991; and the taxation of up to 85 percent of social security and equivalent tier 1 railroad retirement benefits, beginning with 1994.

(11.) For the data for capital gains, see Mark Ledbetter, "Comparison of BEA Estimates of Personal Income and IRS Estimates of Adjusted Gross Income," SURVEY 84 (April 2004): 10-11; tables 2 and 3, line 12.

(12.) Tax liabilities for 2001 were affected by increases to personal exemptions, to standard deductions, to the alternative minimum tax exemption, and to the child tax credit and by decreases in the capital gains tax rate, in the luxury excise tax rate, and in the tax on investment income of minors under 14.

(13.) For the estimates of total income tax, see figure A in "Individual Income Tax Returns, 2001," 9; for estimates of the assessments, see Data Book 2003; the estimate for fiduciary income taxes was supplied to BEA from the IRS; and "other miscellaneous items" is a BEA estimate.

(14.) These revisions were incorporated into the NIPAs as part of the 2003 comprehensive revision.