Criticism of CBPP Pension Analysis Rests on Selective Use of Data
And Leaves Misleading Impressions

by Peter R. Orszag, Iris Lav, and Robert Greenstein (1)

Table of Contents

I.  Criticism #1

II.  Criticism #2

III.  Criticism #3

IV.  Criticism #4

V.  Criticism #5

VI.  Criticism #6

VII.  Criticism #7

VIII.  Criticism #8

IX.  Criticism #9

On October 8, the Center on Budget and Policy Priorities published an analysis of the pension tax preferences included in the tax bill that Congress approved in August and President Clinton later vetoed (H.R. 2488). These provisions, with the exception of the vetoed legislation's IRA changes, also are part of minimum wage legislation recently introduced by Representatives Rick Lazio, Gary Condit and John Shimkus (H.R. 3081). The full House of Representatives is expected to vote on the Lazio package in the next few weeks.

The Center's analysis, by Peter Orszag, Iris Lav and Robert Greenstein, found that the proposed pension provisions would primarily benefit highly paid individuals and business owners; 80 percent of the pension and Individual Retirement Account (IRA) tax breaks in the vetoed tax legislation would accrue to the 20 percent of the population with the highest incomes, while only 3.8 percent of these tax breaks would go to the bottom 60 percent of the population. The analysis also found that these provisions could induce reductions in pension coverage among rank-and-file workers, rather than expansion of such coverage. The Center's findings are consistent with the conclusions of a number of other independent pension experts and with serious concerns the Treasury Department has expressed about these proposals.

On October 19, the Association of Private Pensions and Welfare Plans, a trade association that is lobbying on behalf of these proposals, issued a "Legislative Action" bulletin taking issue with the Center's analysis. This new Center analysis assesses the arguments raised in the APPWP legislative bulletin.

This analysis of the APPWP bulletin finds that its claims often rest on selective use of data. As this paper shows, virtually none of these claims stand up under close scrutiny. The conclusion stands that the changes in the pension provisions of the vetoed tax bill and the Lazio bill would further skew the already-skewed distribution of pension tax benefits toward those at the high end of the income scale and could reduce pension coverage among lower- and moderate-income workers rather than expand such coverage.

The APPWP bulletin lodges nine criticisms against the Center's October 8 analysis. This paper examines each of these criticisms in turn.


Criticism #1: The distribution of pension tax preferences simply mirrors the tax code, and the distribution of actual pension benefits is focused on low- and middle-income workers

APPWP does not dispute the fact that under these legislative proposals, the bulk of the new pension tax benefits would go to individuals in the upper part of the income scale. APPWP argues, however, that since the top 20 percent of taxpayers pay the majority of income taxes, it should not be surprising that a comparable share of the proposed pension tax breaks would go to this group. APPWP also argues that while most pension tax breaks go to people in the top fifth of the income distribution, pension benefits themselves flow heavily to low- and moderate-income workers, with 77 percent of the workers who are currently earning retirement benefits having earnings below $50,000 a year.

Considering the pension provisions of the vetoed legislation alone, separately from the IRA provisions — which APPWP argues is the appropriate way to look at the issue — the tax benefits that would flow to the 20 percent of the population with the highest incomes substantially exceed the share of taxes this group pays. Analyzing these pension changes by themselves is particularly useful now, since the Lazio bill includes all of the pension tax provisions of the vetoed legislation but not the IRA provisions.

The analysis also finds that the top five percent of the population, a group that pays 41 percent of all federal taxes and 54 percent of federal income taxes, would receive 87 percent of these pension tax benefits. The top one percent of the population, which pays 25 percent of all federal taxes and 34 percent of income taxes, would receive 56 percent — more than half — of these tax benefits. By contrast, the bottom 60 percent of the population would receive less than one percent of the pension tax benefits.

As a result, what the APPWP's argument — that new pension tax provisions necessarily must go to high-income people because these individuals pay the majority of income taxes — really suggests is that our ability to bolster retirement security among low- and moderate-income workers by expanding these pension tax preferences may be limited. If it is impossible to deliver a significant portion of new pension tax breaks to the bottom 60 percent of the workforce, it seems unlikely that such tax breaks would serve their ostensible purpose of significantly raising retirement savings for those who have the lowest pension coverage rates.

The APPWP's second point on this matter is that the Center's analysis of the distribution of the proposed pension tax benefits "misses the point," because APPWP contends that pension benefits — as distinguished from pension tax preferences — do flow heavily to low- and moderate-income workers. APPWP states that 77 percent of the workers accruing pension benefits are individuals who earn less than $50,000 a year. Here, APPWP presents data in a somewhat misleading fashion.

To see the flaw in the APPWP's argument, consider a group of three workers, one of whom earns $20,000, another of whom earns $40,000 and the third of whom earns $150,000. All are covered by a pension. It thus could be said that two-thirds of this group of workers with pension coverage earn less than $50,000. But suppose the contribution to the first worker's pension is $600, the contribution to the second worker's pension is $1,400, and the contribution to the high-income worker's pension is $8,000. The worker with $150,000 in earnings would secure 80 percent of the total pension benefits going to this group of workers ($8,000 out of $10,000). The workers with earnings under $50,000 would constitute 67 percent of those earning pension benefits, but only 20 percent of the benefits would accrue to them.

The share of workers earning pension benefits who have incomes below $50,000 (or some similar level) differs substantially from the share of the pension benefits this group of workers actually accrues. APPWP argues that "the important question in evaluating pension policy is where the pension benefits go." But it fails to provide the relevant data on this very matter and misleads by instead providing data on the percentage of those accruing any pension benefit who have incomes below $50,000.

This is much like an argument frequently made by some proponents of large capital gains tax cuts. They often argue that a capital gains tax cut is a middle-class tax break because more than half of those with capital gains income have incomes below $50,000. Those who make this argument usually fail to mention that because those taxpayers with incomes below $50,000 who have capital gains income generally have only small amounts of such income while high-income taxpayers with capital gains income tend to have much larger amounts of it, 75 percent of all capital gains income realized in any year accrues to the highest-income five percent of the population.(4) APPWP's argument is misleading in the same manner that this capital gains argument is.

Finally, APPWP argues that the tax benefits under consideration are relatively small compared to existing pension tax preferences. Even if true, this it is not a reason to adopt the proposed tax preferences. Given the fiscal constraints facing policy-makers and the alternative purposes to which scarce federal resources could be put, the proposed tax preferences do not represent sound policy.


Criticism #2: The legislation would create new retirement benefits that benefit all workers

APPWP's second point is that the pension changes contained in the vetoed tax bill and the Lazio bill would lead key business decision-makers "to start, maintain and improve plans" that would benefit all of their workers. This is the basic rationale for the pension tax preferences in this legislation.

As the analysis the Center published earlier this month explains, however, this argument may make a good sound-bite, but there is no empirical evidence to support it. There is no basis in the research literature for the contention that raising the pension tax benefits available to high-income executives and small business owners will lead to additional pension coverage for lower- and moderate-income workers. To the contrary, as the Center's paper noted, some of the proposed tax preferences in the legislation could result in reductions in such coverage.

At the end of its legislative bulletin, APPWP implies that academics and pension experts support these legislative proposals and support APPWP's views on it. Yet the two studies APPWP cites to back-up this assertion consist of a paper that APPWP itself issued, in conjunction with the principal capital gains lobbing organization in Washington (the American Council of Capital Foundation), and a paper by William Gale of the Brookings Institution which covers a different set of pension issues and does not bear on these pension proposals.(5) In fact, Gale concurs with the Center's finding that there is no empirical evidence showing that increasing the benefit and contribution limits would raise coverage for lower- and moderate-income workers.

Numerous pension experts who are not connected to any lobbying group or special interest have voiced strong concerns that the proposed pension provisions could reduce, rather than increase, coverage for lower- and moderate-income workers and have recommended against adoption of these proposals. These experts include: Professor Daniel Halperin of Harvard Law School; Professor Norman Stein of the University of Alabama Law School; Donald Lubick, former Assistant Secretary of the Treasury for Tax Policy; Dianne Bennett, president of a large law firm in Buffalo and a nationally recognized expert in pension issues; and Karen Ferguson, director of the Pension Rights Center.


Criticism #3: The legislation merely restores contribution and benefit limits to prior levels

APPWP's third criticism is that the proposed increases in pension contribution and benefit limits (provisions that confer most of their benefits on individuals with high incomes) "would only partially restore the limits to where they stood prior to revenue-driven cutbacks in the 1980s and 1990s." This argument is predicated on an assumption that the previous levels were justified and reasonable and that restoring them is therefore warranted. Even the Reagan Administration's Treasury Department noted in the early 1980s, however, that the limits then in place were substantially too high.

In 1982, John E. Chapoton, Assistant Secretary for Tax Policy in the Treasury Department under the Reagan Administration, stated the Administration's position that these limits should be reduced. Chapoton testified: "Treasury agrees that the limits on contributions and benefits in section 415 of the Code should be reduced. Current dollar limits allow an overly generous amount of tax-favored retirement income."(6) The levels in place prior to the reductions of recent decades thus are not a reasonable benchmark against which to judge current reforms.

It should be noted that Chapoton also stated: "We have also heard the argument that reducing the limits may cause some plan terminations....We are not aware, however, of any empirical evidence that reducing the dollar limits to $30,000 and $90,000 will cause a significant number of existing plans to be terminated."

This testimony was delivered in 1982. Seventeen years later, there still is no empirical evidence that the reverse of Chapoton's statement is true — that raising the limits would cause a significant number of new plans to be created. This did not stop interests that lobby for higher limits from advancing that argument in 1982 or recycling the argument in 1999.

Finally, APPWP argues that by "returning the limits to historical levels, H.R. 2488 [the vetoed tax bill] does not offer a boon to the rich but merely helps the pension system to keep up with inflation." This sentence may lead readers to assume the current limits are not adjusted for inflation. That, however, is incorrect; the pension benefit and contribution limits already are indexed to inflation under current law. For example, because of inflation, the maximum amount a worker can contribute to a 401(k) plan will increase from $10,000 in 1999 to $10,500 in 2000, a five percent increase. Similarly, the maximum compensation that can be taken into account for pension purposes will rise from $160,000 to $170,000, a six percent increase. The proposed expansion of pension tax breaks is not needed to enable limits on pension tax preferences to keep pace with inflation, as the APPWP bulletin implies.


Criticism #4: Revenue-driven pension changes in the 1980s and early 1990s harmed the pension system

APPWP states that the reductions in the contribution and benefit limits in the 1980s and 1990s have "resulted in many highly paid employees leaving the private pension system for alternative vehicles such as non-tax-qualified plans." APPWP notes that 74 percent of Fortune 1000 firms offered non-qualified plans in 1995, up substantially from the prior year and from the 1980s. This movement into non-qualified plans has, according to APPWP, reduced the benefit dollars flowing into qualified plans and thereby hurt rank-and-file workers. This argument, too, does not stand up well under examination.

APPWP cites an increase in non-qualified plans among Fortune 1000 companies (particularly following the reduction in the maximum considered compensation limit in 1993) and asserts that such plans are displacing qualified plans. But the evidence suggests that these non-qualified plans may be supplementing qualified plans rather than supplanting them. For example, large firms such as those in the Fortune 1000 firms are the only corporations that would have substantial numbers of participants in single-employer pension plans. Data from the Pension Benefit Guaranty Corporation show that the number of single-employer PBGC-insured defined benefit plans with more than 5,000 participants increased from 1,021 in 1993 to 1,099 in 1998.(8) The number of participants in such plans also climbed.(9) The growth in non-qualified plans at Fortune 1000 firms does not seem to have displaced qualified plans.

In fact, to the extent that non-qualified plans do not displace qualified plans, then providing new tax incentives for higher-income executives — as the vetoed tax bill and the Lazio bill would do — would merely allow such executives to shift assets from non-qualified plans to qualified plans and secure larger tax breaks as a result. Such new tax breaks would then not increase private saving. To the contrary, the expanded tax preferences would grant these executives a tax break for saving that already was occurring.

In addition, if Congress were to extend such tax breaks without offsetting their costs through tax increases or spending cuts, that would reduce public saving because it would make the budget surplus smaller than it otherwise would be. The overall effect of these pension tax benefits thus could be to reduce national saving, since the asset shifting that would occur among executives who move contributions from non-qualified plans to qualified ones would not raise private saving, while public saving would decline due to the revenue loss that the tax breaks would cause.


Criticism #5: Non-discrimination rules ensure that rank-and-file workers will benefit

APPWP's fifth criticism is that the non-discrimination and top-heavy rules of federal pension laws would ensure that higher-income executives could not take advantage of the higher pension limits the legislation would provide without increasing pension benefits for their rank-and-file employees at the same time. APPWP argues that "in order for highly paid employees to take advantage of the higher limits and still pass the non-discrimination tests, many companies will have to provide greater benefits to all other workers."

Here, again, an assertion that may sound convincing at first blush turns out to present only part of the story.

This increase in the compensation limit would allow larger contributions for high-income executives and business ownersand would do so without requiring any increase in benefits for lower-income workers through the non-discrimination rules. Higher-income executives could accrue these larger pension benefits without violating the non-discrimination rules.

In fact, as the Center's earlier paper explained, this increase in the compensation limit would allow small business owners to reduce pension contributions for lower- and moderate-income workers while maintaining their own contribution levels — and while still complying fully with the non-discrimination rules. The APPWP argument also does not apply to 401(k) safe harbor plans, which are exempt from the non-discrimination rules because they provide a specified employer contribution.


Criticism #6: The legislation contains specific measures to aid rank-and-file workers

The APPWP correctly notes that the vetoed tax bill and the Lazio bill contain some specific measures to aid rank-and-file workers. The Center's October 8 analysis does not overlook this point, which is specifically discussed early in that analysis. The Center's paper notes that the proposed legislation includes beneficial reforms such as faster vesting for employer matching contributions to 401(k) plans and easier rollover rules. These changes are beneficial.

These constructive changes should, however, be kept in perspective. They are relatively minor — both in cost and in scope — in comparison to the other pension provisions in the legislation. For example, the Joint Tax Committee estimates that the faster vesting of employer matching contributions would have a negligible revenue effect and that the rollover provisions would cost a total of $106 million over 10 years. The increase in benefit and contribution limits and other potentially dangerous reforms comprise the vast majority of the $15 billion in pension tax breaks and $35 billion in IRA tax breaks in the vetoed tax bill.

It also should be noted that APPWP mischaracterizes one of the changes it cites as directly benefitting rank-and-file workers. APPWP describes the proposed increase in the 25 percent-of-compensation limit on combined employer and employee contributions to defined contribution plans as exclusively benefitting "modest-income savers." In fact, the 25 percent-of-compensation limit now applies to anyone earning less than $120,000 and, under the proposed legislation, would apply to anyone with less than $160,000 in income. (Under current law, combined employer and employee contributions to defined contribution plans can not exceed 25 percent of pay, or $30,000, whichever is lower. The 25 percent-of-pay limit therefore applies only to people earning less than $120,000, since 25 percent of earnings of less than $120,000 is less than $30,000. In addition to eliminating the 25-percent limit, the legislation would raise the dollar limit on combined employer-employee contributions to defined contribution plans from $30,000 to $40,000. If only the increase in the dollar limit from $30,000 to $40,000 were passed, the 25 percent limitation would apply to anyone earning less than $160,000.)

Indeed, individuals who are high-income but earn below $160,000 may be the primary group this provision would benefit. Workers who truly have modest incomes cannot afford to save more than 25 percent of their compensation; their disposable income is typically too limited to meet ongoing needs and have more than 25 percent of income left over. Much of the benefit from increasing this 25 percent limit thus is likely to accrue to individuals who earn more than $100,000 (but less than $160,000), as well as to individuals who earn more moderate salaries themselves but are members of high-income families, such as secondary earners who have high-earning spouses. Because such families have high overall incomes, they can afford to put away more than 25 percent of the lower-earning spouse's wages. But these families cannot be described as moderate-income households. The portrayal of this proposal as exclusively benefitting modest-income workers is thus off the mark.


Criticism #7: Pension policy should not be driven by plans on the fringe

APPWP argues that any potential for coverage to decline as a result of the new tax preferences in the legislation would be limited to a very small set of employers and the Center's analysis is skewed by an emphasis on such firms. This criticism is a curious one. One of the principal rationales for this legislation is the claim by its proponents that it would induce an expansion of pension coverage among this same group of employers; the contention is that by making coverage more generous for small business owners and highly paid executives in these firms, the legislation would induce more of those firms to institute or expand pension coverage. The Center's analysis responds to this contention, showing that there is no empirical basis to support this argument and that the legislation is more likely to induce such firms to scale back coverage than to expand it.

For very large employers (i.e., major corporations), we doubt that the principal provisions of the legislation would have much effect on the incentives to adopt pension plans one way or the other. Put simply, top executives at very large firms are unlikely to base their views of whether the firm should adopt a pension plan on their own individual circumstances. If they did, they would violate their fiduciary responsibility to their shareholders. Furthermore, they typically enjoy such substantial compensation and benefit packages that the proposed changes in the contribution and benefit limits — while large relative to the vast majority of workers' salaries — would not be likely to affect significantly their incentives to adopt or drop a plan.


Criticism #8: More detail is needed on the methodology used to assess the effects of the proposed pension changes on different income groups

APPWP contends that the Center's paper provides too vague a description of the methodology used to analyze the distribution among different income groups of the benefits of the pension and retirement provisions of the vetoed tax bill. Given space constraints, we did not include a highly detailed description of the methodology in the October 8 report. The Center's report specifically explained, however, that the distributional analysis was conducted using the Institute on Taxation and Economic Policy tax model. This is a well-known model that is well-respected by a wide range of tax policy professionals. For this reason, the Center did not think it necessary to include a lengthy, highly technical discussion of the model. For those who are interested, a detailed description of the methodology underlying the model may be found at http://www.ctj.org/itep/itepdesc.htm. We would note that the model contains extensive data on current use of retirement savings plans.

It also appears from the APPWP paper that APPWP misunderstood the model and the methodology underlying the Center's analysis in some respects. Some points raised in this part of the APPWP paper are incorrect or misleading.

APPWP also states that "CBPP focuses on tax expenditures rather than the percentage tax cut provided to different income classes." In fact, the Center's report examined both the average tax cut amounts received by taxpayers in each quintile and the percentage of the total tax cut benefit that each quintile would receive. These are standard ways to analyze tax changes and are reflected in virtually every Treasury and Joint Tax Committee analysis of tax changes.

APPWP argues instead for analysis of the percentage tax cut that each income group would receive. While useful in some respects, such an analysis also can be misleading. Consider, for example, two taxpayers, one of whom is moderate income and pays $200 in federal income tax, and the other of whom is very high income and pays $100,000 in income tax. If a tax subsidy proposal gives a tax cut of $40 to the first taxpayer and $10,000 to the second tax payer, the first taxpayer would get a 20 percent tax cut while the second taxpayers would get a 10 percent tax cut. But while the percentage cut would be greater for the lower-income taxpayer, 99.6 percent of the tax cut benefits would go to subsidize, and lower the taxes of, the high-income taxpayer.

APPWP also asks a few other questions about the methodology of the Center's analysis, including which provisions of the vetoed bill are included in the distributional analysis. The distributional analysis in our October 8 report reflects all the proposed tax changes relating to pensions and IRAs in the vetoed tax legislation. Answers to other APPWP technical questions are provided in the footnote below.(10)


Criticism #9: The study distorts the effects of pension reform by including IRAs

The APPWP's final criticism is that the Center's analysis is distorted because it includes provisions relating to individual retirement accounts (IRAs). But including those provisions, which were an integral part of the vetoed tax bill, is entirely appropriate. Like employer-provided pensions, IRAs serve as a tax-preferred mechanism for retirement savings.

Furthermore, the private pension system is intimately related to the rules governing IRAs. Firms deciding whether to set up a 401(k) plan may consider the alternative of encouraging workers to create IRAs. In addition, as the Center's analysis pointed out, provisions of the vetoed tax bill that would increase the maximum IRA contribution that a high-income couple which does not have access to an employer-sponsored plan could make from $4,000 a year to $10,000 could induce some small business owners to terminate or not to offer pension coverage through their firms (because they now would be able to put away $10,000 for themselves in a conventional IRA without the firm having to bear the expense of making pension contributions for all employees). For these reasons, it would have been inappropriate to exclude the IRA provisions of the vetoed bill from our analysis.

APPWP's criticism in this area is particularly ill-founded for another reason as well. APPWP asserts that including the IRA provisions in the Center's analysis of the distributional effects of the vetoed bill's pension provisions magnifies the extent to which these tax benefits are found to be tilted to those at the top of the income scale. APPWP asserts that if the IRA provisions were removed from the distributional analysis and the pension provisions examined by themselves, the distributional effects would be less skewed to the top.

This is simply wrong. When the IRA provisions are omitted (as the Lazio bill would do), and the pension provisions examined by themselves, the benefits are found to be even more skewed to those at the top. Nearly all of the new tax breaks from the proposed changes in pension tax preferences are found to accrue to those in the top fifth of the income distribution when the IRA provisions are set to the side.

These figures apply to the pension tax provisions of the Lazio bill. That bill includes the pensions tax provisions of the vetoed tax legislation but not the vetoed bill's IRA measures.



The APPWP legislative bulletin that sought to disparage the Center's analysis does not bear up well under scrutiny. The findings of the Center's analysis continue to hold: when examined as a whole, the pension provisions of the vetoed tax bill and the Lazio bill tilt their tax breaks heavily toward higher-income executives and business owners who already have pension coverage. These provisions also risk inducing reductions in pension coverage and retirement security among lower- and moderate-income workers.


1. Citizens for Tax Justice, "Minimum Wage Plan Targets Three-Quarters of Its Tax Breaks to the Top 1%," October 18, 1999, available at http://www.ctj.org.

2. Since our tax system is progressive, disparities in after-tax income are somewhat smaller than disparities in before-tax income. A tax cut that is proportional to current tax burdens would make after-tax income disparities look more like before-tax income disparities, and would therefore increase after-tax income disparities. For a discussion of current income disparities in the United States, see Isaac Shapiro and Robert Greenstein, "The Widening Income Gulf," Center on Budget and Policy Priorities, September 4, 1999.

3. Some historical evidence suggests that pension wealth is distributed somewhat more evenly than non-pension wealth. Given the substantial inequality in wealth holdings, however, — the top 1 percent of the population holds 39 percent of the wealth, while the bottom 80 percent of the populations holds just 16 percent of the wealth — and the increasing concentration of wealth since these historical data were gathered, pension wealth is likely to be skewed toward those whose earnings exceed $50,000. See Ann McDermed, Robert Clark, and Steven Allen, "Pension Wealth, Age-Wealth Profiles, and the Distribution of Net Worth," in Robert Lipsey and Helen Stone Tice, The Measurement of Saving, Investment, and Wealth (Chicago: University of Chicago Press, 1989). Also, for further discussion of wealth distribution and recent trends in this area, see Edward N. Wolff, Top Heavy: A Study of the Increasing Inequality of Wealth in America, Twentieth Century Fund, 1995, and Edward N. Wolff, "Recent Trends in Wealth Ownership," in Benefits and Mechanisms for Spreading Asset Ownership in the United States, forthcoming.

4. Congressional Budget Office, Perspectives on the Ownership of Capital Assets and the Realization of Capital Gains, May 1997.

5. The paper by Gale is relevant to one provision in the vetoed tax bill, a change in the rules under which defined benefit pensions can be funded. The vetoed bill would remove a provision under current law that limits pension funding to 150 percent of "current liability." The legislation would instead allow funding up to the "accrued liability." Current liability and accrued liability differ because of projected wage increases over the career of a worker; current liability does not recognize such projected increases, while accrued liability does. Gale's paper does not discuss the proposed increases in pension contribution and benefit limits, which are the principal source of controversy. While this change in funding rules could be beneficial, it is a relatively minor component of the pension provisions of the vetoed tax legislation and the Lazio bill.

6. Statement of the Honorable John E. Chapoton on H.R. 6410, the Pension Equity Tax Act of 1982.

7. For more detail on these calculations, see Isaac Shapiro and Robert Greenstein, "The Widening Income Gulf," Center on Budget and Policy Priorities, September 4, 1999

8. Pension Benefit Guaranty Corporation, Pension Insurance Data Book 1998, table 547.

9. At least some of the increase in the number of participants and number of plans may be attributable to mergers of smaller plans that therefore moved into the larger-plan category. But even after taking this factor into account, the number of plans or participants at larger firms does not seem to have declined.

10. APPWP asked what method is used to place taxpayers into quintiles. The ITEP model counts a married couple filing jointly as one tax return, or tax unit. An unmarried taxpayer (excluding dependent filers) is similarly counted as one tax return. This is the same approach as the Joint Committee on Taxation uses. (In contrast, Treasury and CBO combine related members of a household into one tax unit. The difference is inconsequential for distributional results.) Thus, part-time teen-age workers are not counted as a separate unit if they are dependents of other taxpayers. Retired taxpayers are counted, so long as they are not dependents of others.

APPWP also asks whether an effort was made to take into account retirement benefits over a lifetime. The ITEP model is not a lifecycle model; here, again, the model follows a similar course to the Treasury and Joint Tax Committee models. It should be noted that the benefits of the tax cut proposals are analyzed on a net basis rather than a gross basis. That is, the analysis takes account of the fact that the tax benefits of 401 (k) plans, traditional IRAs, etc. are tax deferrals rather than tax exemptions. In other words, the model accounts for the extent to which the taxes paid when funds are withdrawn after retirement offset the tax breaks given during the working years.