Revised May 5, 1999
The Archer-Shaw Social Security Proposal
by Robert Greenstein
Revised Version of this Paper Reflects Actuaries' Analysis
This is a substantially revised version of an analysis on the Archer-Shaw proposal that the Center issued on April 29. Since the time the original analysis was issued, the Social Security actuaries have issued an extensive analysis of the fiscal impacts of the Archer-Shaw plan over the next 75 years. This revised Center paper reflects the actuaries' analysis.
On April 28, Reps. Bill Archer and Clay Shaw unveiled their Social Security proposal. On the positive side, the plan is designed to restore long-term (i.e., 75-year) solvency to the Social Security system and guarantee that beneficiaries would receive the benefit levels to which current Social Security law entitles them. Like the Clinton proposal, the Archer-Shaw plan transfers general revenues into Social Security and neither reduces Social Security benefits nor raises Social Security taxes.
While the plan seeks to restore solvency and guarantee Social Security benefit levels, however, it has several major shortcomings.
- It could place too great a strain on the rest of the budget for much of the next half century. If most or all of the non-Social Security surplus is consumed by tax cuts, as the Congressional budget resolution envisions (or by a combination of tax cuts and upward adjustments in discretionary spending levels, as could result from negotiations between Congress and the Administration), there would be only one place from which the plan's individual accounts could initially be funded the Social Security surpluses. This evidently is what the plan's sponsors have in mind.
After about 2012, however, the plan's costs would exceed the Social Security surplus. For several decades after that, financing the individual accounts the plan would establish would result in substantial problems for the rest of the budget and likely lead to large cuts in other programs, increases in taxes, or budget deficits.
The cost of the plan would be substantial in these years. The plan's costs include not only the cost of depositing funds into the individual accounts but also the cost of higher interest payments on the federal debt. (Higher interest payments would have to be made because large sums would have been deposited in the individual accounts rather than used to pay down the debt.) According to the Social Security actuaries, the net costs of the Archer-Shaw plan the costs of the deposits into individual accounts and the higher interest payments on the debt, minus the savings the plan would produce in Social Security costs would run from $300 billion to $600 billion a year each year from 2016 through 2042.(1)
With the Social Security surpluses no longer able to cover such costs and with little, if any, surplus likely to remain in the non-Social Security budget in these years because the baby boomers will be retiring in large numbers and costs for health care programs and some other expenditures will be mounting accordingly financing the individual accounts is likely to entail substantial program cuts or tax increases if policymakers seek to avoid sizeable deficits. Although a transfer of general revenue to Social Security can be a beneficial part of a plan to restore long-term Social Security solvency, the general revenue transfers the Archer-Shaw plan requires are too large, especially during the years of the baby boom generation's retirement.
Eventually, the Archer-Shaw plan is likely to be self-financing. But according to the actuaries' estimates, that would not occur until about 2050. The plan would carry substantial risk to the rest of the budget in the decades before then.
- The Archer-Shaw plan is structured in a way that renders it inefficient and wasteful. The plan would transfer general revenues to individual accounts that would be managed by Wall Street brokerage firms and other private fund managers and enable these firms to take substantial sums out of the accounts in commissions and management fees, only to have nearly all of the proceeds from these individual accounts then transferred back to the Social Security trust funds to help pay Social Security benefits. Based on the actuaries' assumptions regarding these costs, the administrative and management costs that would be paid on these accounts would total approximately $350 billion over the system's first 30 years, equaling $34 billion a year by 2030 and larger amounts in years after that. It makes little sense to incur costs of this magnitude. If the goal is to provide the Social Security trust funds with additional financing so the trust funds can pay full Social Security benefits for at least 75 years, that can be done much less expensively by providing a general revenue transfer to Social Security and allowing an independent, nonpartisan, professional board to invest a portion of the transferred revenues in index funds in the equities market.
- There is a high degree of risk the plan would lead over time to a substantial weakening of support for Social Security. This is one of the plan's most significant weaknesses over time, it could undermine the system it seeks to save. Under the plan, many middle- and upper-middle-income workers would receive only a modest Social Security benefit, which would equal the difference between the annuity payment from their individual account and the Social Security benefit level to which they are entitled. Social Security would appear to provide little in return for the 12.4 percent of wages these workers and their employers pay into the Social Security system.
Moreover, the plan could invite misleading comparisons. Many retirees would likely compare the annuity benefit their individual account would provide which would have been financed with annual deposits equal to two percent of their wages to their Social Security benefit, financed with 12.4 percent of their wages. They could conclude Social Security was a bad deal and the law should be changed to shift large sums from Social Security to individual accounts. As a number of leading Social Security analysts have written, however, such a comparison would be highly misleading; it would ignore the fact that Social Security payroll taxes must finance benefits to previous generations of workers, pay for disability and survivors insurance, and finance the provision of more adequate benefits to low-wage workers and to spouses who spent years out of the labor force raising children, among others. If the same amount of additional funding were provided directly to the Social Security trust funds and allowed to be invested in a similarly diversified manner, the Social Security trust funds would secure a higher rate of return than the Archer-Shaw individual accounts, since the administrative costs of establishing and maintaining 150 million individual accounts would be avoided.
- The plan is not likely to be sustainable over time for another reason as well. The feature of the plan that reduces Social Security benefits one dollar for each dollar in income an individual receives from his or her individual account is not likely to survive politically over time. This aspect of the plan effectively imposes a 100 percent tax rate on most accounts. After being told the individual accounts are their property, American workers would see their accounts entirely taxed away when they retired.
If the Archer-Shaw plan were enacted in its current form and later changed so Social Security benefits were reduced much less than one dollar for each dollar paid from an individual account, Social Security solvency could be jeopardized unless major reductions were made in Social Security benefits. This knotty problem heightens the risk that the plan would lead over time to individual accounts replacing an increasing share of Social Security benefits.
- The plan raises equity concerns. The only group of retirees who could receive an increase in government-funded retirement benefits under the plan would be upper-income workers. Yet a broad array of Americans, including many of average or modest means, might have to absorb cuts in other benefits or services or tax increases to help finance the individual accounts after 2012. In contrast, the Clinton budget would provide increased retirement benefits primarily to low- and middle-income retirees as a result of its USA accounts. (Note: Claims the Archer-Shaw plan provides a progressive tax cut are without merit. The plan contains no reduction in payroll taxes for the next 50 years; workers would continue paying the same amount of payroll tax during this period that they pay today. Nor can the deposits into individual accounts be considered tax cuts since the government would eventually reclaim these accounts to finance Social Security benefits. Most workers would pay the same amount of payroll tax and receive the same amount of Social Security benefits as under current law; the only group of individuals likely to receive an increase in retirement benefits would be those at high income levels. This consequently should be not viewed a progressive tax cut.)
- Finally, the plan could fall short of restoring solvency. In their analysis of the plan, the actuaries assume that stock investments will provide an average real rate of return of nearly seven percent per year over the next 75 years and that investments in corporate bonds will provide a 3.5 percent real rate of return. The actuaries note that if returns are one percent lower than they have assumed, the Social Security trust funds would become insolvent in 2048 under the plan. The plan's success on restoring solvency is heavily dependent upon the performance of the stock market in future decades.
Other Features of the Archer-Shaw Plan
This analysis is not a comprehensive analysis of the Archer-Shaw plan; it does not cover features of the plan that would eliminate the Social Security earnings test, reduce payroll tax rates after 2050 if the financing of the system allows, and preclude use of Social Security surpluses for purposes other than providing retirement benefits or paying down debt. The Center on Budget and Policy Priorities will issue an analysis in the future on the Social Security earnings test.
One other point should be noted. It should not surprise anyone that the plan restores long-term Social Security solvency. The Social Security actuaries have forecast that the program's long-term financing shortfall equals 2.07 percent of wages over 75 years. The Archer-Shaw plan provides a general revenue contribution equal to two percent of wages, invests a portion of these revenues in equities, and reduces Social Security benefits one dollar for each dollar in income the individual accounts generate.
If the same amount of additional funding were provided directly to the Social Security trust funds and the trust funds were allowed to invest a tiny portion of it in equity index funds, solvency would be restored in a more efficient manner.
Several of these issues are discussed in more detail below.
Under the Congressional budget resolution approved a few weeks ago, nearly all of the non-Social Security surplus would be used for tax cuts. As a result, little of the non-Social Security surplus would remain; it would not be available to finance the individual accounts the Archer-Shaw plan would create. These accounts thus initially would be financed from the Social Security surplus. (Under a possible deal later in the year, the non-Social Security surplus may be consumed largely by a combination of tax cuts and an easing of restrictions on discretionary spending, rather than just tax cuts. Either way, little of the non-Social Security surplus is likely to remain available to help finance the Archer-Shaw individual accounts.)
The plan apparently envisions that Treasury would take the surplus revenue the Social Security trust funds receive each year, provide Treasury bonds to the trust funds in return, and use the borrowed Social Security surpluses to make deposits into individual accounts. This essentially is the same type of financial transaction that many Republican Members of Congress criticized as "double counting" when the Clinton Administration proposed it. (Like the Clinton plan, the Archer-Shaw plan would effectively use the Social Security surpluses twice. The two uses under the Archer-Shaw plan would be to fund the provision of Treasury bonds to the Social Security trust funds and then to use the same funds to finance deposits into individual accounts. This is a legitimate financial transaction. Criticizing the Clinton Administration for such an approach while defending its use in the Archer-Shaw proposal, however, entails application of a double standard.(2))
Even with this financial transaction, funding individual accounts out of the Social Security surpluses would work only for a limited number of years.(3) After about 2012, the annual cost of the deposits into the individual accounts, including the increased interest payments the Treasury would have to make on the federal debt (because these funds would be placed in individual accounts rather than used to pay down debt), would exceed the actuaries' projections of the Social Security surplus. After 2017, the cost of the individual accounts would exceed the size of the Social Security surplus even without counting the increased interest payments on the debt.
How would the individual accounts be financed at that time? If the non-Social Security surplus had been consumed by tax cuts, as would be the case under the Congressional budget resolution or by a combination of tax cuts and upward adjustments in discretionary spending levels, which seems more likely no significant surpluses would remain anywhere in the budget. Financing the individual accounts would likely entail cutting programs or raising taxes.
That might not be a serious problem if the amount of financing needed for the accounts were small. But it is not. As noted earlier, the actuaries' estimates show the accounts would have a net cost including increased interest payments on the debt of $300 billion to $600 billion a year between 2016 and 2042. These are the years in which the baby-boom generation will retire in large numbers and Medicare and Medicaid costs will rise substantially as a consequence. CBO projects that deficits will return during this period and begin to swell.(4) Incurring costs of this magnitude in these years, as the Archer-Shaw plan entails, would place too great a strain on the rest of the budget. In a statement on the Archer-Shaw plan issued April 28, Rep. Charles Stenholm observed that the plan would crowd out other budgetary programs, increase pressures on taxes, and threaten to result in a higher national debt.
The plan would eventually be self-financing. According to the actuaries' estimates, that would occur in about 2050. But the plan would substantially exacerbate budgetary pressures and pose fiscal risks in years before then.
Under the Archer-Shaw plan, a portion of the added money the government would channel to the retirement system would be used to cover the administrative costs, management fees, and profits of brokerage houses and other private investment firms involved in handling the tens of millions of individual accounts. Instead of providing the added funds directly to the Social Security trust funds, the Archer-Shaw plan sets up a rather elaborate system to channel the funds out to individual accounts and then back into the Social Security Administration, allowing Wall Street firms to siphon off billions of dollars in commissions and fees in the process. This adds a large element of inefficiency to the plan. (Although the amount of these costs may sound small on an annual basis 25 basis points per year, according to the actuaries' estimates the costs mount substantially over time due to compounding effects. The actuaries estimates' indicate that these costs would consume about $350 billion during the plan's first 30 years and much larger amounts after that.)
As noted earlier, achieving the same effect as the Archer-Shaw plan does i.e., using a general revenue transfer to bolster the ability of the government to finance the Social Security benefit guarantee could be done much more efficiently (and without raising the sustainability and equity concerns the Archer-Shaw plan poses) if a portion of the transfer were invested directly in equity index funds rather than channeled through individual accounts.
It is unlikely this plan would be sustainable over time. Docking retirees a dollar in Social Security benefits for each dollar they receive in income from their individual accounts is likely to be increasingly unpopular as time goes by, generating mounting opposition. This would appear to be a 100 percent tax rate. Ultimately, this aspect of the plan is very unlikely to be sustainable politically.
If the plan were enacted and this 100 percent recapture feature were subsequently weakened as dissatisfaction with it grew, Social Security solvency could be threatened. If Social Security benefits are not reduced one dollar for each dollar received in retirement income from an individual account, the Social Security trust funds may not be able to pay the full Social Security benefits that beneficiaries are due unless Social Security benefits are reduced.
In addition, as explained above, large numbers of retirees could make inappropriate comparisons and incorrectly conclude that private accounts are a much better deal than Social Security and should replace Social Security in full or in substantial part.
Reps. Archer and Shaw may respond to this latter point by proposing that the Social Security Administration combine into a single payment a retiree's Social Security benefit payment and the annuity payment from the retiree's individual account, so that beneficiaries could not compare the two payments. Although this would help, it would not solve the problem. Beneficiaries almost certainly would receive annual or other periodic statements of their private account balances and the annuity value of their accounts, as well as of their accrued Social Security benefits. This inevitably would lead to comparisons. Without extensive consumer education, much of the public probably would not understand why such comparisons were not valid.
The only retirees likely to receive increased retirement income under the Archer-Shaw plan would appear to be some of those with high wages or salaries. The Social Security actuaries' analysis identifies "single workers with very high earnings" as the one group that might receive retirement payments from their individual accounts that exceeded the Social Security benefits to which they otherwise would be entitled.
Social Security benefits equal a higher percentage of the average wages that low- and moderate-income workers have earned than of the average wages that highly paid workers have earned. The benefits that the individual accounts would generate would not have a similarly progressive structure. As a result, unless the stock market produced extraordinary returns that substantially surpassed the average returns of recent decades, it would be nearly impossible for most workers who had not earned high wages to receive an increase in retirement income under the Archer-Shaw plan. The annuity from their individual accounts would not exceed the Social Security benefit to which they are entitled.(5)
If other federal programs are cut to help finance the individual accounts after the Social Security surpluses are no longer large enough to do so themselves and if the programs cut primarily benefit low- and middle-income people, as most federal programs do one effect of the plan could be to shift a modest amount of income up the income scale.
It may be noted that both the Clinton budget and the Archer-Shaw plan add general revenues to bolster government-funded retirement benefits. Under the Clinton plan, those whose retirement incomes would rise the most would be low- and middle-income workers, because the Administration's USA accounts would be tilted toward them. Under the Archer-Shaw plan, government-funded retirement income would rise only for some affluent workers. (A full evaluation of who would gain or lose under the Clinton Social Security proposal cannot be made; to restore solvency for a full 75 years, the Clinton plan envisions changes to be made on a bipartisan basis that have not been specified and hence cannot be assessed.)
1. These figures are in current dollars. See Stephen C. Goss, Deputy Chief Actuary, Social Security Administration, "Long-Range OASDI Financial Effects of the Social Security Guarantee Plan - INFORMATION," April 29, 1999.
2. Some defenders of the plan reportedly are claiming their plan differs from the Administration's with respect to double-counting because the Administration would simply issue more IOUs to the Social Security trust funds while the Archer-Shaw plan would back up its individual accounts with "real assets." This assertion is without merit. Under the Clinton plan, the Social Security trust funds would be provided additional resources, which would be invested in Treasury bonds and equity index funds. Suppose that under the Archer-Shaw plan, account holders purchase Treasury bonds and equity index funds with their account balances. To argue that the Administration's plan entails giving the trust funds IOUs that are not real assets while the Archer-Shaw plan's individual accounts would hold real assets is to argue that Treasury bonds and shares in equity funds have no real value when the Social Security trust funds hold them but gain such value when they are held in the name of individual accounts.
3. Technically, the funds used for deposits into individual accounts would be "scored" as coming from the non-Social Security budget. Under federal budget accounting, when the Treasury borrows the Social Security surpluses and provides the Social Security trust funds with Treasury bonds in return, the borrowed cash becomes part of the non-Social Security budget, and expenditures made with these funds are categorized as "on-budget" expenditures.
4. In its new recent report, CBO notes that in these decades, "the budget will face 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 faster than average wages." CBO also notes that whereas Social Security, Medicare, and Medicaid consume about 40 percent of federal receipts today, "by 2030, according to CBO's projections, they will consume about 70 percent of receipts and leave few budgetary resources available to address other national needs." CBO, An Analysis of the President's Budgetary Proposals for Fiscal Year 2000, April 1999, pp. 27-28.
5. The Archer-Shaw plan also includes another feature that is likely to benefit high-income workers more than those who earn lesser amounts. The plan provides that if a worker dies before reaching retirement age, the worker's account is combined with the account of the worker's spouse (or other survivors if there is no surviving spouse). If a high-wage worker who has built up a substantial account dies shortly before reaching retirement age and the worker's account is transferred to a surviving spouse who also is a high-earner, the surviving spouse's enlarged account may provide an annuity payment when the spouse retires that exceeds the Social Security benefit he or she otherwise would receive. The spouse would experience an increase in income as a result. This is much more likely to occur in the case of well-off two-earner couples than in the case of those with lower earnings.