CRS Examines Swing in Budget From Surpluses to Deficits

Furthermore, some determinants of spending and revenues are not directly controlled by Congress. When economic conditions change, spending and revenues automatically change without any change in law. For example, when economic growth slows, the growth of taxable income slows, so that less revenue is collected at a given tax rate than previously. Likewise, if a slowdown in economic growth causes unemployment to rise, spending on unemployment insurance and other means-tested mandatory spending programs will also rise without any change in law. Economists refer to these changes as "automatic stabilizers" because they automatically cause the deficit to rise when the economy slows, thereby helping to offset the slowdown in growth because of the deficit's expansionary effects on aggregate spending.¹

For a more detailed answer to the question of why the budget moved from surplus to deficit, it is necessary to have some benchmark to which the actual deficit can be compared. One benchmark would be to compare the 2006 deficit to the 2001 surplus, but this approach would be fraught with several difficulties. For one thing, spending and revenues are expected to rise over time because of inflation, economic growth, and so on, so that $1 spent in 2006 is not comparable to $1 spent in 2001. For another, the economy in 2006 is not at the same level of production or in the same position in the business cycle as the economy in 2001, so the economy's effect on the budget is not the same either. Finally, the same law yields different levels of spending or revenue over time. Therefore, comparing spending or revenue levels from one year to the next could give the false impression that policy had changed when it had not.

This report uses a different benchmark: it compares the actual budget balance in the past six years to the January 2001 Congressional Budget Office's (CBO) baseline projection of the surplus for each of those years. CBO produces an updated 10-year budget baseline projection twice every year.² The baseline is a projection of the future path of government spending and revenues under current policy assuming no changes in the law. As can be seen in Table 1, CBO projected at that time that if policy had not changed, the surplus would have grown each year.³ Any year could have been chosen as the benchmark; this report uses their 2001 baseline as the benchmark because their 10-year surpluses peaked in this projection. In subsequent reports, the surplus projections would be continually adjusted downward, as CBO became progressively more pessimistic about the future path of revenues, until the Fall 2004 report. Since then, CBO has become more optimistic about future revenues, although revenues are still far lower today than they were projected to be in 2001. Thus, the results that follow are partly a function of the benchmark chosen.

As seen in Table 1, CBO projected a 2006 budget surplus of $505 billion under the policies in place in January 2001. The actual budget deficit in 2006 turned out to be $248 billion. In other words, the difference between CBO's 2001 projection and the actual outcome was $753 billion. The difference between projections and actual results peaked in 2004, when the deficit equaled $413 billion, compared with a surplus of $397 billion that had been projected in 2001, a shift of $810 billion. As these results suggest, budget projections -- even over relatively short periods of time -- are prone to large errors far beyond effects that can be explained by policy changes.

     Table 1. Differences Between 2001 Baseline Projections and

First, revenues have fallen and spending has risen due to legislative changes, because, as expected, policy has changed since CBO made its projection in 2001. Overall, legislative changes account for about three-fifths of the cumulative shift from surplus to deficit over the past six years. The legislative changes that increased the deficit the most were the tax cuts enacted between 2001 and 2004 and the increase in military spending in Iraq and Afghanistan.⁴ For the cumulative total increase in the deficit between 2001 and 2006, tax cuts accounted for about half of all legislative changes and military spending accounted for about one quarter, with the other quarter being caused by higher mandatory and non-military discretionary spending.⁵ Comparing legislative changes to the baseline surplus projection made in 2001 demonstrates that, even if there had been no economic downturn or any other projection error, legislative changes alone would still have caused a budget deficit in each year from 2003 to 2006. The comparison also demonstrates that the decline in the deficit since 2004 is unrelated to policy changes. Legislative causes of the deficit have grown each year since 2001; it is only because economic changes and technical changes have declined that the deficit has declined since 2004. In 2006, legislative changes were responsible for over four-fifths of the shift to deficit.

These legislative changes helped move federal revenues as a share of gross domestic product (GDP) from a 50-year high in 2000 to a 45-year low in 2004. The subsequent rise in revenues brought revenues slightly above the 50-year average share of GDP in 2006. Meanwhile, federal spending as a share of GDP rose from its lowest level in 35 years in 2000 to a level about equal to the 50-year average in 2004, then rising slightly in 2006.

Of course, an estimate of how much higher spending contributed to the shift to deficit depends on one's baseline definition of spending under current policy. Because discretionary spending is largely determined on an annual basis and not bound by previous year law, there is no obvious definition of current policy for discretionary spending. CBO's mandated definition of current policy is that discretionary spending grows at the same rate as inflation. Many analysts have criticized this definition as being too low because it would have consistently under-predicted spending historically. If a higher rate of spending growth (i.e., a rate equal to GDP growth) was assumed to represent current policy, then CBO's original estimates of future surpluses in 2001 would have been smaller, and higher spending would have subsequently accounted for a smaller proportion of the shift to deficit.

Second, actual economic conditions have differed from CBO's 2001 projection. As a result, revenues and spending are different than projected because of "automatic stabilizers," higher than expected inflation, and lower than expected interest rates. These are shown as economic changes in the table. The economic recession of 2001 was an important cause of the deficit in 2002 and 2003, but after 2003, economic conditions made barely any contribution to the shift to deficit. For the six years as a whole, economic changes accounted for about one-tenth of the shift to deficit. The effect of the economy on the budget balance actually turned out to be slightly better in 2005 and 2006 than was projected in 2001.

Third, after legislative changes and the estimated effect of the economy are accounted for, the remaining difference between the original projection and actual results is classified as technical changes. These changes occur because CBO's projections of revenues and mandatory spending are also based on other technical assumptions unrelated to the state of the economy, and actual results will differ from these assumptions as well. For example, mandatory spending levels will depend on program participation rates that will differ from projected participation rates. Similarly, tax revenues will depend not only on the growth in income, but how quickly tax liability rises as income rises and how income growth is distributed across taxpayers. Most technical changes tend to be on the revenue side rather than the spending side of the budget. Technical changes peaked as a cause of the shift to deficit at $259 billion in 2003, and have declined somewhat since then. For the last six years overall, technical changes accounted for about three-tenths of the shift from surplus to deficit.

Large technical changes point to the significant uncertainty behind budget projections, even over short periods of time. For example, holding policy constant, the 2001 surplus was $72 billion smaller ($37 billion due to economic changes, $35 billion due to technical changes) than CBO's projection made only nine months earlier.

Some technical changes are the result of legislative changes. The cost of the legislative changes listed in Table 1 are ex ante projections of their cost (as scored by CBO and the Joint Committee on Taxation) made at the time the policy was enacted. If policies turned out to be more (less) expensive than the official score, this would appear in the table as a negative (positive) technical change. There has been recent discussion that the 2006 deficit was smaller than anticipated because the tax cuts partly "paid for themselves" through higher economic growth, an effect that was not included in their original score. Table 1 suggests otherwise -- although the 2006 deficit was smaller than anticipated in projections made in 2004 and 2005, when CBO was extremely pessimistic about future revenues, the deficit was much larger than anticipated in projections made before 2004. The 2001 projection of the state of the economy in 2006 -- made before the tax cuts were enacted -- turned out to be extremely close to the actual state of the economy in 2006. In other words, the actual performance of the economy after the tax cuts was almost identical to how CBO expected the economy to perform had there been no tax cuts. For example, CBO projected in 2001 that economic growth (without tax cuts) would average 3.1% between 2003 and 2005 (on a calendar year basis); actual growth in those years equaled 3.2%. Furthermore, there were extremely large technical revisions in the last five years, mostly because tax revenue turned out to be lower than expected (after taking the tax cuts into account) each year from 2002 to 2006. If anything, this would suggest that tax cuts cost more in reality than the original score had anticipated, rather than less.⁶

² For the most recent CBO baseline, see Congressional Budget Office, The Budget and Economic Outlook: An Update, Aug. 2006.

³ See CRS Report RL31414, Baseline Budget Projections: A Discussion of Issues, by Marc Labonte. For an alternative evaluation, see Alan Auerbach et al., "New Estimates of the Budget Outlook," Brookings Institution, Issues in Economic Policy #3, Feb. 2006.

⁴ The cost of revenue changes has declined since 2004 because of the expiration of certain tax provisions, most notably accelerated depreciation for corporate investment.

⁵ Debt service is excluded from this calculation since the debt service shown in the table is the direct result of other legislative changes that increased government borrowing. (Debt service also increased because of economic and technical changes, but that is not broken out separately in the table.)

⁶ See also CRS Report RL32502, What Effects Have the Recent Tax Cuts Had on the Economy? by Marc Labonte; and CRS Report RL33672, Revenue Feedback from the 2001-2004 Tax Cuts, by Jane G. Gravelle.