Reducing the Budget Deficit: Overview of Policy Issues

August 7, 2014 (R43680)





The federal budget deficit was the largest it has been since World War II as a percentage of GDP from 2009 to 2012, peaking at 10.1% of GDP. This occurred because spending reached its highest share of GDP since 1945 and revenues reached their lowest share of GDP since 1950. Since then, the deficit has declined to a projected 2.8% of GDP in 2014, which is still above the 1946 to 2008 average. Over the next 25 years, deficits are projected to become very large again under current law.

The recent decline in the deficit is partly due to improvements in the economy, the expiration of temporary measures taken in response to the recession, and spending cuts (mainly to discretionary spending). Spending was cut by the Budget Control Act of 2011 (BCA; P.L. 112-25) and the reduction in overseas contingency operations (OCO), primarily in Iraq and Afghanistan. Since September 2008, legislative changes to spending have added a cumulative $1.12 trillion to deficits and legislative changes to revenues, mainly the extension of expiring tax provisions, have added $1.75 trillion to deficits, excluding resulting interest costs.

Looking forward, several uncertainties are inherent in the baseline that may lead to different outcomes than projected. There have been large errors to budget projections historically, in part because economic forecasting is subject to large errors. The budget has also proven to be highly sensitive to recessions, and CBO does not project a recession in its 10-year projection. Budget projections also do not assume any significant changes in spending on future wars or disasters. The baseline projection follows current law, assuming that the "doc fix" and tax "extenders" will expire as scheduled. If Congress temporarily extends either, as it has done regularly in the past, the deficit will be larger than projected.

In the long run, legislative changes will be needed to reduce spending or increase taxes to keep the debt on a sustainable path. Postponing action requires larger changes to be made in the long run, and limits the ability to phase in changes gradually. Economists view the debt as currently unsustainable because it is projected to grow faster than gross domestic product (GDP) indefinitely under current policy, causing an ever growing share of national income to be devoted to servicing the debt. The main source of long-term fiscal unsustainability is the growth in elderly entitlement spending. In particular, spending on major health programs, such as Medicare and Medicaid, is assumed to continue to grow faster than GDP, as it has historically. Overall, mandatory spending has grown as a share of GDP, rising from 4.7% of GDP, when data were first compiled in 1962, to a projected 12.3% of GDP in 2014, and is projected to continue rising. By contrast, discretionary spending has fallen from 12.3% of GDP in 1962 to a projected 6.8% of GDP in 2014, and under the baseline it is projected to decline to its lowest share of GDP ever, primarily because of the BCA's statutory caps. Revenues are projected to stay near their historical average over the next 10 years.

Economic theory predicts that deficits have a stimulative effect on the economy during recessions, but harm economic growth by resulting in an increase in interest rates or the trade deficit during economic booms. Thus, the state of the economy is a consideration for the timing of deficit reduction.

Options for deficit reduction on the spending side are constrained by the fact that Social Security, Medicare, net interest, and defense discretionary spending make up almost two-thirds of total spending. On the revenue side, 80% of revenue is raised by income and payroll taxes.

Reducing the Budget Deficit: Overview of Policy Issues


The federal budget has been in deficit (i.e., spending has exceeded revenues) since 2002, and deficits were significantly larger from 2009 to 2012 than in any other year since World War II. As a result, the federal debt held by the public increased from 39% of gross domestic product (GDP) in 2008 to 72% of GDP in 2013, which was its highest share of GDP since after World War II.1

From 1946 to 2008, budget deficits averaged 1.7% of GDP and exceeded 5% of GDP only three times (equaling 7.2% in 1946, 6.0% in 1983, and 5.1% of GDP in 1985). The budget deficit was 10.1% of GDP in 2009, falling to 7% of GDP in 2012.2 The fiscal outlook has improved significantly since then, although the budget is projected to remain in deficit under current policy. In 2014, the baseline deficit is projected to be $492 billion or 2.8% of GDP. Not since the end of World War II has the deficit fallen so much, so quickly. Despite the improvement, the deficit still remains above its historical average.

In the past, persistent deficits have proven sustainable because periods of moderate growth in the debt relative to GDP have been followed by periods when debt fell relative to GDP (typically, because the debt grew more slowly than GDP). Current policy is unsustainable because projected deficits are large enough to cause the federal debt to continuously grow faster than national income. This increase is projected to be gradual—the debt does not increase relative to GDP until 2019, is five percentage points higher by 2024, and reaches 100% of GDP by 2036 or sooner. At some point, spending cuts or revenue increases will be needed to restore fiscal sustainability, although the recent decline in the deficit has provided policymakers more flexibility on the timing of these changes.

Congress has expressed interest in examining policy options to reduce the deficit. The Congressional Research Service (CRS) does not take a position on the best way to reduce the deficit. This report organizes and presents information to help policymakers frame the debate. This report first discusses the size of projected budget deficits, then discusses how much the deficit would need to be reduced to return to long-term sustainability, then analyzes alternative time frames for reducing the deficit, and finally discusses broad policy choices for reducing the deficit. This report assumes a familiarity with basic budget terms and concepts; for an overview, see CRS Report R43472, The Federal Budget: Overview and Issues for FY2015 and Beyond, by [author name scrubbed].

Why Did Deficits Rise and Then Decline?

What is the Baseline?

Absent policy changes, budget deficits would change over time because spending and revenue levels automatically change over time. CBO baseline budget projections are used to provide a benchmark against which budgetary decisions can be made and policy proposals can be compared over the budget window. A baseline is not meant to be the best guess of future outcomes or the most desirable outcome, but a reasonable starting point for comparing policy options. The projected magnitude and path of budget deficits depends greatly on what baseline assumptions are made. These assumptions are not always realistic, but they are set in statute, limiting CBO's flexibility. Discretionary spending is generally determined annually, so various assumptions could be made about what constitutes an extension of current policy. Where discretionary spending is not determined by statutory caps, statute requires the baseline to assume an extension of current levels plus an inflation adjustment in the baseline.

CBO estimates an official 10-year current law baseline, which assumes that certain provisions of law will expire as scheduled, even though many of these provisions have routinely been extended in the past.3 An example is the many tax provisions that regularly get extended in "tax extender" legislation.4 Another example is the Medicare cuts required in law by the sustainable growth physician payment formula; Congress has enacted legislative overrides to prevent those cuts (popularly known as the "doc fix") each year since 2003.5 CBO's current law baseline also assumes that the automatic spending cuts under the Budget Control Act will occur in future years as scheduled, although they have been reduced by subsequent legislation in the first three years they have been in place.6

The budget deficit increased by almost $1 trillion between 2008 and 2009. As shown in Table 1, this increase was not anticipated—two months before FY2009 began, CBO projected that the 2009 baseline deficit would be slightly lower than the 2008 deficit (for more information on the baseline, see the text box below). From 2012 to 2014, deficits declined by a similar amount—the 2014 deficit is currently projected to be about $1 trillion lower than the 2009 deficit. Table 1 explains what factors caused this sharp increase and subsequent decrease in the deficit. Because the same policy generates different levels of spending and revenue in different fiscal years, it is not possible to say what "caused" the deficit to increase or decrease from one year to the next. However, it can be determined with precision what caused the actual deficit in any given fiscal year to be smaller or larger than what CBO had previously projected. The exact shares assigned to each category will be sensitive to the chosen starting point, however. Table 1 uses the September 2008 projection because of the dramatic increase in the deficit relative to the projection that occurred after that date.

Table 1. Contributors to the Difference Between Actual Deficits and Deficits Projected Before the Financial Crisis

(billions of dollars)







2014 (projected)

Total 2009-2014

9/08 Baseline Deficit








Legislative Changes
























—Net Interest
















Economic Changes








Technical Changes








Total Changes








Actual Deficit








Source: CRS calculations based on CBO data.

Notes: Negative sign indicates increase in the budget deficit, positive sign increases decrease in the budget deficit. CBO does not categorize the causes of discrepancies between the actual deficit and its August projection for the current FY; these discrepancies are categorized here as technical changes. CBO classifies the government's decision to transfer funds to the GSEs as a technical change. CRS has reclassified the BCA's automatic spending cuts, which came into effect when the Joint Select Committee on Deficit Reduction failed to propose legislation, from the technical changes category to the legislative changes category.

Actual budgetary outcomes can turn out differently than previously projected for three reasons—the subsequent enactment of legislation ("legislative changes" in Table 1), differences in CBO's projections attributable to economic forecast errors ("economic changes" in Table 1), and all other differences ("technical changes" in Table 1). For example, if more people received a tax credit or entitlement program than predicted for a given level of GDP, absent any legislative change, that would be classified as a technical change that increased the deficit. Because this table is based on the CBO baseline, policies that maintain current policy but change current law, such as extending expiring tax provisions, are classified as legislative changes.

Legislative changes since 2009 can be broken down into three broad categories, with the major changes7 provided for each:

Most financial stabilization outlays in the federal budget occurred within the TARP13 or on transfers to two GSEs, Fannie Mae and Freddie Mac. These programs had a mostly one-time effect on the deficit due to improvements in financial markets and the expiration in 2010 of Treasury authority to enter into new contracts with regard to TARP and the GSEs. In 2009, outlays of $243 billion were recorded for TARP and the GSEs. In 2010 and 2011, transfers to the GSEs declined and, due to budgetary conventions, TARP recorded negative outlays that reduced the deficit.14 CBO included transfers to the GSEs in technical, not legislative changes, although their entry into receivership was a policy decision.

The temporary spike in spending caused by ARRA peaked in 2010 had mostly dissipated by 2012. ARRA was intended to be only a temporary boost to spending to stimulate the economy, and Congress allowed budget authority to return to near pre-ARRA levels in nominal terms.15 (ARRA also temporarily reduced taxes, but those provisions were subsequently extended.)

Extended unemployment insurance was in place from July 2008 through 2013.16 Payroll taxes were temporarily reduced by two percentage points from 2011 to 2012.17 This was the only major legislative change affecting revenues that was allowed to expire.

Budget authority for OCO has fallen from a peak of $187 billion in 2008 to $93 billion in 2013 as overseas operations, primarily in Iraq and Afghanistan, have been scaled back. This has reduced discretionary defense spending relative to the September 2008 baseline, which assumed that such spending would increase at the rate of inflation each year.

For 2012 to 2021, the Budget Control Act (BCA; P.L. 112-25) reduced discretionary spending and mandatory spending relative to the baseline.18 The BCA implemented spending cuts through statutory caps for defense and non-defense discretionary spending for 2012 to 202119 and sequestration for mandatory spending for 2013 to 2021. The BCA exempted most mandatory spending programs from sequestration and capped Medicare cuts at 2%. On the discretionary side, the BCA exempted military OCO, disaster spending, and emergency spending from the spending caps. The BCA did not provide any details as to what discretionary programs would absorb the cuts needed to adhere to the defense and non-defense caps, however, and Congress has struggled to reach a consensus on that issue each year since the BCA was enacted. Through 2014, total discretionary spending has ended up higher than envisioned under the BCA because of subsequent legislation reversing some of the BCA spending cuts and because spending in categories exempt from the caps was higher than envisioned.20

Overall, discretionary spending was $75 billion lower in 2013 and is projected to be $120 billion lower in 2014 than the September 2008 baseline level. The BCA cuts to mandatory spending were not large enough to offset other legislative changes that increased mandatory spending in 2013 or 2014.

Congress temporarily extended expiring tax provisions, including major tax cuts enacted in 2001 and 2003, numerous times. Most provisions from the 2001 and 2003 tax cuts were permanently extended in the American Taxpayer Relief Act (ATRA; P.L. 112-240) in January 2013.22 ATRA also temporarily extended ARRA's tax cuts until 2017 and reduced revenues overall by $3.9 trillion over 10 years. On net, all legislative changes to revenues made since September 2008 have increased the budget deficit by $1,750 billion from 2009 to 2014 compared with the current law baseline. To the extent that these tax provisions involved refundable tax provisions, which are classified as mandatory spending, extending expiring provisions also increased mandatory spending. Compared with a current policy baseline, ATRA reduced the deficit because it allowed the payroll tax cut and certain provisions affecting high-income taxpayers to expire at the end of 2012. OMB estimated that the expiration of provisions affecting high-income taxpayers increased revenues by $618 billion over 10 years compared with a current policy baseline.23

Medicare's sustainable growth rate (SGR) system is intended to constrain Medicare payments to physicians. Congress has temporarily prevented the SGR from reducing physician payments each year since 2003 (called the "doc fix"). If Congress continues the policy that has been in place since 2003, it will permanently increase the deficit, even if there has technically been no permanent policy change so far.

All legislative changes that increase (decrease) the deficit also increase (decrease) net interest payments on the debt in future years. Combined, legislative changes to revenue, mandatory spending, and discretionary spending have increased net interest by $277 billion from 2009 to 2014.

In sum, several major legislative changes were enacted that increased the deficit between 2009 and 2012 by increasing discretionary or mandatory spending. These changes were temporary and had little effect on the deficit after 2012, except for the fact that they increased interest payments on the debt. By contrast, the (mostly permanent) extension of expiring tax cuts increased the deficit relative to current law, by $1,750 billion over six years. Policies enacted during this time that permanently reduced the deficit, notably the tapering of OCO spending and the BCA as modified, did so primarily by reducing discretionary spending. In 2013 and 2014, legislative changes reduced total discretionary spending relative to CBO's September 2008 projection, but the resulting deficit reduction was more than offset by legislative changes to mandatory spending and revenues that increased the deficit.

Economic changes, related to the recession and the sluggish recovery, added around $100 billion each year to the deficit in 2009, 2010, 2011, and 2013, and added smaller amounts in 2012 and 2014.24 Revenues automatically rise and certain mandatory spending automatically fall as the economy moves from recession to full employment. Economists refer to these changes as "automatic stabilizers." Revenues automatically fall in a recession because taxable income falls. Recessions cause spending on certain means-tested programs, such as unemployment insurance, to automatically increase due to increased eligibility. Overall, CBO estimates that automatic stabilizers at their peak added 2.7 percentage points of GDP ($409 billion) to the budget deficit in 2010.25 The role of automatic stabilizers has since diminished, but is still exerting some upward pressure on the deficit.

Technical changes were a major contributor to the actual deficit from 2009 to 2012, but had little effect in 2013 or 2014. These technical changes reflect the uncertainty inherent in the multitude of assumptions underlying budget projections, putting aside economic uncertainty. The joint effects of economic and technical changes illustrates that even if no legislation had been enacted since September 2008, the actual deficit would have turned out much larger than projected each year and would have increased sharply from 2008 to 2009 and remained large in 2010. There was no year between 2009 and 2014 where economic and technical changes combined exceeded legislative changes, however. Overall, the effect of legislative changes on the deficit was more than twice as large as economic and technical changes combined from 2009 to 2014, and would have been larger if assistance to the GSEs was classified as a legislative change.

How Large Are Projected Deficits Under Current Policy? What Might Change These Projections?

In the CBO baseline projection, budget deficits fall from 4.1% in 2013 to 2.8% of GDP in 2014, reaching a low of 2.6% of GDP in 2015. (The decline in 2015 is attributable to baseline assumptions and would not occur under different assumptions.) They then rise each year until reaching 4.0% of GDP in 2022. The debt held by the public declines each year until it reaches 72.4% of GDP in 2017, and then increases each year, reaching 78.1% of GDP in 2024. Long-term projections indicate that budget deficits would eventually become very large relative to GDP outside the projection window under current policy. Under a long-term projection of current policy, CBO projects deficits would exceed 10% of GDP by 2073.26

These projections are highly uncertain, both because of forecast errors and the underlying assumptions made about future spending and revenues. Budget projections are subject to a high degree of uncertainty—based on history, actual outcomes are likely to be much better or worse than projections, even after accounting for policy changes. Projections are often inaccurate even over short-time frames, but small changes in assumptions compound to large changes in budget projections over the long run. For that reason, forecasting errors become larger further into the future. For example, OMB estimates that the absolute average errors for its budget deficit projections are 1.4% of GDP for the next budget year and 3.6% of GDP for five years in the future.27

Were GDP growth lower or interest rates higher than projected, the primary deficit (i.e., deficit before interest payments) would have to be smaller in dollar terms to be sustainable, and vice versa. By historical standards, CBO's projections of economic growth over the next 10 years are relatively modest, but this is mostly because CBO projects that the labor supply will grow much more slowly than it has historically due to the aging of the population. More problematic is the role of the business cycle on budget projections. CBO's economic forecast only accounts for the short-term effects of the business cycle on the budget in the first few years of the budget window. After that, it assumes that the economy will grow at a steady rate. As a result, CBO's current projection assumes that there is no recession at any point in the next 10 years. Given that the economy has been in expansion since 2009, this projection in effect assumes that the economy will experience the longest uninterrupted economic expansion in the history of the United States. The projection would seem to account for the business cycle by assuming a growth rate that averages out the potential ebbs and flows in the business cycle. The budget has proven to be disproportionately affected by past recessions, however. In other words, were a recession to occur in the next 10 years, it would likely increase cumulative budget deficits over the course of the budget window by more than the baseline projection of steady growth, even after accounting for higher than projected growth in expansion years.

CBO's and OMB's projections assume that interest rates will remain at relatively low levels by historical standards over the next 10 years. CBO estimates that if interest rates rose to their average level from 1991 to 2000, the budget deficit would be an average of $144 billion higher per year over the next 10 years. If interest rates rose to their average level from 1981 to 1990, the budget deficit would be an average of $627 billion higher per year over the next 10 years.28

Besides forecast errors, the baseline projection is based on a number of policy assumptions when alternative assumptions could be employed that would be equally valid. Changing these assumptions would increase or decrease deficit projections. The following examples involve ambiguity about how to define current policy (in contrast to projections about future legislative changes, not meant to be captured in the baseline).

Unforeseen circumstances and inaccurate assumptions played a major role in the movement from projected baseline surpluses to actual deficits in the last decade.30 Historically, emergencies and other unforeseen events have typically been deficit financed. The baseline could not anticipate events, such as war in Iraq and Afghanistan, Hurricane Katrina, and two recessions, that were major contributors to last decade's budget deficits. Furthermore, baseline assumptions (some of which are set in statute) that were unrepresentative of past events—such as assumptions that discretionary spending would grow only at the rate of inflation (in years without statutory caps) and tax provisions would expire as scheduled—proved unrepresentative of future events, and this caused baseline deficit projections to repeatedly undershoot. This experience raises the question of the utility of relying on projections, particularly over longer budget windows. Another conclusion is that if budget plans do not use realistic assumptions or build in assumptions about unforeseen contingencies, they are unlikely to prove accurate.

In addition to ambiguity about how to best capture current policy in a baseline, it is useful to remember that, on a conceptual level, the baseline does not attempt to account for policy changes even if they are likely or predictable. For example, Congress has intervened to prevent the full effects of the Budget Control Act's automatic spending cuts from occurring each year (although it has not eliminated them entirely). One might conclude it is not only possible, but probable, that future Congresses may similarly intervene to reduce these spending cuts. Because the baseline measures current law, it assumes that those cuts will be implemented in their entirety.

To account for some of these issues caused by baseline assumptions, CBO also presents an Alternative Fiscal Scenario that assumes that the doc fix and expiring tax provisions do not expire and that the BCA's automatic spending reductions do not come into effect. Under the Alternative Fiscal Scenario, the budget deficit continues to rise and peaks at 5.1% of GDP in 2022 and the debt held by the public rises as a share of GDP each year, reaching 86.7% of GDP by 2024. Deficits reach 10% of GDP in 2032 and the debt held by the public reaches 100% of GDP by 2029.

How Much Deficit Reduction Is Necessary?

There is no clear answer to the question of how much the deficit should be reduced because the targeted amount of deficit reduction depends on the policy goal.

Because deficits are projected to continue to grow in the long run, growing reductions in spending or increases in taxes would be required over time to maintain sustainability. Cutting spending or raising taxes sooner rather than later would reduce the need for future changes. CBO projects that spending would need to be cut or revenues increased by 1.8% of GDP immediately and permanently to stabilize the debt-to-GDP ratio over the next 75 years, or 7.4% of GDP under the alternative fiscal scenario.34 Today's debt levels, high by historical standards, leave policymakers less room to maneuver to cope with future challenges—both anticipated, such as the retirement of the baby boomers, and unanticipated.

How Quickly Should the Deficit Be Reduced?

The state of the economy is an important factor to consider in determining the desired timing of deficit reduction. All else equal, mainstream economic theory predicts that reducing the deficit would have a contractionary effect on GDP in the short run, whether through tax increases or spending reductions (which is a component of GDP). During a period of robust economic growth, that contractionary effect would be more easily offset by other sectors of the economy, and the expansion would likely be sustained. During a period of high unemployment, reducing the budget deficit would be expected to make unemployment higher (or fall more slowly) than would otherwise be the case, all else equal. The economy is now closer to full employment than it has been since the recession, but there is considerable disagreement among economists as to how close and growth has not yet been robust.35

On the other hand, deficit reduction will eventually be needed to achieve sustainability, and given that the effect on the economy is proportional to the size of the deficit reduction, it could be argued that avoiding unwanted contractionary effects suggests a gradual approach to deficit reduction. Furthermore, policy changes can be phased in such a way that they have little contractionary effect in the short run. Waiting until deficits have become larger leaves less latitude for phasing in changes.

The U.S. fiscal outlook is not a purely long-term issue, however—deficits are already above average today, and while projected deficits outside the 10-year budget window are larger than today's deficits, they are also more uncertain. Deficits are also a long-term issue in the sense that most observers believe fundamental reforms to outlays and revenues would be necessary to put the budget on a sustainable path; however, any delay to implementing those changes increases the eventual budgetary cost of returning to a sustainable fiscal path, all else equal.

The deficit is a long-term issue in that any negative economic consequences from running large deficits have been minor to date, but there is the risk that the deficit's effect on the economy could become negative at any time. Although economic theory suggests that larger deficits provided a stimulative boost to the economy during the recent recession by partially offsetting the contraction of private spending, continued deficits would be expected to eventually have a negative effect on the economy. At some point, the economy is expected to return to full employment (i.e., practically all existing labor and capital resources are in use). When it does, government budget deficits are expected to "crowd out," or compete with, private-investment spending in the standard macroeconomic model. Setting aside foreign capital flows for the moment, borrowing can only be financed through saving, and government borrowing competes with business borrowing for the same pool of national saving. By increasing the demands on that pool of national saving, government borrowing pushes up the cost of all borrowing through higher interest rates, causing businesses to finance less capital spending than they otherwise would. Business borrowing finances capital spending on plant and equipment, and lower capital spending results in lower potential gross domestic product, and hence lower future national income, than would otherwise occur.

The financial crisis and Great Recession led to a marked decline in private-investment spending and increase in private saving. Both of these factors reduced the potential for large government deficits to crowd out private-investment spending. Low interest rates since the recession began support the view that the deficit caused little crowding out to occur.

With international capital mobility, borrowing can also be financed by foreign saving. In the standard macroeconomic model with perfect capital mobility, the boost in aggregate spending from the stimulus would cause the trade deficit to rise as foreign capital is attracted to higher domestic interest rates. Net foreign borrowing is equivalent to the trade deficit because one country can borrow from the rest of the world only if it imports more than it exports. The availability of foreign credit would avoid the crowding out of domestic capital investment. But the boost to aggregate spending from the budget deficit would be negated (or "crowded out") by the higher trade deficit (in the form of lower exports, higher imports, or both). The United States relies heavily on foreign borrowing, and this is another reason that large budget deficits could be less effective at stimulating the economy. Since the recession began, the trade deficit has fallen substantially, so a problem of crowding out from the trade deficit is not apparent at this time.

Despite the significant decline in the budget deficit since 2012, crowding out and higher interest rates will likely become a more pressing issue as the economy returns to full employment because of the anticipated increase in business investment back to historically normal levels, particularly if the increase in private saving following the recession proves not to be permanent. If that is the case, the decline in the trade deficit may reverse. Interest rates have already increased modestly as the economy has strengthened, although they remain low compared with those prevailing in recent decades.

Policy Options for Deficit Reduction

Budget deficits can be reduced through cuts in spending, higher taxes, or a combination of both.36 Plans to reduce the deficit that are narrowly targeted require commensurately larger spending cuts or tax increases to targeted programs or provisions. To that end, deficit reduction proposals can start with the observation that Social Security, Medicare, net interest (which cannot be changed), and defense discretionary make up almost two-thirds of total spending. Figure 1 compares the projected deficit to overall spending and revenues in 2014. It shows that the deficit is about two-fifths the size of all non-interest spending outside of Social Security, Medicare, and defense discretionary. It is about one-sixth the size of total revenues.

Figure 1. Selected Data from 2014 Federal Budget

(billions of dollars)

Source: CRS calculations based on CBO data.

Note: Budget deficit, spending, and revenues for 2014 are projected based on the CBO baseline.

Budget deficits are the result of the shortfall between spending and revenue. In 2009, spending reached its highest share of GDP since 1945 and revenues reached their lowest share of GDP since 1950. As seen in Figure 2, from 1946 to 2008, outlays averaged 19.6% of GDP and were generally below 20% of GDP until 1975, above 20% of GDP from 1975 to 1996, and below 20% of GDP from 1997 to 2005. From 2009 to 2011, outlays averaged 24.1% of GDP. From 1946 to 2008, revenues averaged 17.8% of GDP, showing no long-term upward or downward trend from 1952 to 2007. Revenues were at least 17% of GDP in each year during that period except for 1955, 1959, 2003, and 2004, when they were between 16% and 17%. From 2009 to 2012, revenues were below 16% of GDP.

Over the next 10 years, revenues are projected to be near their historical average at 18.1% of GDP. Outlays are projected to be 21.5% of GDP, above their historical average of 19.6% of GDP and comparable to their level from 1975 to 1996. However, the composition of outlays is projected to be significantly different, as discussed below.

Figure 2. Federal Revenues and Outlays, Historical and Baseline Projection


Source: OMB and CBO.

Just as recent deficits are the combination of all past outlay and revenue decisions, returning the budget to balance would be difficult without a combination of outlay and revenue changes. For instance, to return the budget to balance while maintaining baseline revenue levels would require that outlays decline to a share of GDP last seen in the early 1970s. Likewise, to balance the budget while maintaining baseline outlay levels would require revenues to rise to their highest share of GDP ever.

Federal Spending

Total spending has fallen from an average of 24% of GDP from 2009 to 2011 to 21% of GDP in 2013. It is projected to rise to 22% of GDP from 2020 to 2024 under current policy. From 1946 to 2008, there were only three years when outlays were above 23% of GDP. Spending rises under the baseline projection despite discretionary spending falling to its lowest share of GDP since data were first collected because mandatory spending and net interest on the federal debt are projected to grow relative to GDP. The projected increase in net interest is due to the growth in the federal debt and the return to more normal interest rates from the current below-average rates.

Discretionary Spending

Discretionary spending has fallen from 12.3% of GDP in 1962 to a projected 6.8% of GDP in 2014. Discretionary spending was at high levels relative to GDP from 2009 to 2012, but not historically high levels since data were first available in 1962, as seen in Figure 3. Over time, the two components of discretionary spending, defense and non-defense, have followed different paths. Defense discretionary spending was lower from 2009 to 2012 than it was in all but two years from 1962 to 1992 as a share of GDP. From 1963 to 2001, defense discretionary spending generally fell relative to GDP, but rose in nominal dollars. It then began to grow, when overseas military operations expanded. In 2014, defense discretionary spending is projected to fall to its lowest share of GDP since 2002.

Non-defense discretionary spending has shown no long-term upward or downward trend relative to GDP—except for an elevated period from 1975 to 1983, it has always stayed within 3% to 4% of GDP. Over the late 1990s, it fell to its lowest level of GDP since data have been collected, and then rose from that low base in the 2000s. It was above its long-term average from 2009 to 2012, but still below the levels prevalent from 1975 to 1981. Since 2009, much of the growth in non-defense discretionary spending was a result of the 2009 stimulus (ARRA). Most discretionary spending provided under this act was completed by 2011. In 2014, non-defense discretionary spending is projected to fall back to the historically low levels of the 1990s, which is well below its share of GDP before the 1990s.

Figure 3. Discretionary Spending, Historical and Projected in the Baseline


Source: OMB and CBO.

Note: Figure assumes BCA's automatic spending cuts are implemented and current levels of OCO and disaster spending are maintained, adjusted for inflation.

In the CBO baseline, discretionary spending declines significantly relative to GDP over the next 10 years to its lowest share of GDP ever since data were first collected in 1962, as shown in Figure 3. The baseline, based on current law, assumes that discretionary spending will adhere to the levels set in the Budget Control Act (BCA) through 2021 and discretionary spending not subject to the caps (notably, OCO) will rise at the rate of inflation. In long-term projections, CBO assumes discretionary spending will remain at its 2024 share of GDP. Because discretionary spending is held constant, it becomes a smaller share of spending over time and therefore does not contribute to the growth of the long-term budget deficit. If assumed instead that discretionary spending remained at its current share of GDP, non-interest spending would be 1.5 percentage points higher annually and long-term deficits would be even larger.

Discretionary spending fell in nominal terms (i.e., not adjusted for inflation) each year from 2011 to 2014. From 1963 to 2010, discretionary spending fell in nominal terms in only four years, most recently in 1996. Adjusted for inflation, both defense and non-defense spending fell in real terms each year from 2011 to 2013, and they are projected to continue falling each year through 2018. The decline is caused by the decline in the BCA caps in real terms and, through 2014, in OCO spending. This contrasts to the significant growth in discretionary spending that occurred from 2000 to 2010, as shown in Table 2.

Table 2. Average Annual Real Growth Rates of Discretionary Spending Outlays

(actual and CBO baseline, adjusted for inflation)
















Source: CRS calculations based on OMB and CBO data.

Notes: Data for 2000 to 2013 are actual; data for 2014 to 2021 assume that spending cuts under the BCA will occur as scheduled, with OCO spending maintained in real terms at 2014 levels.

One macroeconomic implication of reducing discretionary spending is that most federal spending on physical capital (e.g., infrastructure) and human capital (e.g., education) is located in the discretionary portion of the budget. Economic theory predicts that a lower future capital stock would result in a smaller economy from what it otherwise would be in the long term, all else equal.37

To achieve further deficit reduction through discretionary spending would involve reducing discretionary spending even lower as a share of GDP than the all-time low already achieved under the BCA. By 2030, if discretionary spending were reduced to zero, there would still be a deficit in the extended baseline scenario. For defense, the timing and magnitude of any potential drawdown in overseas military operations could cause military spending to decline relative to the baseline, but traditionally such changes have not been motivated by the desire for deficit reduction. Even if current operations are reduced, future geopolitical events could require military personnel to be deployed elsewhere in the next 10 years, so the baseline does not necessarily overestimate future defense spending.

Mandatory Spending

Unlike discretionary spending, mandatory spending has grown as a share of GDP since 1962. Net of offsetting receipts, it doubled relative to GDP from 1962 to 1975, increasing from 4.7% to 9.4% of GDP. It then fluctuated in a range of 8.7% to 10.4% of GDP from 1975 to 2007, peaking in recession years in part because of automatic stabilizers. It peaked at 14.5% of GDP in 2009, marking its highest share of GDP since data were first compiled in 1962. Mandatory spending has fallen as a share of GDP since, but has remained above its pre-crisis share of GDP.

In contrast to discretionary spending, mandatory spending is projected to continue to grow faster than inflation and exceed 13% of GDP over the next 10 years under current policy. Since 1962, mandatory spending fell in nominal terms in only one year, 2010; outside of TARP and deposit insurance, which recorded negative outlays that year, all other mandatory spending grew in 2010.

Outside of health and retirement programs, outlays on other mandatory programs are projected to decline relative to GDP over the next 10 years. As the economy improves, "automatic stabilizer" spending has declined significantly: spending on income security programs more than doubled in nominal terms between 2007 and 2010, but is projected to decline from $438 billion in 2010 to $219 billion in 2014. Over its 75-year projection, CBO assumes that other mandatory spending will fall from 2.7% of GDP in 2013 to 0.8% of GDP by 2088. Since data were collected in 1974, it has never been lower than 1.9% of GDP.

Over the long term, however, the upward trend in mandatory spending in the past and future is dominated by entitlement spending on the elderly (Social Security, Medicare, and Medicaid), which accounted for about three-quarters of total mandatory spending in 2010. Social Security has risen from 3.7% of GDP in 1974 to 4.9% of GDP in 2013, its highest share of GDP ever. Major mandatory health programs (Medicare, Medicaid, the Children's Health Insurance Program, and health insurance exchange subsidies, net of offsetting receipts) rose from 1% of GDP in 1974 to 5% of GDP in 2011, its highest share of GDP ever, before falling to 4.6% of GDP in 2013.

Figure 4. Mandatory Spending, Historic and Projected


Source: CBO, Long-Term Budget Outlook, Supplementary Data, July 2014.

Notes: Projection using CBO's Extended Baseline.

Within the 10-year baseline, Social Security and major health spending each exceed total discretionary spending for the first time. In long-term budget projections, rising budget deficits are driven primarily by the growth in entitlement programs for the elderly, particularly health spending. Social Security outlays are projected to rise from 4.9% of GDP today to 6.3% of GDP in 2039, and federal health outlays (mainly on Medicare and Medicaid) are projected to rise from 4.9% today to 8.0% of GDP in 2039. By 2066, outlays on Social Security and health programs would exceed projected total revenues in 2014 as a share of GDP.

In long-term projections, health spending per capita is projected to continue to grow faster than GDP per capita because it has historically grown much more quickly. If health spending per capita grew at the same rate as GDP per capita (technically, this is referred to as an excess cost growth rate of zero), one-third of the increase in federal health spending would be avoided through 2039, although spending would still grow somewhat because of demographic changes—namely, the retirement of the baby boomers—that increased the number of recipients. Figure 5 illustrates that if there was zero excess cost growth over the next 75 years, spending on major health programs would decline from 14% of GDP to 7.7% of GDP in 2089. The difference between the two is significantly larger than the baseline primary deficit (2.6% of GDP) in 2089. Therefore, if health spending were to grow at the rate of GDP instead, the budget would be on a sustainable path.

Figure 5. Projected Spending on Major Health Programs:
CBO Baseline vs Zero Excess Cost Growth

(2014 to 2089)

Source: Congressional Budget Office, Long-Term Budget Outlook, Supplementary Data, July 2014.

The growth in elderly entitlement spending makes deficits unsustainably large in the long run. In part, the link between entitlement spending and deficits is driven by the assumptions that go into these projections—other spending is assumed to gradually fall relative to GDP in the long run, revenues are projected to gradually rise relative to GDP because of bracket creep, while health spending per capita is projected to continue to grow faster than GDP per capita. Restraining the future growth rate of elderly entitlement spending would not result in significant deficit reduction in the short run, however. Most proposals to reform elderly entitlement programs would generate significant budgetary savings in the long run, but little budgetary savings in the short run, partly because most proposals exempt current retirees from reform and partly because the savings from these changes would compound over time. For example, immediately reducing excess cost growth for federal health spending to zero would reduce federal spending by 2% of GDP after 20 years, but by 0.3% of GDP after 5 years.

Mandatory spending could be reduced by restricting benefits or eligibility. For retirement programs, it is straightforward to do so through a change in the benefit formula. For health programs, it is less straightforward, and there is a lack of consensus on how excess cost growth, which has persisted for decades, can be effectively minimized. On a positive note, excess cost growth has shown a downward trend over time. However, CBO's projections already assume the decline will continue, and it still results in an unsustainably large budget deficit.


In 2009 and 2010, revenues were at historically low shares of GDP across all major categories—individual income taxes were at their lowest share of GDP since 1950, corporate income taxes were at their lowest share of GDP since the 1930s, social insurance receipts were at their lowest share of GDP since the 1970s, and excise taxes were at their lowest share of GDP since 1934, the first year for which data are available. Total revenues remained at historically low shares of GDP in 2011 and 2012, as shown in Figure 2. In recent years, most revenues have come from the individual income tax and social insurance categories. In 2013, revenue as a share of GDP returned closer to historically average levels because of the economic recovery and because of the expiration of the payroll tax cut and certain provisions from the 2001 and 2003 tax cuts.38

Revenues in the CBO baseline are projected to equal about 18% of GDP from 2015 on, around their historical average. Total revenues would gradually increase relative to GDP over the next 10 years due to the improvement in the economy (initially) and "real bracket creep," returning to their historical average around 2014. "Real bracket creep" refers to the fact that the same tax structure generates more revenue relative to GDP when incomes rise. Bracket creep does not lead to a significant increase in revenues relative to GDP over a 10-year projection, but it does over a 75-year projection. In CBO's Long-Term Budget Outlook, projected revenues increase by 1.8 percentage points to 19.4% of GDP as a result of bracket creep in 2039. By 2089, revenues would reach 23.9% of GDP, which would represent their highest share of GDP ever. Revenues also increase by 0.3 percentage points by 2039 because it is assumed that temporary provisions will be allowed to expire, although many are routinely extended. An argument could be made that current law does not represent current policy in long-term projections, because current law would make revenues reach an unprecedented share of GDP. If argued instead that the revenues equal to current share of GDP represented current policy, then the long-term fiscal gap would be much larger because revenues would not rise in tandem with spending as a share of GDP.

If policymakers decided to increase revenue to reduce the deficit, five broad choices are often discussed:39

Redesigning the tax system or adding new revenue sources could theoretically improve economic efficiency and might be more appealing to some than increasing existing taxes, but in practice, compensating those made worse off from these changes may result in those policies raising little additional revenue. Generally, economists favor eliminating tax expenditures over raising marginal tax rates on efficiency grounds, although some specific expenditures may promote economic efficiency. Tax expenditures have also been criticized on the grounds of equity and complexity.40



All budget data presented in this report are from Congressional Budget Office, Updated Budget Projections, April 2014, Congressional Budget Office, Budget and Economic Outlook, February 2014 (hereafter referred to as "CBO baseline"), Congressional Budget Office, The 2013 Long-Term Budget Outlook, July 2014, or Office of Management and Budget, FY2015 Budget of the U.S. Government, March 2014 (hereafter referred to as "President's Budget"). The budget deficit is the excess of outlays over revenues in a given year, broadly similar to the amount borrowed from the public that year. The publicly held debt is the accumulation of all past borrowing from the public. The gross debt is the publicly held debt and the intragovernmental debt (the debt that one part of the federal government borrows from another part, mainly government trust funds). All references to years are to fiscal years, unless otherwise noted. This report primarily focuses on CBO's projections because of its role for Congress, but this analysis also applies to budget projections made by OMB and others.


Expressing budget data as a percentage of GDP is most appropriate for making comparisons over extended periods of time because it allows the relative size of those amounts to be consistently compared. Dollar figures are not appropriate because of inflation, which means that a dollar has a smaller amount of purchasing power over time, and because as the economy has become larger, there are more resources available to finance spending through borrowing or taxes.


Some expiring mandatory spending programs are assumed to continue under the baseline. According to CBO, "The Deficit Control Act specifies some exceptions. For example, spending programs whose authorizations are set to expire are assumed to continue if they have outlays of more than $50 million in the current year and were established at the time of or before the enactment of the Balanced Budget Act of 1997. Programs established after that law was enacted are not automatically assumed to continue but are considered individually in consultation with the budget committees." Congressional Budget Office, Budget and Economic Outlook, January 2012, p. 11.


For more information, see CRS Report R43124, Expired and Expiring Temporary Tax Provisions ("Tax Extenders"), by [author name scrubbed].


For more information, see CRS Report R40907, Medicare Physician Payment Updates and the Sustainable Growth Rate (SGR) System, by [author name scrubbed].


For more information, see CRS Report R43411, The Budget Control Act of 2011: Legislative Changes to the Law and Their Budgetary Effects, by [author name scrubbed].


The Affordable Care Act (P.L. 111-148) was a major legislative change that permanently increased taxes and mandatory spending, but it is not highlighted below because its effect on the deficit was relatively modest until 2013 and 2014, when CBO estimated it would reduce the deficit. The effect on the deficit was modest until then because these changes were offsetting and gradually phased in.


For more information, see CRS Report R41427, Troubled Asset Relief Program (TARP): Implementation and Status, by [author name scrubbed].


ARRA resulted in a one-time increase in budget authority that was outlayed over several fiscal years. For more information, see CRS Report R40537, American Recovery and Reinvestment Act of 2009 (P.L. 111-5): Summary and Legislative History, by [author name scrubbed] et al.


For more information, see CRS Report R41508, Expiring Unemployment Insurance Provisions, by [author name scrubbed].


For more information, see CRS Report R41648, Social Security: Temporary Payroll Tax Reduction, by [author name scrubbed].


For more information, see CRS Report RL34661, Fannie Mae's and Freddie Mac's Financial Problems, by [author name scrubbed].


TARP was created by the Emergency Economic Stabilization Act (P.L. 110-343).


Reductions in TARP's projected lifetime cost led to negative outlays in 2010 and 2011 to compensate for what proved in hindsight to be too large an estimate of its cost in 2009.


In 2008, before ARRA was enacted, non-defense discretionary budget authority was $494 billion. In 2009, it was $803 billion, but in 2011, it was $511 billion.


For more information, see CRS Report R41508, Expiring Unemployment Insurance Provisions, by [author name scrubbed].


For more information, see CRS Report R41648, Social Security: Temporary Payroll Tax Reduction, by [author name scrubbed].


Discretionary spending is spending controlled by the appropriations process. Mandatory spending is controlled by other laws and includes entitlement spending.


In 2012, there were security and non-security caps instead.


See CRS Report R43411, The Budget Control Act of 2011: Legislative Changes to the Law and Their Budgetary Effects, by [author name scrubbed].


For more information, see CRS Report R40907, Medicare Physician Payment Updates and the Sustainable Growth Rate (SGR) System, by [author name scrubbed].


See CRS Report R42894, An Overview of the Tax Provisions in the American Taxpayer Relief Act of 2012, by [author name scrubbed].


Office of Management and Budget, "American Taxpayer Relief Act Reduces Deficits by $737 billion," blog, January 1, 2013,


In September 2008, CBO was projecting that the economy would grow at a modest rate in 2008 and 2009. Instead, the economy sharply contracted.


Congressional Budget Office, The Effects of Automatic Stabilizers on the Federal Budget as of 2013, March 2013.


Long-term budget data can be found in Congressional Budget Office, The Long-Term Budget Outlook, July 2014.


Office of Management and Budget, FY2015 Budget of the U.S. Government: Analytical Perspectives, March 2014, p. 20.


Congressional Budget Office, How Different Future Interest Rates Would Affect Budget Deficits, Blog post, March 27, 2013, available at


While tax provisions are assumed in the baseline to expire as scheduled, it is assumed that major mandatory spending programs will continue past their expiration date.


For more information, see CRS Report R41134, The Impact of Major Legislation on Budget Deficits: 2001 to 2010, by [author name scrubbed] and [author name scrubbed].


In simple terms, saving is measured by the excess of income over spending. When the government's outlays exceed its revenues, it must borrow from the public to finance the difference, and it has a negative saving rate.


Congressional Budget Office, The Effects of Automatic Stabilizers on the Federal Budget as of 2013, March 2013.


Economic effects of an unsustainable budget deficit are discussed in CRS Report R40770, The Sustainability of the Federal Budget Deficit: Market Confidence and Economic Effects, by [author name scrubbed].


Congressional Budget Office, The Long-Term Budget Outlook, September 2013, p. 117. Economists call this concept the "fiscal gap." Because the debt to GDP ratio rises continuously over long-term projections, a longer time period leads to a larger fiscal gap estimate.


See CRS Report R43476, Returning to Full Employment: What Do the Indicators Tell Us?, by [author name scrubbed].


CBO estimates that the budget would still be in deficit at full employment, so there is a limit to how much higher economic growth could reduce the deficit.


About half of federal spending on physical capital is for defense, and reduced defense capital spending may not have the same economic effects as reduced non-defense capital spending.


See CRS Report R42894, An Overview of the Tax Provisions in the American Taxpayer Relief Act of 2012, by [author name scrubbed].


For a further analysis, see CRS Report R41641, Reducing the Budget Deficit: Tax Policy Options, by [author name scrubbed].


For more information, see CRS Report RL33641, Tax Expenditures: Trends and Critiques, by [author name scrubbed].