Federal Reserve Board
Testimony of Chairman Alan Greenspan
Before the Committee on the Budget, U.S. Senate
January 28, 1999
Mr. Chairman and other members of the committee, as you requested in your letter of invitation, today I will be addressing one of our nation's most pressing public policy challenges: social security.
The dramatic increase in the ratio of retirees to workers that is projected, as the baby boomers move to retirement and enjoy ever greater longevity, makes our current pay-as-you-go social security system unsustainable. Furthermore, the broad support for social security appears destined to fade as the implications of its current form of financing become increasingly apparent. To date, with the ratio of retirees to workers having been relatively low, workers have not considered it a burden to share the goods and services they produce with retirees. The rising birth rate after World War II, which, in due course, lowered the ratio of retirees to workers, helped make the social security program exceptionally popular, even among those paying the taxes to support it.
Indeed, workers perceived it to be a good investment for their own retirement. For those born before World War II, the annuity value of benefits on retirement far exceeded the cumulative sum at the time of retirement of contributions by the worker and his or her employer, plus interest. For example, the implicit real rate of return has been strikingly high for those born in 1920--on average, near 6 percent. The real interest rate on U.S. Treasury bonds, by comparison, has generally been below 3 percent.
But, births flattened after the baby boom, and life expectancy beyond age sixty-five continued to rise. Consequently, the ratio of the number of workers contributing to social security to the number of beneficiaries has declined to the point that maintaining the annuity value of benefits on retirement at a level well in excess of accumulated contributions has become increasingly unlikely. Those born in 1960, for example, are currently calculated to receive a real rate of return, on average, of less than 2 percent on their cumulative contributions. Indeed, even these low rates of return for more recent cohorts likely are being overestimated, because they are based on current law taxes and benefits. In all likelihood, these taxes will have to be raised, or benefits cut, given that the system as a whole is still significantly underfunded, at least according to the intermediate projections of the Old-Age and Survivors Insurance (OASI) actuaries. For the present value of current law benefits over the next 75 years to be fully funded through contributions, social security taxes would have to be raised by about 2.2 percent of taxable payroll; to be fully funded in perpetuity, that is, to ensure that taxes and interest income will always be sufficient to pay benefits, social security taxes would have to be raised much more--perhaps something on the order of between 4 and 5 percent of taxable payroll.
This issue of funding underscores the critical elements in the forthcoming debate on social security reform, because it focuses on the core of any retirement system, private or public. Simply put, enough resources must be set aside over a lifetime of work to fund retirement consumption. At the most rudimentary level, one could envision households saving by actually storing goods purchased during their working years for consumption during retirement. Even better, the resources that would have otherwise gone into the stored goods could be diverted to the production of new capital assets, which would, cumulatively, over a working lifetime, produce an even greater quantity of goods and services to be consumed in retirement.
In this light, increasing our national saving is critical. The President's approach to social security reform supports a large unified budget surplus. This is a major step in the right direction in that it would ensure that the current rise in government's positive contribution to national saving is sustained. The large surpluses projected over the next 15 years, if they actually materialize, can significantly reduce the fiscal pressures created by our changing demographics.
To maximize the benefits of this increased saving, it is crucial that the saving is put to its best use. For productivity and standards of living to grow, financial capital raised in markets or generated from internal cash flow from existing plant and equipment must be continuously directed by firms to its most profitable uses--namely new physical capital facilities perceived as the most efficient in serving consumers' multiple preferences. It is this continuous churning, this so-called creative destruction, that has become so essential to the effective deployment of advanced technologies by this country over recent decades. Indeed, improved productivity of capital probably explains much of why the American economy has done so well despite our comparatively low national saving rate. In addition, the profitability of investment in the United States has attracted saving from abroad which has facilitated the expansion and modernization of our capital stock. Clearly, we use both domestic saving and imported financial capital in a highly efficient manner, apparently more efficiently than many, if not most, other major industrial countries.
Looking forward, the effective application of our capital to its most highly valued use is going to become, if anything, more important, as we strive to increase the resources available to provide for the retirement of the baby boomers without, in the future, significantly reducing the consumption of workers.
Investing a portion of the social security trust fund assets in equities, as the Administration and others have proposed, would arguably put at risk the efficiency of our capital markets and thus, our economy. Even with Herculean efforts, I doubt if it would be feasible to insulate, over the long run, the trust funds from political pressures--direct and indirect--to allocate capital to less than its most productive use.
The experience of public pension funds seems to bear this out. Although relevant comparisons to private plans are difficult to construct, there is evidence that the average rate of return on state and local pension funds tends to be lower than the return realized on comparable private pension funds, other pooled investments, and market indexes. Of course, a significant part of this disparity would be eliminated were these returns adjusted for risk, because public pension plans are often invested more conservatively than private plans. But there is evidence that returns are lower even after accounting for differences in the portfolio allocation between stock and bonds. For example, it has been shown that state pensions plans that are required to direct a portion of their investments in-state and those that make "economically targeted investments" experience lower returns as a result. Similarly, there is evidence that suggests that, the greater the proportion of trustees who are political appointees, the lower the rate of return. A lower risk-adjusted rate of return on financial assets is almost invariably an indication of lower rates of return on the real underlying assets on which they are a claim.
Some have argued that the Federal government has already shown itself capable of investing in equities without political interference, and I have no doubt that the investments of the $60 billion Federal thrift plan and the $6 billion Federal Reserve retirement plan have been made independently. Moreover, the Federal thrift plan has not been an attraction since it is a defined contribution plan, and therefore effectively self-policed by individual contributors. These plans do not reach the asset size threshold to engage the political establishment--but that would not be the case for a multi-trillion dollar social security trust fund. A trust fund invested in U.S. Treasury securities does not appear to be available for politically supported private projects. A fund that can own equity would.
This issue is of particular concern when the pension plan provides a defined benefit. Under these circumstances, the beneficiaries' returns are government guaranteed, and hence they have no incentive to monitor the performance of their invested funds. In sum, because I do not believe that it is politically feasible to insulate such huge funds from governmental influence, investing social security trust fund assets in equities compromises the efficient allocation of our capital--which, as the past few years have demonstrated, is so essential to raising our standards of living.
This risk might be worth taking if having the social security trust fund invest in equities provided real benefits to households. But this is not likely to be the case on average. Having the trust fund invest in private securities most likely will increase its rate of return, although perhaps not on a properly risk-adjusted basis. But, as I have argued previously before this committee, unless new savings are created in the process, the corporate securities that displace Treasury securities in the social security trust funds must be exactly offset by the mirror image displacement of corporate securities by government securities in private portfolios, probably largely in private funds held for retirement. This swap is essentially a zero sum game. To a first approximation, aggregate retirement resources--from both social security and private funds--do not change.
The crucial retirement funding issues center on how to increase the amount of saving, and how to allocate resources between active workers and retirees. It may turn out that the additional new resource requirements, whether mandated savings or additional taxes, to fully fund current benefit levels will prove too burdensome, particularly once future Medicare benefits are accounted for. If so, the level of retirement benefits, funded through social security or private retirement accounts, that is affordable in our economy will remain an important issue. There have been extensive discussions of potential changes, such as extending the age of full retirement benefit entitlement, altering the benefit calculation bend points, and adjusting annual cost-of-living escalation to a more accurate measure. Considerations such as these should not be taken off the table.
While a sharp rise in the number of retirees in about ten years seems almost a certainty, the financial and economic state of the American economy in the early twenty-first century is not. We cannot confidently project large surpluses in our unified budget over the next fifteen years, given the inherent uncertainties of budget forecasting. How can we ignore the fact that virtually all forecasts of the budget balance have been wide of the mark in recent years? For example, as recently as February 1997, OMB projected a deficit for fiscal year 1998 of $121 billion--a $191 billion error. The CBO and others made similar errors. Likewise, in 1983, we confidently projected a solvent social security trust fund through 2057. Our latest estimate with few changes in the program is 2032.
It is possible, as some maintain, that the OASI actuaries are too conservative, and that productivity growth could be far greater than is anticipated in their "intermediate" estimate. If that is, in fact, our prospect, the social security system is not in as much jeopardy as it currently appears. But proper fiscal planning requires that consequences of mistakes in all directions be evaluated. If we move now to shore up the social security program, or replace it, in part or in whole, with a private system, and subsequently find that we had been too pessimistic in our projections, the costs to our society would be few. If we assume more optimistic scenarios and they prove wrong, the imbalances could become overwhelming, and finding a solution would be even more divisive than today's problem.