1994-96 Advisory Council Report

Restoring Security to Our Social Security Retirement Program 

Joan T. Bok, Ann L. Combs, Sylvester J. Schieber, Fidel A. Vargas, and Carolyn L. Weaver 1

I. Introduction

Relatively early in the Advisory Council’s deliberations it became clear that there was strong opposition to, and perhaps no support at all for, a general increase in payroll taxes to deliver the Social Security benefits promised by current law. For most members this meant we needed to consider changes to the benefit side of the program. The problem that we faced as we examined benefit reductions was that such reductions run counter to two of the traditional goals of Social Security. The first of these goals relates to the adequacy of benefits and retirees’ ability to achieve a reasonable standard of living. The problem with many of the benefit reduction packages that we considered was that several Advisory Council members were not confident that Social Security, together with other likely sources of income, would continue to meet broadly accepted measures of income adequacy for future retirees.

The second of these goals relates to the fairness of Social Security in its treatment of participants in the program. Traditionally, this goal has meant that, other things being equal, a worker who contributes more to Social Security over his or her lifetime should receive absolutely higher benefits than one who contributes less, although the one who contributes less should receive relatively higher benefits in comparison to pre-retirement earnings. We looked somewhat more broadly at this goal than past Advisory Councils, in that we considered how Social Security treats workers across time. Most Council members expressed concerns that under current law Social Security was not a "good buy" for current and future generations of workers. Potential benefit reductions would make a "deal" that was already questionable even worse.

In sum, the Advisory Council faced a Social Security system projected to be underfunded by 25 to 30 percent after the turn of the century, a strong reluctance to increase payroll taxes to deliver current-law benefits, and real concerns about cutting benefits sufficiently to rebalance the system. These constraints required that we begin to "think outside" the normal boundaries of public policy that have constrained the development of Social Security policy for more than 60 years. The members of the Advisory Council ultimately split into three groups, each taking a different approach to addressing the issues we faced in trying to restore Social Security’s financing. We believe that the option that we recommend is superior to either of the other two. We come to that conclusion partly because we believe it will ultimately garner more public support. More fundamentally, however, we believe that it will provide greater retirement security for current and future workers than a shored up pay-as-you-go system or either of the other proposals, and that it will be better for our national economy over the long term.

Very likely, this Advisory Council has taken longer to conclude its deliberations than any prior Council. The very length of our deliberations is an indicator of the complex problems we face in restoring solvency and confidence in Social Security, a vital system of retirement income support for our citizens.

II. General Support for Advance Funding and Equity Investment

When the rhetoric is set aside, there was more agreement on this Advisory Council than it might appear. Faced with a serious and pressing financing problem, declining public confidence, and poor rates of return for younger generations, all of us saw the need to improve Social Security’s funding base and reap the gains from the higher returns to equity investment. This was a major point of agreement, implying a fundamental change in the way Social Security is financed and the way surplus funds are invested. Consensus broke down on the questions of whether or not workers should be allowed to reap the benefits and accept the risks of their investment decisions and whether or not an increase in saving and investment, not just a better return for the trust funds or for individual workers, should be a goal of reform.

Even with these disagreements, a majority of Council members recognized: (1) the desirability of moving toward a fully funded component of Social Security with investment in private capital markets, believing that both individuals and the economy as a whole stand to gain from a system built on saving and real capital investment; (2) the potential hazards of centralizing and possibly politicizing investment decisions; and (3) the powerful effect that private ownership could have in building confidence about the future of Social Security. These areas of agreement on important substantive issues led a majority of Council members to support fully-funded personal savings accounts and to include them as a key component of their reform proposals (the Individual Accounts and Personal Security Accounts plans).

There are many ways that personal accounts might be incorporated into Social Security, and the two options contained in this report reflect the breadth of at least some critical aspects of the choices that are available. Most importantly, personal accounts may be a small add-on to Social Security, as in the case of the IA plan, or a larger replacement for a portion of Social Security, as in our PSA plan. In addition, personal accounts may be structured so that workers have limited discretion in their investment decisions and in the way they may withdraw funds from their accounts at retirement, as in the IA plan, or they may be structured so that workers have considerable flexibility, as in the PSA plan. Personal accounts also may be introduced on a mandatory or voluntary basis, although both of our plans would be mandatory for workers.

III. Why We Support Personal Security Accounts

	As we approached our task of reviewing the status of Social Security and trying to develop a proposal that would create a secure, sustainable system into the next century, we quickly realized that there was more to be dealt with than an "actuarial imbalance." While the long-range deficit is, in our view, large and pressing, other problems also raised serious concerns. They include: the growing lack of confidence about the future of Social Security; the growing concerns about the value of Social Security to younger workers; the increasing interest in private alternatives to Social Security, including reforms being undertaken in many countries throughout the world, plus our own sense of the growing interest in these approaches here at home; and, more generally, the growing concerns about the impact of Social Security on the federal budget, national saving, and economic growth. We developed the PSA plan not only to respond to the actuarial deficit, but also to address these broader concerns.

The Need for Advance Funding of Our Retirement Obligations

A central problem with Social Security, in our view, is the way the system is financed, largely on a pay-as-you-go basis. Pay-as-you-go financing amounts to an income transfer system from workers to retirees rather than a retirement saving mechanism for workers. Under this arrangement, workers—and society more generally—forgo the opportunity to invest in real private capital and to earn the higher rate of return it would afford.

According to estimates contained in a Social Security Administration study, the average rate of return on Social Security taxes for the cohort of workers now turning 65 is about 4 percent, net of inflation. 2 In other words, if the taxes of this cohort of workers had been collected in an account yielding a real compound rate of interest of 4 percent annually, the accumulated balance would be just equal to the present value of the Social Security benefits they are projected to receive. For comparison purposes, cohorts of workers retiring before 1970 enjoyed double-digit rates of return. The cohort retiring in 1960, for example, had a real rate of return of 15 percent annually and the cohort of workers retiring in 1980 had a real rate of return of 7 percent.

With pay-as-you-go financing, the maximum rate of return payable to future cohorts is determined by the real rate of growth of taxable wages in the economy, which is generally assumed to be 1 to 2 percent in the long term. This is substantially below the real return to private capital investment, which is estimated to be on the order of 9.3 percent for society as a whole on a pre-tax basis, or 5.4 percent net of corporate income taxes. 3

Under present law, the average real rate of return on Social Security taxes is projected to fall to about 2 percent on average for workers reaching age 65 in 2020, and to level out at 1 to 2 percent on average for younger workers and future generations. Benefit cuts or tax increases designed to close the long-range deficit would depress rates of return to even lower levels—below 1 percent on average—until such changes were fully phased in. 4Individual workers fare better or worse than average depending on their marital status and earnings, among other factors—with single workers and high-wage workers faring worse than average, and low-wage workers and one-earner couples faring better than average. For some workers, the rates of return are projected to be negative. Importantly, these figures ignore any financial or political risks that might alter the financing base of Social Security or the level of Social Security benefits or taxes in the decades ahead.

There are those who are critical of rate of return calculations on the grounds that they are used to evaluate Social Security along the single dimension of "money’s worth," ignoring other important functions of social insurance. Money’s worth is only one of the standards against which a Social Security system should be evaluated, but it is an important one. It is hard to dismiss the money’s worth issue when discussing a program whose primary function is to deliver retirement benefits to workers in direct relation to their past earnings and thus to their tax contributions. Rates of return are a useful measure of performance and they confirm the simple fact that, as Social Security has matured, it has become relatively less effective at delivering retirement benefits in excess of taxes paid. With the population—and Social Security—aging, Social Security’s ability to generate windfall gains to retirees, or benefits well in excess of taxes paid, is evaporating and wealth losses are now in the offing. To maintain political support in the decades ahead, Social Security will have to offer better value to younger generations for their very substantial tax contributions.

From an economic perspective, the fact that Social Security offers younger workers and future generations a rate of return on their taxes that is much lower than the real return to private capital has two important implications: First, there is less saving and investment than there otherwise would be and workers’ return on their "investments" in Social Security is lower. This translates into significant losses of real income and wealth, not just for individual workers, but for society as a whole. In addition, Social Security amounts to a net tax on wages for the typical worker, causing distortions in labor market outcomes and in workers’ desired form of compensation. This translates into additional significant losses of income and wealth.

Studies by economists across the ideological spectrum suggest that the economic gains attributable to moving away from Social Security’s pay-as-you-go method of finance—a low-yielding system of income transfers—toward a system that is significantly more advance-funded would be very large. Virtually everyone on the Advisory Council has recognized the benefit of more advance funding, witnessed by the nature of the proposals they helped design and now support. The PSA plan has two clear advantages over the other proposals: First, it calls for greater advance funding than either of the other proposals; and second, it is designed in a way that we believe is vital to actually accomplish such prefunding.

Our plan would fully advance fund about one-half of the Social Security retirement program leaving a base program financed on a pay-as-you-go basis; the fully-funded portion would be a system of privately managed, individually-owned retirement accounts. We refer to these accounts as Personal Security Account (PSAs), similar to individual retirement accounts or 401(k) plans in the private sector, except that they would be mandatory.

In general terms, the benefits of the PSA plan exceed those offered by the current system for a number of reasons. The portion of Social Security that becomes fully funded through individual accounts would:

• replace a system of unfunded benefit promises that shifts the cost of benefits for the elderly to workers and future generations with a fully-funded system in which each generation saves for its own retirement;
• create a direct link between the tax contributions workers make and the benefits to which they are ultimately entitled, eliminating much of the complexity of the current system and alleviating labor market distortions;
• replace a system whose solvency is highly sensitive to demographic developments—both foreseen and unforeseen—with a system that basically runs on automatic pilot—personal accounts vary in value with changes in investment choices and performance but remain fully funded at all times;
• eliminate the retirement earnings test, which penalizes people who have invested mainly in human capital and derive their incomes mainly from wage earnings;
• significantly reduce other work disincentives that encourage many workers to retire when they are still extremely productive;
• avoid the double taxation that applies to ordinary savings;
• allow individual workers and families to be directly involved in the investment decisions that will vitally influence their future wealth and income; and
• give workers real ownership claims over the contributions to and the proceeds of their accounts, thus substantially reducing the political uncertainty surrounding the size and cost of future benefits.

For the portion of the system that remains a tax-transfer mechanism, the PSA plan would:

• provide a basic floor of protection for all full-career workers, which is fully cost-of-living adjusted for people in retirement and keeps pace with earnings’ growth prior to retirement;
• intentionally redistribute benefits toward low-wage workers; and
• in combination with the funded portion of the proposal, distribute the financial risks associated with a funded retirement program away from low-wage workers who are least able to absorb such risks and toward high-wage workers who are most able to absorb them.

The floor of protection in the PSA proposal would continue to be supplemented by the nation’s elderly public assistance program, Supplemental Security Income (SSI).

How the PSA Plan Works

Our proposal would very gradually transform Social Security into a system that offers workers two kinds of benefits: first, a basic flat benefit for full-career workers, which is scaled to years of work and financed on a pay-as-you-go basis, and second, the proceeds of Personal Security Accounts, which are fully advanced funded with workers’ tax contributions and investment earnings. PSAs would be funded with 5 percent of the current payroll tax (which is 12.4 percent, excluding Medicare), or about half the tax devoted to the retirement portion of Social Security. The first tier benefit and benefits to disabled workers and survivors would be financed with the remaining 7.4 percent. Five percentage points of the tax would be rebated for investment directly to workers or to the financial institutions of their choice.

Workers would own their PSAs and the interest thereon; they would be free to invest them in a wide range of investments and financial institutions; accounts would be held and managed by private financial institutions; individuals could begin making tax-free withdrawals at 62—regardless of their income or work status; annuitization of PSA balances would be an option but would not be forced on individuals who wished to control the timing and amount of their own distributions; and any balances remaining in an account upon death would be included in an individual’s estate. The PSA proposal would retain a spouse benefit. Unless a woman has earned a higher tier-one benefit on her own work record, she would still be eligible for 50 percent of her husband’s benefit. Widow’s and widower’s benefits would continue as well, only at a higher percentage amount than under current law. Young survivors would receive benefits as under current law but with actuarial reductions equivalent to benefits payable at age 65.

;It is important to note that this proposal does not, as suggested by one group of Council members, "abolish the present Social Security system"-- words apparently designed to arouse grave concern among people of all ages, but especially among senior citizens and older workers. The new two-tiered system affects no one older than age 55. It is phased in very slowly, becoming fully effective only for workers under age 25, and the projected benefits these workers would receive under the PSA plan are superior to those projected under current law. People from ages 25 through 54 would receive a portion of their first tier benefits based on what they had accrued under the current system, wage indexed until retirement, plus a portion of the new flat benefit based on the number of years worked under the new system; in addition, they would own their PSA accumulations.

Also, while the flat benefit for full career workers would be set initially at the equivalent of $410 monthly in 1996, the amount future retirees actually would receive would be considerably higher than this. The reason is that it takes many years for the new system to be phased in and, during the intervening years, the benefit would be increased annually to keep pace with average wage growth in the economy. The flat benefit for a full career worker retiring in 2041, the first year someone could retire at normal retirement age who has spent a lifetime under the new two-tiered system, is projected to be $664 monthly in constant 1996 dollars, 62 percent higher in real terms than the initial benefit amount. Once an individual is on the benefit rolls, the benefit would be indexed to the cost of living, as under present law.

 This reform would turn the vast majority of Social Security taxpayers into investors and, in the next decade alone, would release literally hundreds of billions of dollars of payroll taxes for investment in the private sector. As an indication of the magnitudes involved, taxable payroll in the U.S. is now about $3 trillion, 5 percent of which equals $150 billion annually, with the amount of additional revenues available for investment each year growing at the rate of growth of total wages in the economy. With workers assumed to allocate half of their contributions to equities and half to U.S. government securities, the SSA actuaries project that the total accumulation of assets in personal accounts will be close to $6 trillion in 2020 and $16 trillion in 2040 in constant 1996 dollars.

The tier-one benefit would serve as a back stop to the PSA accumulations. Set initially at a level equal to about two-thirds of the poverty level, this benefit would, as indicated, gradually rise over time, providing a minimum level of protection for all workers, but especially low-wage workers and others who, for whatever reason, accumulate relatively small PSA balances.5Of course, Social Security would continue to be backed up by SSI, which provides a guaranteed floor of protection to the elderly and disabled poor.

How Workers Fare Under the PSA Plan

Workers of various ages and at various earnings levels are projected to fare well under the PSA plan in comparison to the two alternative proposals considered by the Advisory Council. The PSA plan clearly dominates the others for younger workers and future generations.

In trying to evaluate how people fare, it is important to be aware of the assumptions underlying the actuaries' projections. Apart from the host of economic and demographic assumptions which follow the 1995 Trustees' intermediate assumptions with a modified CPI assumption, the actuaries assume that workers invest in stocks and bonds in a particular way, that they earn a particular interest rate on their investments, and that they pay a particular amount in fees to the financial institutions that manage their personal accounts.

For the estimates shown in the "Comparison of Plans," the real yield on stocks is assumed to be 7 percent, which is consistent with historical returns.6The real yield on bonds is assumed to be 2.3 percent, which is the actuaries' long-range assumption for the real yield on long-term U.S. government bonds.7In addition, the actuaries assume that administrative fees on personal accounts would be 100 basis points, or 1 percentage point. Together, these assumptions imply that workers who hold a 50-50 mix of stocks and bonds would have an average real yield of 4.65 percent, or 3.65 net of administrative fees. In addition, workers are assumed to hold the same mix of stocks and bonds that is observed to be held in 401(k) plans. Workers under 40 are thus assumed to hold 55 percent of their PSA balances in equities yielding a net return on their PSA accounts of 3.9 percent, whereas workers aged 60 to 69 are assumed to hold 43 percent of their balances in equities yielding a net return of 3.3 percent on their PSA accounts.

Of course, workers who invest more heavily in equities or buy funds with lower investment management fees, such as index funds, would have higher expected returns. Workers who invest more heavily in bonds or spend even more on administrative fees would have lower expected returns. The actuarial memo in Appendix II of this report shows estimates for higher and lower yield portfolios.

The other two plans, it should be noted, are assumed to have much lower administrative fees—10.5 basis points in the case of the IA plan and 0.5 basis points in the case of the MB plan. This, together with the assumption that under the MB plan the government will invest up to 40 percent of trust fund reserves in private equity, at a real return of 7 percent annually, helps account for the inflated estimates for the MB plan.

We raise these details for two reasons. First, since we cannot know what the future holds, assumptions must be made and the resulting estimates are no better or worse than the underlying assumptions. In this regard, we believe the interest rate and asset mix assumptions are generally reasonable. However, we believe that the gap between the administrative fees on the PSA plan and the MB plan is too large. We would expect money managers to compete aggressively to reduce fees and we would expect workers to search out low cost investment options such as no load mutual funds and index funds. Some mutual fund companies currently offer index funds for retail management fees under 30 basis points per year.

In addition, we do not believe it is appropriate to evaluate the MB plan as if its proponents actually endorsed centralized investment, which, in the end, 8 Removing this from the evaluation of their plan creates a large funding gap and forces the need for additional tax increases or benefit cuts, either of which would worsen the estimated performance of the plan.

Second, when evaluating the effect of a change in Social Security, it is crucial to know what the baseline point of comparison is. Since Social Security is substantially under financed after the turn of the century, present law does not provide a point of comparison. The MB plan does not provide one either, in our view, now that its proponents have withdrawn their endorsement for centralized investment. In either case, benefits will have to be lower or taxes will have to be higher than now projected. In our discussion, we consider a baseline in which the payroll tax is maintained at its current level of 12.4 percent (excluding Medicare), and benefits are gradually scaled back to restore actuarial balance.9We compare the PSA plan to this and to the IA plan.

In addition, we refer to internal rates of return rather than to money's worth ratios. We do this simply because rate of return calculations are somewhat easier to interpret since they are not affected by historical movements in interest rates and make no implicit assumptions about the appropriate interest rate for determining who does and does not get their "money's worth." Data on replacement rates and benefit levels, which ignore taxes, are contained in the actuarial memo in Appendix II and are discussed in the "Comparison of Plans" section of the report.

As revealed in the Appendix tables prepared by the Social Security actuaries, workers at all earnings levels, ranging from low to maximum, are projected to earn higher rates of return under the PSA plan than under the IA plan or the baseline. 10The younger the worker, the greater the benefits to the PSA plan, the reason being the relatively longer period of time over which the younger workers can accumulate personal accounts at the higher rate of return.

For single male workers with low earnings, rates of return under the PSA plan vary from 2.46 percent for the 1943 birth cohort—i.e., those now aged 53—to 2.99 percent for the 2004 birth cohort; this compares to 2.34 percent to 1.93 percent under the baseline. For the 1964 birth cohort, now age 33, the rate of return under the PSA plan is nearly double the baseline—1.40 percent compared to 0.73 percent—for the average wage worker and even higher for workers at higher wage levels. For single females, rates of return under the baseline are somewhat higher than for males, making the gains for females under the PSA plan somewhat smaller but nevertheless significant.

For two-earner couples, which will be the dominant type of couple in future decades, the PSA plan also dominates the alternatives shown. Here again, since couples tend to fare better than singles under the baseline and under present law, the gains are not as large as for singles but are nevertheless significant. Considering the couple with average earnings who was born in 1964, the PSA plan is projected to offer a real rate of return of 2.08 percent as compared to 1.64 percent under the baseline.

To appreciate the significance of a half or full point difference in real rates of return, when compounded over many years, consider the following example. If a couple invests a lump sum of, say, $20,000 today at age 30, by the time they reach age 65, this amount would have grown to $41,111 with a 2.08 percent return as compared to $35,342 with a 1.64 percent return. With a full percentage point increase in rate of return, the accumulation would have grown to $49,787.

As compared to the IA plan, the PSA plan outperforms the IA plan for workers at all ages and earnings levels, whether single or in two-earner couples. The differences are significantly smaller, however, than when the PSA plan is compared to the baseline.

These findings are especially significant in light of the fact that the PSA plan includes a transition payroll tax supplement and the cost of servicing the debt, both of which are included in these estimates. In other words, these estimates are net real rates of return. Unfortunately, the estimates do not extend far enough into the future to show how workers would fare who enter the work force once the supplemental transition tax is repealed, who naturally would earn even higher rates of return. 11These workers also stand to gain the most from any savings-induced increase in national income.

Judgments must be made about the level of various risks under the different options. The relatively greater share of benefits coming about through wealth accumulation in individually-owned accounts under the PSA plan, for example, exposes workers to more financial risk but shields relatively more of their retirement incomes from political risk. As noted elsewhere, we believe that the political risks attached to government benefit promises 20, 30, or 40 years down the road far outweigh the financial risks of a well-diversified portfolio.

IV. Transition Financing and Government Obligations

As with any proposal to move toward fully-funded personal accounts or simply to increase the funding basis of the current Social Security system, there is a transition cost of getting from where we are to where we would like to be. The reason is Social Security's pay-as-you-go method of finance, which implicitly is a form of deficit finance. In contrast to a fully-funded system, in which each generation of workers makes contributions that are saved and invested to fund their own future benefits, such as in a 401(k) plan in the private sector, a pay-as-you-go system shifts the cost of benefits for current elderly generations to current and future generations of workers. Whereas a fully funded system always has assets equal to outstanding liabilities, and thus operates with no unfunded liabilities, a mature pay-as-you-go system operates with very large unfunded liabilities. These liabilities are ignored in federal budget accounts.

According to the SSA actuaries, in present value terms, Social Security is scheduled to pay about $18.6 trillion in benefits to current workers and retirees over the next 75 years. This compares to OASDI's reserve fund plus projected tax income from these workers, which totals $9.8 trillion. Social security's unfunded liability is thus about $9 trillion in present value terms. It is the existence of this unfunded liability—an implicit or "off the books" debt of the federal government—that creates transition costs. Moving toward a fully funded component of Social Security, without reneging on expected benefits, requires that this implicit debt to retirees and workers be officially recognized, and gradually met, at the same time workers' taxes are being deposited in personal accounts to help fund a portion of their retirement benefits.

The SSA actuaries estimate that the long-range transition cost of our proposal is the equivalent of 1.52 percent of taxable payroll in the economy. We recommend spreading this cost over a 70 to 75 year period so as not to concentrate the burden too heavily on today's workers. Further, we recommend that this cost be met through a payroll tax supplement and, because of the concentration of these costs in the early decades of the transition, an increase in federal borrowing. In particular, we propose a payroll tax supplement of 1.52 percent—employee and employer shares combined—coupled with new, explicit government borrowing during the early years of the transition. This borrowing would be fully repaid with the proceeds of the tax during the later years of the transition, at which point the tax supplement would be repealed.

We would like to stress that none of us favor a tax increase to fund the current system. We favor a tax increase only to fund the transition to personal accounts. In addition, among taxes, none of us favor a payroll tax supplement. We would prefer a broad-based consumption tax, which would be paid by a broader segment of the population, including the elderly who have fared so well under the current pay-as-you-go system. Such a tax would also create fewer labor market distortions and be consistent with our more general goal of boosting national saving. However, since the U.S. does not presently have such a tax, we concluded that the costs of setting up the administrative apparatus and layering this tax on top of the existing income tax structure would outweigh the gains. Should the U.S. tax system move in the direction of a consumption base, we would regard this as a highly preferable means of meeting part of the cost of transition.

 We also would favor general spending reductions coupled with any tax increase. However, there was concern expressed that targeting specific federal programs for reductions would be outside the charge of this Advisory Council. In light of this, we proposed a tax supplement in order to deal forthrightly with the issue of transition costs.

We do believe that debt-financing part of the transition is desirable. Issuing new government debt just as we are about to rebalance our federal books may seem like an unsavory prospect. This debt, however, helps spread the burden to future generations, who stand to gain so much from these reforms, rather than concentrating it purely on current workers.

Why Are the Transition Costs So Large?

Two questions that have been raised about our proposal are why the transition costs are so large and why we chose to meet these costs the way we did, rather than the way Senator Bob Kerrey (D-NE) and former Senator Alan Simpson (R-WY) proposed in the legislation they introduced. The Kerrey-Simpson bill (S.825) would involve smaller transition costs and, in addition, would not require a tax increase to fund the transition.

The Kerrey-Simpson bill would create a system of personal accounts that is fully funded with 2 percent of the payroll tax, about one-fifth of the overall tax devoted to the retirement program. The PSA proposal, on the other hand, would create such a system with 5 percent of the payroll tax, about one-half of the tax devoted to the retirement program. The larger cost of our proposal reflects the greater extent to which we move toward fully-funded personal accounts and make explicit and begin to pay off Social Security's unfunded liability.

;In our view, it is highly desirable to move toward larger personal accounts than in the Kerrey-Simpson bill. Larger accounts offer workers the potential for higher retirement incomes by allowing them the opportunity to invest more of their taxes in higher-yielding stocks and bonds, and on net should result in significantly larger economic benefits. In addition, larger accounts would give workers keener incentives to make informed investment decisions and to monitor the performance of their investments. Larger accounts also would be relatively less costly for financial institutions to administer.

As for the means of financing the transition, the Kerrey-Simpson bill provides for what might be called "internal financing." Social Security benefits are scaled back sufficiently to close the long-range deficit and, in addition, cover the cost of the 2 percent personal accounts. This cost, moreover, is met by a combination of reductions in future benefits for current recipients and for future recipients. As analyzed by SSA, our proposal finances the transition "externally"—i.e., through explicit government borrowing and a general tax increase or equivalent spending reductions. Actually, the PSA proposal includes internal financing as well since, in recognition of the fact that individuals would be accumulating sizable personal savings accounts, Social Security benefits are scaled back more than necessary to close the long-range deficit. These benefit reductions would apply almost exclusively to future retirees. As a result, under the Kerrey-Simpson bill, relatively more of the cost of transition is borne by current older generations. While there are advantages to this, in particular, reversing some of the enormous wealth transfers that have taken place from younger to older generations, there also are advantages to spreading the burden to future generations who stand to gain the most from reform. The approach taken in the Kerrey-Simpson bill also effectively limits the extent to which personal accounts can substitute for a portion of Social Security.

From a practical political standpoint, there was little support on the Advisory Council for modifying the cost-of-living adjustment beyond the adjustments that will flow from measurement changes made by the Bureau of Labor Statistics. 12 Also, there was no support in our group for increasing the share of benefits subject to taxation—in fact, we propose reducing the share of benefits subject to taxation to 50 percent to be consistent with the tax treatment of most benefits provided through employer-sponsored retirement plans. As a result, there were few practical means of spreading the burden to current older generations except by way of a consumption-based tax. The one change affecting the elderly that our group did support, as did the other members of the Advisory Council, was a phase-out of the thresholds used for determining how much of an individual's benefits are taxable. Because of other features of the tax code that limit income tax liabilities for the low-income elderly, this change is expected to affect only a small share of beneficiaries.

It also should be noted that, with the exception of a few program changes that apply generally, such as the increase in the retirement age, our proposal would pay all benefits "earned" to date, with that amount indexed to wage growth until the time of retirement. Among the various options for transitioning to a new system, this is quite generous. Other less costly options include price indexing or freezing accrued benefits, or meeting less than full accrued benefits for younger workers or even canceling those benefits for workers under a certain age, such as 35. These options have been suggested by some in recognition of the greater risk younger people attach to receiving scheduled benefits and the larger gains they can expect under the reformed system of personal accounts. As a group, we felt that the benefits that would flow from the PSA plan made the cost of our transition worthwhile.

Finally, the fact that our present system is significantly underfunded on a pay-as-you-go basis makes the transition appear more costly than it actually is. In present value terms, fully $3.1 trillion is required over the next 75 years simply to close the deficit. 13To make ends meet and also allow for personal accounts, our proposal includes a number of changes to reduce the ongoing cost of the program, such as raising the retirement age to 67 and then indexing it to longevity, and creating the new tier one benefit.

The burden of the transition, of course, depends on the size of the economy in future years, which is discussed below.

Are We Really Proposing a 1.52 Percent Payroll Tax Hike?

Another question voiced about our proposal is whether we really are advocating a payroll tax increase comparable to the increase in the IA plan. While the 1.52 percent payroll tax supplement in our proposal is similar in magnitude to the 1.6 percent payroll tax increase contained in the IA plan, the similarities stop there. Under our proposal, the payroll tax supplement is a means to pay off accrued liabilities and to make possible the transition to fully-funded personal accounts that ultimately comprise half the retirement program. When the transition is completed, in about 70 years, there is no continuing tax liability. In the case of the IA plan, by contrast, 1.6 percent would be established in the law as the increase in the payroll tax used to fund the individual accounts. These accounts and the related tax are permanent add-ons to Social Security, in contrast to our plan in which the accounts are a substitute for a portion of the program and the tax is transitional. We refer to the tax supplement as a "Liberty Tax" since it frees workers of a significant portion of the debt they have been passed from older generations.

We also want to stress that the size of the tax supplement is only an estimate, and it is based on the actuaries’ assumptions that the long-term rate of economic growth will be 1.4 percent to 1.5 percent annually and that our proposal will result in no savings-induced increase in the capital stock or per capita income. In fact, under the actuaries' methodology, each of the three reform plans offered in this report and, indeed, any reform is assumed to have the same effect on national saving and economic growth—precisely none. We feel strongly, however, that the reforms we propose will be very beneficial to the economy. If this is the case, or if the rate of economic growth is simply higher than now projected, the actual rate of taxation would be lower. This does not follow with the IA plan.

Putting the Borrowing into Perspective

As to the concern about proposing to issue so much new debt, particularly in the current budget climate, the first point we would make is that the "new debt" being issued would basically amount to making explicit a portion of a debt that already exists—in the form of outstanding, unfunded benefit promises—but is not officially recognized in national debt figures. The second point is that the amount of new borrowing is not large by recent historical standards and, more importantly perhaps, is ultimately dwarfed by the accumulation of assets in personal accounts. With earnings and labor force growth, net new capital investment, made possible through growing contributions to personal savings accounts, could be undertaken even during the early years of the transition when the government is issuing bonds.

Consider first the increase in annual federal borrowing from the public under the PSA plan. As evaluated by the Social Security actuaries, this borrowing is projected to rise from 1.23 percent of GDP in 1999 to a peak of 1.93 percent of GDP in 2007, and then to fall gradually to 1.0 percent of GDP in 2020 and to zero by 2030. Beyond 2030, Social Security improves the overall federal budget relative to present law. This borrowing has two components: first, direct borrowing specified in the PSA plan to help fund the transition and, second, what might be called indirect borrowing which results from the loss of projected Social Security surplus funds and the more rapid draw-down of trust fund reserves than under present law. While the former debt and its draw on private resources are clearly attributable to and specified in the PSA plan, the latter is assumed to result, meaning the federal government is assumed to take no offsetting actions in the rest of the budget. Certainly, the government could respond to the loss of surplus funds and the need to redeem trust fund reserves in the early years of the reform by cutting other federal spending or raising other taxes, in which case the indirect debt would be reduced or eliminated.

The indirect borrowing assumed to result from the PSA plan is heaviest in the first few years of implementation, at 1.2 to 1.3 percent of GDP, but drops quickly, falling by half by 2005 and to zero by 2016. The direct borrowing from the PSA plan is never as large, ranging from 0.6 to 0.8 percent of GDP in the period 2002 to 2009, and falling gradually to 0.29 percent in 2015 and to zero in 2034.

Some sense can be made of these figures by noting that the PSA plan envisions moving back toward pay-as-you-go financing of the government program. As a result, trust fund assets are redeemed for cash only until 2015, the point at which pay-as-you-go financing has been restored and reserves are maintained at 100 to 150 percent of annual outgo thereafter. By contrast, under the assumptions used in this report, the current system begins running cash flow deficits in 2015 and trust fund reserves are drawn down steadily thereafter, until completely exhausted. Thus, beginning in 2016, the financial condition of the trust funds begins improving under the PSA plan. Direct borrowing to help finance the transition ceases in 2034, and repayment of this borrowing begins in 2035. The overall impact of the PSA plan, including direct and indirect borrowing, on the unified federal budget is projected to be positive relative to present law by 2030, with the improvement in the federal budget balance growing thereafter. 

In relation to annual contributions to the PSA accumulations, the amount of new borrowing is at its peak in the first five years, amounting to 69 percent of PSA contributions in 1999. The increment to federal borrowing relative to PSA contributions falls to 44 percent in 2010 and to zero in 2017. Recognizing that the government may offset some or all of the indirect borrowing through spending reductions or revenue increases in the rest of the budget, the lower figures for direct borrowing are also relevant. In this case, new borrowing peaks at 43 percent of PSA contributions in 2003, then falls to 21 percent in 2010, to 13 percent in 2020, and to 4 percent in 2030.

Figure 1 shows, in constant 1995 dollars, the total amount of federal borrowing relative to total PSA accumulations under the proposal. Again, we distinguish between direct borrowing and that stemming from the loss of projected trust fund monies that, in the first analysis, is assumed to increase federal borrowing. As shown, total PSA accumulations rise quickly from about $1.0 trillion in 2003 to $4.2 trillion in 2015 and to $8.0 trillion in 2025 in constant 1995 dollars. Total borrowing peaks at around $2.4 trillion in the period 2019 to 2024. In 2020, total borrowing equals 40.0 percent of PSA accumulations, but is declining at a rate of about 2 percent per year relative to aggregate PSA balances—i.e., by 2025 it will be down to 29.3 percent of PSA balances. Total direct borrowing reaches $1.9 trillion in 2034, and, with interest continuing to accumulate, peaks at $2.1 trillion in 2039, or about 15 percent of total PSA accumulations in that year.

Figure 1 

PSA Balances, Total Federal Transition Borrowing, and Direct OASDI Borrowing from the Treasury in 1995 Dollars 

Considering recent historical experience, between the end of fiscal year 1980 and the end of fiscal year 1994, the amount of federal debt held by the public rose from $1.3 trillion to $3.5 trillion in constant 1995 dollars, for an increase of $2.2 trillion in constant 1995 dollars. The increase in debt under our proposal, accumulated over 35 to 40 years, is about equal to the amount of federal debt absorbed by financial markets in a period less than half as long. By the standards of recent historical experience, that is, the incremental growth of the federal debt would not be large.

Could financial markets absorb this increase in explicit government debt? Presumably there would be some upward pressure on the interest rates on government securities relative to private securities, which could affect the cost of servicing the debt and the net returns to personal accounts. Given recent historical experience, however, together with our general understanding of the relationship between interest rates and the size of the national debt (which, empirically, is not strongly positive), we conclude that these effects would not be large. We also presume that the Federal government would recognize the desirability of tightening its belt in the rest of the budget, ideally moving toward a surplus position in the next century, in anticipation of major structural reform of Social Security.

Furthermore, if the non-Social Security portion of the federal budget were brought into balance over the period during which the total value of new federal debt generated by the PSA plan, relative to present law, was growing in real terms (the period from 1998 to 2022), the total federal debt as a percent of GDP would decline throughout the transition period. It would decline in absolute terms beyond 2028. In developing this estimate, shown in Figure 2, we assumed that our national leaders accomplish their stated goal of balancing the budget after 2002. We assumed that deficits in fiscal 1996 and 1997 would equal $120 billion and then decline at a rate of $20 billion per year until achieving balance thereafter. Figure 2 shows estimated federal debt that would be held by the public if the federal budget were balanced after 2002 except for the debt associated with a transition to the PSA plan. Over the full transition period, the federal debt claim on the economy would shrink under the PSA transition plan, unless the efforts to otherwise balance the budget fail.

No One Right Way to Fund the Transition

Although we believe the financing mechanism we offer has real merit, we recognize that there is no single "right way" to fund the transition to a system of personal accounts. Transition costs can be met in any number of ways—through reductions in federal spending, through increases in the payroll tax or other federal taxes, through the sale of federal assets, through new explicit borrowing from the public, or through some combination of the above.

Figure 2 

PSA Balances and Estimated Federal Debt Held by the Public under the PSA Proposal with Transition Borrowing as a Percent of GDP from 1998 to 2030

Source: Office of the Actuary, Social Security Administration, 1996 and calculations by the authors as described in the text.

As noted, under the Kerrey-Simpson bill, transition costs are met by scaling back the growth of future Social Security benefits. This imposes relatively more of the cost of transition on current older generations. Our proposal, by contrast, finances the transition outside of the Social Security system, accommodating much larger personal accounts and spreading the burden to future generations.

Chile provides an example of yet another approach. In moving toward full-scale personal accounts funded with 10 percent of workers' wages, effective in 1981, it took a multi-pronged approach: through spending reductions and tax increases in the 1970s and early 1980s, the central government ran budget surpluses in anticipation of the social security reforms; benefit commitments under the old system were trimmed somewhat as indexing was modified and the retirement age was increased and part of the payroll tax continued to be used to meet outstanding liabilities; and formal government bonds, known as "recognition bonds," are to be deposited in workers’ accounts at retirement, reflecting part of the system's implicit debt—that part equal roughly to the unfunded past service liabilities for workers who chose to switch to the new system. These bonds, which carry a real interest rate of 4 percent annually, are redeemed with revenues from the general fund of the Treasury at the time workers retire.

How a nation decides to finance the transition is important. It determines how the burden is distributed across generations and income classes and what the net economic benefits of reform are likely to be during the transition. Which method is chosen, however, is less important than the fact that, on a going forward basis, the economic gains to a fully-funded, privately managed system of personal accounts are likely to be very large.

V. Our Response to Criticisms of the PSA Plan

The PSA proposal has evoked a number of criticisms by members of the Advisory Council advocating other proposals and by other public policy analysts. We believe that few of these criticisms have merit, and that those that do can be dealt with through plan design or regulation. Concerns have also been raised which we hope to allay by further clarification.

Workers Can Manage Their Own PSA Investments

One concern that has been raised about the PSA plan is that many Americans are "unsophisticated" about making investment decisions and that a plan based on personal choice is either impractical or must be accompanied by some type of guarantees that would ensure workers a minimum rate of return or a minimum fund accumulation.

When evaluating our proposal for personal accounts, we believe it is crucial to appreciate the enormous strides that have been made in the last couple of decades in financial markets and in individuals' participation in them. With the introduction of IRAs, 401(k) plans, and other self-directed investment vehicles, millions of workers and retirees have gained an enormous amount of experience with making investment decisions. In addition, with the explosion of mutual funds and, in particular, equity index funds, ordinary working men and women do not need to "play the market" -- incurring large transactions costs and exposing themselves to excessive risk -- in order to reap the benefits of stock market participation. And, no doubt owing to the tremendous competition for new customers and new funds, there is a wealth of financial information available about alternative investment strategies and institutions, and performance ratings are widely available. In our view, workers have never been better positioned to make sound financial decisions.

Admittedly, there are people in our society with little knowledge or understanding about how to invest money. While no one has attempted to determine who these people are or why they are ill-informed, we believe that among workers, the problem is concentrated among those who have never had any money to invest. Why would people take the time to learn about something that has no practical value for them, knowledge that would likely only frustrate them? Under the PSA plan, with a portion of their Social Security taxes rebated for investment in accounts they own, workers at all income levels would have keen incentives to make sound financial decisions, either by acquiring the needed expertise or seeking out those who have it.

This is not to say that with personal accounts everyone will make the best financial decisions, reaping the best possible rates of return. Some workers will take on too much risk; some workers will not take enough. We were in general agreement, however, that workers would fare better, ex ante, under this option than under the present inadequately-financed system, a shored up pay-as-you-go system, or either of the other two options developed by the Advisory Council. 14

In our deliberations about how to structure personal accounts, we were presented with no hard evidence that workers did not—or could not, with experience—make sound financial decisions. Good decisions appear to come with education and information, and with experience and learning—all of which would be gained rapidly under a system in which workers were making regular contributions to personal accounts offered by competing financial institutions. The investment decisions required of people would not be like other major investment decisions that are made very infrequently. For example, for most people purchasing a new home is a relatively rare event so little experience is gained that might improve future decision making about home purchases. Another kind of decision might be deciding whether or not to undergo some major surgery that involves an element of urgency that precludes the acquisition of appropriate information. For workers investing PSA balances, market returns on their accounts would provide steady information on investment performance. The relative success of competing financial institutions would provide valuable information on alternative investment options.

Our proposal contains only one proviso: that personal accounts be invested in regulated financial instruments widely available in financial markets. Some supporters of the PSA proposal would prefer to see limited restrictions placed on certain categories of investments that would not be permitted in the early years of the PSA program while the general public becomes more educated about financial risk and return. However, none of us want to restrict workers ability to invest in a highly diversified set of financial instruments widely available in the marketplace. 15While we recognize the government's—i.e., the taxpayers'—potential interest in limiting excessive risk taking, there was no consensus about the kinds of restrictions that might be needed or be found cost-effective. And indeed, the concern expressed most frequently, in financial news and other coverage of retirement income planning issues, is that workers do not take enough risk. We also recognize the possibility that with some investment options offered by some institutions, administrative fees could be high in relation to investment returns. Proposals to cap the fees that could be charged and to require special licensing of financial institutions handling personal accounts, however, were without general support.

The problem here, as in so many areas of government regulation, is making sure that there is a problem worthy of federal intervention, that there is a regulatory solution well-tailored to the problem, and that the regulations are likely to result in net economic gains. For example, we were well aware of the concern that workers err on the side of taking too little risk in their investments for retirement and thus may not generate adequate retirement incomes. However, we were presented with evidence, based on experience with a sample of 401(k) plans, suggesting that this concern may be overstated because of the failure to disaggregate the data on the basis of workers' ages. The data suggested that asset allocation decisions are appropriately related to age: the older the worker, the smaller the share of assets allocated to equity and the larger the share allocated to fixed-income investments. The relatively high overall share of assets in fixed-income investments in these plans where workers manage the investment of their own retirement money derives, at least in part, from the fact that older workers tend to have accumulated larger balances than younger ones.

More generally, in considering the population as a whole and the kinds of regulations that might be appropriate, it is unclear what benchmark one would use to determine whether workers were taking too much or too little risk. Certainly, the "right" way to allocate investments depends not only on one's age but also on the size and risk-return profile of non-pension assets, among other factors.

In general, we envision a regulatory environment consistent with a wide range of choices for workers—for example, a range of options comparable to that now available to workers through 401(k) plans—offered by a wide array of financial institutions competing for workers' PSA balances. We felt that it was beyond the scope of our charter to develop a full-blown regulatory framework. We do feel, however, that the experience with 401(k) and other self-directed retirement savings plans indicates that a regulatory environment can be structured to give workers a wide range of options with which to meet their own needs and desires. We also believe that concerns about "unsophisticated investors" can be rectified most effectively by an educational effort, not by significantly restricting investment choices or by substituting government decisions for individual decisions.

We believe that new financial institutions and arrangements will continue to emerge, as they have in the past. One logical development would be group purchasing arrangements through employers, trade associations, and other such places where people cluster together during their working lives. In cases where employers are already offering 401(k) investment options, it would be quite simple for the same investment vendors to simply offer their funds for PSA investments.

As to the question of whether there would be any guarantees of a "reasonable" return on workers' investments, the short answer is no—and neither are there any such guarantees under the present system. Many middle-aged and younger workers are projected to earn negative rates of return on their Social Security taxes—and this is before factoring in the political risk that future benefits for middle- and high-wage workers might be scaled back 25 percent or more; that the cost-of-living adjustment might be capped, reduced, or eliminated altogether, even if only temporarily; that benefit taxation may be increased appreciably; or that benefits may one day be means-tested. One of the features we find most appealing about personal accounts is that workers would own their accounts, and the retirement savings they embody, and are thereby exposed to much less political risk than under the present system—political risks that, over the next 20, 30, or 40 years could easily dwarf the financial risks of a well-diversified portfolio.

Having said this, when discussing the adequacy of benefits, the first tier of our proposal cannot be ignored. Fully half of the Social Security retirement program financing would continue to be accomplished through the central defined benefit system. The first tier embodies a high degree of redistribution from high to low wage workers—indeed, two workers with identical years of work, one of whom earns the minimum wage and the other of whom earns four times that, and pays four times the taxes, would get the same benefit. This redistribution is not hidden in complex benefit formulae and eligibility criteria, but is a straightforward result of moving toward a flat benefit for full-career workers, prorated only for years of work. As under present law, this benefit is fully cost-of-living adjusted.

The first tier benefit is designed to ensure that, together with second tier accumulations, all full-career workers, regardless of income level, can expect to receive a minimally adequate retirement income from Social Security. By minimally adequate we mean enough so that even low-wage workers—and even workers who invest in relatively low-yielding assets—should not have to resort to means-tested poverty assistance. Our expectation is that workers will do considerably better than this.

Of course, tier-one benefits are no more "guaranteed" than are Social Security benefits today. They may be more secure, however, by virtue of the fact that they can be financed at a significantly lower projected payroll tax on workers—7.4 percent versus 12.4 percent.

Annuitization of PSA Balances at Retirement Should Not Be Mandatory

Another concern we have heard expressed is that workers would not be required to annuitize their account balances at retirement. The fear is that workers would have so much freedom to decide how to use their PSA accumulations at retirement that they may withdraw their entire balance, spend it, and end up penniless—or worse, a burden on public assistance programs. A less extreme version of this concern is that workers will simply spend their accumulated savings too quickly and then be unable to maintain their consumption throughout their retirement years. The answer, some say, is to force workers to purchase an annuity at retirement, spreading their payments out evenly over their retirement years. Under our plan, workers would be free to purchase an annuity with some or all of their accumulated savings, if they so wished, but they would not be required to do so.

While we question the premise that large numbers of workers would make decisions that would leave them worse off in a lifetime sense, we are aware of the risk of over-consumption by low-wage workers due to the array of federal safety net programs that tie eligibility to low assets. We considered several options for mitigating this potential problem. For example, if annuitization is seen as the only way to assure that retirees do not exhaust their resources prematurely, it might make sense to require retirees to show that they have annuity incomes from Social Security and private savings that would be equal to the poverty level income or some multiple such as 1.5 times the poverty line—over their life expectancy. For those who did not have such annuity incomes, it would be possible to require that they purchase an annuity with a portion of their PSA accumulation up to the target level. In addition, or alternatively, full redemption of PSA balances in a single withdrawal could be restricted.

One of the issues we grappled with in our own deliberations on whether to require annuitization was how to determine the appropriate target level of such annuitization. Under current law, Social Security benefits are only about 85 percent of the federal poverty line for the hypothetical "low-wage" worker retiring at age 65 in 1996. The flat benefit alone under the PSA plan for such a worker would grow to that level by 2020. Low-wage workers retiring in 2020, however, will still be getting three-fourths of their benefit determined under the current formula under the transition that we propose. By the time the full transition is completed, the flat benefit for normal retirement will exceed the poverty line.

While we understand the desire to restrict people from spending down their assets in order to qualify for means-tested transfer programs, we believe that forcing people to convert their PSA balances into annuity incomes beyond levels that protect against such behavior goes beyond the reasonable province of the public interest. Mandatory annuitization forces individual to purchase annuities upon retirement, regardless of the current market price for annuities. Annuities are highly sensitive to interest rates and, depending on one’s retirement date, the interest rate at the time of purchase will determine the size of the annuity payments. Optional annuitization, on the other hand, would allow individuals to determine when, if ever, they wish to annuitize all or part of their benefits. Retirees who want to want to live a more thrifty lifestyle than the general public might deem appropriate, in order to preserve some of their PSA accumulations to pass on to their heirs, should be able to do so without the interference of the government. In addition, forced annuitization discriminates against populations with low life expectancies, such as certain minority groups, depriving them of the right to pass along a portion of their Social Security wealth to their heirs. There is also some concern that between Social Security and Medicare, the nation's elderly have too much of their wealth annuitized, leading to higher levels of consumption and smaller bequests than they would deem desirable.

In the end, we concluded that the tier-one benefit provided a significant degree of forced annuitization, and that any further requirement was unnecessary.

Benefits in Retirement Should be Taxed Appropriately

Proponents of the MB plan have asserted in their supplementary statement that our proposed tax treatment of benefits is without parallel in the federal tax treatment of retirement benefits. While this is technically correct, it is substantively wrong. In order to explain the rationale behind our proposal, it is necessary to explain how the income tax treats various forms of retirement savings. Throughout this discussion, we assume that workers save some portion of their earnings during their working career to help provide for their retirement needs. To keep the example as simple as possible, we are only going to track the effects of alternative tax treatments on a single year’s savings. If we tracked the effects on multiple year’s savings, it would complicate the analysis, but have no real bearing on any of the conclusions.

If a worker saves through a regular savings account, the contributions to the account are made with earnings that have already been taxed. For example, assume that a worker has a marginal federal income tax rate of 28 percent and that, in preparing for her retirement, she wants to save the after-tax income from $2,000 of earnings, or $1,440 — i.e., $2,000 x (1 - 0.28). For simplicity, assume that the savings are deposited 

Table 1

Relative Value of Money in a Regular Savings Account Paying an Interest Rate Equivalent to a 5 Percent Inflation Rate and Subject to a 28 Percent Tax Rate 

at the beginning of the year in which the saving is done in an account that earns 5 percent interest annually, and further assume that the rate of inflation in the economy is 5 percent per year—i.e., the real rate of interest is zero under these assumptions. In the first year, the 5 percent interest rate would generate $72.00 in interest on the savings. Since this is interest income in a regular savings account, however, it would be taxed at the worker’s 28 percent marginal tax rate, leaving $51.84 cents in net interest after taxes. Assume that the original deposit and the net interest income are left to accumulate in the account until the worker retires. Table 1 shows the practical effects of the income tax treatment of regular retirement savings under the set of assumptions used here.

Column 2 in the table shows how the original $2,000 in earnings would have to increase in nominal terms to maintain constant purchasing power over time. Column 3 shows the worker’s accumulated savings balance in nominal terms. The accumulated value of these savings as a percent of the real purchasing power of the original earnings is shown in Column 6, which is Column 3 divided by Column 2. The net effects of the tax system on regular savings is shown in Column 7, which is 100 percent minus the value in Column 6.

Referring to the last column in the table, the picture is clear: the federal income tax system places an ever higher claim on the earnings of workers who defer consumption from their working years to their retirement years if that consumption is deferred through a regular savings account. Recognizing that it is desirable for workers to save for their retirement so they might have a reasonable standard of living in their old age, the federal tax code has made provision for taxing retirement savings less punitively than it taxes regular savings. It does so by allowing tax-preferred saving through tax-qualified retirement programs sponsored by employers and through IRAs.

Table 2 shows how benefits accumulate in a tax-preferred savings account, such as a 401(k) plan or an IRA, relative to how they accrue in a regular savings account, using the same assumptions as in Table 1. Under the tax-preferred account the earnings are not taxed at the point they are earned so the full $2,000 of earnings goes into the account and accumulates interest until withdrawn. Assuming the total accumulation is withdrawn at once, the original savings plus accumulated interest is taxed at the assumed 28 percent tax rate.

Column 2 in Table 2 corresponds with Column 3 in Table 1, and shows the nominal value of accumulated savings in a regular savings account net of federal income taxes. Column 3 in Table 2 shows the annual tax liability on the original earnings and the tax liability on interest income each year it is earned. The tax liability in the year the wages are earned is the difference in original earnings ($2,000) in Table 1 and the net amount actually invested ($1,440). The tax liability on interest income shown in Table 2 is the difference in the gross interest and net interest in Table 1. Column 4 in Table 2 shows the nominal tax paid in each year stated in dollar terms at the end of the 10 years, and is the product of the nominal tax liability from Column 3 multiplied by the equivalent dollar multiplier in Column 7. For example, the $560 in taxes paid on the original earnings would have a nominal value of $912.18 stated in current dollars 10 years later. This is equal to $560 times 1.62889 (1.63 rounded) or $912.18.

The table shows that under the tax-preferred account the disposable income from the original savings plus the net interest it has earned is 14.4 percent higher than for the regular savings account—i.e., $2,345.61 versus $2,050.97. The effective tax rate on the original earnings adjusted to account for inflation is 28 percent—i.e., the $2,345.61 is 72 percent of the original $2,000 adjusted by inflation to the date the benefit is distributed. In this case, there is no tax penalty on the worker for deferring consumption from the period when wages are earned to the retirement period.

Table 2

Taxes and Benefit Accumulations under Regular Savings and Tax-Deferred Accounts at 28 Percent Tax Rat with Interest and Inflation Rates at 5 Percent Per Year 

The federal tax code treats workers’ voluntary contributions to IRAs, 401(k)s, 403(b)s, and section 457 plans on a tax-preferred basis, like the tax-preferred contributions in Table 2. It also treats employer contributions to tax-qualified plans in the same fashion, in that the employer makes contributions on a pretax basis and interest on the contributions are not taxed until benefits are distributed. The tax code, however, does not allow workers’ mandatory contributions to these plans to be made on a pretax basis. Thus, to the extent that an employer might require that workers make contributions to a defined benefit plan, the tax code requires that these contributions be made on a post-tax basis. When benefits are distributed from the plan, any benefits over and above the original worker contributions are taxable. In this regard, these kinds of plans treat workers’ contributions much more like a regular savings account than a tax-qualified account. Most employers have realized the punitive nature of this tax treatment of employee contributions to tax-qualified plans. As a result, with a small number of notable exceptions, virtually all private sector employers sponsoring tax-qualified defined-benefit plans today fund them solely with employer contributions.

In the case of tax-qualified plans, most contributions are made on a pretax basis. In the case of Social Security, employer contributions are made on a pretax basis and worker contributions are made on a post-tax basis. We were mindful in designing the PSA proposal that if employee contributions to the PSAs were made on a pretax basis it would result in reduced federal tax collections in the early years of the program. As a result, we have recommended a tax treatment of benefits that minimizes the short-term effects on federal tax revenues. Specifically, we have proposed that employee contributions be made on a post-tax basis as is presently the case with Social Security payroll taxes, but that benefits not be taxed when distributed.

Table 3 compares the tax payments and benefit distributions for the hypothetical worker considered in the development of the prior two tables under two tax regimes. The first regime is one where pretax contributions finance the benefit which is taxed at distribution. The second regime assumes that post-tax contributions finance the benefit which is not taxed at distribution. The table shows that when the amounts are stated in equivalent dollars, the results of the two cases are identical from both the government’s and the worker’s perspective.

When we considered the tax treatment of benefits under the PSA proposal in the context of this analysis, we felt that benefits that were financed with pre-tax employer contributions should be taxable at distribution. We felt that the benefits that were financed with workers’ post-tax dollars should not be taxable.

One issue that might be raised by this analysis is our underlying assumption that the rate of return on the savings in our examples is assumed to be equivalent to the general rate of inflation. If this assumption is relaxed so the rate of return on assets is higher than the general inflation rate, the tax treatment of employee contributions to the PSAs would be somewhat more favorable than the tax treatment accorded tax-qualified plans. Referring back to the analysis of internal rates of return in the "Comparison of Plans" section of this report, however, expected real rates of return under Social Security for middle- and higher-wage earners are extremely low. Only if policymakers are concerned that potential rates of return would become to high, which we find unlikely, would it make sense to allow workers to contribute to their PSAs on a pretax basis and to tax benefits. 

Table 3 

Tax Collections and Retirement Accumulations under Alternative Tax Regimes for Tax-Deferred Accounts Assuming 28 Percent Tax Rates and 5 Percent Rates of Return