Skip Navigation

Mr. Chairman and Members of the Committee, we are pleased to be here today to discuss the framework for Social Security in the President's budget.

The President's budget is, and has always been, a policy document. Budget folklore holds that President Lyndon Johnson created the President's Commission on Budget Concepts in part to mask the costs of the Vietnam War. Creating the unified budget moved the Social Security trust funds and their surpluses on-budget. The most recent President's budget continues this long-standing, bipartisan tradition--observed by the executive branch and the Congress alike--of using trust fund accounting to facilitate policy and political objectives that are often unrelated to the trust funds or the programs with which they are associated. In the process, clarity and ease of understanding are often sacrificed.

Our presentation today is based solely on the President's budget. The Congressional Budget Office (CBO) is now working on a reestimate of the President's budget, the results of which should be available in a few weeks. Our reestimate will change the numbers somewhat, but it will not change much of what we say today.

The President's budget does not contain details on a number of relevant and potentially significant policy proposals that the President or members of his Cabinet have announced or endorsed. For example, the changes in Social Security benefits that the President mentioned in his State of the Union address are included in the text of the budget but are not incorporated into any of the numerical presentations.

Those changes, if made without offsetting reductions in benefits, would result in expenditures from the Social Security trust funds that are greater than current law--that is, greater than the baseline CBO presented to you last month. The same is true of the proposal to add a pharmaceutical benefit to Medicare. A recently announced regulatory change in the Disability Insurance program will also add costs. The Administration has not indicated how the additional obligations would be financed.

The President's budget also lacks details on the full 15-year estimates included in many of the tables. We expect that some of those details will be forthcoming. Today we will take you through the accounting as presented in the budget, discuss some of the history behind those conventions, and offer selected comparisons with current scoring practices and the baseline.

After you have thought about it for a while, you may well conclude, as the President has, that you should change the accounting for this and other trust funds. The connection between taxes and benefits is unique to those programs. Perhaps more important, "accounting" up to now has allowed the impression that there are assets in those trust funds--assets that can be subsequently sold to help defray the cost of benefits. But that accounting ignores what has happened in the rest of budget. On net, we have debt, not assets.
 

PRINCIPLES FOR REFORM

Trust fund accounting is an imperfect means of assessing the viability of a program. It is even less appropriate for assessing the impact of reforms. For a more complete view of the President's proposal and its effects on the real economy, it is necessary to step back and look at a broader range of issues. I will suggest several questions that might guide our analysis of the President's proposal as well as others. Other experts, such as the Comptroller General, have made similar suggestions.

Judging the desirability of reform depends critically on several related questions:

• Can the reform help economic growth?
  
  
• Does the reform improve the long-term fiscal balance of the program?
  
  
• Does the reform enhance equity or fairness?
  
  
• Can the reform reasonably be expected to work?

The first question is critical. It is the size of the economy that ultimately determines the nation's ability to support a growing elderly population with fewer workers. In the macro sense, which mechanism is used to transfer resources from the working population to retirees matters little. What matters most is how much the working population creates--how big the pie is relative to the piece devoted to retirees.

What does that mean in the context of Social Security reform? A point of agreement among many economists involved in the Social Security debate is that increasing national saving raises productivity and increases economic growth. Given the broad agreement among economists that paying down the debt has a positive impact on saving and economic growth, perhaps that should be the standard against which all other proposals are measured. That is, do proposals to increase federal spending, reduce taxes, or purchase equities raise production and economic growth more than paying down the debt?

Few generalities apply, and not all alternatives are equal. The impact of any alternative to paying down the debt would have to be carefully examined. The details matter, as does the current state of the U.S. and international economies. Some programs that appear on the spending side of the federal budget might help productivity, as might a number of tax measures. The analysis must be based on interest rates, for example, that are determined in world markets.

Second, what is the reform's impact on the long-term outlook for the program? The Social Security and Medicare programs face long-term pressures from demographic changes and rising health care costs, although the buoyant outlook over the near term will help delay the onset of serious fiscal problems. The large and rising surpluses projected for the next 10 years will be replaced with mounting pressures as the baby-boom generation begins to draw benefits from Social Security and Medicare, the average life span increases, and the costs per beneficiary of federal health care programs continue to rise. Put another way, does the President's proposal for Social Security actually contain reforms--changes to the program's structure so that it is on sound fiscal footing?

Third, what are the implications of reform for the perceived equity of the program? Admittedly, many concepts and measures could be used to assess fairness. One that is commonly understood in this context is the relationship between what people pay in payroll taxes and what they receive in benefits.

Addressing the fourth question--whether reform can be expected to work--includes considering practicality, ease and cost of administration, protection against severe losses, and the extent of regulation. Is the program feasible? Does the government have the wherewithal to administer it? Will it be subject to increased fraud? Will it involve more invasion of privacy? Can it be insulated from political influence, as Chairman Greenspan noted?
 

PROGRAMMATIC ISSUES

The Administration's framework for Social Security also raises several important questions about the structure of the Social Security program and the federal budget.

First, would breaking the link between payroll taxes and benefits eliminate an important mechanism of program discipline? In the past, the projected depletion of trust fund balances has often provided the impetus for taking painful steps to increase taxes or scale back scheduled benefits. The imminent exhaustion of the Social Security trust funds spurred action in 1983, and shortfalls in the Hospital Insurance Trust Fund served a similar function in 1997. Although such deadlines may be artificial from an economic point of view, they can have real consequences. The President essentially substitutes general fund solutions for programmatic solutions. I understand that the Comptroller General will have more to say about that.

Second, would a massive infusion of nonpayroll taxes in Social Security significantly change the way the program is viewed? Heretofore, using payroll taxes has been considered integral to maintaining Social Security as a social insurance program. The program is financed by a nearly universal tax on earnings, and a person's benefits depend on the earnings on which taxes were paid. The use of general revenues could potentially undercut one or more elements of that carefully balanced system.

Third, is accumulating balances in the Social Security trust funds a potentially effective way of encouraging more saving by the government? Put another way, will these changes in accounting prevent the funds from being used for other purposes? Whether this approach would work for long is open to question. The recent past in the United States and the experience in other countries is not promising in this regard.

Last, is the direct purchase of equities by the federal government appropriate?
 

TRUST FUND ACCOUNTING

However valid the reasons may have been to establish the accounting conventions the federal government currently uses for trust funds, those conventions confuse almost everyone: the Congress, the media, government officials, and most of all, the public.

I believe the main source of that confusion is the fact that the federal government's trust funds are not trust funds in the traditional sense; that is, they do not set aside current income for future use. Excess income over outgo for any given trust fund is invested, in a certain sense, in special Treasury securities, which are as safe and secure as all other Treasury debt. But the Treasury securities held by federal trust funds are nothing more than the government's IOUs to itself. Look at it this way: if the government had truly invested trust fund net income for future use, the Treasury would currently be holding hundreds of billions of dollars of real assets that could be liquidated in the future to pay for future obligations. But the Treasury does not hold any net assets; in fact, all that remains from the so-called investment of trust fund surpluses is net debt to the public of $3.7 trillion.

Although there is no money in the Treasury to pay for future obligations, the obligations to people eligible for Social Security benefits are real. And most important, those obligations are a direct result of federal law, not a consequence of whatever may or may not be credited to the trust funds. In particular, the size of the balances in the Social Security trust funds--be it $2 trillion, $10 trillion, or zero--does not affect the obligations that the federal government has to the program's beneficiaries. Nor does it affect the government's ability to pay those benefits.

This fact is explicitly recognized in the President's budget for fiscal year 2000 in the same words used in previous budgets. To quote page 337 of the Analytical Perspectives volume: "The existence of large trust fund balances, therefore, does not, by itself, have any impact on the government's ability to pay benefits." The fact that trust fund balances are unrelated to the government's obligation or ability to pay benefits needs to be recognized before any proposals to address the Social Security and Medicare trust funds can be analyzed. In other words, look first to the impact of proposals on increasing national saving and raising real growth and then to the impact on paying down the debt held by the public.

Let me apply those principles to the Social Security trust funds. In their most recent report, the Social Security trustees estimate that the trust funds will not be exhausted until 2032. However, the report also includes the fact that starting in 2013, Social Security taxes will not be sufficient to meet obligations. If the Social Security trust funds were trust funds in the traditional sense, their assets could be sold to cover the shortfall. However, as stated above, the surpluses in the trust funds have been loaned to the federal government, and although special bonds have been issued to indemnify the funds, the bonds are nothing more than the federal government's IOUs to itself. Starting in 2013, the program's expenditures will exceed payroll taxes, and the government will eventually have to go further in debt, raise taxes, cut spending, or infuse more general revenues to be able to send out Social Security checks. We must look beyond the balances in the trust funds to be able to properly evaluate any proposal.
 

MAJOR ELEMENTS OF THE ADMINISTRATION'S PROPOSALS

Both the Administration and the Congressional Budget Office project that under current laws and policies, the federal budget will record a total surplus of more than $100 billion in fiscal year 2000 and more than $200 billion in 2004. Excluding the framework for Social Security, the proposals in the President's budget are intended to leave the surplus unaffected. That is, the cost of proposals to increase spending or lower taxes is said to be fully offset by spending cuts or tax increases elsewhere in the budget. Although we will have more to say about that after we have finished reestimating the President's budget, by using only the data available in that budget and including the Social Security framework, the surplus is smaller in all years.

Proposed Budget Surpluses and Reduction of Debt Held by the Public

A primary feature of the Administration's Social Security framework is that it attempts (under scoring by the Office of Management and Budget) to keep the total federal budget in surplus under the traditional accounting. Excluding its Social Security framework, the Administration projects cumulative total budget surpluses of $828 billion over the 2000-2004 period. That is, the federal government will collect a total of $828 billion more from the public than it will spend in transactions with the public. That figure comprises Social Security surpluses of $719 billion, a Postal Service deficit of $5 billion, and on-budget surpluses of $114 billion (see Table 1).
 

In its Social Security framework, the Administration proposes to use $258 billion, or 31 percent, of the projected total surpluses over the next five years for new spending on goods and services ($138 billion), financial assets ($96 billion for the new Universal Savings Accounts), and additional debt-service costs ($24 billion). Taking those items into account, the remaining surplus would total $569 billion. Table 1 shows that the President's proposals clearly lower the debt held by the public by $380 billion in fiscal year 2004 and by more than $2.5 trillion in 2014 relative to where the debt was expected to be at the end of 1999.

But under the Administration's current-services baseline projections, debt held by the public would drop from $3,670 billion at the end of 1999 to $2,927 billion at the end of 2004 and would be totally paid off by about 2014. Under that computation, as Table 2 shows, debt held by the public under the Administration's policies would be $3,290 billion in 2004, or $362 billion higher than it would be if the policies in the budget for fiscal year 2000, including the proposed Social Security framework, were not implemented. So the President's proposals lower the debt relative to where it is now but increase it relative to no new action at all. In addition, as Table 3 shows, relative to saving the Social Security surpluses over the period, the President's proposals increase the debt held by the public--that is, some of the Social Security surpluses are used for non-Social Security spending--by $334 billion in 2004.
 

General Revenue Payments

The second major element of the Administration's framework consists of general revenue payments from the Treasury to the Social Security and Medicare Hospital Insurance trust funds. Over the 2000-2004 period, the Administration would credit an additional $445 billion to Social Security and $124 billion to Hospital Insurance, above and beyond the payroll taxes, interest, and other income that would be credited under current law. Social Security currently receives hardly any general revenues; income taxes on Social Security benefits represent less than 2 percent of the program's income. Medicare's Supplementary Medical Insurance Trust Fund, however, gets three-quarters of its income from general revenues. Even if general revenue payments were made in an amount sufficient to make the Social Security trust funds actuarially sound for everyone over 15 years old today--$8.4 trillion--those payments would have no effect on the surplus or deficit in any year.

Although the Administration describes the proposed general revenue payments as a use of the budget surplus, those payments would not alter the total surplus as traditionally measured. In fact, they would not affect the surplus no matter how large or small they were. General revenue payments are purely intragovernmental--a transaction between one government account and another. The general revenue payments to Social Security would move the government's on-budget accounts from surplus into deficit over the 2000-2004 period, but they and the payments to Medicare would not affect federal transactions with the public and would therefore have no effect on the economy.

The Administration's proposal further confuses the situation by treating the general revenue payments as a reduction in the total budget surplus, although not as a net outlay to the public. That approach can be viewed as an attempt to protect the surplus by making it seem to disappear, but it is not consistent with the principles of federal budgeting that were set forth by the President's Commission on Budget Concepts and that have been followed for the past 30 years.

That approach has, however, been proposed before. In the context of the fiscal year 1991 budget, President Bush asked for a similar change for a similar reason--to "save" the "surplus" of a trust fund. The Congress did not approve the request; the change to budget accounting was criticized by Members of both parties at that time.

Some observers have worried that the proposed general revenue payments, plus interest, would substantially increase gross federal debt and debt subject to statutory limit (see Table 4). That concern, however, is misplaced. The increase in the amount of debt held by the Social Security trust funds would be merely a bookkeeping transaction and would not represent an increase in the net liabilities of the federal government. The government's liability for Social Security and Medicare is the obligation to pay future benefits, and, as stated above, those benefits--and therefore the government's liability--would be unaffected by the proposed payments of general revenues and unaffected by any "balance" in the trust fund.
 

Purchase of Equities

As a third element of its framework, the Administration proposes that one-fifth of the general revenues credited to Social Security be used to purchase corporate equities or other private financial instruments. Like the proposed general revenue contributions, this element of the Administration's framework is designed to increase the balances in the Social Security trust funds, but it, too, would have little economic effect. For each dollar to be invested in equities, the federal government would have to borrow an additional dollar from the public. After the transaction, the private sector would hold fewer equities and more debt, but total national wealth and national saving would not be appreciably affected. Moreover, the technical problems cited by Chairman Greenspan might turn the President's economically neutral proposal into one that could harm the economy.

The CBO staff and I will be happy to provide additional analysis of the Administration's proposal, Mr. Chairman, as more details become available.