February 6, 2001

Projections of a Large Budget Surplus Are Surrounded
by a High Degree of Uncertainty
By Richard Kogan

View PDF version

If you cannot access the file through the link, right-click on the underlined text, click "Save Link As," download to your directory, and open the document in Adobe Acrobat Reader.

The Congressional Budget Office (CBO) recently projected sizable and growing surpluses over the next decade — surpluses totaling $2.7 trillion outside of the Social Security and Medicare Hospital Insurance trust funds. Some consequently believe there is room for a wide variety of costly initiatives, from expensive tax cuts to a Medicare prescription drug benefit to a new generation of military hardware.

Yet CBO itself, joined by the Office of Management and Budget (OMB) and Federal Reserve Chairman Alan Greenspan, has stressed that the surplus projections are highly uncertain. Even without tax cuts or any other changes in budget policy, there is a noticeable chance the government will run deficits outside of Social Security and Medicare, rather than surpluses. Based on its own track record, CBO concludes that "the estimated surpluses could be off in one direction or the other, on average, by about $52 billion in 2001, $120 billion in 2002, and $412 billion in 2006."(1)

History suggests it is far easier to enact additional tax cuts and budget increases in coming years if the projections hold up or improve, while it is far more difficult to enact tax increases or budget cuts in future years if the projections prove to have been too optimistic.

The possibility that the current projections may substantially overstate projected surpluses is especially troubling because CBO, OMB, GAO, and independent analysts agree that the long-term budget picture shows deficits returning during the course of the baby boomers' retirement and ultimately rising to unsustainable levels unless future generations raise taxes or cut federal programs. The need to prepare for the baby boomers' retirement suggests that policymakers should show caution in committing for other purposes surpluses that may not materialize. This memorandum briefly reviews the degree to which CBO is uncertain about its own projections.


CBO and Three Measures of Budget Uncertainty

In its recent baseline report, CBO devotes a chapter to examining the uncertainty surrounding its own projections. It measures uncertainty in three ways, but each method produces the same result — a reasonable possibility of budget deficits rather than surpluses, excluding Social Security and Medicare.

CBO's Track Record: The Average Error.

Figure 1All projections are wrong, but the question is by how much. CBO found that, on average, its own projections of the budget deficit or surplus for the coming fiscal year have been too high or too low by an average of 1.1 percent of the economy (Gross Domestic Product, or GDP).(2) Its projections five years out have been too high or low by an average of 3.1 percent of GDP. Therefore, if CBO's current projection is no better or worse than the its past projections, the unified budget surplus CBO projects for 2006 could be understated or overstated by $412 billion (3.1 percent of projected GDP in 2006).(3) The average range of uncertainty encompasses the possibility of noticeable deficits excluding Social Security and Medicare; the lower line on the graph shows such deficits occurring as soon as this coming year and becoming deeper as time passes. In fact, the graph shows that, based on the average size of CBO's past errors, the budget results in 2006 could range from a surplus of $573 billion to a deficit of $127 billion, excluding Social Security and Medicare.

CBO has not produced ten-year forecasts frequently enough to produce meaningful averages of its ten-year projection errors. However, CBO's projections are necessarily more uncertain as it looks farther into the future; if uncertainty continues to grow in years six through ten at the same rate it has proven to grow in years one through five, CBO's expected surplus in 2011, excluding Social Security and Medicare, would be expressed as "$524 billion, plus or minus $800 billion." Moreover, 71% of the projected surplus outside of Social Security and Medicare occurs in the second half of the ten-year projection, the period more subject to estimating error.

CBO's Track Record: The Probability of Deficits.

CBO has also calculated the odds that the surplus will fall above or below any particular point given its current projection and its historical track record. For instance, CBO thinks it equally likely that the surplus excluding Social Security and Medicare will be above its 2006 projection of $223 billion as below it. CBO's analysis suggests about a 20 percent chance that, under current law, the budget excluding Social Security and Medicare will be in deficit in each year from 2002 through 2006.

Can We Rely on the Track Record?

CBO fears that this January's projections might be even more subject to error than indicated by its historical track record. It notes:

[T]he short term outlook for the economy, and hence for the budget, may be particularly uncertain when the business cycle may be approaching a turning point. The longer term outlook is also unusually hard to discern at present. ... CBO's projections, like those of other forecasters, are based on very limited information about just a few years increased growth of productivity and strong investment in information technology. Projections of those recent changes as far as five or 10 years into the future are bound to be highly uncertain. ... not enough time has elapsed for analysts to be sure that those changes really represent a new trend in the economy rather than a temporary deviation.

Chairman Greenspan made the same point a few days earlier, when he said, "These impressive upward revisions [to projected long-term growth rates] are based on inferences drawn from economic relationships that are different from anything we have considered in recent decades. The resulting budget projections, therefore, are necessarily subject to a relatively wide range of error."(4)

CBO's Alternative Projections.

In addition to using its own track record to calculate its average absolute error and to estimate the probability the surplus would be larger or smaller than projected, CBO also produced two alternative projections based on more and less optimistic forecasts.

OMB Also Sees Great Uncertainty

On January 16, OMB wrote, "These projections are imprecise, and if experience is any guide, they could err by large margins. ... the more distant the projection, the greater the uncertainty. Over the history of five year budget projections, ... every administration has made substantial errors. ... Such enormous uncertainty about budgets just a few years in the future should influence policymakers’ decisions about expensive, long-term commitments on the basis of mere projections." OMB, FY 2002 Economic Outlook, Highlights from FY 1994 to FY 2001, FY 2002 Baseline Projections, January 16, 2001, p. 26

CBO's less optimistic projection assumes the economy reverts to the path it followed in the decades before 1996. Specifically, productivity growth is assumed to equal its historical average of 1.5 percent per year from 1973 through 1995. In addition, Medicare and Medicaid are assumed to grow 1 percentage point per year faster than currently projected. Under this scenario, instead of seeing a ten-year surplus of $2.7 trillion excluding Social Security and Medicare, one sees a ten-year deficit of $0.6 trillion. CBO comments that "[i]f CBO's track record is any guide, both the optimistic and the pessimistic scenarios lie well within the range of uncertainty in the budget projections."

End Notes:

1. In this paper, whenever CBO is cited, the source is Chapter 5 of The Budget and Economic Outlook: Fiscal Years 2002-2011, January 31, 2001, pp 93-103.

2. CBO's analysis of its historical projection errors deliberately exclude changes caused by legislation.

3. Other aspects of CBO's analysis suggest that the error in projecting the surplus excluding Social Security and Medicare might be about 85% as large as the error in projecting the unified budget surplus. Figures in this memorandum are calculated on that basis.

4. Testimony before the Committee on the Budget, U.S. Senate, January 25, 2001.