Henry J. Aaron, Alan S. Blinder, Alicia H. Munnell, and Peter R. Orszag1
October 16, 2000
This report updates our earlier analysis of Governor Bush’s individual account proposal to reflect two recent clarifications.2 First, Governor Bush has indicated that he would finance his individual account proposal out of the Social Security surplus rather than with general revenue transfers.3 Second, according to Bush advisors, the Bush plan may require borrowing by the Social Security Trust Fund (in the form of what the advisors call a “bridge loan” over a span of decades).4
This note highlights two key points:
First, using revenue from the Social Security surplus to finance individual accounts would require cuts of more than 50 percent in traditional Social Security benefits for younger workers. Furthermore, younger single workers would end up with average total retirement benefits (including the reduced traditional Social Security benefits and the expected income from the individual account) about 20 percent below current-law Social Security benefits.
Second, allowing the Trust Fund to borrow for several decades has no effect — positive or negative — on Social Security’s long-run solvency if the loans must be repaid with interest. Such transactions therefore have no effect on Social Security’s ability to pay benefits in the long run. The reason is elementary: In present value terms, the loan proceeds received by the Trust Fund are exactly offset by the repayments it must make.5 Since any borrowing must eventually be repaid with interest, even massive borrowing by the Trust Fund would have virtually no effect on the required benefit reductions within the traditional Social Security system.6
Financing the individual account proposal out of the Social Security surplus
Using the Social Security surplus to finance individual accounts was the primary focus of our original paper.7 By itself, using the Social Security surplus to finance deposits into individual accounts exacerbates Social Security’s projected long-term deficit. The reason is simply that funds that would have been credited to the Social Security Trust Fund are instead diverted into individual accounts, thereby reducing the balance in the Trust Fund and worsening the long-term deficit in Social Security. For example, if individual account contributions of two percentage points of payroll started in 2002, and if no other actions were taken, the Social Security Trust Fund in 2010 would be $1.2 trillion lower than under current-law projections.8
Assuming that the Bush plan does not rely on an accounting gimmick, using the Social Security surplus to finance deposits into individual accounts necessarily implies only two possible outcomes:
The Social Security Trust Fund would become insolvent by 2023, or
Traditional Social Security benefits would have to be reduced substantially.
Given Governor Bush’s stated opposition to increasing the payroll tax or investing the Social Security Trust Fund in private securities, there are simply no other possibilities.
Since accelerating the date of Trust Fund exhaustion did not appear to us to be politically acceptable at the time, our original analysis focused on the only other logical alternative — cuts in traditional benefits. In particular, we asked how large the cuts in traditional benefits would have to be to offset the effects of the diverted payroll and restore long-term solvency within Social Security. We also examined whether the proceeds of investments in individual accounts could compensate for the benefit cuts within the traditional program. The conclusions of that analysis were:
If the reductions in traditional Social Security benefits were phased in under a reasonable schedule, Social Security benefits would have to be cut by 29 percent for those who will be 50 years old in 2002, and by a much larger 54 percent for those who will be 30 years old or younger, to ensure solvency of the Social Security system over the next 75 years.9
Given these reductions in traditional Social Security benefits, the average total retirement benefit — including the expected income from the individual accounts — would fall by 20 percent (relative to current law) for single average earners who will be 30 years of age in 2002.10
Not only would the average combined benefit be 20 percent lower than under current law, but individuals would be exposed to greater financial market risk: For example, if holders of individual accounts suffer from market returns as low as the worst thirty-five-year period since World War II, the total benefit reduction (including the individual account income) for 30-year-old single average earners would be 37 percent rather than 20 percent. On the other hand, if individual account holders enjoy market returns as good as the best thirty-five years since World War II, income for 30-year-old single average earners would be about the same as under current law.
Benefit cuts for lower earners and for married couples with a stay-at-home spouse would be larger than those for single average earners, while cuts for higher earners would be smaller than those for average earners.
All these conclusions are based on assumptions that, if anything, understate the reductions in benefits that are likely to occur under a Bush-like individual account plan.11
Allowing extended Trust Fund borrowing
Governor Bush’s advisors have now clarified that his individual account plan may require massive borrowing by the Social Security system during a lengthy period in which the Trust Fund is technically insolvent.12 However, allowing for such borrowing does not fundamentally change the results of our analysis because, as we explained above, the loans would have to be repaid with interest. Therefore, such loans do not fundamentally alter Social Security’s long-term financial condition and do not significantly affect the magnitude of the required benefit reductions.
Unlike the general revenue transfers proposed in the Archer-Shaw individual account plan or Vice President Gore’s Social Security plan, Governor Bush’s proposed borrowing would not represent a permanent transfer of general revenue to the Social Security Trust Fund. Instead, the Social Security Trust Fund would borrow from the general fund in order to pay benefits for a period of time. The Social Security Trust Fund would then later repay this borrowing with interest, generally after several decades. During the borrowing period, the Social Security Trust Fund would have a negative net financial balance for the first time in its history.13
Allowing the Trust Fund to borrow for a protracted period would not obviate the need for reductions in guaranteed benefits. Diverting two percent of payroll into individual accounts, by itself, would exhaust the Social Security Trust Fund by 2023. If Social Security started taking out a “bridge loan” from this date — and if traditional benefits were never cut — the Trust Fund would fall further and further into debt with no prospect of repaying the loan.
Significant reductions in traditional benefits are inescapable even if the Trust Fund borrows for extended periods, as several individual account proposals illustrate. For example, although the individual account plan proposed by Representative John Kasich includes a substantial “bridge loan” it still cuts guaranteed benefits by more than 50 percent for workers retiring after 2070.14 In addition, Martin Feldstein of Harvard University and Andrew Samwick of Dartmouth University have proposed a plan that involves borrowing by the Social Security Trust Fund between 2031 and 2052. However, their plan also involves a permanent general revenue transfer to the Social Security Trust Fund, with a net present value of $3 trillion between 2001 and 2075. That permanent general revenue transfer is the key factor that allows Feldstein and Samwick to avoid reductions in overall benefits (including the income from the individual accounts) relative to current law.15 Without the $3 trillion in general revenue transfers, Feldstein and Samwick’s plan would also have to reduce traditional benefits substantially, despite the borrowing by the Trust Fund. (As noted above, Governor Bush has apparently ruled out the large general revenue transfers embodied in the Feldstein-Samwick plan. His approach would therefore require significant traditional benefit reductions even if the Trust Fund were allowed to borrow temporarily.) The principle is clear: permanent transfers of general revenue improve the financial condition of the Trust Fund; temporary borrowing does not.
The analysis in our original paper can easily be extended to explore the impact of Trust Fund borrowing. The results suggest that a “bridge loan” would do little to change the magnitude of the required cuts in assured benefits that Social Security provides. For example, we examined a scenario in which Trust Fund borrowing was allowed, but the traditional test of long-run solvency continued to apply.16 In this scenario, the Trust Fund would begin borrowing in 2035 and would be projected to repay the loans, with interest, by 2055. The loan would reach a maximum, in nominal dollars, of $1.7 trillion in 2045 (or more than $350 billion in constant 2000 dollars).
Permitting such borrowing would have only a minor impact on traditional benefit reductions.17 If the Trust Fund were allowed to borrow between 2035 and 2055, the reduction in traditional benefits for those 30 and younger in 2002 could be reduced to 52 percent, rather than 54 percent as in the absence of Trust Fund borrowing. Taking into account the expected income from the individual account for the average worker, the average reduction in total retirement benefits would be 19 percent, rather than 20 percent as in the absence of Trust Fund borrowing (see table).
Impact of Governor Bush’s Individual Account Proposal, With and Without Trust Fund Borrowing, on Single Average Earner (Earning $31,685 in 2000) Retiring at Age 65 in 2037
Without borrowing by Trust Fund
With borrowing by Trust Fund
Current-law Social Security benefit
Required cut in Social Security benefit (as % of current=law benefit)
Average individual account
Total retirment benefit
Percentage change relative to current law
This note amplifies our earlier analysis of Governor Bush’s individual account proposal by highlighting two issues.
First, financing individual accounts out of the Social Security surplus, as Governor Bush has now indicated he would do, would require significant traditional benefit reductions. For workers aged 30 or younger in 2002, the traditional benefit reduction required to eliminate the long-term deficit within Social Security after funds were diverted into individual accounts would amount to more than 50 percent.
Second, allowing the Trust Fund to borrow (assuming the loan must be repaid with interest) does not significantly affect the scale of the required traditional benefit reductions. If two percentage points of payroll were diverted into individual accounts, reductions in traditional benefits of more than 50 percent would still be necessary to restore solvency within Social Security even if the Trust Fund were allowed to borrow over a span of two decades.
1Henry Aaron is the Bruce and Virginia MacLaury Senior Fellow at the Brookings Institution. Alan Blinder is the Gordon S. Rentschler Memorial Professor of Economics at Princeton University. Alicia Munnell is the Peter F. Drucker Professor of Management Sciences at Boston College. Peter Orszag is the President of Sebago Associates, Inc.
2Our earlier paper, “Governor Bush’s Individual Account Proposal: Implications for Retirement Benefits,” Issue Brief No. 11, The Century Foundation, June 2000, is available at http://www.tcf.org.
3The following exchange is taken from the New York Times transcript of the October 3 debate: “MR. GORE: If you cut the amount going in, one out of every six dollars, then you have to cut the value of each check by one out of every six dollars unless you come up with the money from somewhere else. I would like to know from the governor…does that trillion dollars come from the trust fund or does it come from the rest of the budget? MR. BUSH: No. There’s enough money to pay seniors today and the current affairs of Social Security. The trillion comes from the surplus.” In addition, and consistent with our interpretation of the Bush position, the budget released by the Bush campaign does not include any general revenue transfers to Social Security.
4Glenn Kessler, “Bush Walks a Fine Line on Social Security,” Washington Post, September 28, 2000, page E1, and Jacob Schlesinger, “Bush Evades Inevitable Choices in Social Security Plan,” Wall Street Journal, October 10, 2000, page A28.
5This statement assumes that the interest rate the Trust Fund pays on its borrowing is equal to the interest rate used in Social Security actuarial calculations (which itself is equal to the interest rate earned by the Trust Fund on its assets while those assets remain positive).
6As explained below, allowing the Trust Fund to borrow would have a small effect on the required benefit reduction rate if benefit “notches” were not allowed among younger workers (i.e., if the benefit reduction rate were required to be the same for all workers aged 30 or younger in 2002).
7In this update, as in our original analysis, we assume that the individual account contributions are equal to two percentage points of payroll. This assumption is consistent with Secretary Cheney’s discussion of the Bush-Cheney individual account proposal in the October 5 debate. He stated, “The reform we would like to offer is to allow our young people to take a portion of the payroll tax, 2 percent of it, and invest it in a personal retirement account.” A two percent-of-payroll contribution rate is also consistent with examples used by Governor Bush and other examples on the Bush campaign web site. (If the contribution rate were larger, the required traditional benefit reductions would also be larger than calculated in our original paper.) In addition, in the absence of further information about the nature of the voluntary choice offered by Governor Bush, we continue to assume that everyone will participate in the accounts. That assumption is consistent with the trillion dollar cost estimate Governor Bush cited in the first Presidential debate and with cost estimates that have been reported in various news accounts.
8Under the same assumptions, the Social Security surplus excluding interest on the Trust Fund would be eliminated in 2005, and the Social Security surplus including interest on the Trust Fund would be eliminated in 2014. We assume that the contributions to the individual accounts would continue even after the Social Security surplus had been eliminated. It is therefore more precise to state that the contributions would be financed out of existing payroll tax revenue, rather than out of the “Social Security surplus.” Nonetheless, we adopt the latter terminology because it reflects what appears to be the stated position of the Bush campaign.
9If the cuts were equi-proportionate for all generations, the reduction in traditional benefits would be 41 percent. That approach, however, would involve a very uneven pattern of reductions in total retirement benefits: Older workers, who have less time to accumulate individual account balances before retirement, would suffer a much sharper reduction in total retirement benefits than younger workers.
10This 20 percent reduction in total retirement benefits is essentially the same as the cut required to restore Social Security solvency over the next seventy-five years without individual accounts. In contrast, plans that employ general revenue transfers or broader Trust Fund investment options to address part of the long-term financing problem would require smaller benefit cuts.
11Our earlier paper discusses in detail how our assumptions, if anything, understate the reductions in benefits.
12Our original analysis discussed this possibility only briefly in a footnote, since we assumed that an extended period of technical insolvency would not be allowed to occur.
13The OASI Trust Fund borrowed from the DI and HI Trust Funds on a one-time, emergency basis in 1982. The borrowing was repaid by 1986. At no time, however, was the net financial balance of the combined OASDI Trust Funds negative.
14Stephen Goss, “Estimated Long-Range OASDI Financial Effect of Proposal by Representative John Kasich,” Office of the Chief Actuary, Social Security Administration, June 14, 1999.
15Transfers of $3 trillion in net present value without individual accounts would be sufficient to maintain Social Security solvency for 75 years without traditional benefit reductions.
16Long-run solvency is defined as having a Trust Fund ratio of 100 percent at the end of 2074 (i.e., the Trust Fund is sufficient to cover full benefits in 2075).
17Indeed, the only reason the borrowing has any effect on the required benefit reductions in our analysis is that we assume that an individual account proposal would not allow benefit “notches” among younger workers. In particular, we require that the benefit reduction rate for all workers 30 and younger in 2002 would not vary by age, so that a 29-year-old worker could not enjoy a smaller benefit reduction than a 30-year-old worker. The interaction of this assumption and the ability of the Trust Fund to borrow produces a small change (instead of no change) in the required benefit reductions for 30-year-old workers.