Social Security Reform 

Henry J. Aaron

Mr. Chairman:

It is an honor to appear once again before this committee. The manner in which the President, Congress, and this committee, in particular, have dealt with the issue of Social Security reform deserves the highest praise. The issue is complex, politically sensitive, and easily demagoged. Maintaining civility and raising the level of public understanding is both difficult and critically important to the successful outcome of this debate. That is what you have done in previous hearings and what you are continuing to do now.

The president initiated this process by highlighting the importance of Social Security reform and by urging forbearance in using projected budget surpluses to cut taxes or raise spending until the projected long-term deficit in Social Security has been closed. He also summoned the nation to a year of debate and self-education, a prerequisite to action on an issue as complex and potentially divisive as Social Security reform. Members of Congress have been holding meetings around the nation. This and other committees of Congress have held numerous hearings. These are precisely the sorts of steps that had to be taken to give reform its best chance of success.

This process has corrected certain misconceptions. Perhaps the most notable such misunderstanding was that Social Security is in crisis. After a year of examining the facts, most people now understand that Social Security faces a long-run financing problem and that addressing this problem early is highly desirable. But Social Security surpluses are now running at about $100 billion a year and are projected to grow for several years. Revenues and reserves are sufficient to pay all currently promised benefits for the next generation. Most people now understand these facts and realize that reference to the long-run financing problem as a “crisis” is an Orwellian distortion of language. It is a “problem” and one that is best dealt with now. By acting promptly, we can introduce any changes gradually and give workers ample time to adjust to those changes before they reach retirement age. Delay will narrow our options. It is also appropriate to reexamine the structure of a system that is six decades old and that was drafted in a nation quite different from contemporary America. But we should address this task soberly and without a sense of panic.

Despite the educational efforts of the President, Congress, and this committee, a series of myths has persisted. If these myths continue to survive, public understanding will be hindered and a sensible discussion of options for reform will be rendered even more difficult than it would otherwise be. I shall focus on four such myths:

  • Myth 1 Privatization of Social Security would raise the rate of return pensioners receive. This proposition is the reverse of the truth. Privatization would lower the rate of return that pensioners receive.
  • Myth 2 Privatizing Social Security would increase economic efficiency, improving the allocation of labor and other resources. This proposition is false. Whether privatization of Social Security would raise or lower economic efficiency is unclear, but in any event, the effects would be small and would occur only at the price of increased income inequality.
  • Myth 3 It is possible to privatize Social Security and raise benefits without raising taxes, cutting government spending, or increasing the national debt. This “we-can-make-bricks-without-straw” approach to privatizing Social Security rests on false claims. If pension benefits are to be increased, with or without privatizing Social Security, it will be necessary to choose one or more from the unpalatable list of raising taxes, cutting benefits, or increasing the national debt.
  • Myth 4 The Social Security trust fund is a fraud, a collection of meaningless IOUs. This often repeated statement is also false. The fact that Social Security has collected more in taxes than it has paid in benefits means that future workers will have to pay less in payroll taxes to support future pensions than would have been required if these reserves had not been accumulated.

In my testimony, I shall explain in more detail why each of these four widely held beliefs is false. Then I shall show how some of these myths and inaccurate claims stand behind the claim that the county of Galveston, by withdrawing from Social Security, has produced higher benefits at less cost to its employees and at no cost to citizens of the rest of the nation. In fact, taxes are higher, investment returns have been lower, some benefits would be higher but most would be lower, benefits are less secure, and sizeable costs have been imposed on the rest of the nation.

Myth 1

—Privatization will raise the rate of return on taxes paid.
Discussions of individual accounts suffer from a confusion among three separate issues:

  • whether to build reserves to back up future pension obligations (so-called “prefunding”);
  • whether reserves should be invested in a diversified portfolio; and
  • whether to shift management of retirement accounts to the private sector;

Chart 1
Differentiating Privatization, Prefunding, and Diversification Privatization?

  • Current System
  • Diversification?

    No:  Current System
    Yes:  Borrow, invest proceeds in equities for Trust fund
  • Create individual accounts
  • Issue recognition bonds
  • Perpetually roll over
    principal and enough interest to keep debt same as that of unfunded liability under current system
  • Diversification?

    No:  Require individual accounts
    to hold bonds
    Yes:  Permit individual accounts

  • Raise taxes/cut benefits
    to decrease unfunded
  • Diversification?

    No:  Invest trust fund
    in bonds
    Yes:  Invest trust fund
    in equities
  • Create individual accounts
  • Issue recognition bonds
  • Raise taxes/cut benefits to make path of debt lower than path of unfunded liability under current system
  • Diversification?

    No:  Require individual accounts
    to hold bonds
    Yes:  Permit individual accounts

    Source: John Geanakoplos, Olivia S. Mitchell, and Stephen P. Zeldes, “Would a Privatized Social Security System Really Pay a Higher Rate of Return?” in Framing the Social Security Debate: Values, Politics, and Economics, edited by R. Douglas Arnold, Michael J. Graetz, and Alicia Munnell, National Academy of Social Insurance, 1998, p. 140.

    To explain this confusion, I ask you to focus on Chart 1 from a study prepared by John Geanakoplos, Olivia S. Mitchell, and Stephen P. Zeldes for the recent National Academy of Social Insurance study of privatization. As this chart makes clear, we can answer “yes” or “no” to each of these three questions. Under the current system, the answers are “no” to all three, with the qualification that current reserve accumulation is partial, albeit temporary, funding. The practical question is how the rate of return will change with different answers to these three questions.


    Funding. Increased funding to raise pension reserves is possible only with some combination of additional tax revenues, reduced benefits, or increased investment returns from investing in higher yielding assets. Increased funding has two effects. It reduces the taxes that workers will have to pay in the future to support given future pension payouts, because the enlarged trust fund yields more income. And it raises national saving, provided that any induced reductions in private saving or increased government borrowing on operations of government other than Social Security do not fully offset the additions to reserves. Note that these effects are independent of privatization.

    Diversification. Any given reserve can be invested in low-yielding assets, such as government bonds, or in a diversified portfolio that includes higher yielding corporate bonds and stocks. A shift from a “government-bonds-only” policy to a diversified portfolio would raise the returns to pensioners, because the increased yields would enable any given level of taxes to support higher pension benefits.

    How reserves are invested does not directly affect national saving or national income. But it does determine how the returns from the added saving represented by the Trust Fund are distributed among worker-pensioners and others. An important reason for investing Social Security reserves in a diversified portfolio is that returns from such a portfolio more closely match the total economic returns from the added saving that reserve accumulation represents. Since most Social Security pensioners derive most or all of their income during retirement or disability from Social Security, it is only fair to invest the reserves that support those benefits in assets that provide them with the total returns their saving has generated. To limit reserve investments to relatively low-yielding government bonds is fundamentally unfair to such pensioners, since it denies them returns that their saving has generated.

    Once again, note that it is how reserves are invested, not whether the system is privatized, that accounts for differences in the rate of return.

    Privatization. The final question concerns the management of pension accounts and pension reserves. All current proposals to privatize management of Social Security also entail a shift from a defined-benefit plan to a defined-contribution plan. While it would be possible, in principle, to shift management of a national defined-benefit plan to private companies, no one is currently proposing it. The reason, I think, is that everyone understands that such a shift would be disruptive, costly, and pointless.

    Hence, the question of whether to privatize cannot be disentangled from the question of whether to shift from a defined-benefit to a defined contribution plan. I have testified before this committee on why a defined-contribution plan is an inferior instrument for fulfilling the purpose of social insurance, to provide assured basic income. A defined-contribution plan places financial market and other risks squarely on the back of each worker individually. In contrast, a defined-benefit plan spread these risks among workers at each point in time and across generations over time. The increase in individual risk that comes with a defined-contribution plan is flatly inconsistent with the purpose of social insurance—to assure basic income.

    Whether one agrees with this social judgment or not, however, the key point is that the act of privatizing Social Security has no economic consequence whatever, apart from the increase in administrative costs that would result. Whether the increase in administrative expenses is large or small depends on the nature of the privatization plan. But all privatization plans would increase costs and, whatever the increase, it has to come out of the pensions that workers would otherwise earn. All of the potential gain to the economy or to future workers comes from increases in saving or investment of reserves in a diversified portfolio. None comes from privatization.

    Chart 2
    Social Security, 1999

    Source: 1998 Annual Report of the Board of Trustees of the Federal Old-Age and Survivors Insurance and Disability Insurance Trust Funds, April 30, 1998, pp. 102-103.

    The falsity of the proposition that privatization will boost rates of return may be seen in another way. Currently, approximately 80 percent of total Social Security revenues go to cover current expenses (see Chart 2). The other 20 percent of current revenues goes to build reserves. The percentage of revenues required to cover current costs will fall slightly in the next few years and then rise. Eventually, if we do nothing now to close the projected long-term deficit, expenditures will absorb all current revenues and reserves will have to be sold to cover benefits. There is only one way to raise the return that workers can receive on the payroll taxes levied under current law—to raise the rate of return on reserve investments. That requires investing in higher yielding assets, which can be done either through private accounts or diversified investments of the Trust Funds.

    Improving the Management of Social Security Reserves. A higher rate of return can be secured for pensioners without exposing them to the risks of individual accounts. Management of Social Security reserves should be reformed by removing it from the executive branch of government and placing it under independent, non-political, professional management. Under this new management, the investment of Social Security reserves would be diversified, with a portion of Social Security’s reserves placed in index stock equity funds, just as is are the assets of virtually every corporate and public pension fund.

    The management of Social Security reserves would earn the average return generated by common stocks, which has exceeded that on bonds by an average of several percentage points per year. If individuals invested in common stocks, they too would earn the average return on common stocks. But their net return would be reduced by the sizeable administrative costs of managing more than 140 million mostly quite small individual accounts. By comparison, the administrative costs involved in managing investment of Trust Fund reserves in equities would be miniscule. Because administrative costs would be smaller, investment of part of the trust funds in equities would yield higher returns than individual accounts, while protecting beneficiaries from the risks they would bear under a system of individual accounts.

    Under this approach, a substantial share of Social Security reserves would remain in Treasury securities, enabling the system to ride out market downturns without having to sell off substantial holdings when equity prices are low. Under individual accounts, by contrast, people who retire when the market is low or who are unlucky in their investments could suffer reduced incomes until they die.

    The principal arguments against this approach are that politics might intrude on Trust Fund investment decisions and that the government might own too much of private corporations. These concerns are easily addressed, however. Management of Social Security reserves would be placed in the hands of a Social Security Investment Board, modeled on the Federal Reserve. Like the Fed, which has preserved its political independence for decades, the SSIB would be financially independent, covering its expenses from small charges on investment earnings. Its members would be appointed for lengthy, staggered terms and could not be removed for political reasons.

    Investments would be limited to broad, passively managed index funds. Management of shares would be distributed among private funds managers hired on competitive bid, who would be directed to merge Trust Fund investments with other funds they manage for private clients. If considered desirable, voting rights of Trust Fund shares could be sterilized. A cap could be placed on the proportion of shares in any private company held in the Trust Fund’s portfolio.

    To make sure that accumulated reserves are held to back up future pensions and are not used to fatten current benefits, the Senate could vote a change in its own rules to require a super-majority in order to pass any Social Security bill not certified by the Social Security Investment Board to preserve actuarial balance measured over 5 years, 25 years, and 75 years. In addition, it is important for Congress to use the opportunity provided by the emerging forecasts of budget surpluses to adopt the recommendation of Representative (and former House Speaker designate) Bob Livingston to change budget rules so that all budget documents and budget targets are stated exclusive of Social Security.


    • Increased funding can add to national wealth and reduce future tax burdens of workers.
    • Investment in a diversified portfolio will raise returns to pensioners, but will not immediately raise national saving.
    • Privatization, itself, will boost administrative expenses and lower returns to pensioners.

    Myth 2

    —Privatization will improve economic efficiency.

    Some people allege that Social Security reduces economic efficiency because it distorts people’s decisions on how much to work and forces them to hold, indirectly, a portfolio of government bonds backing up their pensions that is different from the diverse portfolios of investments they would and should prefer. They perceive the combination of taxes and subsequent benefits as a net tax, and this tax, like most taxes, reduces economic efficiency.

    To begin with, it is important in interpreting this statement to understand that if a policy lowers economic efficiency, in the sense that economists use the term, that policy does not necessarily reduce work, saving, or national output. It means, instead, that people would perceive themselves as better off if they were not subject to the this policy. They might work or save more. They might work or save less. National income might be higher or lower. But people would be more satisfied with the outcome.

    How might Social Security reduce economic efficiency in this sense?

    Saving. Social Security holds all assets in a single portfolio of government guaranteed bonds. In contrast, people with sizeable asset holdings usually hold a diversified portfolio. People with few assets usually do not find it worth the cost or trouble to diversify their assets and may not even bother to think about the matter. But few people choose voluntarily to invest exclusively in government bonds. Individuals typically choose mixed portfolios and people select varied mixes. That means that no single portfolio can match everyone’s preferences. To use the banal and tired cliché, “one size does not fit all.” Only if people choose their own portfolios can they select the mix of risk and return that most closely conforms with their preferences. This is part of the line of reasoning that underlies the claim made by some analysts that replacing Social Security with private accounts would improve economic efficiency

    Unfortunately, this line of reasoning is irrelevant to the institution of social insurance. The goal of social insurance is to assure basic income. Any arrangement that jeopardizes the achievement or that increases the cost of achieving this objective is inefficient. Selection of different portfolios raises the cost of assuring basic incomes by boosting administrative costs. Furthermore, individually managed defined contribution accounts are inherently incapable of achieving the goal of assuring basic income because each worker must shoulder large financial market and economic risks. Free selection of portfolios is therefore an inefficient instrument for assuring basic income. Furthermore, individually managed accounts could not on the average achieve a better combination of risk and return than can a centrally managed fund invested in a prudently diversified portfolio, as administrative costs would be higher with individual management.

    While individual portfolio selection and management are inefficient ways to achieve the goals of social insurance, they produce ideal incentives to motivate voluntary saving and investment. With appropriate regulatory protections, free financial markets have served the United States admirably and will continue to do so. They are a major factor in making our economy as strong as it is today. But we should not confuse the goals of creating the right incentives for voluntary saving and investment with the goal of social insurance, which is to assure basic income.

    Labor Supply. The creation of private accounts could affect economic efficiency in one other way—by influencing labor supply. According to standard economic theory, economic efficiency is highest if people receive a wage equal to the value of their production. Workers then equate the marginal value of their compensation with the marginal disutility of work. Any policy that increases or lowers the marginal return to work, it is argued, upsets this balancing process, lowers economic efficiency, and reduces economic welfare.

    Voluntary private savings do not alter this balance because the returns just compensate workers for deferring use of their earnings. But mandatory pension plans can alter this balance because workers may value the eventual pension less than the current tax or contribution. In general, any tax system or transfer program that shifts or equalizes income among workers reduces economic efficiency in this sense. Despite such theory, every major economy has mandatory pension systems for paternalistic reasons—people simply do not voluntarily save enough, given current needs and temptations, to prepare adequately for retirement and disability. If one puts aside this fundamental reason for social insurance, Social Security can be said to reduce efficiency in labor supply for two reasons.

    The first reason is historical. Social Security initially paid out benefits far larger than workers’ payroll taxes (and the employers’ taxes paid on their behalf) could have justified. This generosity produced the so-called “unfunded liability,” which current and future workers will have to pay off at no direct benefit to themselves. Furthermore, workers must pay off that liability whether or not we retain the current system, privatize, or do some of both. I shall not debate whether paying relatively generous benefits in the past was a good or a bad idea. The simple fact is that we did, an unfunded liability exists, and we must deal with it. Current and future workers must pay extra taxes to pay off the liability. These taxes reduce the efficiency of labor supply today, in the sense described above. Privatizing Social Security in no way lowers this inefficiency.

    The second reason that replacing Social Security with private accounts might affect economic efficiency is that Social Security is the largest and most powerful instrument for reducing economic inequality managed by the federal government. Each year it lifts the incomes of approximately 16 million people above official poverty thresholds, twice as many as all income-tested assistance in cash and in kind combined.

    Most taxes or transfers that equalize income reduce efficiency.

    Social Security accomplishes this anti-poverty function because it raises revenues with a tax that is proportional to earnings but provides benefits that are proportionately larger for low earners than for high earners. In contrast, individual accounts would provide proportionately lower benefits to low earners than to high earners. They would do so for two reasons.

    • Administrative costs rise less than proportionately with the size of accounts and hence reduce returns more for low earners, whose accounts will be small, than for high earners, whose accounts will be comparatively large.
    • Abundant experience with 401k plans shows that people with small accounts invest in lower-yielding assets than do people with larger accounts.

    Replacing Social Security in part with private accounts would therefore replace a program that lowers inequality with one that increases it. Whether such a swap would raise or lower efficiency of labor supply is not clear, but any such gains would assuredly come only at the price of increased income inequality.

    Myth 3

    —It is possible to privatize Social Security and raise benefits without raising taxes, cutting government benefits or other spending, or increasing the national debt.

    Revenues are necessary to build private accounts. These revenues could come from taxes or from the deposit in private accounts of new government bonds. The deposit of new government bonds would directly raise the national debt. Such bonds would be assets of the private accounts, but would eventually have to be paid off by the general tax payer or would generate a stream of interest payments with the same discounted present value. The diversion of payroll taxes from Social Security to private accounts would deepen the projected long term deficit and necessitate larger benefit cuts or tax increases than would otherwise be necessary.

    The final way to fund individual accounts is through the appropriation of general revenues. This approach also requires one or more of the following adjustments: general tax increases that would not otherwise be necessary; cuts in general government spending that would not otherwise be required; or increases in government borrowing and interest payments above those that would be necessary if general revenues were used to reduce government debt held by the public. These alternatives represent simple accounting arithmetic.

    The prospect that the budget for federal government operations other than Social Security may at some time run a surplus in no way modifies this arithmetic. The budget, apart from Social Security, remains in deficit, although the Congressional Budget Office forecasts that surpluses will emerge in a few years. Those surpluses are projected to last several years and then vanish when retiring baby-boomers push up health care costs. They may never appear and will vanish much sooner if Congress raises spending or cuts taxes now. A recession could cause projected surpluses to evaporate as fast as water droplets on a hot griddle.

    Some plans to use general revenues to fund individual accounts actually raise the total pension obligations to the elderly. I believe that such plans are rash. With the imminent retirement of the baby-boom generation, pension costs will increase dramatically, even if benefits are not increased. Health care costs through Medicare and Medicaid are projected to increase by at least twice as much as Social Security. Without allocating more revenues to health programs, it will be quite impossible to assure the elderly, disabled, and the poor even basic health care. In addition, we have not yet even begun to design a sensible set of institutions, private or public, to pay for the costs of long-term care. Furthermore, current and future Congresses and presidents just might consider it important to cut taxes, promote education, underwrite research, build roads, strengthen national defense, or seek other goals. Committing large amounts of federal revenues to funding private accounts would put the future Congresses in a fiscal straightjacket. Under these circumstances, measures to increase pension obligations to the elderly would be irresponsible and short-sighted.

    Myth 4

    —The Social Security trust fund is a fraud, a collection of meaningless IOUs.

    To see why the assertion that the Trust Fund is meaningless is false, it will be helpful to look at the actual expenditures, revenues, and net asset position of Social Security and the rest of the federal government for fiscal year 1999, as projected by CBO in August 1998. These are shown in the upper half of table 1. The bottom of half of table 1 presents business and pension operations of a hypothetical corporation. The numerical values of this corporation’s operations happen to be the same as those for Social Security, but the report is silent on whether the company is reporting in dollars, cents, or some other unit of currency. In both cases, it surely seems that a pension fund exists, with a value of 853. Are there circumstances under which one could say that this appearance is misleading?

    Table 1
    Operations of Social Security and a Hypothetical Corporation

    (in billions of dollars)

    Social Security

    Outlays Revenues Difference Cumulative
    [Surplus (+)
    or Debt (-)]
    Other Operations 1,396 1,359 – 37 – 4,508
    Social Security 325 442 + 117 + 853
    Total 1,721 1,801 +  80 –  3,655

    Private Corporations

    Outlays Revenues Difference Cumulative
    [Surplus (+)
    or Debt (-)]
    Corporate Activities 1,396 1,359 –  37 –  4,508
    Pension 325