Skip to main content

Should the Budget Be Changed to Exclude the Cost of Individual Accounts?

Executive Summary

Proposals to replace part of Social Security with individual accounts are now a focus of attention, with the President expressing a strong desire to push forward on creating individual accounts within Social Security. This paper considers the appropriate budgetary treatment of proposals to create such accounts.

Under proposals to replace part of Social Security with individual accounts, the federal government generally would contribute to individual accounts, using either a portion of payroll tax revenues or general revenues. Such contributions would normally be recorded as increases in federal expenditures, or “outlays.” Unless the new expenditures were offset through concurrent reductions in other programs or increases in other taxes, the contributions to the individual accounts would require additional borrowing and enlarge the debt held by the public.

Accordingly, the Congressional Budget Office, the Social Security Administration, President Bush’s Council of Economic Advisers, and the President’s Commission to Strengthen Social Security all have treated contributions to individual accounts as expenditures that increase the deficit in the unified budget. Estimates from Social Security actuaries, for example, show that the principal proposal put forward by the President’s Social Security Commission (the proposal known as “Model 2” and introduced in legislative form by Senator Lindsey Graham) would increase federal borrowing and deficits by between $1.4 trillion and $2.2 trillion over the next decade and substantial additional amounts for several decades after that. Other individual account proposals would raise borrowing by as much as $5 trillion over the first ten years.

Such increases in deficits are not necessary, however, to reform Social Security or to create individual accounts. Long-term balance can be restored to Social Security through modest revenue and benefit adjustments that entail no borrowing. Replacing part of Social Security with individual accounts is not necessary to restore Social Security solvency.

Nor does the creation of individual accounts require large-scale borrowing. Individual accounts can be financed through additional worker contributions or by making concurrent adjustments in other federal taxes or spending (i.e., by reducing other programs or raising other taxes to generate the funds to finance the switch to individual accounts). There is no requirement to increase federal borrowing to fix Social Security or to establish individual accounts.[2]

In the past few years, however, a number of individual account plans have emerged that eschew financing the accounts through new worker contributions or changes elsewhere in the budget. This new generation of individual account plans relies heavily on deficit financing instead. To facilitate this type of deficit-financed plan, the Bush Administration and some Congressional leaders have recently indicated they are considering a dramatic shift in federal budget rules that would mask the impact of individual account plans on the deficit. Under the proposed shift, the large increases in borrowing associated with deficit-financed individual accounts would be omitted from the budget and would not show up as an increase in the deficit. [3]

Those who favor this approach note that individual account proposals typically combine the creation of individual accounts today with a reduction in Social Security benefits in the future. They seek to obtain immediate budgetary “credit” for the future benefit reductions. They argue that as long as the future benefit reductions would ostensibly offset the cost of the upfront borrowing, the borrowing costs should be excluded from the budget. (Because of the link between increased borrowing today and reduced benefits in the future, borrowing to finance individual accounts is sometimes referred to as a “transition cost.” This cost, however, is likely to last for several decades and to involve several trillion dollars of borrowing. The proposed change in budget rules would effectively exclude the transition costs from calculations of the budget deficit.)

This paper finds that such a sharp departure from the established budget rules would be ill-advised and fiscally irresponsible. Federal borrowing to finance individual accounts should be included in the budget, and the borrowing costs should be reflected in the deficit. The proposed change in long-established budget rules should not be adopted for four reasons.

  • The proposed borrowing of several trillion dollars would require the government to go much more heavily into private credit markets over the next few decades and seek much larger amounts from domestic and foreign creditors. This should not be hidden through an accounting maneuver.
  • Leaving the costs of borrowing for individual accounts out of the budget would open the door to “free lunch” Social Security plans, which hold a natural appeal for politicians but would undermine key underlying goals of Social Security reform, such as increasing national savings.
  • Proponents of leaving these large borrowing costs out of the budget argue that the borrowing would merely exchange future government debt for current government debt and thus would not affect the government’s overall financial condition. This claim is not correct: Transforming potential future debt into current actual debt could worsen the nation’s fiscal outlook and would reduce the government’s fiscal flexibility.
  • Bending long-established budget rules so that borrowing for individual accounts can be omitted from the budget would establish a dangerous precedent and could lead to increased gimmickry in other parts of the budget.

Large-scale Borrowing Should Not be Hidden

The public debt is already projected to grow from a level of 34 percent of the Gross Domestic Product (the basic measure of the size of the U.S. economy) in 2000 to nearly 70 percent of GDP by 2030. The borrowing called for under the main plan advanced by the President’s Social Security Commission would raise the debt to nearly 100 percent of GDP by 2030. Under some other individual account plans, the debt would be raised to even higher levels. These elevated levels of debt would increase the risk of a crisis in which the government faces difficulty paying the interest on this debt or issuing new debt in the bond market. The borrowing that would create such a fiscal situation should not be obscured in, or omitted from, the federal budget.

Opening the Door to “Free Lunch” Plans

Leaving borrowing for individual accounts out of the budget, and treating it as having no effect on the deficit, would increase the attractiveness of “free lunch” Social Security proposals. Such proposals purport to restore Social Security solvency without raising payroll taxes or reducing benefits.

Such plans accomplish this on paper by pouring in massive amounts of borrowed money and by ignoring the higher degree of risk associated with stock-market investments. Free-lunch plans, which purport to solve Social Security’s financing problems without making hard choices, are likely to widen — rather than narrow — the government’s overall long-term fiscal imbalance.

Under the existing budget rules, the Achilles heel of free-lunch plans is that they substantially increase the budget deficit. The proposed change in the rules, which would conceal the fiscal impact of massive borrowing to fund individual accounts by leaving it out of the budget, would create a carte blanche for such plans. It would remove the clearest and most readily understandable marker of why these plans are fiscally irresponsible. That could be particularly dangerous, since free-lunch plans hold a natural appeal for elected officials.

Indeed, leaving borrowing costs for individual accounts out of the budget would virtually guarantee that the government would borrow all of the money to fund the accounts. The result would be a lost opportunity to increase national saving.

Individual Accounts and National Saving

One of the primary goals of Social Security reform is supposed to be to increase national savings, and thereby to increase investment and economic growth and make it easier to meet our obligations to future generations. Until now, there has been consensus that this is an essential part of Social Security reform. But individual accounts fail to increase national savings if they are deficit-financed (i.e., financed through borrowing). If the money saved in individual accounts is money that has been borrowed, then total national saving is unchanged at best, since the new saving and the new borrowing cancel each other out.

Moreover, if people conclude that having an individual account means they can safely reduce other retirement savings — a conclusion that many current savers may well reach — then individual accounts financed by government borrowing would actually reduce national savings, since the amount the government borrows would exceed the net amount of new saving. A decline in national savings would worsen the nation’s long-term fiscal and economic prospects.

Explicit Debt and Implicit Debt

Advocates of leaving borrowing for individual accounts out of the budget argue that such borrowing would merely create “explicit debt” today (in the form of new Treasury bonds) in exchange for “implicit debt” that the federal government has already incurred (in the form of benefit promises to future Social Security beneficiaries that will exceed future Social Security revenues). They argue that these two types of debt — “implicit debt” and “explicit debt” — are essentially the same, and that converting implicit debt to explicit debt is not an increase in federal liabilities and need not be reflected in the budget.

This argument is seriously flawed. These two types of debt are decidedly not the same. Converting implicit debt to explicit debt could worsen the nation’s fiscal outlook and would reduce the government’s fiscal flexibility.

The “explicit debt” that the government would incur as a result of large-scale borrowing to finance individual accounts would have to be purchased by creditors in financial markets. (In other words, the federal government would have to float more bonds.) The Treasury would have to borrow much more in financial markets over the next few decades than it otherwise would do. When the additional debt matured, it would have to be paid off or rolled over.

By contrast, the “implicit debt” associated with future Social Security benefit promises does not have to be financed in financial markets now. It also might not have to be financed at a later date, because the implicit debt could — and likely would — be reduced through future policy changes. Implicit debt is essentially potential debt that can be reduced through policy changes before that debt is actually floated in financial markets. Explicit debt is different; it is debt that already has been purchased by creditors.

In 1983, for example, Social Security faced a large implicit debt; benefits would soon exceed the revenues to pay them and would continue to do so indefinitely. Congress and the President acted — they changed Social Security benefits and taxes and did so without borrowing new money — and the implicit debt was substantially reduced. The same is likely to occur with regard to future unfunded Social Security promises.

By contrast, once explicit debt is incurred and Treasury bonds have been issued to cover it, the government is stuck with the debt unless it can shrink or eliminate the debt by raising taxes or cutting programs immediately. A government with a large explicit debt has less room for maneuver and is more vulnerable to a lessening of confidence on the part of the financial markets than a government with a large implicit debt.

Finally, despite their proponents’ claims, individual account plans might not substantially reduce the implicit debt in future decades. Under individual account plans that rely on large-scale borrowing, the borrowing is generally assumed to be “paid for” by substantial reductions in Social Security benefits that are slated to take effect (or to take full effect) decades into the future. When those changes are about to bite, however, political pressures could build to undo them. If future Congresses succumbed to the pressures and scaled back the future budget cuts before they took effect, much of the implicit debt that an individual account plan was supposed to eliminate could persist, and the increased expenditures incurred in establishing the accounts would not be offset.

The history of the past decade is instructive in this regard. Over the past decade, at least three major program reductions enacted into law — reductions in farm price supports, reductions in certain Medicare provider payments, and reductions in military retirement benefits — were reversed in whole or substantial part before they took effect. The reversal of these measures increased deficits and the debt by tens of billions of dollars.

If several trillion dollars are borrowed to establish individual accounts in exchange for Social Security benefit reductions that are slated to take effect decades from now, but those benefit reductions are scaled back by future Congresses, the net result could be an increase in the government’s liabilities. If that occurred, Social Security “reform” would have made the government’s already dismal long-term fiscal outlook worse.

Creating a Precedent for More Budget Gimmickry

There is no shortage of spending and tax proposals in other parts of the budget that are promoted as providing economic or other payoffs that yield budgetary benefits over the long term. Bending the budget rules to make it look as through borrowing for individual accounts would have no effect on deficits, on the grounds that the cost will be offset by savings in future decades, would set a dangerous precedent. It could lead over time to the use of other, comparable budgetary maneuvers.

For example, a large tax cut could be coupled with an implausibly large increase in taxes designed to take effect in future decades or coupled with unspecified steep reductions in future discretionary spending. Proponents of the tax cut could argue it had no net cost over time because of the subsequent offsets, even if the offsetting changes would not take effect for many years and it was questionable whether they ultimately would materialize, since future Congresses might reverse them.

Unsettling Financial Markets

Finally, establishing individual accounts while changing the budget rules to facilitate massive government borrowing would not only be likely to lead to unsound policymaking, but would also have the potential to unsettle financial markets. Borrowing trillions of dollars in private credit markets while failing to include the borrowing in the principal and most prominent measure of the federal budget, taking action that might lead to a reduction in national saving, and setting a precedent for future budget gimmickry could lessen financial markets’ confidence in the reliability of federal budget reporting and the soundness of the nation’s fiscal policy course.

The Responsible Approach: Do Not Bend the Budget Rules

Accordingly, the best approach is not to bend the budget rules in a politically convenient fashion and instead to continue adhering to the current, well-founded rules. Under those rules, federal borrowing to finance individual accounts would be treated like any other federal borrowing, including borrowing to finance investments in other programs or policies that might have long-term budgetary effects.

If the federal government must borrow in private credit markets for individual accounts, the borrowing should be treated as what it is: an increase in the deficit. Policymakers can supplement the basic measure of the deficit with additional benchmarks, such as projections of the deficit over longer time periods, the 75-year actuarial deficit in Social Security, and accrual-based budget measures. The interest in some quarters in creating individual accounts financed through higher deficits, however, does not warrant a change in the way the government keeps its books.

The conclusion that the government should not bend the budget rules to give an advantage to deficit-financed individual-account plans is further underscored by the fact that deficit financing is not necessary either to reform Social Security or to establish individual accounts. Long-term balance can be restored to Social Security through modest revenue and benefit adjustments that start to reduce the deficit within the next 10 years. Several Social Security plans that would accomplish this have been put forward, such as a plan designed by Peter Diamond and Peter Orszag and a plan by former Social Security Commissioner Robert Ball.

Nor does the substitution of individual accounts for part of Social Security necessitate large-scale borrowing and big increases in current deficits. Individual accounts could be financed through new worker contributions, as would be done under a plan developed by economist Edward Gramlich (currently a member of the Federal Reserve’s Board of Governors and previously the chair of the 1994-6 Advisory Council on Social Security), or by making concurrent adjustments in other federal taxes or spending (i.e., by raising other taxes or cutting other programs to provide the funds to transfer to the individual accounts).

Indeed, it is the rejection of such approaches that is now leading the Administration and a number of other individual-account proponents to propose massive government borrowing, accompanied by an effort to mask the effects of that borrowing by leaving it out of the budget.

The remainder of this paper examines these issues in further detail.

The Current Budget Treatment of Individual Accounts

The unified budget is the most prominent measure of balance between government revenues and expenditures. It is measured almost entirely on a cash basis. Government cash outlays to buy anything — whether a physical asset or a service — are scored as outlays and increase the unified budget deficit (or reduce the unified surplus). Thus, the unified deficit (or surplus) shows the extent to which the federal government borrows (or lends) in credit markets during the year.[4]

Some have argued that policymakers would be better served by replacing or supplementing the current budget with an “accrual budget” that measures changes in the government’s overall assets and liabilities or with “generational accounts” that record each generation’s net contribution to the budget.[5] This paper does not discuss which budget presentation is the most appropriate focus for policymakers; some of us have addressed that issue elsewhere. [6] Our purpose here is to ask whether making a major exception to the budgeting rules for individual accounts — and exempting individual accounts from those rules — is likely to lead to fiscally responsible policymaking.

The Budget Rules As they Now Stand

It should be noted at the outset that the standard budgetary treatment of individual accounts and Social Security plans has been embraced by the President’s Commission to Strengthen Social Security and, until now, by the Bush Administration. The Commission and the Administration have presented individual account plans as increasing the unified deficit in coming decades. For example, the Commission report stated, “The three reform models outlined here are therefore transparently scored in terms of plan provisions, effects on workers’ expected costs and benefits, and effects on Trust Fund operations as well as the unified federal budget.” [7]

In the same vein, the analysis of Social Security individual accounts conducted by the President’s Council of Economic Advisers and presented in the Economic Report of the President 2004 concluded that “Personal retirement accounts widen the deficit by design — they refund payroll tax revenues to workers in the near term while lowering benefit payments from the pay-as-you-go system in later years.” [8]

Key government budget and accounting agencies also have determined it appropriate and reasonable to treat government borrowing for individual accounts in the same manner as they treat government borrowing for other purposes. CBO studied the budgetary treatment of individual accounts, which it has called Personal Retirement Accounts, or PRAs, and concluded: “The budget should also record as outlays any payroll or income taxes that workers direct to privately owned PRAs… Recording the payments as outlays would increase total federal outlays and reduce any budget surplus.” [9] CBO analyses of specific individual account plans follow this approach.[10]

Likewise, the Office of the Actuary at the Social Security Administration treats government contributions to individual accounts as budgetary expenditures. To the extent that a Social Security reform plan would not offset such outlays through other, concurrent policy changes, the actuaries show the plan as causing an increase in federal borrowing and in the unified budget deficit.

When the Office of the Actuary produces an estimate of the effects of a Social Security reform plan, it includes estimates of the changes that would occur in annual budget deficits or surpluses under the plan. The actuaries have reported that plans with individual accounts would increase federal borrowing, and the unified budget deficit, by anywhere from $1 trillion to $5 trillion in the first ten years (see Table 1). Additional borrowing and resulting increases in the public debt would continue for decades after that.

Table 1:
Ten-Year Costs of Social Security Proposals to Restore Long-term Social Security Solvency


10-Year Cost (FY2006 - FY15)

Non-individual Accounts Plans Reduce Short-run Deficit

Diamond-Orszag plan

-$0.6 trillion

Ball plan

-$0.3 trillion

Individual Accounts Plans Increase Short-run Deficit

Kolbe-Stenholm plan

$1.0 trillion

President’s Commission Model 2 (assuming 66.7% participation)

$1.4 trillion

President’s Commission Model 2 (assuming 100% participation)

$2.2 trillion

Ryan-Sununu bill (based on Ferrara plan)

$5.3 trillion

Note: Costs based on memoranda from the Office of the Actuary, Social Security Administration, available at The actuary’s estimates are converted from constant dollars to current dollars using the Social Security Trustees CPI projections.

The Push to Change the Rules

Despite this consensus, however, some advocates of individual accounts — apparently including Senator Judd Gregg, the incoming chairman of the Senate Budget Committee — have recently suggested altering the rules so that federal contributions to individual accounts would be left out of the budget, rather than shown as outlays. This would mean that the potentially massive increases in federal borrowing to fund individual accounts would have no effect on the reported budget deficit.

The argument advanced by those who call for this rather radical change in the budget rules is that as long as the cost of diverting federal revenue into individual accounts is offset eventually by subsequent reductions in Social Security benefits — even if those benefit reductions are not slated to take full effect for decades — the net effect on the budget should be considered a wash. They argue that the borrowing that would occur over the next several decades should not be shown in the budget if those costs would ostensibly be offset by benefit reductions decades later. They thus argue for omitting from the budget as much as several trillion dollars in government borrowing.

The remainder of this analysis considers in greater detail whether exempting individual accounts from long-established budget rules, and leaving up to several trillion dollars of borrowing out of the budget, is justified — and whether it is likely to lead to more fiscally responsible or less fiscally responsible policymaking.

Excluding Transition Costs Is More Likely to Lead to “Free Lunch” Plans

The ratio of workers to retirees will be lower in the future than it is today. One way of dealing with this shift is to boost national saving today. Greater national saving today would increase the capital stock owned by future generations, thereby increasing future income and thus the nation’s ability to finance future retirement and other costs.

Does Borrowing $2 Trillion Really Save $10 Trillion?

Administration officials have downplayed the significance of the $2 trillion in transition costs required by some individual accounts plans. They have compared that cost to the unfunded liability in Social Security over an infinite time horizon, which totals more than $10 trillion. For example, on September 6, White House Press Secretary Scott McClellan responded to a question about how the White House would pay for the $2 trillion transition cost by arguing “It’s a savings, because the cost is $10 trillion of doing nothing, and this will actually be a savings from that cost of doing nothing.”

This argument, however, is misleading. First, the $10 trillion figure is likely to create a mistaken impression of the magnitude of the Social Security shortfall. The $10 trillion figure refers to the Social Security shortfall not over 75 years, but into eternity. While Social Security does face a long-term deficit, the deficit is, in fact, relatively modest as a share of the economy. Most important, borrowing $2 trillion to fund individual accounts does nothing to reduce Social Security’s long-term deficit. Most individual account plans that would eliminate the long-term deficit in Social Security, such as the principal plan the President’s Social Security Commission proposed, do so entirely by reducing future Social Security benefits, not because of borrowing.

The $10 Trillion Figure

When using the $10 trillion figure, Administration officials have not explained that it reflects Social Security’s imbalance not over 75 years — the period normally used to evaluate Social Security’s finances — or even over centuries, but into infinity (or “over an infinite horizon”).

According to the Social Security actuaries, the deficit in Social Security over the next 75 years is 0.7 percent of GDP (or $3.7 trillion). According to the Congressional Budget Office, the deficit over 75 years is 0.4 percent of GDP. (Over an infinite horizon, the deficit is 1.2 percent of GDP, according to the actuaries’ projections.) By way of comparison, if the tax cuts enacted in 2001 and 2003 are made permanent (and not eroded by the Alternative Minimum Tax), their cost over the next 75 years will be approximately 2 percent of GDP — or three to five times larger than the Social Security shortfall.

In December 2003, the American Academy of Actuaries, the nation’s leading professional organization of actuaries, stated that estimates of Social Security’s shortfall over an “infinite horizon” should not be used in policy discussions. The Academy warned that infinite-horizon projections “provide little if any useful information about the program’s long-term finances and indeed are likely to mislead anyone lacking technical expertise in the demographic, economic, and actuarial aspects of the program’s finances into believing that the program is in far worse financial shape than is actually indicated.”[1]

The Academy stated that the problems with this measure are such that the $10 trillion figure should not even be printed in the annual Trustees’ report and that including the measure in the report “is, on balance, a detriment to the Trustees’ charge to provide a meaningful and balanced presentation of the financial status of the program.”

Borrowing $2 Trillion Would Not Eliminate the Long-term Shortfall

Furthermore, the notion that borrowing $2 trillion now will save $10 trillion over time is simply incorrect. The basic flaw in this notion is seen by examining the principal plan that the President’s Social Security Commission proposed, often referred to as “Model 2.”

The individual accounts in the Model 2 plan would create a new financing hole for Social Security, which would be filled with more than $2 trillion in transfers from the rest of the budget to Social Security. To be sure, Model 2 would eliminate the long-term deficit in Social Security, which, as noted amounts to more than $10 trillion in present value over an “infinite horizon” (if this figure is used despite the problems that the American Academy of Actuaries warned of). But the individual accounts in Model 2 play no role in eliminating the long-term deficit. The $2 trillion cost associated with the individual account component of the plan is not the “price” of obtaining the long-term savings.

Model 2 contains three key components. It first restores long-term balance to Social Security and does so entirely through Social Security benefit reductions. These benefit reductions would be very large and would affect all beneficiaries, including disabled beneficiaries, surviving spouses and children of deceased workers, and even beneficiaries who do not elect private accounts. These benefit reductions would themselves more than eliminate the long-term deficit in Social Security. They — and not the borrowing of $2 trillion — are why Model 2 saves more than $10 trillion over an infinite horizon.

Second, Model 2 would replace part of the scaled-back Social Security system that would remain (after these large benefit reductions were instituted) with a system of private accounts. Those who chose the individual accounts would have some of their payroll taxes diverted from Social Security to the accounts; in return, their Social Security benefits would be reduced further. But that would do nothing to close Social Security’s shortfall. The amount that Social Security would lose because of the diversion of payroll tax revenues to the accounts would exceed the additional Social Security benefit reductions to which these beneficiaries would be subject. (This would be the case on a permanent basis, not just during a transition period.) In addition, the individual accounts would create a cash flow problem for Social Security because funds would be diverted from Social Security decades before a worker’s Social Security benefits would be reduced in return. The private accounts, by themselves, would actually push Social Security back into insolvency and permanently worsen the program’s financial condition.

To avoid insolvency and restore long-term balance, the plan’s third component consists of the transfer of large sums from the rest of the budget to make up for the losses that Social Security would bear because of the private accounts. These transfers would exceed $2 trillion.

Such transfers are not needed to address the long-term imbalance in Social Security; they would be necessitated by the introduction of the individual accounts, not by the need to close Social Security’s deficit. As noted, the accounts themselves would do nothing to address the deficit.

As a result, comparing the long-term deficit under Social Security to the cost of borrowing money to establish individual accounts, as the Administration has done, is a comparison of apples and oranges. The comparison is not valid.

Raising national savings has been a central goal of most Social Security reform plans. The 1994-96 Social Security Advisory Council unanimously agreed that this objective should be a feature of any Social Security reform plan, [11] and many subsequent Social Security reform efforts have been designed to raise national saving. The President’s Social Security Commission wrote, “This Commission agrees with the unanimous finding of the 1994-96 Social Security Advisory Council that partial advance funding of Social Security benefits is desirable. Advance funding raises national saving, increasing the nation’s capital stock and productive capacity and reducing Social Security’s financial burden on future generations.” [12]

A related goal for Social Security reform is to reduce pressures on the federal budget, which are expected to rise dramatically over coming decades. Most past Social Security reform plans have reflected this objective, as well. They have done so by reducing future Social Security costs, increasing future Social Security revenue, or both.

The Emergence of “Free Lunch” Plans

In the last few years, however, a growing number of “free lunch” Social Security plans have emerged that would not raise national saving and would exacerbate, rather that relieve, budget pressures in the future. Such “free lunch” plans propose diverting Social Security revenue into individual accounts and then relying on two major gimmicks to “pay for” the costs. First, they propose transferring trillions of dollars from the rest of the budget to Social Security, thereby increasing the already large deficits projected outside Social Security. Second, they assume that stocks are no riskier than bonds, despite the overwhelming evidence to the contrary. As the leading White House staffer on Social Security observed in a 2000 book, “The degree to which we simply divert payroll taxes into personal accounts and do not change contribution or benefit levels does not add to aggregate saving.”[13]

A notable example of this approach is a plan developed by Peter Ferrara and introduced as legislation by Rep. Paul Ryan and Senator John Sununu. The Social Security actuaries have reported that this plan would transfer $7 trillion in revenue from the rest of the budget to Social Security over the next 75 years, an amount nearly double the projected $3.7 trillion Social Security shortfall over this period. [14] The Ryan-Sununu plan ostensibly finances its massive transfers in large part by assuming that government spending can be reduced (relative to what it would be with no policy changes), but it provides no credible mechanism for achieving the necessary reductions. The plan should thus be viewed as predicated on a massive magic asterisk, in which trillions of dollars are simply assumed to be forthcoming from the rest of the budget.

As a result of such gimmicks, “free lunch” plans neither raise national saving nor address the nation’s long-term budgetary problems. A change in the budget rules undertaken so borrowing to fund individual accounts can be left out of the budget would make “free lunch” plans more politically attractive, as it would cause part or all of their massive short-term borrowing costs to disappear.

Individual Accounts and National Saving

The key point here is that adding individual accounts to a Social Security reform plan will increase national savings only to the extent that the contributions to the accounts are paid for by reducing other government spending or raising additional revenues. If that occurred, national savings would increase, because some money that otherwise would have been spent on consumption items by taxpayers or the government would be saved in private accounts instead. The resulting increase in national saving would lead to a reduction in future budgetary pressures. [15]

Some prominent individual-account proponents have themselves made the point that including the cost of borrowing for individual accounts in the budget is critical to ensuring that the plans contribute to national savings. According to two prominent supporters of individual accounts, Martin Feldstein and Andrew Samwick, the increase in deficits from government borrowing to fund individual accounts “reduces the likelihood that future Congresses and administrations would use those funds to finance additional government spending or additional tax cuts that finance private spending.” [16] In other words, Feldstein and Samwick argue that the enlarged deficits that would result from borrowing to establish individual accounts would inhibit spending increases and tax cuts and thereby result in lower spending and higher revenues than otherwise would be the case.

The crucial issue is that if the costs of the accounts are not offset by other immediate policy changes, preventing a decline in national saving would require that private saving rise by as much as the deficit increased, so that these two effects would offset each other and national savings would not diminish. Unfortunately, it seems more likely that in such a circumstance, a reduction in national saving would occur. Some individuals would likely scale back the amounts that they save through 401(k)s, IRAs, or other savings mechanisms, because they now would have their own individual accounts.[17] To the extent that this occurred, private saving would rise by less than the total amount placed in individual accounts — and hence by less than the increase in the budget deficit. The result would be a decline in total national savings, since the amount the government would borrow to finance the individual accounts would exceed the amount of new savings. (If withdrawals were allowed from the accounts before retirement, the reduction in national savings would be larger still.) In other words, once behavioral effects are included, national savings may not only fail to rise, but may actually fall.

Economists at the Federal Reserve Board and Harvard University have estimated that if policymakers do not offset the borrowing associated with establishing individual accounts by making concurrent reductions in other programs or raising more revenue, a plan similar to the principal plan (Model 2) that the President’s Social Security Commission developed would reduce the nation’s capital stock by 14 percent to 41 percent by 2070. [18]

In sum, leaving borrowing to establish individual accounts out of the budget would facilitate free-lunch plans and likely lead to a perverse outcome — Social Security “reform” would fail to raise national saving and might well reduce it. Such a Social Security “reform” would leave the nation no better prepared — and possibly worse prepared — than it is today for the demographic challenges and resulting budgetary pressures that lie ahead.

Converting Implicit Debt to Explicit Debt Would Reduce Fiscal Flexibility and Could Worsen the Long-term Fiscal Outlook

The previous section of this paper explained that funding individual accounts through borrowing would not increase national saving and could reduce it by leading people to reduce their personal retirement savings. This section finds that, even if national savings were unaffected, the process of converting future Social Security promises into public debt would reduce the government’s fiscal flexibility and could well worsen the long-run fiscal outlook.

Some economists and policymakers closely allied with the Administration have recently attempted to make the opposite case. Glenn Hubbard, former chair of President Bush’s Council of Economic Advisers and currently dean of ColumbiaBusinessSchool, has compared individual accounts to “prepaying a mortgage.” (Hubbard’s analogy is misleading. Borrowing to fund individual accounts is like prepaying a mortgage by running up a credit card debt.) Hubbard argues that “If the transition costs are borrowed, the resulting higher explicit federal debt in the near term is offset by lower implicit debt (Social Security obligations) in the longer run.” [19] Converting implicit debt into explicit debt, the argument goes, has no impact on the government’s long-term financing and thus should not be recorded in budget accounts.

The argument that implicit debt is equivalent to explicit debt, however, is mistaken. First, the trillions of dollars in “explicit debt” that the government would incur today would have to be financed in financial markets. That debt must be rolled over when it matures. The magnitudes involved are substantial. Table 2 shows what would happen to debt held by the public over the short-to-medium run under two of the leading individual account plans.

Table 2:
Debt Held by the Public Under Individual Account Plans (Percent of GDP)






CBO Baseline





Commission Model 2





Ferrara / Ryan-Sununu





Calculations based on CBO, 12/2003, The Long-Term Budget Outlook, Scenario 2 and memoranda from the Office of the Actuary, Social Security Administration.

In the absence of individual accounts, the debt held by the public is projected to rise from 34 percent of GDP in 2000 to 69 percent of GDP in 2030, based on conservative CBO projections. The Commission Model 2 plan would add to the debt an amount equivalent to 27 percent of GDP, bringing the total debt held by the public to nearly 100 percent of the size of the U.S. economy. (The Ferrara / Ryan-Sununu plan would more than double the debt held by the public by 2030, raising it to 144 percent of GDP, or nearly 1½ times the size of the economy.) To put this in perspective, the fiscal requirements associated with the European Monetary Union require debt levels below 60 percent of GDP.

These increases in explicit debt reduce the government’s flexibility, because the explicit government debt must be financed in private credit markets. The “implicit debt” associated with future Social Security benefit promises, however, does not need to be purchased now by bond market investors, and it does not need to be rolled over, as explicit debt does when it matures. Other countries have experienced fiscal crises when they were unable to roll over their explicit debt in financial markets. We are unaware of any country that has experienced a fiscal crisis solely because of its implicit debt. In other words, transforming explicit debt into implicit debt increases the government’s vulnerability to a drop in confidence in financial markets.

Relatedly, the implicit debt associated with future benefit promises can be reduced through future tax increases or future benefit reductions. Explicit debt, however, can be reduced only through current tax increases, current program cuts, or an unprecedented federal default.[20] Thus, replacing implicit debt with explicit debt limits the time-flexibility of the policy choices the federal government faces. The “harder” existing debt that would replace unfunded Social Security benefit promises is another manifestation of the reduced room for maneuver associated with transforming implicit debt into explicit debt.

Furthermore, borrowing to fund individual accounts could worsen the federal government’s long-term balance sheet, in addition to reducing its policy flexibility. Much of the projected future savings incorporated into individual accounts plans may never materialize. The argument for leaving several trillion dollars of “transition costs” out of the budget rests on a shaky assumption — that benefit reductions in an individual-account plan that are not slated to take full effect for many decades can be counted upon to offset large upfront borrowing costs. Once those future benefit reductions begin to bite, however, political pressures may build that ultimately cause the reductions to be reversed in whole or in part.

It also should be noted that if the stock market plunged at some future point, markedly reducing the value of the individual accounts, there could be tremendous political pressure for the federal government to bail out Social Security beneficiaries who hold the accounts. The market can, and probably would, factor in the cost of a potential bailout, even if that cost were not included in the official scoring of a Social Security individual accounts plan.

As a result, under a plan that rests on large-scale borrowing to establish individual accounts, the upfront borrowing costs would be certain to be incurred while some of the offsetting savings assumed in later decades could fail to materialize. This type of outcome has occurred a number of times before in other policy areas. (See box.) As a result, there is a distinct possibility that an individual account plan that rests on large-scale borrowing could end up worsening the federal balance sheet over the long term.

Three Examples of Program Reductions that Were Repealed Before They Took Effect

In 1985, an increase in “readiness” funds for the military was paid for by a reduction in future military retirement benefits that was slated to take effect in the latter part of the 1990s. (The future benefit reduction was treated as reducing the current “accrual” payments that the Pentagon needed to make to the military retirement trust fund.) But the reduction in the retirement benefits was later repealed, shortly before it was to take effect. Similarly, the “Freedom to Farm” Act of 1996 increased farm price support payments initially in return for a significant reduction in those payments in subsequent years. The reductions, however, were undone by later Congresses. Finally, the 1997 Balanced Budget Act paid for tax cuts partly by reducing Medicare provider payments, but the provider reductions have repeatedly been delayed or scaled back.

In short, the argument for leaving the borrowing to fund individual accounts out of the budget rests on the mistaken assumption that such a budgetary maneuver merely exchanges one type of debt for another and would be fiscally sound. In reality, such a maneuver would be unsound, as it would reduce the government’s fiscal flexibility and could lead to an increase in its overall liabilities.

Creating a New Loophole Could Set a Precedent for Future Budget Gimmickry

Allowing the budget to ignore large-scale federal borrowing undertaken in conjunction with individual accounts would create incentives for policymakers to enact individual account plans that fail to increase national saving, reduce the government’s long-term fiscal flexibility, and actually worsen the long-term fiscal outlook. Compounding this problem, omitting borrowing for individual accounts from the budget — and acting as though it would not increase the deficit — could have far-reaching implications for federal budgeting as a whole. It could establish a precedent that could encourage new levels of budget gimmickry.

For example, the same logic would suggest that a large tax cut or a program increase such as an expansion of the Medicare drug benefit could be “paid for” by raising taxes or cutting programs in the future, even if such tax increases or programs reductions were not slated to take effect until far in the future and were so implausibly large that they likely would be repealed by a future Congress. Under such an approach, the short-run costs of the tax cut or program increase would be left out of the budget. Furthermore, the larger and hence the more unrealistic the future tax increases or program reductions were, the more that taxes could be cut or other programs increased now, with no apparent budget “cost.”

These considerations are similar in many ways to the logic that has led policymakers to reject capital budgeting, based on concerns that it would defer the consideration of costs in a way that would make it easier to avoid tough budgetary choices. According to an analysis included in the most recent Bush Administration budget:

There have been a number of proposals to change the basis for measuring capital investment in the budget. Many of these would undermine effective consideration and control of costs by spreading the real cost of the project over time… This could be several years after the initial expenditure, in which case the budget would record no expenses at all in the budget year or several years thereafter, even though the Government is obligated to buy the asset... Control can only be exercised up front when the Government commits itself to the full sunk cost. [21]

The Proposed Budget Shift Could Unsettle Financial Markets

Taken together, the problems posed by establishing individual accounts and bending the budget rules in a way that virtually guarantees the accounts will be funded by large-scale borrowing could unsettle financial markets.

Financial markets are likely to believe that a good part of the projected “implicit debt” in Social Security will never materialize because, as insolvency approaches, Congress will take steps to ensure that Social Security does not become insolvent, as it did in 1983. As a result, capital markets are likely to assume that the long-term Social Security debt actually will be significantly lower than the implicit debt projected when current program parameters are assumed to remain in effect forever. (This is an extension of the late economist Herbert Stein’s famous dictum that if a trend can’t be sustained forever, it won’t be.) According to the Washington Post, even Glenn Hubbard has essentially acknowledged that the capital markets believe the often-cited figures for Social Security’s “implicit debt” overstate the problem “because few international lenders believe future retirees will get all of their scheduled benefits.” [22]

In addition, as emphasized above, transforming implicit debt into explicit debt increases the vulnerability of the government to a fiscal crisis in which it faces trouble financing or rolling over its public debt. Financial markets are likely to recognize that the possibility of a crisis is higher with explicit debt than implicit debt; unfortunately, that recognition itself makes a crisis more likely. It also should be noted that to the degree that national saving falls (as a result of people reducing other retirement saving in response to individual accounts funded by government borrowing), both the current account deficit and interest rates would rise.

Finally, leaving the borrowing costs for individual accounts out of the budget could weaken financial markets’ confidence in the reliability of federal budget reporting and the soundness of the nation’s fiscal policy course. Failing to record the costs associated with individual accounts in the budget would reduce the credibility of the federal budget and thus of statements the nation’s leading public officials would presumably make in the future regarding the deficit and the nation’s fiscal well-being. Markets might also assume that further budget gimmicks along the same lines would be employed in the future, resulting in additional deficits and borrowing.

Given these factors, it seems highly unlikely that funding individual accounts through large-scale borrowing would have no effect on financial markets. Indeed, such borrowing could provide a spark that ignites broader fiscal troubles.


Treating federal borrowing for individual accounts differently from federal borrowing for other purposes would create a large loophole in longstanding budget rules that is unwarranted and inconsistent with other federal budgeting practice. Of particular concern, it would likely lead to less fiscally responsible policy outcomes.

The measure of the unified budget deficit provides a valuable benchmark of the federal government’s impact on capital markets. It also provides a useful discipline on policymakers.[23] This measure should not be compromised, with potentially disruptive effects on financial markets, in order to make it easier to pass a particular piece of legislation.

Moreover, this change is entirely unnecessary. Policymakers can and should examine the impact that various proposals would have on the unified budget beyond the first five or ten years. But that does not require altering the budget rules. Instead, policymakers should simply analyze the impact of policies on the unified budget over a longer time period, while exercising appropriate judgment concerning the inherent uncertainties of such projections. Other benchmarks, including solvency projections for Social Security and Medicare and other accrual-based measures of government liabilities, also are of use in Social Security and other budget deliberations. Treating these measures separately would allow policymakers to apply their judgment to questions such as whether future policy changes are likely to materialize or to be reversed before they occur.

Given the fact that the nation faces substantial near-term deficits that are projected to enlarge significantly over time, the appropriate goal for policy should be to lower the unified deficit in the near term and to restore the fiscal health of Social Security and Medicare in a fiscally responsible manner over the long term. Traditional Social Security reforms, like the reform plan designed by former Social Security Commissioner Robert Ball (who helped fashion the 1983 Greenspan Commission Social Security plan) and a plan designed by economists Peter Diamond and Peter Orszag, contain those features. Individual account proposals can have these features, as well, if the substantial cost of creating such accounts is offset by reductions in other current government spending or increases in current revenues, rather than financed by borrowing.

Manipulating the budget rules to ignore the borrowing costs of setting up individual accounts would weaken incentives for policymakers to achieve these essential goals. Doing so also would endanger the nation’s credit standing at a precarious time and likely exacerbate the serious long-term fiscal problems the nation faces.


End Notes

[1] Jason Furman has served as a staff economist on the President’s Council of Economic Advisers and Special Assistant to the President for Economic Policy (during the Clinton administration) and as a lecturer at Columbia and Yale Universities; William G. Gale is the Arjay and Frances Fearing Miller Chair in Federal Economic Policy at the Brookings Institution and Co-Director of the Urban Institute-Brookings Institution Tax Policy Center; Peter R. Orszag is the Joseph A. Pechman Senior Fellow at Brookings and Co-Director of the Tax Policy Center. The authors thank Henry Aaron, Alan Auerbach, Alan Blinder, Peter Diamond, Douglas Elmendorf, Edward Gramlich, Robert Greenstein, David Kamin, Richard Kogan, Jeff Liebman, and David Wilcox for helpful discussions or comments.

[2] Although we believe there are significant problems with individual accounts (see, for example, Diamond and Orszag, 2003, and Aaron and Reischauer, 2001), this paper is not intended to be a criticism of individual accounts per se. We are concerned here only with the appropriate budgetary treatment of such accounts.

[3] Richard Stevenson, “Vast Borrowing Seen in Altering Social Security,” New York Times, November 28, 2004, and Jonathan Weisman, “Republicans Finding Ways To Account For Overhaul,” Washington Post, November 23, 2004.

[4] The major exception to the current cash-based measure of the unified budget deficit is the credit scoring rules adopted in the Credit Reform Act of 1990. Those rules specify that direct loans are not counted as an outlay on a cash basis; instead, only the subsidy value of the loan is charged to the unified budget. In addition, the expected costs of loan guarantees are charged to the unified budget when they accrue, without waiting for them to materialize on a cash basis. The arguments for making an exception for credit scoring do not apply, however, to individual accounts. First, loans are a contract and therefore a financial asset of the government. In contrast, future Social Security benefits are not a contract and can be altered by the government. Second, our goal in this paper is to ask whether making an exception to current scoring rules for individual accounts is likely to lead to more responsible or less responsible policymaking. Our conclusion is that it would lead to less responsible policymaking, principally because ignoring the cost of individual accounts reduces the incentives to ensure that the establishment of such accounts contributes to national savings and reduces the government’s overall liabilities. The issues in credit scoring are very different, and they generally do not involve solving long-term fiscal problems. Finally, the amount of money involved in individual accounts is unprecedented. The accounts thus are much more likely to have a macroeconomic impact on American capital markets than direct lending has.

[5] For example, Jagadeesh Gokhale and Kent Smetters, Fiscal and Generational Imbalances (AEI Press: Washington, DC) American Enterprise Institute, 2003.

[6] Alan Auerbach, William Gale, Peter Orszag, and Samara Potter, “Budget Blues: the Fiscal Outlook and Options for Reform,” in Henry Aaron, James Lindsay, and Pietro Nivola (eds.) Agenda for the Nation, (Washington, DC: Brookings) and Peter Diamond and Peter Orszag, “Comment: Accrual Accounting for Social Security,” Harvard Journal on Legislation, 41. These papers generally conclude that annual accrual accounting would be a useful addition to presentations of the federal budget outlook but should not become the central accounting approach used in policy discussions or in budget scoring decisions. The issue is not whether accrual accounting information is helpful. Rather, the issue is whether recasting the political process to concentrate primarily on accrual accounting measures rather than the current budget accounting measures would be more likely to generate sound, responsible decision-making. Although accrual measures are more comprehensive in their estimates of choices over a long time horizon, they also contain much more uncertainty and are much more sensitive to small changes in assumptions than is often recognized.

[7] President’s Commission to Strengthen Social Security, 2001, Final Report.

[8] Council of Economic Advisers, Economic Report of the President (U.S. Government Printing Office, Washington: 2004), pp. 143-144.

[9] CBO notes that shifting a portion of payroll tax contributions to individual accounts would not affect the unified deficit if the accounts were effectively owned by the government, so that the contribution of payroll taxes into the account represented a shift from one government account to another. As CBO has explained, “For example, if the government had complete control over the use and disposition of PRA balances, the accounts should be included in the budget. In that event, outlays recorded when funds were deposited into the PRAs would represent a transfer of money to a federal fund. The fund would also record all other transactions of the accounts.” In this case, the transfer would generate an outlay in one part of the budget and an offsetting receipt in another part of the budget, with no net effect on federal borrowing and hence no effect on the unified budget.

[10] For example, see CBO, “Long-term Analysis of Plan 2 of the President’s Commission to Strengthen Social Security,” July 21, 2004, and CBO, “Long-term Analysis of H.R. 3821, the Bipartisan Retirement Security Act of 2004,” July 21, 2004.

[11] Advisory Council on Social Security, Report of the 1994-1996 Advisory Council on Social Security, 1997.

[12] President’s Commission to Strengthen Social Security, Final Report, 2001.

[13] Charles P. Blahous, Reforming Social Security (Praeger, 2000), page 133.

[14] These figures are presented in present value: the amount today that, with interest, would exactly cover these future costs.

[15] National saving is not the only consideration in evaluating individual accounts proposals. A proposal might boost national saving but be undesirable for other reasons, such as unnecessary administrative costs, increased risk, or the undesirability of the offsetting changes made to finance the individual accounts. Such issues are beyond the scope of this paper. It is safe to say, however, that there is virtually no justification for an individual account proposal that does not raise national savings.

[16] Martin Feldstein and Andrew Samwick, 2000, “Allocating Payroll Tax Revenue to Personal Retirement Accounts to Maintain Social Security Benefits and the Payroll Tax Rate,” NBER Working Paper No. 7767.

[17] Some supporters of individual accounts argue that private saving would rise because the accounts would demonstrate to households the benefits of saving. We are skeptical that this effect, even if it exists, would be sufficient in the aggregate to offset the negative effect from households reducing their 401(k) and IRA saving in response to the creation of individual accounts within Social Security.

[18] Douglas Elmendorf and Jeffrey Liebman, 2000, “Social Security Reform and National Saving in an Era of Budget Surpluses,” Brookings Papers on Economic Activity.

[19] Glenn Hubbard, “How Bush’s Plan Would Secure Social Security,” Business Week, November 29, 2004.

[20] We assume that the Federal Reserve would not allow policy-makers to partially default through unexpected bouts of high inflation.

[21] Office of Management and Budget, Analytical Perspectives, Fiscal Year 2005, February 2004.

[22] Jonathan Weisman, “Theoretically, Tax Reform Should Fly,” Washington Post, December 3, 2004.

[23] This does not mean that the unified deficit is a perfect measure or the only measure that policymakers should use to gauge the impact of their policies.