Carl E. Walsh

FRBSF Economic Letter 1999-27 | September 10, 1999

After 15 years of federal budget deficits that overwhelmed every discussion of fiscal policy, the United States now faces the prospect of huge budget surpluses for the foreseeable future–that is, if recent projections by the Clinton administration and the Congressional Budget Office can be believed.

But can they?

After 15 years of federal budget deficits that overwhelmed every discussion of fiscal policy, the United States now faces the prospect of huge budget surpluses for the foreseeable future–that is, if recent projections by the Clinton administration and the Congressional Budget Office can be believed.

But can they? During the 1980s, projections of future deficits were notoriously inaccurate as forecasts of actual deficits, especially for projections far out into the future. The last two years have seen enormous revisions in the projection surpluses, and future years are likely to see similarly large revisions. This Economic Letter discusses the nature of the budget projections, the sources of the revisions, and the appropriate interpretation of the projections.

The budget revisions

Each year, the Congressional Budget Office (CBO) produces an analysis of the Federal government budget looking out ten years. Figure 1 illustrates how the budget outlook has changed dramatically over the past four years. Each dashed line shows the projected path of the deficit or surplus made at the time indicated next to each line. Each projection starts from the actual deficit at the time of the projection, represented by the points on the solid line.

In January 1997, the actual budget deficit of $107 billion in 1996 was projected to grow to $124 billion in 1997 and swell to $278 billion by 2007. The CBO’s projection for the 1997 deficit turned out to be off by over $100 billion–the projected $124 billion deficit turned into an actual deficit of $22 billion. This was just the first evidence that the budget outlook was about to take a huge swing.

The real change in the outlook for the federal budget shows up in the CBO’s 1998 report. Rather than a continuation of budget deficits, the CBO projected a balanced budget through the year 2000 with rising surpluses thereafter. Looking further out, the revisions between the January 1997 and January 1998 CBO reports were enormous. The 2007 projection shifted from a deficit of $278 billion to a projected surplus of $129 billion, a swing of over $400 billion!

The budget picture continued to improve–the $5 billion deficit projected for 1998 turned out to be off by $75 billion, with the federal government actually running a surplus of $70 billion, its first since 1969. By January 1999, the 2007 surplus had been revised up again, this time to $333 billion, an increase of over $200 billion. Just over two years ago, the CBO was projecting a cumulative deficit between 1999 and 2007 of $1.9 trillion; today it is projecting a $2.2 trillion surplus over those same years.

These large projected surpluses have been the focus of much debate in Washington. The turnaround in the projections in such a short period of time raises a number of questions. First, how should we interpret these projections? Are they forecasts? Or are they something else? How “good” are the projections? And what sorts of assumptions lie behind them?

Conceptual issues

Perhaps the first aspect to clarify is that projections are not forecasts. A forecast is the best guess today of the outcome of some future event. Making a forecast of the future surplus would require forecasting the likely path of government expenditures and receipts, and answering a question like: “What is the most likely value of the surplus for 2001?” That is not what the CBO does. Rather, it tries to answer a question like: “Under current expenditure and tax revenues programs, what is the likely value of the surplus in 2001?”

These two questions are quite different. For example, it is clear that, faced with projected surpluses, the President and Congress will not leave current expenditure and tax revenue programs unchanged. Both houses of Congress have already passed large tax cuts that would reduce the projected surpluses if signed into law. Expenditures also are likely to rise. As a result, the actual surplus the federal government will have in the future will be significantly below the levels currently being projected. If the government raises spending enough, or cuts taxes enough, the budget might turn out to be balanced in future years and no actual surplus will ever occur. In fact, the Clinton administration’s budget report shows a zero surplus for future years by assuming all extra funds will be set aside pending Social Security and Medicare reforms. This won’t mean the projections were wrong, it is just that the projections are based on current policies, and the projections will cause those very policies to be changed in ways that alter the budget.

What caused the big revisions?

Because projections are based on current policies as well as forecasts about economic development, three factors lead to revisions. First, government policies change. Second, forecasts about the economy change. Third, estimates of tax collections and spending may change even if policies and economic forecasts remain unchanged. While each of these factors has played a role in accounting for the marked change in budget projections, the two primary changes affecting the budget projections were the policy changes included in The Balanced Budget and Taxpayer Relief Act of 1997 and the continued strong growth of the U.S. economy.

The 1997 budget act is estimated to have cut the deficit by $127 billion over the 1998-2002 period, with most of the savings resulting from slowing the growth of Medicare spending. Caps on future discretionary spending also were lowered; these caps now require that the dollar value of discretionary spending remain constant between 1999 and 2002. (Congress can override these caps by passing legislation for emergency spending, as it did for expenses related to the war in Kosovo.) Constant nominal expenditures translate into a real decline in discretionary spending. For the period 1999 to 2007, these policy changes added over $600 billion to the surplus projections.

The continued strong performance of the U.S. economy has had an even larger effect on the projected surpluses. In January 1998, the CBO was forecasting 2.7% real GDP growth for 1998; actual growth came in a full percentage point higher, at 3.7%. By January 1999, the CBO had revised its estimate of average real growth for the 1999-2008 period from 2.1% per year to almost 2.3%. These upward revisions in expected growth add to the surplus by raising projected revenues and lowering expenditures. These effects can be quite large.

Revisions in the outlook for inflation and interest rates also have led to improvements in the budget outlook. Between January 1998 and January 1999, the CBO reduced its forecast for average CPI inflation over the 1999-2000 period from 2.8% per year to 2.6%. Lower inflation reduces the cost-of-living adjustments to Social Security, leading to a larger projected surplus. Forecasts of lower interest rates also improve the budget picture by reducing interest costs on the government’s debt.

To sum up, while the CBO was projecting the policy changes in the 1997 budget act would add $600 billion to the 1999-2007 surplus, it also changed its economic assumptions, which added $1 trillion to the surplus projections. In just the three months between September 1997 and January 1998, the CBO increased the projected surplus for 1998 alone by $22 billion and for 1999 by $28 billion due to changes in their economic assumptions. Between January 1998 and January 1999, similar changes added a further $270 billion to the projected surpluses over the six years from 1999 to 2004.

From policy assumptions to forecasts

The assumptions about government expenditures that lie behind the budget projections have come under heavy criticism. Expenditures projections are based on current policies, and these include caps on discretionary spending (spending on items other than mandatory spending, such as entitlement programs, and net interest) that were part of the 1997 budget act. These caps expire in 2002. The CBO’s projections make two controversial assumptions–that the spending caps will be met and that, after they expire, discretionary spending will increase only enough to keep pace with inflation.

Under the spending caps, discretionary spending for 2000 is limited to $587 billion. Simply freezing dollar expenditures at 1999’s level (excluding 1999 emergency spending) would still level discretionary spending $13 billion over the cap for 2000. Allowing spending to rise to reflect inflation so that real discretionary spending remained frozen would put spending $24 billion over the caps next year. Congress and the president would need to agree on $24 billion in expenditure cuts for next year to remain consistent with the spending assumptions that are built into the projections. It seems fair to be skeptical that they will cut existing programs in the face of huge projected surpluses. Actual expenditures are likely, therefore, to exceed the levels incorporated into the projections

Current projections assume federal outlays will fall from 19.5% of GDP in 1999 to just over 17% in 2009, while receipts will hover around 20%. Figure 2 shows how rarely the percent of GDP devoted to outlays has ever been so low or that devoted to taxes so high..

Economic assumptions

Changes in the economy also have the potential to alter the budget outlook drastically. A downward revision in forecasts for economic growth would lower future tax revenues and alter the projections. For example, the $1.2 trillion projected surplus over 1999-2004 would be reduced by over $300 billion if the economy were to grow 1% slower than assumed. Given the difficulties in forecasting future economic developments, the budget projections are subject to great uncertainty even if government policies remain constant.

A final factor to keep in mind is that changes in policy also will affect the economic forecasts. If government expenditures rise, or taxes are cut, national savings will be lower and interest rates will rise. This, in turn, will alter the projections for future interest expenditures.

Will the surpluses actually occur?

By their very nature, budget projections are likely to be wrong. Projections of large deficits, for example, should lead Congress and the President to change course to head off ballooning deficits. If the projections serve their purpose in leading to policy changes, the projected deficits will not occur. So one interpretation of projection revisions is simply that the initial projections lead to the policy changes that invalidate the projections. Similarly, the current projected surpluses are triggering changes in spending and revenue policies, changes that mean actual surpluses will be much smaller than current projections show.

Most economists, while opposing any requirement that the federal government balance its budget every year, do accept that notion that the budget should balance over longer time horizons. This requires that periods of budget deficits, such as those of the 1980s and most of the 1990s, be balanced by a period of surpluses. The U.S. struggled for fifteen years to eliminate the federal deficit; current proposals by Congress and the administration would eliminate the surplus in much less time.

Carl E. Walsh
Professor of Economics, UC Santa Cruz,
and Visiting Scholar, FRBSF

Opinions expressed in FRBSF Economic Letter do not necessarily reflect the views of the management of the Federal Reserve Bank of San Francisco or of the Board of Governors of the Federal Reserve System. This publication is edited by Anita Todd and Karen Barnes. Permission to reprint portions of articles or whole articles must be obtained in writing. Please send editorial comments and requests for reprint permission to