Committee on Ways and Means, 1-21-99 Testimony

Statement by Alicia H. Munnell, Peter F. Drucker Professor of Management Sciences,
Boston College Carroll School of Management

Testimony Before the House Committee on Ways and Means

Hearing on Preserving and Strengthening Social Security

January 21, 1999

Mr. Chairman and Members of the Committee, I am delighted to have the opportunity to appear before you today to discuss proposals to preserve and strengthen Social Security. In my view, the best way to ensure that all Americans have an adequate basic retirement income is to maintain the current Social Security program, which pays benefits based on lifetime earnings, and not to move toward a system of personal savings accounts. Let me provide a brief summary of the reasoning behind that conclusion and, in the process, answer the three questions you raised in the notice for this hearing.

I. Social Security is not facing a crisis and does not need major reform. The projected increase in Social Security spending due to the aging of the population is neither enormous nor unprecedented.

The cost of the program is projected to rise by 2 percent of GDP. Budget changes equal to 2 percent of GDP are not uncommon; defense spending increased by 5 percent of GDP at the start of the cold war and declined by 2 percent between 1991 and 1998. The financing shortfall is manageable and does not require radical change in the program.

Much of the projected shortfall can be eliminated with good policy options. For example, extending coverage to new state and local workers, increasing the maximum taxable earnings base, and reflecting corrections to the CPI in the COLA are all consistent with the goals of the program and will help close the financing gap. Investing the trust funds in private stocks and bonds will increase the return on fund reserves and close much of the remaining gap.

II. Replacing all or part of Social Security's current defined benefit plan with personal savings accounts is risky, costly, and will hurt the disabled and women.

The whole point of having a Social Security system is to provide workers with a predictable retirement benefit. Social Security benefits are quite modest; the average worker retiring at age 62 last year got $780 per month. That modest benefit should be an amount that people can count on and to which they can add income from private pensions and other sources. It should not depend on investment decisions in a volatile stock market.
Another risk with personal savings accounts is that they are unlikely to be kept until retirement. As in the case of IRAs and 401(k) plans people will insist on access to their accounts in order to cover emergencies or to meet expenses. No matter how good the case for withdrawals prior to retirement might be, those withdrawals will mean inadequate retirement income.
A third risk with personal savings accounts is that people may outlive their savings. In contrast, Social Security provides retirees with inflation indexed annuities, so that it guarantees benefits that keep pace with inflation for as long as a person lives.
Personal savings accounts are likely to be expensive to administer. Studies show that administrative costs could well equal a 20-percent cut in benefits. Data from the U.K. and Chile, countries that have adopted personal saving accounts, suggest that the costs could be even higher. Annuitizing individual accumulations in the private market reduces benefits by another 10 percent. Social Security keeps administrative costs low by pooling investments, and low administrative costs ensure a higher net return to workers.
Disabled workers are likely to receive sharply lower benefits with personal savings accounts. They will not have time to build up adequate reserves under a system of personal saving accounts. In contrast, Social Security provides full benefits for disabled workers.
Finally, women have the most to lose from moving to a system of personal savings accounts. To the extent that personal accounts are substituted for Social Security, they would lose the progressive benefit formula that provides proportionately higher benefits for low earners than for high earners; women are more likely to be low earners. They would lose spouses' and widows' benefits, which help support women who spend time out of the labor force taking care of their families. They would lose the guarantee of inflation indexed monthly benefits for life, which is particularly valuable to women who on average live longer than men. Personal savings accounts would put these protections at risk.

III. Instead of personal savings accounts, we can increase national savings and broaden investment options for workers -- changes that have been used to justify personal savings accounts -- within the structure of the current program.

The federal government can accumulate reserves. The non-Social-Security portion of the budget is headed for balance in 2002, probably sooner. We can keep it there and build up reserves in the Social Security trust funds. The states do it for their pension funds; the federal government should be able to do it for its major retirement system.

Investing a portion of the Social Security trust funds in private stocks and bonds is both desirable and feasible. We know how to prevent interference in private sector activity: set up an independent investment board, invest in a broad index, and delegate voting rights to independent fund managers.

In short, there is no reason to move toward a system of personal savings accounts; we can increase national saving and improve returns within the context of the existing Social Security program. Social Security's current defined benefit arrangement, where benefits are based on lifetime earnings, is the best way to provide basic retirement income. Social Security has served us well for nearly sixty years; let's modernize its financing but keep its benefit structure in place. Social Security is not broken; it just needs fine-tuning.

I. Social Security is Not Facing a Financing Crisis

Social Security is not facing a financial crisis. Rather, the current projections show a financing gap in the long run unless remedial action is taken, as it almost certainly will be. According to the Trustees' 1998 Report (intermediate assumptions), between now and 2013 the Social Security system will bring in more tax revenues than it pays out. From 2013 to 2021, adding interest on trust fund assets to tax receipts produces enough revenues to cover benefit payments. After 2021, annual income will fall short of annual benefit payments, but the government can meet the benefit commitments by drawing down trust fund assets until the funds are exhausted in 2032. It is important to remember that the exhaustion of the trust funds does not mean the program ends in 2032, and nothing is left. Even if no tax or benefit changes were made, current payroll tax rates and benefit taxation would provide enough money to cover roughly 75 percent of benefits thereafter. It is this long-run gap between 75 and 100 percent that needs to be filled.

Over the next 75 years, Social Security's long-run deficit is projected to equal 2.19 percent of covered payroll earnings. That figure means that if the payroll tax rate were raised immediately by 2.19 percentage points--roughly 1.1 percentage point each for the employee and the employer--the government would be able to pay the current package of benefits for everyone who reaches retirement age at least through 2075. While such a tax increase is neither necessary nor desirable, it provides a useful way to gauge the size of the problem.

It is also useful to look at the program as a percent of GDP. The cost of the program is projected to rise from 4.6 percent of GDP today to 6.8 percent of GDP in 2030, where it is projected to remain. This increase is due primarily to the aging of the population. A 2-percent-of-GDP increase in Social Security costs is significant, but hardly qualifies as a "demographic time bomb."

Although Social Security's financing problems are manageable and do not require radical changes in the system, two considerations are receiving more attention today than in 1983 when Congress last passed major financing legislation. The first is the so-called "money's worth" issue. Unlike earlier generations that received large benefits relative to the taxes they paid, today's workers can expect to receive relatively low returns on their payroll tax contributions. Since raising taxes or reducing benefits will only worsen returns, almost all reform plans involve trying to increase returns through equity investment in one form or another. The second factor influencing the Social Security reform debate is increasing concern about our low levels of national saving. This concern along with the desire to avoid high pay-as-you-go tax rates in the future has led to considerable interest in some prefunding.

Almost all proposals to restore financial balance to Social Security respond to concerns about rate of return and national saving. Both proposals to maintain Social Security's existing defined benefit plan and proposals to institute personal saving accounts involve a substantial accumulation of assets. Similarly, most proposals provide that those covered by Social Security should have access to the higher returns associated with equity investment either through investments in personal savings accounts or through broadening the investment options available to the trust funds. Because it is possible to have equivalent amounts of funding in the Social Security program and in a system of personal savings accounts and because equity investment is possible in either scenario, the question comes down to which arrangement is better for people's basic retirement income.

II. Personal Savings Accounts are Risky, Costly, and Hurt the Disabled and Women

Here the economics are clear: the current Social Security program, where benefits are based on lifetime earnings, is the best way to provide the basic retirement pension. Personal savings accounts are risky, costly, and likely to hurt the disabled and women.

Personal Savings Accounts Are Risky. Personal savings accounts expose workers to three risks: market risk, the risk of using their savings before retirement, and the risk of outliving their resources.

With personal savings accounts, individuals' basic benefits would depend, at least in part, on their investment decisions. What stocks did they buy? When did they buy them? When did they sell? Uncertain outcomes may be perfectly appropriate for supplementary retirement benefits, but not for the basic guarantee. Herb Stein, Chairman of the Council of Economic Advisers under President Nixon, summarized the argument best.

"If there is no social interest in the income people have at retirement, there is no justification for the Social Security tax. If there is such an interest, there is a need for policies that will assure that the intended amount of income is always forthcoming. It is not sufficient to say that some people who are very smart or very lucky in the management of their funds will have high incomes and those who are not will have low incomes and that everything averages out."

Retirement income that depends on one's skills as an investor is not consistent with the goals of a mandatory Social Security program. Remember that Social Security is the major source of income for two-thirds of the 65-and-over population and virtually the only source for the poorest 30 percent. The dollar amounts are not very large: the benefit for a low-wage worker who retired at age 62 in 1998 was only $473 per month or $5676 per year and for a worker with a history of average wages was $780 per month or $9360 per year. Does it really make sense to put these dollar amounts at risk?

Personal saving accounts also create a very real political risk that account holders would pressure Congress for access to these accounts, albeit for worthy purposes such as medical expenses, education, or home purchase. Although most plans prohibit such withdrawals, our experience with existing IRAs and 401(k)s suggests that holding the line is unlikely. To the extent that Congress acquiesces and allows early access - no matter how worthy the purpose -- retirees will end up with inadequate retirement income.

Another risk is that individuals stand a good chance of outliving their savings, unless the money accumulated in their personal savings accounts is transformed into annuities. But few people purchase private annuities and costs are high in the private annuity market. The reason for the high costs is adverse selection: people who think that they will live for a long time purchase annuities, whereas those with, say, a serious illness keep their cash. Private insurers have to raise premiums to address the adverse selection problem, and this makes the purchase of annuities very expensive for the average person. Moreover, the private annuity market would have a hard time providing inflation adjusted benefits. In contrast, by keeping participants together and forcing them to convert their funds into annuities, Social Security avoids adverse selection and is in a good position to provide inflation-adjusted benefits.

III. Personal Savings Accounts Would Be Costly. In addition to making the basic retirement benefit dependent on one's investment skills, personal savings accounts would be costly. The 1994-96 Social Security Advisory Council estimates that the administrative costs for an IRA-type individual account would amount to 100 basis points per year. A 100-basis point annual charge sounds benign, but estimates by Peter Diamond of MIT show that it would reduce total accumulations by roughly 20 percent over a 40-year work life. That means benefits would be 20 percent lower than they would have been in the absence of the transaction costs. Moreover, while the 100-basis-point estimate includes the cost of marketing, tracking, and maintaining the account, it does not include brokerage fees. If the individual does not select an index fund, then transaction costs may be twice as high. Indeed, costs actually experienced in the United Kingdom, which has a system of personal saving accounts, have been considerably higher than the Advisory Council estimate. Finally, because these transaction costs involve a large flat charge per account, they will be considerably more burdensome for low-income participants than for those with higher incomes.

In addition to costs, a recent study by the Employee Benefit Research Institute (EBRI) has raised real questions about the ability, in anything like the near term, to administer a system of personal savings accounts in a satisfactory way. Unlike the current Social Security program that deals with the reporting of wage credits, a system of personal accounts would involve the transfer of real money. It is only reasonable that participants would care about every dollar, and therefore employer errors in account names and numbers that arise under the current program would create enormous public relations problems under a system of individual accounts.

Personal Savings Accounts Would Hurt the Disabled and Women. Most proposals that move toward personal savings accounts involve a cut in benefits for disabled workers. These proposals typically involve a reduction in Social Security benefit levels for both disabled and retirees that, in theory, will be made up through the accumulations in their personal savings accounts. Thus, projections for the various reform proposals generally show that the combined payment from the personal saving account and the reduced Social Security program equals the benefit currently promised under Social Security for the average retiree. Unlike retirees, however, disabled workers will not have time before their disability to build up any significant reserves in their personal saving account to finance a full supplementary benefit. As a result, disabled workers are likely to experience a substantial reduction in benefits.

For different reasons, personal savings accounts would also likely hurt women and low-earners generally. Although Social Security's benefit rules are gender-neutral, they are particularly helpful for women. First, the progressive benefit formula provides proportionately higher benefits for low earners than for high earners, and women are more likely to be low earners. Second, for women who spend their careers taking care of their families, Social Security provides retirement benefits equal to 50 percent of their husbands' benefits. Divorced homemakers (married least 10 years) can also get these benefits. Third, for older women whose husbands die, Social Security provides widows' benefits equal to 100 percent of their husbands' benefits. This is important because women tend to outlive their husbands. Fourth, if children are getting survivors' benefits, younger widows who stay home to care for them also receive benefits.

Even with more women in the labor force, these family benefits remain important. In 1996 the majority (63 percent) of women beneficiaries aged 62 and older were receiving wives' or widows' benefits; 37 percent had no earnings history and were entitled only as a wife or widow, and 26 percent had a higher benefit as a wife or widow than as an earner.

In addition to a progressive benefit structure and family benefits, Social Security has two other features that help women. First, Social Security pays monthly benefits for life, which is particularly valuable to women who on average live longer than men. Second, Social Security adjusts benefits annually for changes in the cost of living to protect their buying power against inflation. This protection matters more for women than for men because women live longer.

All the protections of the current program would be put at risk if reform moved toward personal savings accounts. First, unless special provisions were enacted, a woman's retirement benefit would depend -- at least in part -- on her contributions into her personal account and the earnings on those contributions. Because women tend to have lower wages and less time in the labor force, their contributions and earnings would, on average, produce low retirement benefits. Second, many of the family benefits currently provided by Social Security would likely be cut back. Third, with individual accounts the money is not automatically converted to a lifetime annuity or protected against inflation. If people get their money back in a lump sum, they could use it up before they die and leave nothing for their widow. This risk is compounded by the absence of inflation protection. In short, the present Social Security system offers a range of protections that are of great importance to women and are not duplicated by any of the proposals to privatize the system.

What then is the best approach?

III. Fund Social Security and Invest in Private Stocks and Bonds

The alternative to personal savings accounts is to accumulate reserves in the Social Security trust funds and invest part of those reserves in private stocks and bonds. This approach offers many of the advantages of personal saving accounts without the risks and costs. It has the potential to increase national saving and offers participants the higher risk/higher returns associated with equity investment. But, unlike personal saving accounts, a partially funded Social Security program with equity investments ensures predictable retirement incomes by maintaining a defined benefit structure that enables the system to spread risks across the population and over generations. In addition, pooling investments keeps transaction and reporting costs to a minimum, producing higher net returns than personal saving accounts.

Accumulating Reserves. Would it really be possible for the federal government to accumulate reserves? A key requirement for success is separating Social Security completely from the rest of the budget. To date, increasing saving through accumulations in the Social Security trust funds has produced ambiguous results. Critics contend that the existence of Social Security surpluses encourages either taxes to be lower or non-Social-Security spending to be higher than it would have been otherwise. Although little evidence exists to support this contention, a unified budget and large deficits have blurred the picture to date. But the fiscal outlook is changing; the unified budget is in surplus and the Congressional Budget Office projects that the non-Social-Security portion of the budget will be balanced by 2002, if not sooner.

Revising the presentation of government accounts to separate Social Security completely from the rest of the budget also would clarify the extent to which the system is adding to national capital accumulation. Technically, the Social Security Amendments of 1983 already have placed the Social Security trust funds "off-budget." This legislation reversed the reliance on the concept of the unified budget first used by Lyndon Johnson in FY1969. The difficulty is that, while Social Security is exempt from most enforcement procedures, budget targets are always stated in terms of the unified budget and the budget numbers reported by the Administration, Congress, and the press always include the balances in the trust funds. Thus, separating Social Security from the rest of the budget requires changing culture more than changing legal requirements.

Is it realistic to evaluate the budget without Social Security? Comparisons of the federal government with the states are always tricky, but states have been successful in this endeavor. They accumulate reserves to fund their pension obligations but generally present their budgets excluding the retirement systems. Their non-retirement budget balance has remained positive, while annual surpluses in their retirement funds have been hovering recently around 1 percent of GDP. Thus, states are clearly adding to national saving through the accumulation of pension reserves. With a commitment to balance the non-Social-Security portion of the budget, the same should be achievable at the federal level.

Investing in Equities. Equity investment for Social Security is also a feasible option, and a partially funded Social Security program with private stocks and bonds is the realistic alternative to personal saving accounts. Everyone involved in the debate recognizes that having the federal government in the business of picking winners and losers and voting on corporate proposals is undesirable. Thus, it is essential to establish mechanisms to ensure that the government does not interfere in private sector decisions, and we know how to do that.

For example, trust fund equity investments would be indexed to a broad market average, and the goal of investment neutrality be established in law. An expert investment board, similar to the Federal Retirement Thrift Investment Board that administers the Thrift Savings Plan for federal employees, would be responsible for selecting a broad market index, such as the Russell 3000 or the Wilshire 5000, for trust fund investments. This board would also be responsible for choosing, through competitive bidding, several portfolio managers to manage the accounts, and for monitoring the performance of these managers. To ensure that government ownership does not disrupt corporate governance, the investment board would be required to delegate voting on proxy issues to the individual portfolio managers. Caps on the holdings in any individual company can be introduced to ensure that Social Security does not disrupt financial markets. Moreover, the investment in stocks would occur gradually over a 10- or 15-year period.

Even though equity investment by Social Security would not disrupt the markets, some critics still worry that it could have a substantial effect on relative rates of return, perhaps driving up government borrowing costs. The portfolio restructuring would be expected to have some effect on relative returns. The equity premium would decline to reflect the increased efficiency of risk bearing in the economy. Some movement would also be expected in interest rates. One study that has estimated the effect on relative returns concluded that the shift to equities in the trust funds would lower the equity premium by 10 basis points and raise the interest on Treasury securities by roughly the same amount (Bohn 1998). With current levels of federal debt, this increase in Treasury rates should have a relatively small effect on the federal budget. As the economy grows and the debt declines, the effect should be negligible.

While Social Security investment in equities is unlikely to disrupt financial markets or cause major shifts in rates of return, many people are concerned that Social Security investment in equities could lead to government interference with the allocation of capital in the economy and with corporate activity.

In the Social Security debate, both supporters and opponents of trust fund investment in equities point to the performance of public pension funds to argue their case. Supporters cite the success of federal plans, particularly the federal Thrift Savings Plan (TSP). The TSP has established a highly efficient stock index fund and has steered clear of any issues of social investing. TSP designers insulated investment decisions by setting up an independent investment board, narrowing investment choices, and requiring strict fiduciary duties. The TSP also operates in a political culture of noninterference. Its creators made clear from the beginning that economic, not social or political, goals were to be the sole purpose of the investment board. The TSP has perpetuated this norm by refusing to yield to early pressure to invest in "economically targeted investments" or to avoid companies doing business in South Africa or Northern Ireland. It has avoided government interference with private corporations by pushing proxy decisions down to independent portfolio managers.

Opponents of trust fund investment in equities point to state and local pension funds. They contend that state and local pensions often undertake investments that achieve political or social goals, divest stocks to demonstrate that they do not support some perceived immoral or unethical behavior, and interfere with corporate activity by voting proxies and other activities. Opponents charge that if the investment options are broadened at the federal level, Congress is likely to use trust fund money for similar unproductive activities.

My view is that the social investing activity of state and local pension plans has been grossly exaggerated, and that any such activity would be even less likely to occur at the federal level. For example, using a very comprehensive definition, a 1993 study for Goldman Sachs reported that economically targeted investment totaled less than 2 percent of total state and local pension fund holdings. Similarly, most of the divestiture activity, which centered on firms doing business in South Africa, ended in 1994. Proxy voting activities would not occur at all in the case of Social Security, since all advocates support the notion of delegating voting to the independent pension fund managers.

In short, a partially funded defined benefit plan with equity investment is feasible and can do everything that privatized accounts can do but at lower costs, thus yielding higher net returns. A recent GAO report did not identify any insurmountable hurdles with direct trust fund investment in equities. Canada should provide some good information about the feasibility of this type of equity investment since is in the process of setting up a board that will oversee the investment of its Social Security trust funds in equities.

IV. Conclusion

Let me conclude. Most plans being discussed today involve both prefunding and equity investment. In economic terms, the goals of prefunding and broadening the portfolio can be achieved either within the context of Social Security's defined benefit program or in personal saving accounts. The question becomes which is the best benefit structure for people's basic retirement income. Here the economics are clear. A defined benefit plan allows for better risk spreading, lower costs, and better protection for disabled workers and women than personal saving accounts.

Once balance is restored to the existing program, it is possible to consider changes that would improve the likelihood that future retirees will have adequate incomes. One option is to introduce voluntary supplemental personal saving accounts coordinated with Social Security for those who would like to set aside more money. Thus, the debate is not about whether personal saving accounts are good or bad in general. With people assured basic retirement protection, personal saving accounts may be a perfectly reasonable addition. What opponents of personal saving accounts object to in the context of Social Security reform is cutting back on existing basic Social Security benefits and replacing those benefits with a risky and costly alternative which leaves a lot of people behind. Introducing personal saving accounts as an add-on to Social Security is a good idea; substituting personal saving accounts for existing Social Security benefits, which needlessly undermines long-standing basic protections, is a bad idea.

REFERENCES

Advisory Council on Social Security. 1997. Report of the 1994-1996 Advisory Council on Social Security (Washington: Government Printing Office).

Bohn, Henning. 1998. "Social Security Reform and Financial Markets "Social Security Reform and Financial Markets," in Steven Sass and Robert Triest eds. Social Security Reform: Links to Saving, Investment, and Growth (Boston, MA: Federal Reserve Bank of Boston).

Diamond, Peter A.. 1997. "Macroeconomic Aspects of Social Security Reform," Brookings Papers on Economic Activity, 2.

Diamond, Peter A..1998. "Economics of Social Security Reform: An Overview," in A. Douglas Arnold, Michael Graetz, and Alicia H. Munnell eds. Framing the Social Security Debate: Values, Politics and Economics (Washington, D.C.: National Academy of Social Insurance and the Brookings Institution).

Employee Benefits Research Institute. 1998. "Beyond Ideology: Are Individual Social Security Accounts Feasible?" EBRI-ERF Policy Forum, December 2.

Hammond, P. Brett and Mark J. Warshawsky, "Investing in Social Security Funds in Stocks," Benefits Quarterly, Third Quarter 1997, 52-65.

Munnell, Alicia H. and Pierluigi Balduzzi. 1998. "Investing the Trust Funds in Equities" (Washington, D.C.: Public Policy Institute, American Association of Retired Persons).

Stein, Herbert. 1997. "Social Security and the Single Investor" Wall Street Journal (February 5,1997)

United States General Accounting Office. 1998. Social Security Investing: Implications of Government Stock Investing for the Trust Fund, the Federal Budget, and the Economy (Washington, D.C.: Government Printing Office)