Why Treasury's Numbers Don't Add Up

Brief Analyses | Taxes

No. 234
Friday, June 27, 1997
by Bruce R. Bartlett

One of the most important factors in evaluating tax legislation is the distributional impact of the tax changes. The Treasury Department produces tables showing the effects of tax cuts and tax increases on people with different incomes. However, Treasury's distribution tables bear no relationship to reality, and fail to convey to policymakers any sense of how people are actually affected by proposed tax changes. They make some people appear to be much wealthier than they actually are and others poorer. Most taxpayers who look at these tables will derive a very distorted picture of how the proposed tax changes will affect them.

In recent days, Treasury has alleged that the benefits of the tax legislation approved by the House Ways and Means Committee and the Senate Finance Committee are skewed heavily toward the rich. According to the Treasury analysis, 67.9 percent of the Ways and Means bill and 65.5 percent of the Finance Committee bill benefits will go to the richest 20 percent of families.

Treasury's Strange Concept of Income. Problems with the Treasury analysis, however, cast grave doubt on its validity. Many of the problems relate to concepts of income. Treasury uses a concept called Family Economic Income (FEI), which has little relationship to income as ordinary people understand it or even to income as people compute it for their tax returns. For this reason, the Treasury analysis is very misleading.

Most people are familiar with the basic concept of Adjusted Gross Income (AGI), which the Internal Revenue Service uses to determine tax payments. AGI includes wages, salaries, taxable interest, dividends, alimony, realized capital gains, business income, pensions and other familiar forms of income. Treasury starts with AGI but adds to it many forms of income that are not included on tax returns and that most taxpayers would not consider income at all. For example, in Treasury's calculations:

  • All contributions to pensions, IRAs and Keogh plans, Social Security benefits, fringe benefits and tax-exempt interest are treated as if they were taxable income.
  • All taxpayers who own homes are considered to pay rent to themselves, and this imputed rent is counted as income.
  • Unrealized capital gains are counted as if they were realized annually, even if no assets were actually sold.
  • Owners of corporate stock are assumed to receive 100 percent of corporate profits, even though much of that profit is never paid in the form of dividends but is retained by the corporation.
As Figure I shows, the result of all these changes is to portray most taxpayers as 50 percent richer than their tax returns say they are and thus to make many taxpayers of relatively modest means appear rich in Treasury's distribution table.

Omitting Income But Not Taxes. On the other hand, the Treasury method excludes much income that taxpayers do find familiar. For example, since pension contributions and all corporate profits are already attributed to taxpayers, actual pensions and dividends received by taxpayers are not treated as income. Thus:

  • Retired people living on pensions and dividends pay taxes on such income currently, but under Treasury's distribution table their income completely disappears.
  • However, since their tax liability is unchanged, they appear to be paying an extremely high effective tax rate when they are not.
  • Thus FEI not only makes many people with modest incomes appear rich, it also make many with middle incomes appear poor!
Another anomaly is that capital gains on corporate stock are excluded from individual income to prevent double-counting, but the capital gains tax is counted. The effect is to make people realizing capital gains appear much more heavily taxed than they actually are.

Finally, although Treasury includes imputed rent as homeowners' income it does not make the same adjustment for public housing, and excludes all noncash welfare benefits except food stamps, making many poor people appear utterly destitute.

Wrong Assumptions about Effective Dates. Treasury's distribution tables make the tax bills appear largely to benefit the rich because of its definition of income. But there are also other reasons. For example, the Treasury assumes that the tax bill will be fully effective in 1998, even though many provisions do not take effect for years.

  • Thus Treasury's distribution table assumes a tax cut of $71.2 billion in the case of the Ways and Means Committee bill and $60.8 billion in the case of the Finance Committee.
  • In fact, the Finance Committee bill would actually increase federal revenue slightly in 1998; even in the year 2007, when the tax cut is fully phased in, it would lower federal revenues by only $40.2 billion.
  • Thus Treasury's distribution table implies phantom tax cuts even in years when tax payments actually rise.
Capital Gains: Reality vs. Treasury. A major reason for this anomaly is capital gains. Reductions in the capital gains tax rate in 1978 and 1981, as well as the rate increase in 1986, had enormous effects on asset sales and thus on revenues. Even Treasury admits that lowering the capital gains tax rate as proposed in both congressional tax bills would temporarily increase federal revenues by increasing capital gains realizations. Yet Treasury's distribution table shows owners of capital assets getting a big tax cut. In effect, Treasury assumes that all capital gains - including those only induced by the lower tax rate - would have been realized anyway!

Congress's Joint Committee on Taxation (JCT) uses the same methodology, which has the effect of making those paying more in capital gains taxes appear to be paying less. Professor Michael Graetz of Yale Law School points out that in 1990 the JCT's distribution table showed President Bush's proposed cut in the capital gains tax giving taxpayers a $15.9 billion tax cut, although the JCT's own estimate showed that federal revenues would be lower by at most $4.3 billion. Based on this contradiction, Graetz constructed the chart shown in Figure II. As one can see, those with incomes above $200,000 appear to be getting a tax cut four times larger than is possible.

Professor Graetz believes the methodology for creating distribution tables is so deeply flawed that the tables should be abandoned altogether during the legislative process. If distributional tables are produced, it should only be after the fact and should show the true impact of a tax change on taxpayers. Another reason to abandon distributional tables is that during the legislative process they tend to overwhelm sound principles of tax policy. As a way of conveying to taxpayers how tax bills would actually affect them, the distribution tables are utterly worthless.

This Brief Analysis was prepared by NCPA Senior Fellow Bruce R. Bartlett.

Read Article as PDF