Budget resolution targets, adopted by both Houses of Congress in most years, specify proposed levels of revenues and spending for the upcoming fiscal year. The targets in the 2004 concurrent budget resolution, adopted in April 2003, yielded a proposed budget deficit of $385 billion. However, the deficit for fiscal year 2004 was $412 billion--$27 billion more than the deficit that the budget resolution anticipated.

In 2004, revenues were $1,880 billion, only about $3 billion lower than expected for the year. Total outlays, at $2,292 billion, ended up being $24 billion higher than anticipated, primarily because of outlays from supplemental appropriations that were not contemplated in the budget resolution.
 

Elements of the Analysis

The budget resolution--which consists of targets for revenues, spending, the deficit or surplus, and debt held by the public--is a concurrent resolution adopted by both Houses of Congress that sets forth the Congressional budget plan over five or more fiscal years. The resolution does not itself become law; instead, it is implemented through subsequent legislation. That legislation includes appropriation laws that are intended to adhere to limits set for discretionary spending, as well as changes in the laws that affect revenues and spending. Those changes are sometimes in response to reconciliation instructions in the resolution, as was the case in 2004.

For this analysis, the differences between the levels specified in the budget resolution and the actual outcomes are allocated among three categories: policy, economic, and technical. Although those categories help explain the discrepancies, the divisions are inexact and necessarily somewhat arbitrary.

Differences attributed to policy derive from enacted legislation not anticipated in the resolution or enacted legislation that was estimated to cost a different amount than the resolution originally assumed. Differences attributed to policy may also reflect lawmakers' decisions not to enact legislation that the budget resolution assumed would pass. To identify such differences arising from legislation, the Congressional Budget Office (CBO) normally uses the cost estimates that it prepared at the time the legislation was enacted. (To the extent that the actual budgetary impact is different from what CBO estimated, that difference is characterized as a technical change.)

A key element in preparing the budget resolution is forecasting how the economy will perform in the upcoming fiscal year. Since 1992, the Congress has adopted the most recent economic assumptions published by CBO.⁽¹⁾ CBO's economic forecast for the budget resolution is usually made more than nine months before the fiscal year begins. Forecasting the economy is an uncertain endeavor, and, almost invariably, the economy's actual performance differs from the forecast. Nevertheless, every resolution is based on assumptions about numerous economic variables--mainly, gross domestic product (GDP), taxable income, unemployment, inflation, and interest rates. Those assumptions are used to estimate revenues, spending for benefit programs, and net interest. In CBO's analysis, differences that can be linked directly to the agency's economic forecast are labeled economic.

Technical differences between the budget resolution targets and actual outcomes are those variations that do not arise directly from legislative or economic sources as categorized. In the case of revenues, technical differences stem from a variety of factors, including changes in administrative tax rules, differences in the sources of taxable income that are not captured by the economic forecast, and changes in the amounts of income taxed at the various rates. In the case of many benefit programs, factors such as an unanticipated change in the number of beneficiaries, unforeseen utilization of health care services, changes in farm commodity prices, or new regulations can produce technical differences.
 

Comparing the Budget Resolution and Actual Outcomes for Fiscal Year 2004

The budget resolution for 2004 adopted the economic assumptions that CBO published in January 2003, which also underpinned CBO's March 2003 baseline prepared in conjunction with the agency's analysis of the President's 2004 Budget. Using those assumptions and incorporating planned policy changes, the resolution established the following targets for the year: total revenues of $1,883 billion, outlays of $2,268 billion, and a deficit of $385 billion (see Table C-1). Ultimately, revenues were lower by $3 billion and outlays were higher by $24 billion, resulting in a deficit that was $27 billion higher than the one anticipated in the resolution. Policy differences--primarily in the form of unanticipated discretionary outlays--raised the deficit by $44 billion relative to the target (see Table C-2). Conversely, a stronger-than-expected economy lowered the deficit by $27 billion compared to the target. Technical factors, mostly on the revenue side, accounted for the remainder of the difference (raising the deficit by $10 billion).

Table C-1.

Table C-2.

Differences Arising from Policy Changes

Of the many proposals incorporated in the budget resolution--some from the President's budget for 2004 and some originating in the Congress--a portion were eventually enacted (although sometimes in a different form than originally envisioned), and a portion were not. In addition, some legislation was enacted that was not envisioned in the resolution. In total, policy actions taken (or assumed but not taken) after the budget resolution targets were established increased the deficit by about $44 billion from the total assumed in the resolution. That net amount reflects $9 billion more in revenues and $53 billion more in outlays than the resolution assumed.

The resolution adopted most of the President's proposed tax cuts, including an economic growth package assumed to reduce revenues by $136 billion in 2004 and by $543 billion over the 2004-2013 period. When enacted, the Jobs and Growth Tax Relief Reconciliation Act of 2003, or JGTRRA, was estimated to lower 2004 revenues by roughly that amount. However, several other pieces of legislation expected to further reduce revenues were not enacted.

The resolution assumed that discretionary outlays in 2004 would total $861 billion--consistent with the level of budget authority in the President's request, adjusted for expected outlays from the April 2003 supplemental appropriations for operations in Iraq and Afghanistan. In fact, new supplementals drove discretionary budget authority $117 billion higher than anticipated in the resolution. Most of that amount stemmed from additional costs of the ongoing operations in Iraq and Afghanistan, which were funded in supplemental appropriation laws in November 2003 (Public Law 108-106) and August 2004 (P.L. 108-287). Supplemental spending thus accounted for much of the $47 billion overage in discretionary outlays attributable to legislation.

Differences arising from policy changes accounted for $8 billion of the mandatory outlays not anticipated in the resolution for 2004. Most important, mandatory spending was altered by legislation not contemplated in the budget resolution. The Unemployment Compensation Amendments of 2003 (P.L. 108-26), which further extended emergency unemployment benefits for recipients whose regular benefits would be exhausted before the end of December 2003, were enacted in May of 2003--at an estimated cost of $5 billion. In addition, JGTRRA included $10 billion in fiscal assistance to the states, with $5 billion of that amount for 2004 (and the first $5 billion in 2003).

The resolution's largest proposal for mandatory spending--albeit in years beyond 2004--was a prescription drug benefit for Medicare recipients. The budget resolution allowed for a program with costs totaling $400 billion over the decade, including $7 billion in 2004 to implement the proposal. Enacted in late 2003, the Medicare Prescription Drug, Improvement, and Modernization Act of 2003 in fact boosted outlays in 2004 by an estimated $4 billion, or $3 billion less than anticipated by the budget resolution.

Other policy proposals assumed in the budget resolution were not enacted. For example, a proposal for health insurance tax credits that would have increased outlays by $50 billion from 2004 through 2013 was incorporated into the resolution but did not pass. (The effect in 2004 would have been small.)

Differences Arising from Economic Factors

Overall, the economic assumptions underlying the 2004 budget resolution proved to be reasonably accurate. Small deviations from the forecast led to revenues that turned out to be $8 billion (0.4 percent) higher than presumed and outlays that were $19 billion (about 1 percent) lower.

The resolution assumed that real GDP would grow by 2.4 percent in 2003 and by 3.4 percent in 2004, but, in actuality, GDP growth for those years was 2.5 percent and 4.6 percent, respectively. The stronger-than-anticipated recovery raised the level of nominal GDP compared to that anticipated by the resolution. Corporate profits were higher and personal incomes were lower than anticipated.

Mandatory spending is also sensitive to changes in the economic forecast. Although such spending flows from the provisions of permanent laws, the spending for many mandatory programs is keyed to the economy. As a result, mandatory outlays for programs such as unemployment insurance and the refundable portion of the earned income tax credit decreased as the economy gained strength. Higher-than-expected inflation caused some offsetting increases in certain programs pegged to those indicators, but, overall, for economic reasons, mandatory outlays turned out to be $4 billion lower than the level assumed by the resolution.

Lower-than-anticipated interest rates drove projected outlays for net interest payments below the level assumed in the budget resolution. Most significantly, the resolution assumed that short-term (91-day Treasury) interest rates would average 2.9 percent in 2004; however, as a result of actions by the Federal Reserve, those rates averaged just 1.1 percent for the year. Consequently, outlays for net interest were $14 billion less than anticipated in the budget resolution.

Differences Arising from Technical Factors

Differences arising from technical factors--that is, differences between budget resolution targets and actual outcomes that cannot be traced to legislation or CBO's economic forecast--lowered revenues by $20 billion (about 1 percent) and outlays by $10 billion (0.4 percent) relative to the target levels. On balance, because of technical factors, the deficit was $10 billion higher than anticipated in the budget resolution.

Some of the decrease in anticipated revenues may have been related to economic factors (for example, decreased capital gains realizations) or may have resulted from economic factors that will be revealed in future revisions to economic variables; however, a full analysis of the 2004 results cannot be done now because information about sources of individual income typically lags behind the tax year by a couple of years.

The decrease in outlays attributable to technical differences resulted from slower-than-expected discretionary spending, as well as slightly slower spending in a host of mandatory programs. Debt-service costs were higher than expected, mostly because of the technical factors that reduced projected revenues.
 

Comparing Budget Resolutions and Actual Outcomes for Fiscal Years 1980 Through 2004

At the end of each fiscal year, actual revenues and outlays have always differed to varying degrees from budget resolution targets for that fiscal year. Over the 1980-1992 period, the deficit consistently exceeded the target in the resolution by amounts ranging from $4 billion in 1984 to $119 billion in 1990 (see Table C-3). That pattern changed in 1993, in part because spending for deposit insurance was substantially lower than expected. From 1994 through 2000, actual outcomes continued to be more favorable than the targets (with the exception of 1999, when there was no conference agreement on a budget resolution). However, in 2001, lower-than-expected revenues and higher-than-anticipated outlays combined to reduce the surplus to less than what was envisioned in the resolution. In 2002, those trends continued and caused very large differences from the resolution's envisioned surplus, resulting in a deficit of $158 billion that year. In 2003, there was no conference agreement for a budget resolution. In 2004, lower-than-expected revenues and higher-than-anticipated outlays caused the deficit to be larger than planned, but the difference was relatively small.

Table C-3.

Differences Arising from Policy Changes

From 1980 through 2004, policy action or inaction (for example, the failure to achieve savings called for in a budget resolution) decreased the surplus or increased the deficit by an average of $19 billion a year compared with the target. In only four of those years did policymakers trim the deficit more, or add to it less, than the resolution provided. The largest differences attributable to policy changes occurred in three consecutive years, decreasing the surplus by $61 billion in 2000, $95 billion in 2001, and $56 billion in 2002 in comparison with the targets. In 2004, as described, policy changes increased the deficit by $44 billion. (By contrast, from 1980 through 1998, the differences ascribed to policy changes averaged less than $10 billion a year.)

Most of the impact stemming from legislation over the period was on the outlay side of the budget. On average, policy decisions added about $17 billion a year to the spending totals. In fact, 1988 and 1991 were the only years in which legislative action held outlays below the budget resolution targets. The biggest difference due to policy changes was in 2000, when the effects of legislation increased outlays by about $65 billion, mostly from higher-than-expected discretionary appropriations and unanticipated assistance to farmers and agricultural producers. The difference in 2004 was second largest: a $53 billion increase in outlays, primarily resulting from the unanticipated discretionary spending discussed above. On the revenue side of the budget, the largest difference arising from policy changes occurred in 2001, when the Economic Growth and Tax Relief Reconciliation Act reduced taxes by $65 billion more than was anticipated by the resolution. By contrast, in 2002 and 2004 that difference was, respectively, a $9 billion reduction and a $9 billion increase.

Differences Arising from Economic Factors

Inaccuracies in the economic forecast over the 1980-2004 period had a small net effect on the cumulative variation between resolution targets and actual outcomes. However, large differences occurred in many years--deviations that were mostly negative before 1994 and positive more recently (except for 2002). Until 1993, budget resolutions tended to use short-term economic assumptions that proved optimistic. The largest overestimates of deficits in the 1980s and early 1990s, not surprisingly, were in years marked by recession or the early stages of recovery--namely, in 1982 and 1983, and over the 1990-1992 period. In 2002, the economic assumptions were again too optimistic, resulting in a $119 billion difference between the budget resolution target and actual outcome--contributing to that year's deficit, despite the fact that the resolution had envisioned a surplus. In contrast, the improving economy during this past year meant that the economic assumptions underlying the 2004 resolution were not optimistic enough: as a result, economic factors pulled the deficit $27 billion lower than what was assumed in the budget resolution.

In absolute terms (disregarding whether the errors were positive or negative), the typical difference in the surplus or deficit attributable to incorrect economic assumptions was about $33 billion a year over the 1980-2004 period. Regardless of the direction of the errors in the forecasts, differences between the resolutions' assumptions and what happened in the economy primarily affected revenues.

Differences Arising from Technical Factors

Technical factors accounted for differences between budget resolution targets and actual surpluses or deficits that averaged $5 billion a year over the past 25 years. In absolute terms, however, such differences caused the targets to be off by $41 billion, on average. Overall, those deviations were about equally represented on the revenue and outlay sides of the budget.

The magnitude and causes of the differences ascribed to technical factors have varied over the years. On the revenue side, technical misestimates were generally not very large through 1990, but the budget resolutions significantly overestimated revenues in 1991 and 1992, when tax collections were weaker than economic data suggested. From 1997 through 2001, revenues were much higher than the budget resolution targets, but in 2002, the resolution again overestimated tax collections by $183 billion. Technical factors lowered revenues in 2004 by $20 billion compared to the amount anticipated in the resolution.

Misestimates arising from technical factors have also shown up on the outlay side of the budget. Through the mid-1980s, discrepancies in estimating receipts from offshore oil leases and spending on farm price supports, defense, and entitlement programs were the dominant technical differences. In addition, in the early 1990s, during the savings and loan crisis, outlays for deposit insurance were a major source of discrepancies attributable to technical factors. In recent years, technical differences between the resolutions' estimates of outlays and actual outlays have been spread among a variety of programs. In 2004, the difference was a relatively small $10 billion.
 

Differences as a Percentage of Actual Revenues or Outlays

Because the federal budget has grown considerably since 1980, differences between the revenue and spending levels in the budget resolutions and actual outcomes over the 1980-2004 period may be best compared as a percentage of total revenues or outlays. The total difference for revenues for 2004, at 0.2 percent below the budget resolution target, was much smaller than the absolute average of 4.2 percent over the 25-year period. Outlays in 2004 were 1.0 percent above the budget resolution target--also lower than the 3.1 percent absolute average difference for the years 1980-2004.

The size of the total difference between actual deficits or surpluses and the deficits or surpluses anticipated in budget resolutions depends in large part on whether the differences in revenues and outlays offset each other. For years in which the discrepancies in revenues and outlays affected the surplus or deficit in opposite ways, the total difference dropped to as little as 0.5 percent of actual outlays. But in other years, the discrepancies for both revenues and outlays affected the surplus or deficit in the same way. From 1980 to 2004, the differences between estimates of revenues and outlays in the budget resolutions and the actual amounts went in the same direction relative to the deficit or surplus in 14 of the 25 years. Although the 2004 outcomes exhibit the same pattern, the magnitude of the differences is much smaller.