Financing Private Accounts in the Aftermath of the Tax Bill, 8/21/01

FINANCING PRIVATE ACCOUNTS IN THE AFTERMATH OF THE TAX BILL:
The Challenge Facing the Social Security Commission and the Administration
by Peter R. Orszag and Robert Greenstein⁽¹⁾

On May 2, President Bush appointed a commission on Social Security reform, which has now held several meetings. The commission faces a daunting challenge: how, in the aftermath of the enactment of the recent tax cut, to finance the individual accounts the President favors.

Contributions to private accounts could be financed either by transfers from the non-Social Security budget or by diverting revenues from the Social Security Trust Fund. The new tax law, in combination with other priorities reflected in the Congressional budget resolution, has consumed essentially all of the previously projected surpluses in the non-Social Security budget. As a result, any attempt to make payments from the non-Social Security budget to create "add on" accounts (that are added on top of existing Social Security revenue) would cause large deficits outside of Social Security and Medicare and thus does not seem feasible politically.

The alternative is to finance "carve out" accounts from existing Social Security revenue. This approach seems most consistent with President Bush's Social Security principles. Such an approach, however, would either exacerbate the long-term deficit in Social Security or require deep reductions in traditional Social Security benefits. Neither of those outcomes is likely to be viable politically either.

The tax cut thus leaves the commission and policymakers in general with politically unappealing choices regarding the financing of individual accounts. Policymakers might be tempted to sidestep these unappealing choices by using budget gimmicks. This analysis explores the dilemma in which policymakers find themselves following the tax cut and includes an examination of three types of accounting gimmicks that policymakers could be tempted to consider to finance private accounts:

The Social Security Trustees project that under current law, Social Security faces an imbalance over the next 75 years equal to 1.86 percent of taxable payroll, or 12 percent of projected expenditures. According to current projections, annual Social Security tax revenue (which does not include interest on the bonds the Trust Fund holds) will fall below Social Security benefit expenditures in 2016. In 2025, total Social Security revenue (including interest earned by the Social Security Trust Fund) will fall below benefit costs, and the Trust Fund will begin to redeem its bonds. The Social Security Trust Fund is then projected to be exhausted in 2038.

To improve Social Security's financial condition, four basic options exist: increase Social Security tax revenue, reduce Social Security benefits, raise the returns the Social Security Trust Fund earns on its reserves, or transfer funds from the rest of the budget to Social Security. Private accounts, in and of themselves, do nothing along these four dimensions and therefore do not directly improve Social Security's financial condition. Rather, the typical argument is that the income from the private accounts would permit reductions in traditional Social Security benefits without causing significant hardship for beneficiaries. The implicit argument behind private accounts is thus that the long-term deficit within Social Security could be eliminated through traditional benefit reductions, but that the income from private accounts would provide a replacement for those benefit reductions.

As explained below, however, carve-out private accounts financed out of Social Security revenue would result in substantial declines in expected total retirement income, including the income from the private accounts, relative to the current Social Security benefit formula. The reductions would be sufficiently large that financing private accounts entirely out of current Social Security revenue is unlikely to be politically viable. The more auspicious alternative therefore would seem to be to finance contributions to private accounts out of the non-Social Security surplus. But as the next section of this analysis explains, the large tax cut precludes financing add-on accounts on top of Social Security without creating deficits outside the Social Security and Medicare Hospital Insurance trust funds. (It should be noted that with or without private accounts, a significant reduction in total retirement income will be necessary to restore long-term Social Security solvency in the absence of payroll tax increases or general revenue transfers.)

In short, carve-out accounts would involve politically unappealing reductions in traditional Social Security benefits. Add-on accounts might have proven to be more politically attractive, but given the recently enacted tax cut, they would create large deficits in the non-Social Security budget. The basic problem is that non-Social Security revenue is needed to make a Social Security reform package politically viable (whether or not it includes private accounts), but the tax cut undermines such an approach.

Add-on accounts would be financed from the non-Social Security budget. In the aftermath of enactment of the tax cut, however, financing accounts in this way would cause deficits outside of the Social Security and Medicare Hospital Insurance Trust Funds.

According to the budget estimates that the Congressional Budget Office issued earlier this year, the projected surplus over the next 10 years outside of Social Security and Medicare Hospital Insurance amounted to $2.74 trillion prior to enactment of the new tax legislation.⁽²⁾ Taking into account the new tax law and the other initiatives in the Congressional budget resolution, such as a prescription drug benefit, the remaining surplus outside the Social Security and Medicare Hospital Insurance Trust Funds appears on paper to be about $500 billion between 2002 and 2011.

Between 2002 and 2011, contributions to private accounts equal to two percent of wages would amount to more than $1 trillion. Such contributions consequently could not be financed from the non-Social Security, non-Medicare part of the budget without creating deficits there.

Furthermore, the situation is even less promising than this simple comparison would suggest, because the official budget forecast embodies several highly unrealistic assumptions. As a result, not even $500 billion is likely to be available for individual account contributions. The unrealistic assumptions include:

After making an adjustment just for the extension of the expiring tax credits (the second item above), the available non-Social Security, non-Medicare surplus vanishes for the years from 2003 to 2006.⁽³⁾ This adjustment does not reflect likely defense spending increases or realistic levels for non-defense discretionary programs. Nor does it include any funds for relief from natural disasters — such as hurricanes, floods, and earthquakes — which have been averaging $5.6 billion a year.

**Table 1**
**Additional Costs Beyond Those Assumed in Congressional Budget Plan**
(ten-year totals in billions of dollars)
Continuing the provisions of the tax cut that are due to expire before 2011, including AMT relief^a	$390
Continuing existing tax credits due to expire during the decade^b	$100
Accommodating the President's FY 2002 defense increase and allowing the resulting level of defense funding to grow only with inflation in subsequent years	$190
Holding non-defense appropriations constant, with adjustment for inflation and population growth	$210
Higher interest payments (because the previous four items shrink the amount of debt reduction that would otherwise occur)	$160
TOTAL	$1,050
Note: may not add due to rounding ^a These figures reflect estimates of the Joint Committee on Taxation. AMT relief is assumed at the level needed to keep the new tax law from increasing the number of filers who will be subject to the AMT. Under this assumption, the number of tax filers subject to the AMT still would rise from 1.4 million this year to 20.5 million in 2011 (which is the number of filers that would have been subject to the AMT in 2011 under the law in place prior to enactment of the recent tax cut). Further AMT relief to prevent this large an increase in the number of filers subject to the AMT would entail additional costs not shown in this table. ^b Source: Joint Committee on Taxation

Finally, these figures do not reflect the impact of the recent economic slowdown on the budget forecast. Revised CBO and OMB budget forecasts that will be released in late August are expected to include a significant downward adjustment in the surplus for this and related reasons, as well as to reflect the impact of the recently enacted tax cut.

When realistic assumptions are made, there is no surplus over the next 10 years left to transfer to Social Security or private accounts. In fact, as Table 1 shows, the cost of actions that policymakers appear likely to take significantly exceeds the $500 billion of surpluses supposedly remaining outside the Social Security and Medicare Hospital Insurance trust funds. The upshot is that the tax cut and the other provisions reflected in the budget resolution preclude add-on private accounts — and general revenue transfers directly to the Social Security system — without creating deficits outside the Social Security and Medicare trust funds.

An alternative approach would be to create carve-out accounts financed out of existing Social Security revenue. Over the next 10 years, Social Security is expected to run cash-flow surpluses totaling approximately $2.5 trillion. Over the longer term, however, Social Security is expected to run a significant deficit, amounting to about 1.9 percent of taxable payroll over 75 years. The Social Security Trustees project that the Social Security Trust Fund will become insolvent in 2038.

**Table 2**
Impact of diverting revenue into private accounts on Social Security solvency
	Current law	Divert two percent of payroll into private accounts, starting in 2002 (with no other changes)
Payroll taxes less than benefits	2016	2007
Total income (payroll taxes plus interest on Trust Fund) less than benefits	2025	2015
Trust Fund exhaustion date	2038	2024

President Bush could finance his individual account proposal by diverting part of the projected Social Security cash-flow surpluses into private accounts. That approach seems most consistent with language he has used to describe his proposal. By itself, however, diverting revenue from Social Security to private accounts would exacerbate Social Security's long-term financial imbalance because it would reduce the revenue available to the Social Security system (see Table 2). For example, if individual account contributions amounted to two percent of payroll, the Social Security Trust Fund would be exhausted in 2024 rather than in 2038 if no other changes in Social Security were made (i.e., if Social Security benefits were not reduced or payroll taxes raised).

To improve Social Security's long-term financial situation while diverting revenue into private accounts requires some combination of higher payroll taxes, investing part of the Social Security Trust Fund in higher-yielding assets (such as stocks), transfers to Social Security from the rest of the budget, and reductions in Social Security benefits. The Administration's principles for Social Security reform rule out raising the payroll tax or having a portion of the Trust Fund reserves invested in equities. Furthermore, as already explained, transfers from the non-Social Security surpluses to the Social Security system would cause substantial deficits outside the Social Security and Medicare Hospital Insurance Trust Funds, given the tax cut. Consequently, the only remaining choice to prevent Social Security's projected long-term deficit from worsening if funds are diverted from the Trust Fund to private accounts would be to reduce Social Security benefits.

Why Guaranteed Benefits Would Have to be Reduced More than 40 Percent

To see why guaranteed benefit reductions averaging 40 percent or more would be required if revenue equal to two percent of payroll were diverted from Social Security into private accounts, note that the Social Security actuaries project that maintaining current-law benefits would cost an average of 15.4 percent of payroll over the next 75 years. The benefits that the President has said he would protect — benefits for current retirees, near retirees, and the disabled — amount to about 6 percent of payroll. This leaves 9.4 percent (15.4 percent minus 6 percent) available to be reduced to reach long-term balance. The projected long-term deficit in Social Security currently equals just under 2 percent of payroll. Diverting two percentage points of payroll into private accounts would raise the long-term deficit to almost 4 percent of payroll. Eliminating a deficit of almost 4 percent of payroll when the available guaranteed benefits that could be reduced amount to 9.4 percent of payroll requires a reduction in guaranteed benefits averaging roughly 40 percent (since 4 percent equals about 40 percent of 9.4 percent).

Henry Aaron, Alan Blinder, Alicia Munnell, and one of the authors of this analysis (Peter Orszag) have examined the size of the benefit reductions that would be needed to restore Social Security solvency while diverting payroll tax revenues equal to two percent of wages into private accounts starting in 2002.⁽⁴⁾ President Bush has promised to protect benefits for current retirees, older workers, survivors, and the disabled. To protect benefits for these people while diverting payroll tax revenue equal to two percent of payroll into private accounts and also eliminating the long-term deficit within Social Security would require reductions in Social Security benefits (relative to the benefits that would be provided under the current-law benefit formula) of more than 40 percent (see box above).

Such an across-the-board reduction in traditional benefits would impose greater burdens on older workers (who would have less time to build up their private accounts) than younger workers. To avoid a sharp reduction in total retirement income for older workers, which would not be politically viable, the reductions in Social Security benefits would have to be phased in over an extended period of time. Because less would be saved in early years as a result of this phase-in of the benefit reductions, the reductions for workers who are young today would have to be larger to ensure that Social Security remains solvent over 75 years. The Century Foundation analysis found that under one plausible approach to phasing in the benefit reductions, Social Security benefits would have to be reduced by 29 percent for those who are aged 50 in 2002 and by 54 percent for those now aged 30 or younger, relative to the Social Security benefit levels under the current benefit structure.

To be sure, income from private accounts would offset some of these benefit reductions. But a large net benefit reduction would occur. The Century Foundation analysis estimated that the expected combined retirement income from Social Security and private accounts for single individuals with average earnings who are 30 years old in 2002 (and who retire at age 65 several decades from now) would be 20 percent below the level they would receive from Social Security under the current benefit structure (see Table 3). For one-earner married couples with average earnings who are 30 years old in 2002 and retire at age 65, the expected combined income from Social Security and private accounts would be 33 percent below the level these couples would get from Social Security under the current benefit structure.

**Table 3**
Impact of Two-percent Individual Account Contributions Financed out of Social Security Revenue (2000 inflation-adjusted dollars)
	30-year-old Single Earner with Average Wages	30-year-old One-Earner Married Couple with Average Wages
Annual Social Security benefit under current-law benefit structure	$15,877	$23,816
- 54% reduction	-$8,510	-$12,771
+ Individual account retirement income	+$5,305	+$4,928
Total combined benefit	$12,672	$15,973
Relative to current law	-20%	-33%
Source: Henry Aaron, Alan Blinder, Alicia Munnell, and Peter Orszag, "Governor Bush's Individual Account Proposal: Implications for Retirement Benefits," The Century Foundation, June 2000.

These figures are based on assumptions about private accounts that are favorable to private accounts and are more likely to understate than overstate the benefit reductions that would be required. These assumptions generally are identical to those made by Professor Martin Feldstein, a leading advocate of private accounts. These assumptions were used to show that even under assumptions favorable to private accounts, the required benefit reductions would be substantial.

In summary, financing private accounts out of projected Social Security reserves would involve either a significant deterioration in Social Security's long-term financial health or substantial reductions in Social Security benefits. Either choice would be likely to encounter stiff opposition both on and off Capitol Hill. (Indeed, a recent National Journal article reports that a number of conservative groups oppose linking private accounts to benefit reductions in Social Security. "If you want private accounts," the article quotes Stephen Moore, president of the Club for Growth, as saying, "linking them with benefit reductions or tax increases is pure poison." The article cites Moore as declaring that the conservative groups with which he is working are unanimous in opposing "any type of benefit reduction as part of personal accounts" and would work to "torpedo" any plan that included one.⁽⁵⁾)

This reality is why many analysts believe that serious Social Security reform will require additional funds from outside of Social Security to be politically viable. Nearly every major plan in recent years to restore long-term Social Security solvency, regardless of whether it creates private accounts, has relied in part on transferring some of the surpluses projected in the non-Social Security budget to the Social Security Trust Fund. For example, President Clinton proposed transfers of non-Social Security surpluses to Social Security. So did the Social Security plan advanced in the last Congress by then-Ways and Means Committee chairman Bill Archer and Social Security subcommittee chairman Clay Shaw. In the aftermath of the tax cut, however, no such funds are available.

As the analysis above shows, if an amount equal to two percent of payroll is diverted into private accounts — and 75-year solvency is to be restored without transfers from the rest of the budget and while protecting certain benefits as President Bush has promised — then traditional Social Security benefits must be cut by an average of more than 40 percent. To avoid presenting a Social Security plan with benefit reductions of this magnitude, policymakers may be tempted to use an accounting gimmick.

Indeed, if a plan shifting two percent of payroll into private accounts has no transfers from the rest of the budget, contains no payroll tax increases, does not entail a 40 percent benefit reduction and still appears to restore long-term solvency, it must contain some sort of gimmick. The purpose of such a gimmick would be to conceal the difficult trade-offs involved in financing the private accounts without transfers from the rest of the budget. Below we explore three potential gimmicks that various supporters of private accounts have proposed using.

To avoid the appearance of either undermining the long-term health of Social Security or necessitating larger reductions in Social Security benefits than would otherwise be necessary, an individual account proposal could credit the same funds to both Social Security and private accounts, while changing the budget accounting rules that the Congressional Budget Office and the Office of Management and Budget use to mask the cost that such an approach entails.

Under this approach, the annual Social Security cash-flow surpluses would be credited to the Social Security Trust Fund. Rather than being used to reduce public debt, these funds then would also be used to make contributions into private accounts. In other words, the same funds would both be credited to Social Security and deposited in private accounts. Peter Ferrara, a leading advocate of private accounts, has urged in a recent memorandum addressed "to the conservative movement" that this approach be used, writing: "When the [Social Security] surplus funds are used for personal accounts, the government should again provide government bonds to the Social Security trust funds in return. The account option would then not reduce the Social Security trust funds in any way, or reduce at all the funds available to pay current Social Security benefits."⁽⁶⁾

This use of the same funds both to credit the Social Security trust fund (and thereby enlarge its reserves) and to finance another program — in this case, private accounts — is sometimes referred to as "double counting." It represents a type of budget policy that was common until the mid-1990s but has been eschewed in recent years. A return to this type of budget policy, which entails running deficits outside Social Security and the Medicare Hospital Insurance program, may be necessary for a temporary period if the nation enters a recession or an extended period of abnormally slow growth, but a broad consensus has developed among policy-makers against such a budget policy except in periods when the economy is weak.

To see how such an approach would work, imagine a dollar of surplus Social Security payroll tax that flows into the Social Security Trust Fund under current law. Since the dollar is not needed to pay Social Security benefits now, the Treasury provides the Social Security Trust Fund with Treasury bonds in exchange for these surplus revenues. The Treasury now uses these funds to pay down the publicly held debt. The reduction in public debt contributes to national saving and also reduces the interest payments the federal government must pay on the debt in the future. It thus reduces future budget pressures outside Social Security, which makes it somewhat easier for the government in the future to honor its promises to pay Social Security benefits.

Now consider what would happen under the double-counting approach that Mr. Ferrara advocates. A dollar of Social Security payroll tax revenue would flow into the Social Security Trust Fund, and the Treasury would issue a bond in exchange for the surplus Social Security revenue. Under the budget accounting rules that the Congressional Budget Office (CBO) and Office of Management and Budget (OMB) have long used, once the Treasury issues this bond to the Social Security Trust Fund, anything the Treasury does with the money borrowed from the Social Security Trust Fund (other than paying down debt) is considered an expenditure from the non-Social Security side of the budget. Therefore, if the funds were used to finance private accounts, that would count as an increase in federal expenditures outside Social Security. Such an increase would result in a return of budget deficits outside Social Security and Medicare Hospital Insurance, since (as explained above) no room exists in the projected surpluses outside Social Security and Medicare to finance individual account contributions.⁽⁷⁾

To avoid showing a deficit outside of Social Security and Medicare, the budget accounting rules could be altered through use of an egregious budget gimmick. The Administration and Congress could seek to mask the resulting deficits in the non-Social Security budget by directing CBO and OMB to reverse their existing budget rules and to fail to count the use of these funds to finance private accounts as an expenditure.

Such a gimmick would camouflage the cost of private accounts financed with Social Security revenue. It would mask the cost to the Social Security system by continuing to credit the Trust Fund with the surplus Social Security revenue, while camouflaging the cost to the budget by failing to record the deposits in private accounts as an expenditure. The costs of the accounts must, however, be a cost to either the Social Security system or the rest of the budget. Under this gimmick, costs would not be charged either against the Social Security Trust Fund (since the Trust fund would still receive Treasury bonds in exchange for its temporary excess revenue) or against the rest of the budget. Using a gimmick to make it appear as though this maneuver entails no cost to either Social Security or the rest of the budget would constitute dishonest budgeting and be fiscally irresponsible.

A second accounting gimmick would involve loans from the general fund to the Social Security Trust Fund. To see how this would work, assume that $1 in payroll taxes is diverted from the Social Security Trust Fund into private accounts. All else equal, that would reduce the funds credited to the Trust Fund and cause the Trust Fund to become insolvent at an earlier date. To avoid exhaustion of the Trust Fund earlier than under current law, the general fund would — under this approach — lend funds to the Social Security Trust Fund. Such loans would be repaid in the future, with interest, from the Trust Fund back to the general fund. The Social Security reform plan proposed by then-Representative John Kasich in 1999 employed such a mechanism.

Under current budget rules, loans from the general fund to the Social Security Trust Fund would not be scored as non-Social Security expenditures — because they are loans that will be repaid with interest — and thus would not cause a deficit in the non-Social Security, non-Medicare budget. Such loans, however, also would do nothing to improve Social Security's long-term deficit, since they would have to be repaid with interest. The loan repayments the Trust Fund eventually would have to make would be equal in present value to the loans the Trust Fund earlier received, with the result that there would be no effect on Social Security's long-term imbalance.⁽⁸⁾

Here is where this approach can be turned into a flagrant gimmick. Such loans could be used to reduce the 75-year Social Security deficit (as opposed to the program's permanent deficit) if the loan repayments are deferred for such a long period of time that they extend beyond the next 75 years. This is precisely the approach the Kasich plan used; under that plan, the loan repayments would not even begin until 2060 and would be stretched out for many years after 2075. Placing some of the repayment outside the 75-year window has no effect on the permanent imbalance within Social Security, but reduces the apparent imbalance over the 75 years traditionally used to measure Social Security's long-term solvency.

This gimmick thus can be used to help finance private accounts and make restoring long-term Social Security solvency while establishing such accounts look like a "free lunch." The loans would not show up in documents presenting the status of the non-Social Security budget, but the loan proceeds would show up in Social Security accounts as infusing trillions of dollars into Social Security over the next 75 years.

In short, loans from the general fund to the Social Security Trust Fund could be employed as a multi-trillion dollar budget gimmick to avoid hard choices and make it look as though 75-year Social Security solvency could be restored and private accounts implemented without either reducing Social Security benefits or causing deficits outside Social Security. Once again, a gimmick would be used to conceal the costs involved.

One reason this gimmick might be available for policymakers to pursue is that the Congressional Budget Office and the Social Security Administration have not encountered this issue before, and it appears that the accounting rules that apply to transactions of this type have not been thoroughly examined. The gimmick essentially relies on inconsistencies in the treatment of loans between the lender (the budget) and the borrower (the Social Security system). In particular, the gimmick relies on the fact that the loan is not counted as an expenditure from the general fund but is counted as income to the Social Security system over the 75-year period. To avoid such a misleading result, one of two changes should be adopted:

Either approach would eliminate the potential for this gimmick to be employed. We prefer the first alternative, under which the Social Security actuaries would evaluate solvency on the basis of the net financial position of the Trust Fund and the current budget scoring rules for loans would be maintained. That approach appears the most analytically sound.

A final possible accounting gimmick involves a form of budget accounting that Martin Feldstein and Andrew Samwick have suggested but that has been rejected by most budget analysts and the Congressional Budget Office.⁽⁹⁾ Feldstein and Samwick would undertake a form of "dynamic scoring," in which assumptions are made about the macroeconomic effects of Social Security proposals and these assumptions are then reflected in the cost estimates for those proposals. (Using dynamic scoring for tax proposals, for example, entails making assumptions about the impact of proposed tax policy changes on economic growth and on the resulting change in revenues that is assumed to arise from the change in economic growth.) This approach has been rejected by CBO, OMB, and policy-makers of both parties because there is little agreement among economists about the response of the economy to changes in tax and budget policies, which makes this approach particularly susceptible to manipulation.

Despite the lack of support for dynamic scoring among budget experts, Feldstein and Samwick would transfer to the Social Security Trust Fund the extra corporate revenue they assume would result from creating private accounts. Corporate income tax collections would be higher, they claim, because a Bush-style individual account plan would boost national saving, investment, and GDP. The amount of additional tax collections they assume over the next 75 years is very large — $3 trillion in present value. (The present value is the amount today that, with interest, would equal over the next 75 years the additional tax collections that Feldstein and Samwick assume.) In other words, their proposal entails transferring the immediate equivalent of $3 trillion from the rest of the budget to Social Security.

Feldstein and Samwick implicitly argue that these $3 trillion worth of transfers to the Social Security Trust Fund would impose no additional burdens on the rest of the budget, because the transfers would reflect net additional revenue from higher corporate income tax collections. Their key assumption is that carve-out private accounts would raise national saving and that higher national saving would then boost corporate income and therefore corporate tax revenue.

This assumption of a dramatic increase in national saving, however, is implausible. First, it rests upon another assumption: that in the absence of private accounts, the Social Security surpluses will not be used to pay down debt — which increases national saving — but instead will be used to finance increases in other government programs, despite pledges by the President and both parties to wall off these surpluses. If the Social Security surplus is walled off, the Feldstein-Samwick argument falls apart; Social Security surpluses already would be committed to national saving by paying down the public debt, so shifting these funds from debt repayment to the financing of private accounts would not increase national saving. In fact, shifting the Social Security surpluses from debt repayment to private accounts could reduce national saving, since individuals may be more likely to reduce their own saving in response to the establishment of private accounts in their names than in response to a reduction in the government's debt.

In other words, the Feldstein-Samwick argument — that private accounts would boost national saving very substantially, which then would increase economic activity and boost corporate tax collections to such an extent that the added revenue would effectively pay for the individual account contributions — collapses once the suspect assumption is removed that the Social Security Trust Fund surpluses would otherwise be fully expended.

Furthermore, the magnitude of the increases in national saving and corporate income tax revenues that Feldstein and Samwick assume is so large as to strain credulity. They assume that carve-out private accounts would result in an increase in corporate income taxes equal to 1.7 percent of projected GDP in 2075; by comparison, the entire corporate income tax amounts to only about 2 percent of GDP even in extremely good economic times. They thus assume corporate tax revenue would nearly double as a share of GDP as a consequence of private accounts.

The dynamic scoring approach is essentially another accounting gimmick, as it is predicated on the dubious assumption that national saving would increase dramatically from the creation of carve-out private accounts and result in massive increases in corporate tax revenues. If the additional corporate revenue failed to materialize, as would likely be the case, the financing of private accounts under the Feldstein-Samwick approach would either produce large-scale budget deficits or require substantial program cuts or tax increases in the rest of the budget, to make up for the $3 trillion being transferred from the rest of the budget to the Social Security Trust Fund.

The size of the recently enacted tax cut requires policymakers to employ one of two unappealing mechanisms for financing private accounts:

The only other alternative is some form of accounting gimmick. In essence, the large size of the tax cut requires policymakers to adopt one, or some combination, of three alternatives for financing private accounts: a deficit outside Social Security and Medicare Hospital Insurance (or substantial budget cuts or tax increases to avert such deficits), large reductions in traditional Social Security benefits, or an accounting gimmick.

Since all of these approaches are politically problematic, the recently enacted tax cut considerably weakens the chances that the President's commission will succeed in designing a credible plan to restore long-term Social Security solvency. In a fundamental sense, the tax cut has undermined the opportunity that the nation had, and could potentially recapture, to use the projected budget surpluses as a substantial downpayment on the longer-term budgetary pressures facing the nation.

1. Peter Orszag is a senior fellow in economic studies at the Brookings Institution. Robert Greenstein is the executive director of the Center on Budget and Policy Priorities.

2. Such projections are highly uncertain, as CBO itself emphasizes. This estimate is likely to be revised downward on August 28, when CBO will issue a new budget and economic forecast.

8. Loans do, however, allow the Trust Fund to avoid liquidity (as opposed to solvency) problems: They ensure that the Trust Fund does not become negative for a temporary period, as they otherwise would under many proposed plans, such as the one proposed by Martin Feldstein and Andrew Samwick. See Martin Feldstein and Andrew Samwick, "Allocating Payroll Tax Revenue to Personal Retirement Accounts," Tax Notes, June 19, 2000, and as NBER Working Paper 7767, http://www.nber.org.

9. See Martin Feldstein and Andrew Samwick, "Allocating Payroll Tax Revenue to Personal Retirement Accounts," Tax Notes, June 19, 2000, and as NBER Working Paper 7767, http://www.nber.org.