Revised April 6, 1998

A Response to the Tax Foundation's Response


On March 26, the Tax Foundation released a response to a report the Center on Budget and Policy Priorities issued two weeks earlier entitled, The Debate Over Tax Levels: How Much Does a Typical Family Pay? The Center's report was highly critical of the way in which the Tax Foundation computed and presented what it described as "...the tax burden on America's median-income families...." The Tax Foundation response to the Center's report, defending the methodology the Center found flawed, characterizes several of the Center's criticisms in misleading ways.

The following discussion examines the Tax Foundation response in more detail.


"Typical" Families and Their Income

The Center's analysis of the federal tax liability of a typical, median-income family relies heavily on analysis by the non-partisan Congressional Budget Office. The Center's report stated:

"The Tax Foundation's statement that taxes absorb 38.2 percent of a median two-earner family's income includes an estimate that this median family pays 26.2 percent of income in federal taxes....The respected, non-partisan Congressional Budget Office has reported that the federal tax burden on families in the middle fifth of the income distribution equaled 19.7 percent of income before the effects of the 1997 Taxpayers Relief Act are taken into account..."

The Tax Foundation claims the discrepancy the Center highlights in the above quotation from the Center's report is in part due to an "error" the Center made in using a Congressional Budget Office table on the tax burdens of families at different income levels. According to the Tax Foundation response, "...the CBPP misread the table and pulled off the wrong figure." That is incorrect.

As the Center's analysis made clear, the Center was looking at the tax burden for a family in the middle 20 percent of the income distribution — effectively the tax burden of a median-income family. By contrast, the Tax Foundation report highlighted the tax burden of a family at the median income for two-earner families. The Center's report pointed out that the income of a family at the median for two-earner families is nearly 50 percent higher than the median income for all families.

The Center noted that the Tax Foundation's figures are usually cited by policymakers, pundits, and journalists as representing the typical or average family — not a typical two-earner family — and that neither those citing the report nor the Tax Foundation point out that the family which the Tax Foundation has used here falls in the top third of the income distribution of all families and has income nearly 50 percent above the overall median. The part of the CBO table the Tax Foundation suggests the Center should have used shows that more than two-thirds of all families have incomes lower than the family the Tax Foundation considered.

In our progressive tax system, the percentage of income that families pay in taxes rises as income climbs. Thus, the Tax Foundation's two-earner family would have somewhat higher taxes than a family whose income placed it in the middle of the income scale for all families.

Far from erring in misreading the CBO table and taking the wrong figure from it, the Center took the CBO figure for families in the middle of the income distribution — that is, the type of family that those who cite the Tax Foundation's figures generally present its figures as describing. The Center's report clearly stated that it was highlighting the difference between the taxes of the Tax Foundation's two-earner family and the taxes of a median-income family. A family in the upper third of the income distribution is not what most people think of when they hear a reference to the "typical" family.

To its credit, the Tax Foundation did not say its calculations represented anything but a two-earner family. Unfortunately, however, that is not how the Tax Foundation's report is generally used. The Center's report provided a series of examples of ways in which the Tax Foundation's statement that the two-earner family pays 38 percent of income in taxes has been used to claim this is the tax burden of the typical or average family. The Center cited the following statements from Majority leader Trent Lott, Speaker Newt Gingrich, and Steve Forbes:

The 38 percent tax burden figure has moved into common usage as representing the taxes of the "typical" family, without any qualification limiting its applicability to two-earner families. We are not aware of any efforts by the Tax Foundation to object to this misuse of its data or to try to prevent such misuse. We would be happy to hear if we are wrong in that respect.

The Center's report states that if one wants to talk about the tax burden of a "typical" family, one should examine the taxes the median-income family pays. The Center's analysis did not seek to recalculate the tax burden of the dual-earner family in the upper third of the income distribution, as the Tax Foundation response mistakenly portrays our analysis report as attempting to do. Our analysis examined the taxes the typical, median-income family pays.

Tax Foundation Underestimates Other Sources of Difference

The Tax Foundation claims much of the difference between its tax burden estimate and the Center's would disappear if the Center had used the CBO tax burden for a family with income in the $50,000 to $75,000 range, the range into which the Tax Foundation's two-earner family falls. The Tax Foundation's response says: "After correcting for the fundamental error in the CBPP analysis of placing the median-income dual-earner family in the wrong income class [this is the incorrect characterization of the Center's report just discussed], there is a difference of only 2.7 percentage points between the CBO estimate of the total federal tax burden on a family and the Tax Foundation's calculation." This may leave the impression that the CBO income-class data in some way validates the Tax Foundation's calculations.

Even leaving aside the issue that a dual-earner family in the top third of the income distribution is not typical of families in general, however, the Tax Foundation methodology remains seriously flawed. In gauging the validity of the Tax Foundation's estimates, it is instructive to look at the one-earner family the Tax Foundation report also examines. The Tax Foundation used the same methodology to compute the taxes of both one-earner and two-earner families.

According to the Tax Foundation, a one-earner family with income of $28,808 (the median income for a one-earner family) pays federal taxes amounting to 24.4 percent of income. According to the very CBO table the Tax Foundation cites in its own defense in its rejoinder to the Center, however, a family at this income level would pay 15.3 percent of income in federal taxes. The difference the Tax Foundation touts as being "only 2.7 percentage points" for the two-earner family is a whopping 9.1 percentage points for the single-earner family. This underscores the point that far more is wrong with the Tax Foundation's calculations than the level of family income used in examining tax burdens.

The Tax Foundation has created a methodology and adopted a set of assumptions that all push in the same direction, toward the overstatement of the tax burden of lower- and middle-income families. The Center's report highlighted in detail a number of problems with this methodology.


Calculation and Attribution of Federal Corporate Income Taxes

The Center criticized the Tax Foundation for assuming that all families, regardless of income, pay the same percentage of income in federal corporate income taxes and estate taxes. The Tax Foundation methodology assumes a median-income two-earner family with an income of about $55,000 pays the same percentage of income in corporate income taxes and estate taxes as a wealthy family with income of $5 million and extensive investments. This is not the case.

The bulk of corporate income taxes and virtually all estate taxes are paid by higher-income taxpayers. Estate taxes are paid only on the estates of the wealthiest two percent of decedents; the other 98 percent of estates are exempt. In addition, the majority of economists, including those at the Congressional Budget Office, believe both that the federal tax on corporate profits ultimately is paid by individuals who hold capital assets such as stocks, bonds, or other forms of capital and that ownership of capital is concentrated among higher-income taxpayers. The Center pointed out that while middle-income families generally have some capital assets, the assets held by middle-income families are a modest fraction of the total capital assets in the economy.

In the Tax Foundation's response to the Center's report, it insists that "...economists do not widely agree on the ultimate incidence of corporate taxes." The Tax Foundation assumes that a significant share of the corporate income tax reduces wages and thus affects workers at all income levels. The Tax Foundation claims its methodology on this matter leads to an estimate that is a "mid-range figure" and that "CBPP has chosen an extreme [assumption]."

In calling the Center's assumption "extreme," the Tax Foundation is effectively saying that the Congressional Budget Office has adopted an "extreme" assumption. The Center simply used CBO's methodology. It is rare, if not unheard of, for CBO to adopt extreme assumptions; it is a highly cautious, non-partisan institution that has managed to retain the respect of both parties through a number of politically-charged debates in part by holding to mainstream assumptions.

CBO is far from alone in assuming that the federal corporate profits tax is paid by individuals who own capital assets. The Treasury Department uses the same assumption. In addition, the classic, widely-used public finance textbook by Richard and Peggy Musgrave states that "...most economists view the burden of the corporation tax as falling on capital, in line with the competitive model."(1)

These assumptions concerning the incidence of the corporate income tax stem from the work of University of Chicago (and now UCLA) economist Arnold Harberger in the 1960s. In a recent National Tax Journal article Jane Gravelle, Senior Specialist in Economic Policy at the Congressional Research Service and a widely recognized expert in tax policy, discusses the extent to which Harberger's work still governs the way economists view the corporate tax. She writes: "The important insight of the Harberger model is that, given reasonable assumptions about the behavioral relationships, the tax is generally borne by capital....Harberger's original finding of tax incidence has stood up surprisingly well to a variety of challenges. Some variations in the modeling of the tax, such as the addition of sectors or allowing monopoly profits, do not alter the general expectation that the tax would fall on capital rather than labor income."(2)

Gravelle notes that recent research suggests monopoly conditions do not necessarily alter the conclusion about who pays federal corporate income taxes. A somewhat older view, reflected in the Musgraves' textbook, holds that in circumstances when markets are not operating competitively because of monopoly behavior by either corporations or labor, some portion of the tax on corporate taxes can be shifted to prices or wages. When the Tax Foundation makes its assumption that one-third of the federal corporate profits tax falls on prices and another one-third on wages, it is implicitly assuming that market competition is weak and market forces work extremely poorly in the U.S. business sector. In this respect, it is that Tax Foundation assumption which is extreme, not the widely accepted assumptions the Congressional Budget Office and the Center on Budget and Policy Priorities use.

Were the Tax Foundation following the CBO assumption, its families would not have the amount of capital income necessary to pay the amount of taxes attributed to them. The Tax Foundation assumes that families pay 2.78 percent of their income in federal corporate income taxes regardless of their income level. It assumes this is true of its dual-earner family with total income of $55,000, some 96 percent of which it says is derived from wages, and also of its single-earner family with income of less than $29,000, of which it says 97 percent is derived from wages. Some of the three percent to four percent of income that comes from non-wage sources might be income from capital, but some also must come from other sources such as rental income or disability or veterans benefits for a member of the household. If most or all of the corporate income tax falls on the owners of capital, as CBO and most economists hold, neither of these families have sufficient capital income to warrant that level of taxation.

Even following its own assumptions on this matter, the Tax Foundation's figures are internally inconsistent. If the Tax Foundation wants to assume that one-third of the corporate income tax falls on workers in the form of lower wages than otherwise would be paid, it must follow the standard procedure it uses elsewhere in its report when it attributes the employer's share of the Social Security tax to families — it must add to family income the corporate income tax that it assumes falls on workers in the form of lower wages (i.e., it must figure their income as if their wages had not been reduced), as well as adding to its tax burden figures the corporate income tax it assumes these workers pay through this wage reduction. The Tax Foundation fails to do this. Thus, whether the assumption is that the corporate income tax falls on owners of capital or on wages, the Tax Foundation's computations do not stand up.


Counting State and Local Receipts that are Not Taxes

The Center's report also criticized the Tax Foundation for counting as state and local taxes the category identified as "contributions for social insurance" in the Commerce Department data series on state and local government receipts. In the similarly named category of federal receipts, a portion of the category is tax revenue; it consists of Social Security, Medicare, and unemployment insurance taxes. But the other portion of the federal category and the "contributions for social insurance" in the state and local section of the data on government receipts are not tax revenue. They consist of the contributions that federal and state and local government employees make to their own pension plans and the contributions that the governments make to those plans in their role as employers.

In no other sector of the economy are contributions to a pension plan made by an employee or an employer counted as a tax. For example, the contribution that General Motors or any other company makes to a pension plan for its employees is not considered a tax; it, along with wages, is part of employee compensation. Moreover, employee contributions (as distinguished from employer contributions) are classified by the Commerce Department as personal savings. It is a quirk of the Commerce Department data series, which is not intended to measure tax burdens, that the transfer of funds to federal or state and local government pension accounts is lumped in with government receipts.

These Commerce Department data are rarely used as a source of federal tax statistics, because more appropriate sources of information are available. Analysts of state and local finances who use this Commerce Department data source (because it is more timely than other data sources) are aware of this quirk and exclude the category "contributions for social insurance" in estimates of state and local taxes. But whether unaware of this convention or for some other reason, the Tax Foundation engages in a clear case of double-counting when it counts the employer share of the pension contributions as a tax. The taxes that governments collect from the public to finance government expenditures, including expenditures for pension contributions, are appropriately counted as taxes. But when these funds are then transferred from general government accounts to pension accounts, the Tax Foundation counts these funds as a tax again. This causes the same dollars to be counted as taxes twice.

The Tax Foundation says these contributions should be counted as taxes because they are like Social Security payroll taxes, but the Tax Foundation's analogy to Social Security taxes is incorrect. State and local pension contributions are most like the contributions the federal government makes to the civil service retirement systems or the contributions a private employer makes to a pension plan. The Tax Foundation does not count either of those as taxes in its report. Moreover, in most states, pension contributions for state and local employees are in addition to Social Security taxes, just as contributions to federal civil service pensions or private employer pension plans are.


Center Criticisms Not Addressed in Tax Foundation Response

The Tax Foundation's response ignores a number of the Center's methodological criticisms. For example, although the Tax Foundation response defended at length its attribution of corporate taxes to the families it examined, it did not defend its attribution of more than $150 a year in estate tax payments to its dual-earner family with income of $55,000. Estate taxes are paid solely on the estates of the wealthiest two percent of decedents; the federal estate tax exemption eliminates taxation for the other 98 percent of estates. The Tax Foundation takes the estate taxes paid by a small number of families and spreads the amount of these taxes over families at all income levels, even though the vast majority of Americans are not touched by the estate tax. It is extremely unlikely that a family with total income of $55,000, some 96 percent of which the Tax Foundation assumes to come from wages, would amass sufficient assets to rank among the top two percent of all estates and be subject to this tax.

The Tax Foundation also failed to respond to the Center's criticism that in computing the state and local tax burden, it counted capital gains taxes but left out the capital gains income on which these taxes are based. As the Center noted in its report, most states — and some cities such as New York City — levy personal income taxes that, like the federal income tax, include taxation of capital gains income. The problem here is that state and local capital gains taxes are included in the receipts component of the Commerce Department data series the Tax Foundation uses, but the capital gains income on which these taxes are levied is not included in the income component of this data series.

As the Center's report noted, avoiding this type of error is particularly important in today's economy. Most states as well as the federal government are reporting surges in taxes on capital gains income as a result of the unusually strong performance of the stock market and the newly lowered federal capital gains tax rate that took effect last year. If total taxes, including surging capital gains taxes, are divided by a measure of income that fails to include capital gains income, tax rates will appear both to be higher and to be rising faster than is actually the case.

Finally, the Center had chided the Tax Foundation for including as state and local taxes such non-tax items as rent that private entities pay for space a state or local government owns and fines that businesses pay for violating local codes or regulations, such as building codes. The Tax Foundation response is silent on this matter.

End Notes

1. Richard A. Musgrave and Peggy B. Musgrave, Public Finance in Theory and Practice, fifth ed. 1989, p. 387. Corporations do business in ways that maximize before-tax profits. Taxes are then paid out of profits. This might appear to suggest that the taxes are borne only by the owners of the corporations. Since capital is free to move between taxed and untaxed sectors, however, the tax effectively is spread over all owners of capital assets.

2. Jane G. Gravelle, "The Corporate Income Tax: Economic Issues and Policy Options," National Tax Journal, June 1995, p. 269.