Tax Cuts and the Budget Outlook

Alan J. Auerbach and
Alan Auerbach Headshot
Alan J. Auerbach Robert D. Burch Professor of Economics and Law - Economics Department, UC-Berkeley, Director - Robert D. Burch Center for Tax Policy and Public Finance
William G. Gale
William G. Gale The Arjay and Frances Fearing Miller Chair in Federal Economic Policy, Senior Fellow - Economic Studies, Co-Director - Urban-Brookings Tax Policy Center

April 2, 2001

Although the current projected budget surplus is $5.6 trillion over the next ten years, basic corrections to incorporate responsible approaches to budgeting and realistic forecasts of current policies reduce the available surplus to between $1.0 and $1.7 trillion. A full recognition of the growth of unfunded entitlement programs makes the surplus vanish entirely. These findings suggest that President Bush’s proposed tax cut—which is billed as a $1.6 trillion proposal but which would in fact cost well over $2 trillion—is excessively large. In addition, the president’s tax proposal is poorly designed to stimulate the economy, and claims that tax cuts are needed to stop wasteful government spending and to avoid fully paying off the government’s debt are flawed.


The Budget Outlook Over the Next Ten Years

The most recent Congressional Budget Office (CBO) baseline forecast, released in January 2001, projects cumulative unified budget surpluses of $5.6 trillion between 2002 and 2011, including $2.5 trillion in the Social Security trust fund (the “off-budget” surplus) and $3.1 trillion in the rest of the budget (the “on-budget” surplus). Although the CBO baseline provides a common and visible benchmark, it is limited in several crucial ways and, by itself, does not provide sufficient information to assess various policy options. The baseline includes trust fund accumulations for Social Security, Medicare, and government pensions, but ignores the associated accruing liabilities. The baseline also uses conventions about spending and taxes that are inconsistent with likely outcomes. Our goal in adjusting the 10-year surplus figures is to derive an estimate of the resources that policymakers may reasonably choose to allocate, using responsible budget practices and realistic assumptions.

Retirement trust funds

In the past, President Clinton and congressional leaders of both political parties agreed that accruing Social Security trust fund balances should contribute to improving that program’s long-term financial viability, and should not be used to finance tax cuts or other spending programs. Indeed, as discussed below, the current Social Security trust fund accumulations are far outstripped by the accruing liabilities that they are meant to help offset. A full accounting of the Social Security system would show it substantially in deficit when long-term liabilities are considered. At the very least, though, if the budget ignores the liabilities, it should certainly not count the assets as available for other purposes. This logic applies to other trust funds as well.

Medicare pays for health care for the elderly in two parts. Part A, hospital insurance, covers hospital costs and is financed by payroll taxes. Part A is very similar in structure to Social Security. Workers contribute payroll taxes to a trust fund while working, and receive promised benefits when they are elderly. Part B, supplementary medical insurance, is financed by a combination of user fees and general revenues. Over the next ten years, CBO projects the Medicare trust fund, which covers Part A, will run surpluses totaling $392 billion. Although Medicare is officially part of the on-budget surplus, both houses of Congress voted last year to support measures that protected the trust fund from being used to finance tax cuts or programs other than Part A. The House of Representatives approved the measure by a vote of 420-2. The Senate passed two separate measures; 98 senators voted in favor of one or both. The House also passed a similar resolution earlier this year by 407-2. The strong votes demonstrated overwhelming congressional support for preserving the Medicare trust fund.

While the Social Security and Medicare trust funds have received significant attention in the budget debate, a third set of retirement funds has not. Trust funds holding pension reserves for federal military and civilian employees are projected to accrue surpluses of $419 billion over the next ten years. Under current budget procedures, these surpluses are a component of the on-budget surplus. Like Social Security and Medicare, however, these trust funds represent current accumulations intended to provide retirement benefits to future workers. Thus, they should be protected as well. Many states, in fact, already separate their pension reserves from funds available for tax cuts and other spending.

The definition of current policy

The baseline forecast projects current policy subject to a variety of statutory requirements, which limit the scope of the forecast’s underlying assumptions and time horizons, and can be at variance with reasonable expectations. Thus, one should not confuse the CBO baseline with a forecast of what is most likely to occur.

Mandatory spending—e.g., entitlements, such as Social Security—is generally assumed to continue as it is currently structured in the law. Discretionary spending, however, poses problems in defining “current policy.” Unlike mandatory spending, discretionary programs—e.g., defense, education, the environment, or infrastructure—are not automatically included in the annual budget and thus require annual appropriations from Congress. No consensus exists about how to project current policy for discretionary programs, and CBO simply assumes that real discretionary spending authority will remain constant at fiscal year 2001 levels.

This assumption is clear, but may not be very reasonable. Discretionary spending totaled 6.3 percent of Gross Domestic Product (GDP) in 1999 and 2000, the lowest share since at least 1962. Under CBO’s 10-year baseline forecast, discretionary spending would fall to 5.1 percent of GDP. It would also fall by more than 10 percent in per capita terms. In a growing economy with growing defense needs and other concerns, this seems a particularly unrealistic projection.

At the very least, it would be more reasonable to have real discretionary spending grow at the same rate as the population (about 1 percent per year). Indeed, President Bush has made precisely the same point, arguing during the campaign that an “honest comparison” of spending growth should take inflation and population growth into account. Incorporating this baseline would raise discretionary spending by $359 billion and, counting the added interest payments on federal debt that would be required, would reduce available surpluses by about $418 billion.

To put these figures in perspective, during the campaign President Bush proposed new spending programs totaling $475 billion, along with cuts in government spending of $196 billion, for a net spending increase of $279 billion between 2001 and 2010. This is virtually identical to the cost of having real discretionary spending grow by 1 percent over the same period (rather than over 2002-2011). Thus, this suggests that having real discretionary spending grow by 1 percent is a lower bound for the likely path of discretionary spending. Congressional Democrats and Republicans are likely to have proposals of their own over the next ten years, President Bush may have more proposals for spending—especially on defense—after his initial round of proposals, and emergencies or other contingencies will inevitably arise.

An alternative, and perhaps more realistic baseline would have discretionary spending grow at the same rate as nominal GDP, thus keeping the ratio of discretionary spending to GDP constant relative to the baseline forecast, this would increase spending by $905 billion and reduce the available surplus by $1,055 billion between 2002 and 2011.

At least two aspects of current policy toward taxation merit consideration. The first concerns the alternative minimum tax (AMT), one of the most complex areas of tax law. The AMT was created to exact some taxes from the small number of taxpayers who are considered too aggressive in creating shelters and claiming deductions.

In practice, the AMT has affected few taxpayers. In 2000, for example, only 1.3 million taxpayers faced the levy. Under current law, however, that figure will rise to almost 21 million by 2011. The main reason is that the AMT exemption is not indexed for inflation. CBO’s surplus forecasts assume that the dramatic rise in AMT taxpayers will occur. The increase, however, would be fought fiercely by the affected groups. Indeed, the problem has already received significant attention, even though only a small portion of taxpayers currently face the tax.

“Current policy” would be better represented by indexing the AMT for inflation. This would keep the number of taxpayers on the AMT limited to about 1.9 percent by 2010. The lost tax revenue from this policy would total $113 billion over the next ten years. Counting the added interest, the net cost would be $130 billion.

A second tax issue relates to temporary tax provisions, many of which are scheduled to expire over the next decade. For all taxes other than excise taxes dedicated to trust funds, CBO assumes that legislated expirations occur as scheduled. In the past, however, the temporary provisions have typically been extended another few years when the expiration dates approached. In light of this practice, current policy is more aptly viewed as assuming that these so-called “extenders” will be granted a continuance. Extending the provisions—except the one relating to AMT, since the AMT is addressed above—through the 10-year horizon would cost a net of $69 billion in lost revenues plus an additional $13 billion in interest payments.

Implications for the available surplus

Adjustments for responsible budgeting practices and realistic policy assessments profoundly affect the amount that should be considered available for tax cuts or new spending (Table 1). Almost 60 percent of the projected 10-year baseline surplus—$5.6 trillion—is due to the retirement trust funds. Removing Social Security generates an “on-budget” surplus of $3.1 trillion. Removing Medicare reduces this figure to $2.7 trillion. Protecting government pension funds reduces the available surpluses to $2.3 trillion.

Adjusting for the AMT and expiring tax provisions reduces the available surplus to $2.1 trillion. If real discretionary spending were held constant on a per capita basis, the available surplus for other programs would be $1.7 trillion. If discretionary spending were held constant as a share of GDP, the remaining surplus would be about $1 trillion.

Looking Beyond the Ten-Year Horizon

As indicated, the Social Security and Medicare trust fund balances fall far short of what would be needed to meet liabilities under current policy. This imbalance is masked by the asymmetry in unified budget accounting practice that counts assets for these programs but not liabilities, a practice only slightly improved on by placing the assets off-budget. Were this problem corrected, the 10-year surpluses just reported would be negative. An alternative way of recognizing these entitlement liabilities is to extend the planning horizon to include the future years in which the liabilities come due and thus can no longer be ignored, even under the cash accounting method.

The use of long-term planning horizons is now standard for Social Security and Medicare. In the context of an aging population and rapidly rising medical care expenditures, this is the only way to get an accurate picture of the fiscal balance of these programs, and hence the government’s budget as a whole.

To take these and other factors into account, we estimate the long-term “fiscal gap” under different policies. The fiscal gap is the size of the permanent increase in taxes or reductions in non-interest expenditures (as a constant share of GDP) that would be required now to keep the long-term ratio of government debt to GDP at its current level. Over an infinite planning horizon, this requirement is equivalent to assuming that the debt-GDP ratio will not explode. The fiscal gap gives a sense of the current budgetary status of the government, taking into account long-term influences. To calculate the fiscal gap, we use the most recent CBO 10-year and long-term forecasts.

Under the CBO baseline assumptions about discretionary spending—constant in real terms until 2011 and constant as a share of GDP thereafter—the fiscal gap through 2070 is projected to be 0.67 percent (line 1, table 2). That implies that a permanent tax increase or spending cut of 0.67 percent of GDP, which would currently be about $67 billion per year, would be required to achieve the same national debt-GDP ratio in 2070 as currently exists.

The fiscal gap on a permanent basis—what would be needed to prevent the national debt from exploding in the long run—is currently 3.33 percent of GDP. The permanent gap is so much larger because the budget is projected to be in substantial deficit during the years approaching 2070 (and those that follow). Allowing discretionary spending outlays to remain constant as a share of GDP from now on, rather than beginning only in 2011, raises the fiscal gap further, to 1.45 percent of GDP over the next 70 years and 4.14 percent on a permanent basis.

Given the long-term imbalance, the fiscal climate may be more troubling now than previously. In the 1980s and early 1990s, when the country faced both short- and long-term deficits, the short-term deficits helped focus attention in a way that also helped reduce long-term gaps. Today, the United States still faces a trade-off between current and future generations, and still confronts a long-term shortfall. But the current policy discussion focuses on policies that likely would exacerbate the long-term situation.


All of the estimates above are subject to considerable uncertainty. Although CBO’s underlying economic assumptions do not appear to be unreasonable, there is massive uncertainty regarding the evolution of the economy and the surplus. Long-term estimates are inherently more uncertain than short-term estimates. But the added uncertainty should not lead us to ignore long-term issues. Indeed, the serious consequences of a relatively bad long-term outcome should spur policymakers to take precautions now. Also, note that the sources of uncertainty differ in the long and short runs. Over the next ten years, the primary factor affecting surpluses will be the economy. Over the longer term, the demographic pressures of an aging population will become dominant. While the magnitude of this demographic shift is uncertain, its occurrence is not.

President Bush’s Budget and Tax Proposals

The administration’s underlying baseline forecast is similar to CBO’s in many respects. The administration baseline forecasts a 10-year surplus of $5.6 trillion, a Social Security trust fund surplus of about $2.6 trillion, and a Medicare Part A trust fund surplus of $526 billion. Thus, the projected non-Medicare, non-Social Security surplus is $2.5 trillion. The administration’s budget does not provide resources for fixing the AMT or granting permanent continuances to expiring tax credits.

The administration proposes significant but subtle changes to the budgetary treatment of Social Security and Medicare. Taken together, these changes could strip more than $1 trillion from the Social Security and Medicare trust funds, threaten recent fiscal accomplishments, and undermine the bipartisan consensus discussed above to protect the Social Security and Medicare trust funds.

The administration claims that only $2 trillion of debt can be repaid during the budget period, so that about $600 billion of the Social Security surplus is left over. The administration would not invest these funds in private assets. Thus, by definition, the funds would be left in the budget and hence would be used to finance some combination of tax cuts above those already proposed or other spending above amounts already called for in the budget. The latter possibility includes Social Security privatization, as hinted in the budget. Of course, if the funds are eventually used for anything other than debt reduction, Social Security benefits under the existing program, or government investment in assets that serve as a reserve from which to pay future Social Security benefits under the existing system, the administration will have effectively siphoned off funds from the Social Security trust fund to use for other purposes.

A similar policy is proposed explicitly for Medicare. The Part A trust fund is projected to run a surplus of $526 billion over the decade. Part B—supplemental medical insurance (SMI)—is stipulated by statute to be funded from general revenues and from medical insurance premiums. The administration, however, claims that Medicare as a whole is running a deficit over the next ten years, because the general revenue financing requirements of Part B exceed the trust fund accumulations of Part A. Thus, the administration proposes to wipe out the $526 billion Part A cash flow surplus by spending it on Part B. There are numerous problems with this policy. The taxes paid into Part A, and hence the surpluses there, are intended to finance future hospital insurance, not current SMI. The Part A trust fund also faces a long-term deficit, as described above, so removing the funds from Part A exacerbates that situation. In effect, the administration is trying to use Part A revenues to pay for SMI in order to release general revenues—which normally pay for SMI—to finance tax cuts.

The president’s tax cut is the centerpiece of his budget. The major elements include: reducing the highest income tax rates; abolishing the estate tax; creating a 10 percent tax bracket; expanding the child credit to high-income households, reducing the phase-out rate, and doubling the credit amount; and allowing a second-earner deduction. In his budget submission of February 28, the president included additional tax incentives, mostly small targeted credits or subsidies.

Issues in the Tax Cut Debate

Many of the arguments put forth in favor of tax cuts are tired refrains that are easily rejected. One cites the historically high share of GDP being absorbed by federal taxes as evidence that tax burdens have risen too high and need to be reduced. Aside from whether such a move would be fiscally responsible, this argument ignores that tax burdens on most families, as a share of income, actually have fallen during the past several years. The increase in taxes as a share of GDP is attributable to a rise in effective tax rates at the upper end of the income distribution and, more importantly, the increasing share of income going to these individuals, and therefore subject to higher marginal tax rates. A second argument, that the surplus is the American people’s money and should be given back to them, simply ignores the fact that, with a long-term fiscal imbalance, giving money to some individuals amounts to taking even more from others. Government is already committed to very large social programs that the American public has no apparent desire to curtail.

Is the president’s tax cut too big?

Considering the size of the president’s tax cut is more complicated than it might first appear. Recently, HR 3—which was passed by the House on March 8—included the creation of a 10 percent bracket and reduction in marginal tax rates, with the former accelerated by one year relative to the president’s original plan. The Joint Committee on Taxation has estimated that if HR 3 were enacted, approximately 36 million taxpayers would face, or be affected by, the AMT by 2011. This is 15 million more than the 21 million that would be placed under the AMT under current law. The administration has acknowledged that the AMT creates a problem for the proposed tax cut. Indeed, the tax program on the Bush campaign web site (where voters could calculate how much of a tax cut they would receive under Bush’s plan) did not allow for the AMT to reduce anyone’s tax cut, and thus implicitly assumed that an AMT offset was de facto a part of the Bush plan. For these reasons, it is arguable that the necessary AMT adjustments should be considered part of the Bush plan. But that argument is semantic. Even if they are not considered a part of the plan per se, the required AMT adjustments are a problem created by the plan that will need to be addressed. To be clear, the adjustments would merely undo the increase in the number of AMT taxpayers due to the Bush plan. They would not address the increase in the number of AMT taxpayers expected to occur under current law even in the absence of tax changes.

Table 3 breaks down the revenue costs into several components. The provisions of HR 3, the other components of the plan the president submitted to the Congress on February 8, and the additional tax incentives in the president’s budget would cost $1.8 trillion through 2011. AMT adjustments would total about $300 billion. The added interest payments on federal debt caused by the reduction in federal revenues would cost roughly $400 billion. Thus, although the proposal is often referred to as a $1.6 trillion tax cut, the real cost of the proposal now comes to $2.5 trillion over the next ten years. Even if the AMT adjustments are ignored, the plan would cost almost $2.2 trillion over the next decade.

Ultimately, assessments of whether Bush’s proposed tax cut is too large are in the eyes of the beholder. Several analytic perspectives, however, suggest the tax cut is excessive. First, the magnitude of the tax cut exceeds the “available surplus” estimates listed in Table 1 by between $500 billion and $1.5 trillion. Second, the surplus excluding retirement trust funds equals $2.3 trillion. Thus, enacting the administration’s tax cut (even without the AMT adjustment) would imply that virtually no other policy priorities could be met unless Congress were willing to finance the programs with balances in the retirement trust funds. Third, the Bush tax cut with AMT adjustment is equal in size to the administration’s entire projected non-Medicare, non-Social Security surplus.

Some have argued that official revenue estimates, in omitting potentially important behavioral responses, such as increases in labor supply or shifts of income from tax preferred to taxable forms, overstate the revenue cost of the tax cut, and that incorporating such responses could reduce the costs by one-quarter. While there is no doubt that official revenue estimates omit some types of behavioral responses, there is significant controversy over how large those responses are. Moreover, analyses of the economic responses typically do not consider the further effects on revenue of the reduction in national saving that would occur along with the cut in taxes. The reduction in national saving would reduce economic growth.

A fourth perspective on the size of the cut can be gleaned from Table 2: the proposed tax cut would substantially increase the long-term fiscal gap. Even if discretionary spending were held constant in real terms through 2011, so that it fell continuously as a share of GDP, Bush’s tax cut would triple the long-term fiscal gap through 2070 and raise it by 150 percent on a permanent basis. This would significantly worsen the long-term fiscal problem the government faces, which should not be surprising. By 2011, the administration’s tax cut, including the AMT adjustment, would reduce income tax revenues by almost 19 percent on a permanent basis.

Fifth, to respond to the criticism that the tax cut is irresponsibly large, some advocates have argued that Bush’s proposals are much smaller relative to the economy than the 1981 tax cut signed by President Reagan. Reagan’s tax cut has been estimated to have reduced revenues by between 4.2 percent and 5.6 percent of GDP, much larger than the estimates for the Bush plan. However, although the tax system is largely indexed for inflation now, it was not in 1981. CBO estimates that at least 40 percent of the Reagan tax cut was simply an elimination of tax increases that would have occurred because of inflation, suggesting a real tax cut of between 2.4 and 3.3 percent of GDP. Bush’s proposed tax cut, on the other hand, may appear smaller than it actually is because it would be phased in slowly. Thus, estimates of the 10-year average cost of the plan relative to GDP are misleading. A better estimate is the annual cost after the plan has been fully phased in. In 2011, for example, the cost of the tax cut and the AMT adjustments would be about 1.95 percent of GDP.

In addition, the 1981 tax cuts were immediately countered by income tax increases in 1982 and 1984, with further increases in 1990 and 1993, reflecting a realization of how damaging the original plan had been to the government’s fiscal balance. Using the 1981 tax cut as a benchmark for responsibility, then, is somewhat problematic.

Is the president’s tax cut a good way to stimulate the economy?

The Bush administration has argued that the case for its tax cut is strengthened by the possibility that the economy is entering a recession. Even if there were currently a recession, and even if tax cuts could in general fight recessions effectively—both problematic assumptions—the president’s tax proposal is poorly designed to fight an economic downturn. First, in the president’s original proposal, there is no tax cut proposed for fiscal 2001 (which ends on September 30, 2001). It is hard to see how a tax cut could boost the economy now if it is not providing any tax cuts now. In addition, the president’s plan would only provide tax cuts of $21 billion (about $75 per person or 0.2 percent of GDP) in fiscal 2002. Thus, in the next 18 months, the plan would provide virtually no stimulus at all. HR 3 would provide about $6 billion in tax cuts in 2001, but in a $10 trillion economy, this adjustment is trivial.

In addition, an anti-recession plan would be aimed at increasing current spending by government, business, or households. The president’s plan is aimed at the private sector, and provides no incentives for business investment. This leaves consumption as the main vehicle for increasing aggregate demand. But the plan targets the bulk of individual tax reductions at high-income households, who are less likely to be the short-horizon consumers who would spend a significant share of increases in after-tax income. To affect the economy sooner, the president’s tax reduction could be accelerated, but this would raise its revenue cost and further crowd out other spending priorities. It would also increase pressure on interest rates, which could hurt the economy.

In an effort to stimulate the economy, recent proposals by both Republicans and Democrats would vastly increase the tax cut in the first year to around $60 billion. This bipartisan effort shows more clearly than any analytical argument that the administration’s claim that its initial tax cut proposal would provide a current stimulus is specious.

Is the president’s tax cut the best way to restrain government spending?

Another argument claims that we need tax cuts because surpluses lead to bloated government spending. There is an element of truth to this claim. For example, Congress exceeded the caps on discretionary spending by increasing amounts in each of the past three years, and it may well be that the pressure of large deficits helps restrain spending. But the argument is flawed in important respects.

First, before concluding that government will spend all available funds, it is important to note that the vast portion of recent surpluses have been allowed to accrue. Government spending is at its lowest level relative to GDP in 35 years. Second, even if large surpluses do encourage undesirable spending, there is a much simpler and less disruptive way to restrict spending, namely, to reform the current method of government accounting that counts trust fund assets but not liabilities. Admitting that we do not have the money surely is preferable to giving away funds that we do not have in the first place.

Is the possibility of paying off all of the public debt a good reason for tax cuts?

An argument put forth recently by Federal Reserve Board Chairman Alan Greenspan, and quickly repeated by tax cut advocates, is that under current surplus projections, the government will pay off all available government debt by around 2006. At that point, with more surpluses pouring in every year, the government would have to start accumulating private assets. Greenspan and others argue that having the federal government hold such assets would raise the specter of government interference in the operation of private companies, and to avoid this it would be preferable to cut taxes now, rather than be forced to either cut taxes or raise spending massively in the future.

However, it is possible for government trust funds currently holding marketable government debt to hold other assets instead. Designing mechanisms to protect against the hazards of government involvement in private securities markets is a serious issue that we do not address here. But it is worth noting that state and local government pension funds already hold $3 trillion in private assets—including $2 trillion in equities—earn competitive returns, do not appear to disrupt financial markets, and—even with virtually none of the safeguards that would be imposed on federal investments—invest only a tiny fraction of their funds in “targeted” investments. Moreover, other investment possibilities are available as well, including but not limited to the president’s own proposal to exchange the government debt being held by the Social Security trust fund for the assets held in the individual accounts of a privatized Social Security system.

How government would handle investment in private assets and how the Social Security system should be reformed are important questions to address. But the complex problems associated with any approach to making more government debt available to the public do not justify adopting an irresponsible fiscal policy, whether through large tax cuts or spending increases.


The current fiscal surpluses are a significant accomplishment, and should not be taken for granted. But the underlying budget situation is bleaker than the official numbers suggest, and proposed tax cuts would cost far more than advertised. Most of the other arguments given in favor of the president’s proposed tax cuts are demonstrably weak. Taken together, these factors suggest that tax cuts that consume virtually all of the available surplus over the next ten years and significantly worsen the long-term fiscal outlook remain a poor idea.

For a more extensive analysis of these issues, including citations for all of the facts in this policy brief, see “Tax Cuts and the Budget.”