The Good News About Social Security
Malcolm Mitchell - August 2002

(Click here for the full version [68 KB] in Microsoft Word)

       "The economic achievements, our increasing social coherence and cohesion, our increasing efforts to reduce poverty among our youngest children -- all of them are threatened by the looming fiscal crisis in Social Security." President Clinton, February 1998.
       "One option, of course, is simply to cut benefits. . . . The older widow scraping by on $600 per month would receive only $450." Schieber & Shoven, The Real Deal, 1999.
       "Today, young workers who pay into Social Security might as well be saving their money in their mattresses. That's how low the return is on their contributions. And the return will only decline further -- maybe even below zero -- if we do not proceed with reform." President Bush, May 2001.
       "I won't say I told you so, but if anyone is in the market for a never-been-used lockbox, they should see me afterwards." Al Gore, April 2002.

The hubbub is nothing new. Since the original Social Security Act was passed in 1935, Americans have heard repeated political attacks on the structure of the system, and relentless warnings of its impending collapse. Yet through it all, Social Security has kept its promises to beneficiaries for 65 years. Today, 46 million retired wage-earners receive monthly benefit checks, and the system is financially stronger than it has ever been.

The attacks and warnings nonetheless take their toll on wage-earners' confidence. Only 13% of them believe that Social Security will be their "most important source of retirement income." The reality for retirees is quite different; 44% of them say that Social Security is their "most important source of retirement income."(1)

This article is not likely to silence the attackers. Nor is it our intention to guess at their objectives or hidden agendas. In order to assure wage-earners and retirees, we will look at published data from the Social Security Administration and the Trustees, which demonstrate the financial soundness of the system today. We'll then deal with arguments that ignore this data and fuel the political clamor. And finally we'll propose an approach to managing Social Security's finances that includes familiar remedies, if long-term financial problems arise.

I. The Good News Today

Let's think about how to measure the current financial condition of Social Security. This is a program that delivers no services or products. It differs from all other government programs, including unemployment insurance and health insurance, in that it mandates annual payouts to every participant, at a time and a level determined by pre-established, although changeable, formulas. Its function, therefore, is purely actuarial -- to manage the balance of contributions and payouts over time.

One financial test is to apply a meaningful measure at specific points over Social Security's entire history. For example, we could look at the system's total assets at the beginning of each year in relation to expected payouts during that year. This "payout coverage" measures the number of months or years the system can continue to make promised payments, even if there are no further receipts.

Some readers may find it clearer to think of these assets as "reserves" -- or as the surplus that accumulates in the system when revenues are greater than payouts. Both descriptions are accurate. In this article we'll refer to the measure as coverage. As we'll see, the Trustees of the Social Security Trust Funds also refer to coverage in their analyses of the program's future.

On this measure, what is Social Security's financial condition today, compared to its successful past? Let's go back to the beginning. For a dozen years starting in 1937 -- the first year in which there were both receipts and payouts -- annual receipts far exceeded payouts. The Trust Fund's assets therefore grew rapidly; the coverage in some early years reached more than 20 times that year's payouts.(2) At the beginning of 1949, assets were 15 times payouts. Over the following years annual expenditures grew more than 8-fold; yet at the beginning of 1957 coverage was still 3 times that year's payouts.

During 1957, two things happened. For the first time, Social Security receipts were less than payouts -- $7.4 billion versus $7.5 billion -- and the program suffered its first year-to-year decline in accumulated assets.

The second thing that happened in 1957 was that benefits were extended to wage-earners who become disabled before retirement age and can no longer work. What was called the Old-Age and Survivors Insurance Trust Fund (OAS) became known in 1957 as the Old-Age, Survivors, and Disability Insurance Trust Funds (OASDI). While the Social Security Administration maintains separate accounting for OAS and DI, public discussion is normally based on OASDI, and so is the data we use in this article.

Let's return to the coverage measure. For most years between 1957 and 1965 the system paid out more than it received annually. As a result, assets in the OASDI Trust Funds at the beginning of 1966 failed to cover even that year's payouts of $20.9 billion. Over the next nine years, amendments to receipt and expenditure formulas enabled annual balances to remain slightly positive. Coverage continued to hover around one year.

The real pinch began in 1975. The annual balance was again in deficit and it continued in deficit for the following six years. Assets dropped steadily in relation to payouts, while total expenditures rose rapidly. At the beginning of 1982, accumulated assets of just $24.5 billion covered less than 2 months of that year's $160 billion in payouts.

This precarious situation continued for several years, despite the 1983 amendments to the Social Security law. In 1986, accumulated assets still covered less than 3 months of payouts. Finally, by 1992, several years of surplus brought the Trust Funds' assets back to one year's coverage.

How has Social Security fared by this measure in the last ten years? At the end of 2001, accumulated assets in the OASDI Trust Funds totaled $1.2 trillion. In their April 2002 report, the Trustees give this year's expenditures as $465 billion.(3) The Funds thus have assets today equal to 2 years and 7 months of current payouts. This is a level of coverage that the Funds have not enjoyed since 1958.

If anything, Social Security has been growing financially stronger. There were some tight spots over the past 40 years, but the program came through intact -- even as the proportion of Americans covered was steadily increasing. In 1957 there were 10.3 million recipients of Social Security benefits, about 6% of the population. Today, the 46 million Americans receiving OASDI benefits represent about 17% of the total population.

II. The Good News for Tomorrow

Can we apply the same measure to Social Security's future? In fact, the Trustees have already done that. They project that accumulated assets at the beginning of 2011 will be "447 percent of expenditures" that year. That's coverage of 4.47 times payouts -- nearly twice today's coverage. Using "intermediate assumptions" (which we'll explain in the next section), the Trustees conclude that without any changes in current laws, OASDI is "expected to be adequately financed over the next 10 years, with large and increasing annual surpluses."(4)

Does the good news fade after that? According to the Trustees, at some time after 2010, "OASDI costs will increase rapidly due to the retirement of the large baby-boom generation." However, receipts are expected to continue to exceed expenditures for several years more. Not until 2017 is the balance of receipts and expenditures expected to move into deficit, causing accumulated assets to begin shrinking, just as they did in the late 1970s.

So do we fall off a cliff after 2017? Actually, no. Remember that our measure of coverage after 2011 will be much larger than it is today. So what will happen after 2017? Receipts will continue to be generated at the level mandated by today's laws. Benefits promised under today's laws will continue to be paid. What will happen is that the coverage will decline -- but not too fast. The Trustees expect that as late as 2033 the coverage will still be two years, or about what it is today. In 2037, the coverage will be one year, a level that, according to the Trustees, "satisfies the test of short-range financial adequacy."(5)

Here's the good news for tomorrow in a nutshell. If we change absolutely nothing in the current laws that govern Social Security receipts and expenditures, the Trustees' "best estimates of future experience" show that for at least the next 30 years, every promised benefit will be paid in every year to every retiree, and the Funds' accumulated assets at that point will still cover more than one year's payouts.

Does that sound like a program in crisis today?

III. So What's the Fuss About?

Given the clarity of the published facts, continued warnings about what's wrong with Social Security seem to us purposeful attacks on a popular social program that works. Attackers know that by ignoring the facts, they are making American wage-earners and retirees fearful that the program will fail them. They then cite participants' fears as evidence of the problem.

The attacks fall into four categories:

  • Social Security is referred to as welfare (by some of its friends as well), with all the negative connotations of that word. We'll show that Franklin D. Roosevelt and his staff created a unique structure for Social Security that was specifically intended to separate it from government welfare.

  • It is argued that Social Security assets aren't really "there" -- they've been spent, or stolen, or removed by sleight-of-hand -- and that to fix the resulting problems, the payroll tax rate will have to be raised significantly. Besides being simply untrue, as we'll show, this argument strikes us as most peculiar, since if you believe the money contributed disappears, why would you suggest increasing it?

  • Those who do recognize that Social Security assets are really there want to treat them as a form of national savings and investment. They urge reforms to the program that will "help to maintain or reduce national deficits relative to GDP and improve economic efficiency and prospects for growth."(6) We'll argue that an important reason for Social Security's popularity is that it stands apart from any other national economic considerations. Furthermore, suggested reforms that drag Social Security into those considerations invariably omit the guarantee of benefits that is the core of Social Security's strength.

  • Trustees' long-term projections of Social Security finances -- even 75-year projections -- are quoted as simple truths about the future, rather than the wide range of uncertain estimates that they are. The Trustees themselves warn of their "degree of uncertainty" in making long-range projections by asking us to imagine "how difficult it would have been in 1925 to project the world of 1930, much less that of 2000."(7)
    We'll deal with each kind of attack in turn.

    Is Social Security welfare?

    To appreciate the unique structure of the Social Security system, we need to recall the economic context in which it was born in 1935. Falling agricultural prices in the 1920s, the stock market crash of 1929, and a string of bank failures, especially in rural areas, plunged the U.S. into the 1930s depression. By the time Franklin D. Roosevelt was inaugurated as President on March 4, 1933, employment had collapsed. "No one knows how many [unemployed] there were on Inauguration Day -- at least 12 to 15 million, over a quarter of the labor force."(8)

    At the time, there were no federal programs that provided direct aid to the unemployed or needy. Individual states disbursed all support under welfare laws that many states (but not all) had passed in the 1920s. What federal help there was came in the form of grants to the states. For example, federal grants to states were introduced in the 1920s to support vocational rehabilitation programs and for maternal and child welfare.(9)

    With the deepening depression, states were swamped by demands for relief. "In some states 40 percent of the people were on relief; in some counties the figure rose to 80 or 90 percent. Everywhere [state] funds were running out."(10)

    Roosevelt's immediate concerns in March 1933 were to stop the financial bleeding and to turn national pessimism into optimism. He closed the banks to let fear dissipate and later reopened them with federal support for those in danger of failing. He established support for farmers to help pay their home mortgages and end a stream of foreclosures -- which was a principal cause of bank failures in the first place. Most famously, he established a series of federal job programs that put millions of Americans to work throughout the 1930s.

    When he turned to direct relief, Roosevelt first proposed a law that maintained the policy of funneling money through the states. The Federal Emergency Relief bill provided $500 million for grants-in-aid to states. It was passed on March 30, 1933, but it proved insufficient. "In mid-1934, about one out of every seven Americans -- over 18 million people -- were still receiving relief; and the total would rise again to 20 million by winter."(11)

    Two additional pieces of federal relief were clearly needed. One was unemployment insurance and the other was some kind of aid for the elderly. In January 1935, Roosevelt and his staff dealt with both needs by proposing what we know as the Social Security Act; it was passed by Congress in August. The act did three things: it provided grants to the states to help pay old-age assistance; it established a federal-state system of unemployment insurance; and it created Social Security, then called OAI. Of the three, "OAI was the only one . . . which was to be a wholly federal program."(12) Indeed, it was the first program that linked the federal government directly to Americans' retirement planning.

    OAI was also unique in its funding and payout methods. Roosevelt insisted that contributions into Social Security should come only from the participants themselves and be paid out in full to those same participants. This structure differed from every other social program in the U.S. or, for that matter, anywhere else in the world. Rather than welfare, it was a self-contained, self-funded system, intentionally isolated from other government programs.

    Roosevelt seems to have stood virtually alone in insisting on this structure. Here is one knowledgeable commentator, writing in 1991:
      Roosevelt wanted old-age insurance to be completely self-supporting. All of the money for the program would have to come from payroll taxes, none from general revenues. When Roosevelt discovered the "government subsidy" in the [draft] plan . . . he summoned staff director Edwin Witte to the White House. In Witte's presence, Roosevelt crossed out the word "subsidy" . . . [and] replaced the word with "deficit," and told Witte to rework the tax rates.(13)
    Another writer treats him less kindly:
      In many respects, the law was an astonishingly inept and conservative piece of legislation. In no other welfare system in the world did the state shirk all responsibility for old-age indigence and insist that funds be taken out of the current earnings of workers. By relying on regressive taxation and withdrawing vast sums to build up reserves, the act did untold economic mischief.(14)
    As for Roosevelt's own explanation of the structure, here is an often-quoted passage from the same writer:
      "I guess you're right on the economics," Roosevelt conceded when told that the employee contributions were a mistake, "but [objections to] those taxes were never [based on] economics. They are politics all the way through. We put those payroll contributions there so as to give the contributors a legal, moral, and political right to collect their pensions. . . . With those taxes in there, no damn politician can ever scrap my social security program."(15)
    The boldness of the proposed legislation generated the expected opposition. "In an unsuccessful attempt to strike out old-age insurance . . . Senator Daniel O. Hastings of Delaware predicted that it might 'end the progress of a great country and bring its people to the level of the average European.' Congressman Daniel Reed of New York predicted 'The lash of the dictator will be felt.'"(16)

    Opposition continued after the law was passed. "Social security became a major issue in the presidential campaign of 1936 when the Republican candidate, Alfred M. Landon, charged it was 'a fraud on the working man. The savings it forces on our workers is a cruel hoax.'"(17)

    Nor was the opposition purely political. "The Brookings Institution, the American Federation of Labor, the Chamber of Commerce, and The New York Times all agreed that the reserve financing plan was a bad idea."(18) When the first payroll deductions began to appear in January 1937, "Some workers found a note in their pay envelopes that equated the new Social Security taxes with theft."(19)

    And what did the American people think? Roosevelt carried his message into the 1936 presidential election and received over 60% of the popular vote. "Roosevelt carried every state but Maine and Vermont, winning the most overwhelming electoral majority since Monroe's victory in 1820. . . . The election gave the Democrats majorities of 76-16 in the Senate and 331-89 in the House."(20)

    Polls showed that the vast majority of Americans supported Social Security, not only in that election but later. "Support for OAI . . . was 68 percent of a national sample in a November 1936 Gallup poll, increasing to an incredible 97 percent in a National Opinion Research Center poll taken in 1943."(21) Bear in mind that wage-earners paying into the system at that time outnumbered beneficiaries by at least 40 to 1, and it would be years before most of them came to collect anything. Yet over 1 million retirees were receiving Social Security benefits, and Americans saw the solid guarantee that a self-funded system offered.

    Here is a typical description of the popularity of Social Security:
      Almost alone among social welfare programs, Social Security came to enjoy a universal following, free of the occupational and class divisions that had plagued previous efforts at social policy . . . [and became] the most popular social welfare program in the nation's history.(22)
    The same writer explains why:
      Neither welfare nor insurance, neither socialism nor capitalism, [Social Security] represented a pragmatic middle ground, a consensus on which social policy could rest. It exemplified what some observed as America's singular ability to rise above ideology and solve social problems.(23)
    In other words, one of the reasons for Social Security's continuing popularity is precisely that, being completely self-funded, it is not welfare. In giving all contributors to Social Security "a legal, moral, and political right" to collect their promised benefits, Roosevelt had planted well. Scores of millions of Americans have since reaped those benefits. If participants use their voting power to make sure that "no damn politician" interferes, they will continue to reap their benefits in the future.

    Are Social Security assets really there?

    For anyone who has taken the trouble to look into the data we've described, the heated rhetoric over Social Security assets seems incomprehensible. Here is Senator Bob Kerrey, a Democratic from Nebraska, at a Senate Finance Committee hearing in August 1998:
      We are not prefunding . . . Are we holding the money in reserve someplace? We are not prefunding! The idea in 1983 was that we would prefund the baby boomers. We began to use it immediately for the expenditures of general government. We didn't prefund anything. What we are doing is asking people who get paid by the hour to shoulder a disproportionate share of deficit reduction. That's what we're doing!(24)
    Kerrey's shrill rhetoric continues to be heard through a self-styled "nonpartisan, grassroots" organization called The Concord Coalition, which he co-chairs with Warren Rudman. "The Social Security trust fund is primarily an accounting device. Social Security's 'assets' consist of Treasury IOU's . . . [and] represent future general fund liabilities -- not a cash reserve."(25)

    Nor does rhetoric that appears to be more rational produce greater clarity. Here are a few comments from President Clinton:
      On Feb. 9, 1998: Today, we're actually taking in a lot more money from Social Security taxes enacted in 1983 than we're spending. Because we've run deficits, none of that money has been saved for Social Security.
      On April 13, 2000: There is a consensus in Congress that we should use all the Social Security surplus for debt reduction, and that is a good thing. . . . My budget locks away the interest savings from the Social Security surplus to lengthen the life of Social Security.
      On October 25, 2000: We're not spending the Social Security taxes now, which we did from 1983 until a couple of years ago. We're not spending the Social Security taxes now, so they're contributing to debt reduction.(26)
    And of course the infamous lockbox is being readied for another round of electioneering, with commentary from gurus:
      Al Gore spoke to Florida's Democrats in early April [2002], he joked: "I won't say I told you so, but if anyone is in the market for a 'never-been-used' lockbox, they should see me afterwards." The laughter was loud and immediate. . . . The lockbox will live on because it is a metaphor for the most important political act: the need to make choices.(27)
    These comments do violence to the realities of Social Security, because they are based on misunderstandings, or misrepresentations, of what actually happens. It seems to us, therefore, that we can best deal with them by describing the realities.

    One preliminary explanation, however, will clear up any honest confusion that arises over the role of the U.S. Treasury in the Social Security system. The Treasury's main function is to act as the bank for the U.S. Government. The Treasury receives all tax revenues and other payments and writes checks for all government outlays, including interest on bonds.

    The Treasury is also the government's investment banker, raising money when the government needs to borrow by designing and selling Treasury bonds. An important constraint in this process is that by law, these bonds can be sold only to the private sector (individuals, corporations, private pension funds, etc.) or to state or municipal funds -- but not to any other federal government entity.

    (For the cognoscenti, we're aware that the actual banking function is performed by the Federal Reserve Bank and designated private banks; that distinction, however, would complicate our description and does not affect our point.)

    Where confusion might arise is over the fact that the Treasury also acts as the banker for the Social Security system, receiving all contributions and sending Treasury checks for benefit payments. However, the two functions -- acting as banker to the government and as banker for Social Security -- are entirely separate. When the Treasury receives income tax payments, or the proceeds from a bond sale, that money belongs to the government and becomes part of the government's "general fund." When the Treasury receives Social Security taxes, that money belong to the Social Security Trust Funds, or more precisely, to the participants in the Social Security system.

    It is true that the Treasury puts all incoming money into a single account and makes separate "accounting" entries for the general fund and for Social Security -- just as a private bank puts Mr. Doe's money and Mrs. Smith's money into the general reserves of the bank and gives each of them an individual accounting. The legal ownership of the two pools of money is not in any way affected by the fact that they are held in the same bank.

    We can now set down some facts about Social Security's assets:

    1. Assets in the Social Security Trust Funds are no less real than assets accumulated by any private citizen or corporation from a surplus of income over expenditure. In the case of Social Security, wage-earners and employers are required to participate at fixed levels, but the assets once accumulated are not for that reason different from any other private accumulated assets.

    2. Social Security's accumulated assets are managed by the Trustees of the OASDI Trust Funds. Since they are private assets, they can be invested in U.S. Treasury bonds, the safest interest-bearing assets in the world. The bonds held in the Trust Funds cannot be sold before they come due, but they are not different in any other way from Treasury bonds owned by millions of private investors, corporations, pension funds, and other organizations, both in the U.S. and around the world. There is no difference in the interest rates on the bonds, as we'll see below.

    3. Social Security assets are no more and no less IOUs from the U.S. Government than are other private assets invested in Treasury bonds. As we explain below, there are $6.4 trillion of such IOUs outstanding, including Social Security's $1.2 trillion, and all of them are backed by the full faith and credit of the United States.

    4. In the 15 years between 1987 -- when as we saw Social Security had less than 3 months of payouts on hand -- and the end of 2001, total receipts collected by OASDI were $5.98 trillion; total payouts were $4.81 trillion.(28)
       The surplus in these 15 years was thus $1.17 trillion. All of it was invested in U.S. Treasury bonds. These bonds are listed individually, by yield and maturity date, in the Trustees' 2002 Report. For example, the OASDI Trust Funds own about $42 billion of 8.75% bonds which will come due in the next three years -- not a bad holding given today's much lower interest rates. They own about $170 billion worth of bonds paying over 7%.

    5. Here's the bottom line. Social Security assets begin life as private assets paid into the Trust Funds by participants in the system. The assets are invested in safe, interest-bearing bonds, and paid out to participants as private citizens when they retire.
       It is true that Social Security participants also pay income taxes into the government's general revenues. Like other American individuals and corporations, they may have complaints about the level of total government debt and the uses to which the government puts the taxes it collects and the money it borrows.
       As participants in Social Security, however, Americans can be happy with the way the program's assets have been managed in the past. If they are able to keep politicians' hands away from those assets, participants can be confident about the future as well.
    Are contributions to Social Security part of a larger scheme of national savings and investment?

    As Social Security assets have grown, they have become a favorite target of economists in their debates about national finances. Suggestions of impending financial crisis in Social Security finances prompt proposals for reform -- whether the crisis is shown to be real or not -- and then debates over those proposals take on a life of their own. For at least several years, the American Economic Review has been carrying articles devoted to the implications of "Social Security's Long-Term Funding Crisis."(29) These articles cover topics ranging from pension reforms in Europe to factors that promote growth in U.S. GDP -- and all of them include proposals for reforming Social Security.

    A 1997 Symposium on "Social Security Reform" sponsored by the Federal Reserve Bank of Boston discussed such things as "How reform would affect labor markets" and "The impact of demographic change and Social Security reform on financial markets and risk-bearing."(30) The editors of the report on the symposium weighed in with an article entitled "Social Security: How Social and Secure Should It Be?" They wrote: "In the United States, and in other industrial nations, we have come to rely on the government to solve much of the old-age income problem"(31) -- as though the unique self-funding element in Social Security didn't exist.

    A related, although separate attack on Social Security focuses on a supposed "investment return" to Social Security participants. In particular, a 2% return has been widely announced, although no cogent analysis exists to explain how that 2% was determined. It seems to us that such an analysis would have to make some aggregate assumptions, and the fact is that the aggregate return to Social Security participants, as we show below, is much higher than 2%.

    In any case, precision is not a hallmark of this kind of attack. Here is President Bush:
      On March 19, 2001. In order to make sure that there is a Social Security system around tomorrow, we've got to get a better rate of return on the money in the Social Security trust. And that means trusting people with their own money, to make wise decisions as to how to get a better rate of return than the 2 percent in the Social Security trust.
      On May 2, 2001. Today, young workers who pay into Social Security might as well be saving their money in their mattresses. That's how low the return is on their contributions. And the return will only decline further -- maybe even below zero -- if we do not proceed with reform.(32)
    Treating Social Security assets as investments, and relating them to a national scheme of savings and investment, may serve some purposes of economists and politicians. As we have come to expect, however, their arguments conflict with the actual structure and goals of Social Security.

    1. By the intent of its creators, Social Security is a closed system. Its assets come solely from wage-earners and are paid out solely to wage-earners upon their retirement. Over time, therefore, participants in aggregate receive exactly as much as they have put in -- plus good interest on the invested surplus (see 6. below) and minus small administrative costs (less than a quarter of one percent).

    2. Whether individual participants get back more or less than they paid in depends on many factors over many years, including how long each one lives after retirement. Trying to determine a typical investment return for individual participants is therefore impossible.

    3. In place of an investment return, individual participants in Social Security are assured of receiving retirement benefits at a level that -- with other assets, including a paid-up home -- can support an acceptable living standard. Some participants think of their promised benefits as an insurance payout.
       In contrast, no proposal for "reform" offers a guarantee of retiree benefits at a level higher than current benefits. The critical word here is "guarantee." It is a word that reformers assiduously avoid.
       This is especially true of those promoting privatization, or, in the new phrase, personal savings accounts. In place of Social Security's pre-determined level of retirement benefits, participants are promised higher benefits, if they can learn "to make wise decisions" with their money. The offer, however, comes not only without a guarantee, but without any historical evidence of the wisdom of individual investors.

    4. The specific purpose of the self-funding structure of Social Security -- what we have called a closed system -- is that it isolates the program from other government programs. It also isolates the program from discussions of national savings and investment and how to reform or improve them. In ensuring that it keeps its promise to future retirees, Social Security may impact the national economy for better or for worse. But the crucial importance of Social Security for so many Americans obliges us to find other means, other than tampering with Social Security, for repairing perceived ills in the economy.

    5. In any case, Social Security is a minor player in the U.S. national economy. Congress this year raised the legal limit on U.S. government debt to $6.4 trillion to permit the government to borrow more money in order to pay its bills and the interest on existing debt. OASDI Trust Funds, as we've seen, own $1.2 trillion of that debt. Over $5 trillion of government debt, therefore, is owned by private individuals, corporations, and other organizations around the world.
       Furthermore, Social Security receipts will be about $600 billion this year; total government receipts will be approximately $2 trillion. It's clear that economic reformers have lots of opportunity within the U.S. economy to ply their trade, without attacking Social Security.

    6. For the record, on the question of an investment return, the Social Security Act provides that "the interest rate on new special obligations will be the average market yield, as of the last business day of a month, on all of the outstanding marketable U.S. obligations that are due or callable more than 4 years in the future."(33) The Trustees report that in 2001, the OASDI Trust Funds "earned interest at an effective annual rate of 6.6%." Furthermore, "The average interest rate on new securities purchased [in 2001] . . . was 5.2 percent."(34)
       By our calculation, the Trust Funds earned roughly $500 billion in interest over the past 15 years -- which means that in aggregate, Social Security participants have $500 billion more available to pay benefits than the total of their contributions.
       For as long as the assets in the Trust Funds are sufficient to cover two years of annual payouts, every dollar contributed will earn annual returns equal to Treasury bond yields as described.

    So how about these long-term projections?

    As we said above, the principal function of the Social Security Administration is actuarial -- that is, to maintain the balance of receipts and promised payouts over time, despite unavoidably imperfect forecasts of the factors that drive both receipts and expenditures. Those forecasts naturally become more difficult as the time frame lengthens. The Trustees themselves explain:
      The future income and outgo of the OASDI program will depend on many demographic, economic, and program-specific factors. . . . Basic assumptions are developed for several of these factors based on analysis of historical trends and conditions, and on expected future conditions. These include fertility, mortality, immigration, marriage, divorce, productivity, inflation, average earnings, unemployment, retirement, and disability incidence and termination. Other factors are projected using methods that reflect historical and expected future relationships to the basic assumptions. These include total population, life expectancy, labor force, gross domestic product, interest rates, and a myriad of program-specific factors. It should be noted that all factors included in any consistent set of assumptions are interrelated directly or indirectly. It is also important to note that these interrelationships can and do change over time.(35)
    Note first that the Trustees do forecast all of the demographic factors that are widely expected to cause financial problems for Social Security in the future: fertility rates, life expectancy, etc. Their estimates of the changes that are likely to occur in these factors over time are included as input for their estimates of the future financial condition of the Trust Funds.

    Note also that the Trustees take into consideration the relationships among all the factors. This creates an additional level of complexity in forecasting. The Trustees admit that "interrelationships can and do change over time," and this makes some commentators particularly wary of very long-term forecasts. For example, most economists have assumed that changing demographics in the early 1970s were partially responsible for a slowdown in national economic growth, and they forecast continuing slow growth in part because of continuing changes in demographics. But some economists question the relationship itself and therefore the forecast. "Continuing slow growth over the next century is not inevitable. . . . An increase in the assumed rate of growth of real wages over the next 75 years would effectively eliminate the [Social Security] funding crisis."(36)

    Let's look at how the Trustees, despite all the difficulties, make their required long-term forecasts. First, they assign "ultimate values" to all the key factors on which the long-term projections are based. These values are arrived at by estimating factors for 10 years out and then assuming that the level of the factor (or its rate of change) will remain constant after that. "These ultimate values generally apply after the first 10 years. Two exceptions are the ultimate fertility rate and the ultimate mortality annual rate of reduction, which are reached in 2026."(37)

    The values that are estimated for all factors are not expressed as a single number, however, but as a range of forecasts. Whether for 20 years or for 75 years, this range is distilled into three forecasts, a "most optimistic," a "most pessimistic," and an "intermediate" forecast. The intermediate forecast is given as the Trustees' "best estimate of the future."(38)

    The range of forecasts can be very wide, reflecting the extraordinary complexity and difficulty of making 75-year forecasts for all the factors and the relationships among them. Let's consider, for example, the range of forecasts for one factor, a critical one for the Trust Funds' financial health -- the ratio of "covered workers" to beneficiaries. The ratio was 3.4 last year, roughly the level it has been at for 25 years.

    The Trustees' intermediate forecast shows the ratio dropping steadily for the next 75 years, to 2.8 in 2015, 2.0 in 2050, and 1.8 in 2080. Their forecast of this factor is the one most often seized upon as proof that Social Security is "unsustainable over the long term."(39)

    In fact, however, the Trustees' uncertainty is evident in the huge disparity between their most pessimistic and most optimistic forecasts. Their most pessimistic forecast for 2080 shows a ratio of just 1.3 wage-earners to beneficiaries. Their most optimistic shows a ratio of 2.4. In other words, taking a 75-year look at a critical Social Security ratio, the Trustees are able to plausibly forecast a high and a low that are nearly 100% apart.(40)

    Their uncertainty is even more evident if we look again at "payout coverage." Indeed, on this measure the difference between the Trustees' most pessimistic and most optimistic forecasts seems to be infinite. Recall that their intermediate forecast shows the coverage still at one year in 2037 -- again, with no changes in current laws and after all promised benefits have been paid. The most pessimistic forecast shows coverage dropping to one year in 2027, a decade earlier. Furthermore, both the intermediate and the pessimistic forecasts show the coverage continuing to drop below one year, until there is zero coverage -- that is, no surplus at all -- in 2041, or worse, in 2030.

    But what about the optimistic forecast? It shows coverage rising to over six times around 2015 and then declining to just over five times around 2035. It then remains over five times until 2080 and, apparently, beyond. In other words, the Trustees envision a possibility that over the next 25 or so years the OASDI Trust Funds may develop a permanent surplus equal to five times annual payouts.(41)

    Is this good news? Hardly. A permanent surplus of such a size in an expanding program would mean that trillions of dollars would have to be invested on a continuous basis. Economists' concerns would then be appropriate; the impact of such an eventuality on markets and on the national economy could be overwhelming. In addition, a large, ongoing surplus would generate questions about the fairness of the Social Security program to current participants. We'll discuss these concerns in the next section.

    So here's the situation regarding the Trustees' most pessimistic and most optimistic forecasts. On the one hand, there may be a crisis in Social Security 20 or 25 years from now because the surplus shrinks too low. On the other hand, there may be a crisis 13 years from now because the surplus rises, and remains, too high.

    And what does this suggest about a 75-year forecast as a reliable picture of Social Security's future? The same uncertainties that create the wide range of forecasts over the next two decades make hash of any attempt to forecast 75 years out. How will inflation, productivity, economic growth and real wages, fertility rates, life expectancy, and total population change between now and 2080? Your guess is as good as anyone's. How about today's concerns over the future price of oil? Who knows whether this single commodity, which has powered the world economy for less than 100 years, will still be of paramount importance in 2080?

    This is not to criticize the Trustees, who are doing what's required of them as well as they can. But as for those who use 75-year forecasts to attack Social Security today and demand reform -- maybe it's time to move on and to consider what's best for Social Security participants themselves.

    IV. A Proposal

    Social Security is financially healthy today. Here's our proposal for keeping it that way. Let's focus on Social Security's payout coverage, and let's manage receipts to maintain 10-year forecasts of coverage at between two and three times annual payouts. Raise receipts when the forecast shows that coverage will drop below the target, and lower receipts when the forecast is for rising coverage.

    For example, if the future unrolls according to the Trustees' most pessimistic forecast, Social Security's payout coverage will increase from today's approximately 2.5 times until it reaches about 3.5 times payouts in 2012. Our proposal to maintain coverage at between two and three years suggests that on this worst-case basis, the receipts formula should be left as it is today.

    Is this risky? Well, we've taken the Trustees' most pessimistic forecast for our example. Their intermediate "best estimate" shows coverage rising much faster, reaching about 4.2 times payouts in 2010 and 4.5 times in 2015. So a recommendation that the receipts formula be left as it is actually creates a greater risk that the system will become overfunded. The Trustees most optimistic forecast is that coverage will soar to more than 5 times payouts in 2011.

    Why is overfunding risky? First, if coverage rises to 5 times in 2011, total Trust Fund assets will reach nearly $4 trillion,(42) an increase of $2.8 trillion over today's assets. That additional $2.8 trillion would be looking to invest in new Treasury bonds over the next nine years. There may be scenarios for the overall U.S. economy in which this massive buying wave will not be disruptive; but there are others in which it would create serious problems.

    The second risk of soaring coverage is in its impact on individual participants' views of their roles within the system. Social Security has become a mature system only in the last 25 years, and each generation's contributions are now thought to be used to pay benefits to living beneficiaries. If coverage rises significantly, and especially if it remains high (as in the Trustee' optimistic long-range forecast), current wage-earners may come to feel that their contributions are for benefits too far in the future, that they are being asked to pay more than is either appropriate or necessary today. What, after all, is a huge surplus for?

    Let's return to our proposal. Managing the coverage to 10-year forecasts does not mean waiting 10 years to act. Each year the Trustees make new forecasts, and thus the previous year's expected trend line becomes clearer. Furthermore, if a 10-year forecast shows coverage moving above or below 2-3 years, a raise or a reduction in the receipts formula actually has 12 years to work at moving the coverage back to where we want it.

    So what would our actual recommendation be today, given the Trustees' intermediate forecasts? Reduce Social Security receipts enough to bring the coverage back from the current 10-year forecast of 4.2 times to a 10-year forecast of 2-3 times. For those who are afraid to act on a 10-year basis and persist in believing that 30-year forecasts are more reliable than 75-year forecasts, we would point out that the Trustees' current best estimate is that coverage will still be at two years or more in 2033. If a reduction in receipts turns out to impact the 30-year forecast negatively, and, we would add, if future 10-year forecasts show coverage dropping too quickly, there will be time to raise receipts. On the other hand, if a first reduction in receipts fails to reduce the coverage expected in the future, then further reductions would be necessary.

    Reducing the formula for Social Security receipts is simple, but we should say a final word about the possible necessity of raising receipts at some point in the future. The reality is that the Social Security tax level has risen in increments since 1937 from 1% for each participant to the current 6.2%. As late as 1983 the tax was just over 5.4%, so it has increased by less than 1% in the past 20 years. A potential further increase on the same scale, if it were to become necessary, hardly represents a radical expansion of costs for a program that pays out every dollar to its participants. For employers, a comparable increase in their Social Security contribution, relative to the growth of business taxes in general, would be no more radical.

    There is another factor in the Social Security receipts formula that could also be revised, if and when necessary. That is the maximum amount of earnings on which Social Security taxes are levied. The maximum covered earnings has also risen in increments over the past 25 years, from $15,000 to $85,000. Current law automatically imposes further annual increases, depending on rises in wages generally. But an additional one-time increase at some point in the future would again not represent a radical expansion of costs for Social Security.

    Indeed, an extra jump in the maximum earnings subject to Social Security taxes would actually move an important ratio closer to what it was 20 years ago. The ratio between total wages and the portion of wages on which Social Security taxes are collected has been falling since 1983. "The ratio of taxable to covered earnings decreased from about 90.2 percent in 1983 to 87.9 percent in 1994. . . . The ratio is estimated to have fallen further to 84.4 percent in 1999, and to 83.4 percent in 2000, due mainly to the very high wage earners capturing a greater proportion of total wages."(43)

    Will any of these methods of raising Social Security receipts be needed in the future? The simple truth is that no one knows. No politician, no economist, no expert knows whether Social Security will be short of money or overwhelmed with too much money in the years ahead. What is certain is that short-term forecasts have a far greater chance of being right than long-term forecasts. And short-term forecasts show us that Social Security is financially healthy today.

    The data provided to the public by the Trustees of the OASDI Trust Funds give participants today every reason for satisfaction with the system's past performance. They also call for confidence in the future.

    And for less hubbub from politicians and experts.


    The American Economic Review. Vol. 90, No. 2, May 2000, contains a series of papers under the heading, "Social Security's Long-Term Funding Crisis." Vol. 92, No. 2, May 2002, contains a series of papers under the heading, "Social Security Reform."

    Edward D. Berkowitz, America's Welfare State: From Roosevelt to Reagan, The Johns Hopkins University Press, Baltimore, 1991.

    The Board of Trustees of the Federal Old-Age and Survivors Insurance and Disability Insurance Trust Funds, The 2002 Annual Report, U.S. Government Printing Office, Washington, 2002.

    Encyclopedia of American History, Harper & Row, New York, Sixth Edition, 1982.

    William E. Leuchtenburg, Franklin D. Roosevelt and the New Deal, 1932-1940, Harper & Row, New York, 1963.

    Gerald D. Nash, Noel H. Pugach, Richard F. Tomasson, eds., Social Security: The First Half-Century, University of New Mexico Press, Albuquerque, 1988.

    Frances Perkins, "The Roots of Social Security," Speech on October 23, 1962. Available at

    Steven A. Sass, Robert K. Triest, eds., Social Security Reform: Links to Saving, Investment, and Growth, Federal Reserve Bank of Boston, 1997.

    Sylvester J. Schieber & John B. Shoven, The Real Deal: The History and Future of Social Security, Yale University Press, New Haven, 1999.

    Arthur M. Schlesinger, Jr., The Age of Roosevelt: The Coming of The New Deal, Houghton Mifflin Company, Boston, 1959.

    The Social Security Administration. A comprehensive electronic library of legal and statistical information on the Social Security Trust Funds is published and regularly updated at


    1. 2002 Retirement Confidence Survey, Mathew Greenwald & Associates, Inc., Washington D.C., with the Employee Benefit Research Institute and the American Savings Education Council.
    2. Data in this section is taken from the tables at and /table4a3.html (case sensitive).
    3. The Board of Trustees, The 2002 Annual Report, p. 3.
    4. Quotations and data in this section are taken from The Board of Trustees, The 2002 Annual Report, pages 3 and following.
    5. Trustees, p. 34.
    6. Cathy E. Minehan, President, Federal Reserve Bank of Boston, in Sass and Triest, Social Security Reform, p. ix.
    7. Trustees, p. 7.
    8. Schlesinger, The Age of Roosevelt, p. 263.
    9. Gerald Nash and Richard Tomasson, in Nash, Pugash, and Tomasson, eds., Social Security, p. 7.
    10. Schlesinger, p. 263.
    11. Ibid., p. 294.
    12. Nash and Tomasson, p. 3.
    13. Berkowitz, America's Welfare State, p. 23.
    14. Leuchtenburg, Franklin D. Roosevelt and the New Deal, p. 132.
    15. Ibid., p. 133.
    16. Nash and Tomasson, p. 9
    17. Ibid., p. 14.
    18. Berkowitz, in Sass and Triest, p. 22.
    19. Ibid.
    20. Encyclopedia of American History, p. 420.
    21. Nash and Tomasson, p. 14.
    22. Berkowitz, in Sass and Triest, p. 27.
    23. Berkowitz, America's Welfare State, p. 64.
    24. Quoted in Schieber & Shoven, The Real Deal, p. 330.
    25. See "Key Questions Voters Should Ask Candidates About the Budget, Social Security and Medicare: Election 2002," on The Concord Coalition web site,
    26. See the web site at:
    27. E. J. Dionne Jr., in The Washington Post National Weekly Edition, May 6-12, 2002, p. 27.
    28. Data in this section are taken from the Social Security web site and from the Trustees' 2002 Report, especially pages 23 and 27.
    29. American Economic Review, May 2000, pages 289 and following, and May 2002, pages 390 and following.
    30. Sass and Triest, page 2.
    31. Ibid., page 29.
    32. See
    33. Trustees, p. 128.
    34. Ibid., p. 6.
    35. Ibid., p. 73.
    36. Diane J. Macunovich, in Sass and Triest, p. 66 and p. 73.
    37. Trustees, p. 8.
    38. Ibid.
    39. The Concord Coalition, "Key Questions."
    40. See Figure II.D4, in Trustees, p. 15.
    41. Ibid.
    42. Trustees, p. 42.
    43. Ibid., p. 107.

  •    For comments and questions,
       please e-mail the editor at