February 9, 1999
TESTIMONY OF ROBERT GREENSTEIN
Center on Budget and Policy Priorities
before the Senate Finance Committee
February 9, 1999
Table of Contents
I. Is There Justification for General Revenue Transfers?
II. Concerns with General Revenue Transfers
III. Courses of Action
I appreciate your invitation to testify on the subject of general revenue financing of Social Security. I am Robert Greenstein, executive director of the Center on Budget and Policy Priorities. The Center is a nonpartisan, nonprofit policy institute that conducts research and analysis on a wide range of issues with particular emphasis on fiscal policy and on issues affecting low- and moderate-income families. We are primarily funded by foundations and receive no federal funding.
The question of whether Social Security should be partly funded with general revenue contributions is an important one. Several types of general revenue contributions to Social Security already exist, including funds transferred from general revenues to reflect the partial taxation of Social Security benefits. So the issue is not entirely new. On the other hand, the President's budget proposes significantly larger general revenue transfers than have existed until now.
In the past 10 days, the issue of whether to make general revenue transfers has become enmeshed with a debate over whether the Administration's proposed revenue transfer rests on an illegitimate double-counting of funds. This is unfortunate, as it is diverting the debate from the serious policy issues the proposed general revenue transfers raise and leading to a focus on much less important technical accounting issues.
So, let me briefly try to set the double-counting issue to the side so as to get to the more serious issues. What has been called double-counting in the President's budget is essentially unified budget accounting, a practice the federal budget has followed for many years. Under unified budget accounting, the Treasury takes the surplus payroll tax revenues the Social Security trust fund is receiving each year and provides the trust fund with Treasury bonds in return. The Treasury then uses the borrowed cash to fund other government operations.
In a sense, these funds show up in the budget twice once as Treasury bonds provided to the trust fund and a second time as expenditures for other government programs. (The tax bill that the House passed last summer followed a similar course, except that surplus Social Security revenues would have been used to finance a tax cut rather than other government operations.)
The Administration's proposal follows along similar lines. The Treasury would borrow the surplus Social Security payroll tax receipts and provide the trust fund with Treasury bonds in return. The only new element is that instead of using the borrowed money to fund other government programs or finance tax cuts, the funds would be put back in the Social Security trust fund and used primarily to buy more bonds for the fund.
All of these approaches entail borrowing funds from the Social Security surplus, crediting the trust fund for the borrowed funds by issuing it Treasury bonds, and using the borrowed funds for a second purpose. That under the Administration's proposal the second use is Social Security rather than another government program or a tax cut does not alter the basic nature of the transaction. If one wants to consider this double-counting, then every budget passed for years including budgets proposed and approved by Presidents and Congresses of both parties have used "double-counting." In all of these budgets, Social Security surpluses were credited to Social Security and also borrowed and used to fund other government operations.
What has led to confusion regarding the Administration's proposal is that in this case, the "second use" of the funds is the same as the first use, namely Social Security. But borrowing funds from the Social Security trust fund in return for Treasury bonds and using the borrowed funds is essentially the same transaction regardless of how the borrowed funds are used and whether they are designated for support of Social Security, another program, or a tax cut. (In fact, placing the borrowed funds in Social Security reserves rather than spending them on other programs or tax cuts should be a more benign use of the borrowed funds, since it does not reduce national saving or prevent the borrowed funds from being used to pay down publicly-held debt, as such other uses of the borrowed funds would do.)
In short, the issue isn't double-counting. It is whether a substantial transfer of general income to the Social Security trust fund, such as that the Administration has proposed, represents good policy.
Is There Justification for General Revenue Transfers?
I believe the answer to this question is that there is justification for transfers, but not on an unlimited basis.
- In Social Security's early years, its designers faced a difficult question should those already retired or nearing retirement age be able to receive benefits? Since the program was in its infancy, these individuals contributed little or nothing to Social Security during their working years. But many of them, including workers who had endured the Depression and fought for the nation in World War I would otherwise face poverty in old age.
Policymakers of that era made the humane decision; they decided to provide, rather than deny, Social Security to these individuals. That decision meant Social Security would primarily be a pay-as-you-go system, with current payroll tax revenues used to fund the benefits of current retirees, rather than a pre-funded system. The establishment of Social Security largely as a pay-as-you-go system also meant these would be an unfunded liability and that when a demographically large generation retired, such as the baby boom generation, financial pressures on the pay-as-you-go system would intensify.
The decision made 60 years ago to provide benefits to retirees of that era who had not paid much into the Social Security system provides a strong justification for a limited-duration infusion of general fund revenue into Social Security today. It makes sense to "reimburse" the Social Security system in some form for bearing the costs of providing benefits to earlier generations of beneficiaries who had paid little into the system because the system was new. That early decision is one of the reasons Social Security faces insolvency in the decades ahead. A general fund transfer of this nature would help reduce the unfunded liability.
- Social Security surpluses are now adding substantially to national saving. Because Social Security is able to purchase so many Treasury bonds, other investors can hold fewer bonds and invest more money in equities, securing the higher average rates of return that equities provide. Robert Reischauer and Henry Aaron of the Brookings Institution have suggested that because Social Security is adding to national saving in this manner, the trust fund should be able to receive its fair share of the higher rates of return that equities provide. They propose this be done by diversifying the trust fund's investments and ultimately placing up to half of trust-fund reserves in equities. This is roughly the same share of reserves as are placed in equities by corporate pension plans and state and local public employee pension funds. The Administration's proposal is much more cautious, placing about 15 percent of trust funds reserves in equities.
To the extent that policymakers are not willing to invest up to 50 percent of trust fund reserves in equities, there is a strong case for a general revenue transfers to compensate the trust fund for the lost income. To the extent that Social Security is artificially confined to lower returns by being restricted to investment in lower-yielding Treasury bonds and private investors are able to secure higher returns general revenue collections will be higher, since investors will pay taxes on the higher returns they secure. A case can be made for transferring a portion of these added general revenues to Social Security.
- If Social Security reserves are saved rather than borrowed and consumed, they will help pay down the debt and add substantially to national saving. Treasury estimates are that the national saving rate would rise significantly under the Administration's plan. That should promote somewhat stronger economic growth. Martin Feldstein and Senator Gramm have argued that the increased saving they estimate their Social Security plans would generate would result in increased corporate income tax revenues. If this is true for their plans, it also would be true for the Administration's. A case can be made that the Social Security trust funds should share, through a general revenue transfer, in the benefits of the stronger economic growth and increased tax revenues made possible by saving the Social Security surpluses.
- Finally, the Administration has said the general revenue transfer it is proposing would be structured in such a way as to "lock-up" budget surpluses so they are saved and used primarily to pay down debt, rather than being available to be consumed by program expansions or tax reductions. Accomplishing this requires a budget process change so the transfers are "scored" as expenditures and placed in a lock-box. (The details on this aspect of the Administration's proposal have not yet been provided.)
Concerns with General Revenue Transfers
Having said this, there also is a significant concern there could be a temptation to rely largely or entirely on general revenue transfers to close Social Security's funding gap.
I believe this would be unwise. If little or no change is made in the Social Security benefits-and-tax structure to help close its long-term financing gap and if the gap is closed solely through general fund transfers the gap in future decades between benefit costs and payroll tax revenues will be too large. The consequent burden on the rest of the budget will be too great. This would not be a prudent course.
I also am concerned by what would occur if Congress were to adopt the part of the President's plan that calls for general fund transfers, but Congress and the President failed to close the other half of Social Security's long-term financing gap through changes in the Social Security benefit-and-tax structure, as the President proposed be done on a bipartisan basis. If we do only the transfer, the gap between benefits and payroll tax revenues will be too great in the 2030's and 2040's, and insolvency will hit in 2055. In my view, the Administration's proposed transfer makes sense in conjunction with structural changes to close the other half of the gap. If the needed structural changes are made, the trust funds will not need to cash in as many Treasury bonds in the 2030's, 2040's and beyond, and the burden on the rest of the government will be much more bearable, especially in light of the substantial anticipated budgetary savings on interest payments on the debt. (In FY 1998, the cost of Social Security benefits plus net interest payments on the debt equaled 7.4 percent of GDP. If we save most of the surplus, interest costs will decline to such a degree that under the Administration's plan, the combined costs of Social Security and net interest payments are projected to remain below 7.4 percent of GDP when the baby boom generation is retired.)
I would note that just as trying to close all of the gap through general fund transfers would be unwise and structural changes are needed, closing the gap mostly or entirely through benefits reductions appears too harsh. Aaron and Reischauer have noted that Social Security benefit levels are more modest than is sometimes appreciated. They note that "A worker retiring in 1996 after a lifetime of year-round work at the minimum wage received a [Social Security benefit] that was slightly under the poverty threshold for a single person; a minimum-wage married retiree received a benefit that was just above the poverty threshold for a couple. Benefits of average earners are less than 1.5 times the poverty threshold if they start drawing benefits at age 65, and are only 16 percent over the poverty threshold if they start benefits at age 62."(1) Aaron and Reischauer also note that for workers with average earnings, the portion of earnings that Social Security benefits replace is less than two-thirds the portion of earnings that is replaced in the French, Dutch, Belgian, Italian, German, and Spanish systems.
Combining a temporary general revenue transfer with structural changes to close the other half of the 75-year gap and keep Social Security solvent after that would strike an appropriate balance.
Finally, we need to maintain the trust fund aspect of Social Security, with a dedicated funding mechanism. Trust fund accounting is important. If we did not have the actuarial estimates projecting an imbalance between trust fund income and trust fund expenditures 33 years from now, we probably would not be having this debate today.
Courses of Action
This suggests several courses of action.
- Consider closing up to half of the current long-term financing gap through general revenue transfers, but not much more than that. It would be preferable if these transfers are temporary, rather than permanent.
- General revenue transfers ought to be accompanied by structural changes and perhaps used as incentives or inducements to get agreement on structural changes (e.g., you get a dollar or some other amount in transfers for each dollar in structural change). I would like to see transfers used in a way that enhances the political system's ability also to make structural changes. Transfers should be in addition to structural changes, not instead of making any structural changes.
- General rules should probably be developed to set parameters on the size of general fund transfers allowed. For example, to the extent that such transfers do not come from on-budget surpluses, they might be limited to the amounts needed to compensate the trust fund for losses due to the investment of less than 50 percent of Social Security reserves in equities. An exception to this limitation is that if and when sufficient progress is made in restoring Medicare solvency, it would be appropriate to transfer to the Social Security trust fund the portion of revenues from the taxation of Social Security benefits that is currently deposited in the Medicare trust fund.
- To enhance fiscal discipline and apply pressure for necessary Social Security reforms, the current pay-as-you-go rules should be maintained until Social Security is in long-term (i.e., 75-year) balance. These rules should be lifted when 75-year balance is secured to the extent there is an on-budget surplus.
We also would like to see the Senate adopt section 13302 of the Budget Act, which now applies only to the House. This rule says that any bill or amendment that weakens the solvency of the Social Security trust fund on a 5-year or 25-year basis should not be considered. (The Senate has an alternative procedure that makes this rule operative for 10 years, but not 75 years.) This rule could be designed to create a point of order against legislation that would violate the rule, with the votes of 60 Senators required to waive the point of order.
The goals of saving rather than consuming most of the surplus, paying down the debt, building national saving, and shoring up Social Security for the long term are the most important domestic policy goals we can pursue. The Administration has designed a plan that uses general revenue transfers as an integral part of advancing these goals; these transfers would be used as part of a plan designed to help ensure surpluses are saved rather than spent and to help shore up Social Security. Overall, the plan would use 77 percent of the unified budget surplus to pay down the debt held by the public or otherwise improve the government's assets-and-liabilities position. It would devote more than 80 percent of the unified budget surplus to building saving rather than for current consumption.
I believe this is a sound course to follow. If we follow this course, we will need to place appropriate limits on the use of general fund transfers so we do not rely on them excessively, and we will need to marry general revenue transfers to structural changes.
1. Henry J. Aaron and Robert D. Reischauer, Countdown to Reform: The Great Social Security Debate, pp. 93-94.
Additional federal budget and income security analyses.