Choosing the Nation's Fiscal Future (2010)

As explained in Chapter 1, examining the budget outlook requires a “baseline” set of projections. A baseline shows what would happen to federal spending and revenues if current policies remained in place indefinitely. It serves as a benchmark for measuring the budget effects of proposed policy changes. It is not intended to be a prediction or forecast.

The Congressional Budget Office (CBO) produces detailed, widely cited, and—in the committee’s view—highly credible baseline projections. The committee used the CBO’s baseline estimates, with a few adjustments, to construct its 75-year baseline projection. This baseline is referred to here and in the body of the report as the “study baseline.” Specifically, the study baseline begins with the CBO June 2009 long-term outlook (Congressional Budget Office, 2009b), which incorporates CBO’s more detailed baseline estimates for the 10 years 2010-2019 published in March 2009 (Congressional Budget Office, 2009c), as a starting point. The committee adjusted those estimates as discussed below to produce the study baseline projection used in the report and shown in Chapter 1.

CBO is required by law to follow certain conventions when constructing its baseline projections.¹ Congress uses this “standard” baseline to “score” the budget effects of bills as they move through the legislative process.

Congress has stipulated the following general rules for constructing the standard baseline for different categories of programs.

• For discretionary programs—which encompass most defense spending, foreign aid, scientific and biomedical research, infrastructure, education, and the many other activities of government that Congress funds annually through appropriation acts: Congress directs CBO to assume the most recent year’s appropriation, adjusted for inflation.

• For revenues and mandatory spending—which includes the “big three” entitlements (Social Security, Medicare, and Medicaid) and plus many smaller entitlements, such as food stamps and unemployment compensation, as well as offsetting receipts—Congress directs CBO to assume that current laws continue unchanged, with two exceptions:

  • Expiring programs² (not to be confused with expiring tax provisions) are generally assumed to continue unchanged. Examples are the programs known as Temporary Assistance for Needy Families (the successor to the former program of Aid to Families with Dependent Children) and the State Children’s Health Insurance Program: such programs come up for periodic reauthorization, and it would not be credible to assume their disappearance.

  • Excise taxes dedicated to a trust fund are assumed to continue. An example is the aviation taxes dedicated to the airport and airways trust fund. Since the CBO baseline assumes that spending under such programs continues, it is reasonable to assume that their taxes continue as well.

• For interest costs, Congress directs use of an estimate of the average interest paid to holders of the levels of borrowing and debt that are assumed in other parts of the baseline, using CBO’s estimate of future interest rates.

It is important to understand several assumptions under the standard rules:

• the expiration of the 2001 and 2003 tax cuts and of other temporary tax breaks;

• the rapid expansion of the Alternative Minimum Tax (AMT) that will apply to more than 40 million tax filers, 10 times today’s number;

• the omission of the likely costs of funding operations in Iraq and Afghanistan in the near term; and

• congressional agreement to allow physician payment rates in Medicare to fall sharply under that program’s “sustainable growth rate” (SGR) mechanism.

These assumptions, as reflected in the baseline, may represent current law, but arguably they fail to represent current policy.

The study committee has adjusted the CBO baseline estimates to make what it considers a more credible projection of spending and revenues on the basis of current policies. In doing so, it has relied where possible on estimates provided by CBO of the effects on its standard baseline of alternative policies. This helps maintain internal consistency among the elements of the baseline.

The study baseline adjusts CBO’s standard baseline policy assumptions in seven areas:

1. extending, beyond its currently scheduled expiration at the end of 2010, the portion of the 2001 and 2003 income tax cuts that affect households earning less than $250,000 a year;

2. indexing the AMT at the 2009 level;

3. extending certain other expiring tax provisions;

4. reducing defense spending to reflect a reduction of troops in Iraq and Afghanistan to 75,000 by 2013;

5. freezing Medicare physician payment rates, in dollar terms;

6. including the administration’s estimated cost of future disasters; and

7. estimating the additional debt and debt service required by the larger deficits that result from the above six adjustments.

These adjustments are intended to produce a more credible projection of current policies; they are summarized in Table B-1. For example, although Congress has not yet decided whether to extend the 2001 and 2003 tax cuts (now scheduled to expire after 2010), it is the committee’s best judgment that those cuts will be extended, as favored by both the Obama Administration and Republican leaders. Similarly, Congress has—with the President’s approval—routinely extended various smaller tax preferences every year or two, making it reasonable to assume future extensions.³ In the same vein, policy makers have consistently adopted temporary “patches” to keep more and more taxpayers from being subject to the AMT. At the administration’s behest, Congress has funded operations in Iraq and Afghanistan chiefly through supplemental appropriations (that is, outside the regular annual defense bill) and it has acted to prevent cuts in physician fees under the Medicare SGR formula.

Deciding how to treat each of these policies in constructing a baseline is obviously a matter of judgment. We have made the adjustments noted for

the study baseline to produce a more realistic projection of the likely budget effects of current policies. The panel believes that this alternative baseline paints a more credible picture of future budgets under current policies.⁴

For its long-term projections, the study has adjusted some CBO projections. The methods used by CBO for its long-term projections vary by category of spending: Social Security, Medicare and Medicaid, defense and other domestic spending, and revenues. This section discusses those methods and the study’s adjustments. For Social Security, the committee uses CBO’s “scheduled benefits” scenario, which assumes payments will be made on the basis of current law, regardless of the balance available in the trust fund. This spending projection implicitly assumes that other sources will be found to continue scheduled benefits, which otherwise by law would drop to the level of current trust fund revenues when it is exhausted, currently projected to occur in 2043. The methods used by CBO are similar to those used by the Social Security program’s trustees, but CBO makes what it terms minor adjustments to the trustees’ methods (for details, see Congressional Budget Office, 2009a).

For Medicare and Medicaid, the study baseline projection is the same as that for CBO’s “alternative fiscal scenario,” which assumes that the programs continue to operate under current law, except (as noted above) that the study assumes that Medicare payments will grow with inflation rather than being reduced under the program’s SGR mechanism.

For defense and other domestic spending, the study baseline assumes that discretionary program spending will grow with inflation through 2019 and that mandatory program spending will grow on the basis of current law, with some exceptions. The study does not assume that the stimulus spending enacted in the American Recovery and Reinvestment Act of 2009 will be a recurring part of the baseline for either spending or revenues. The same is true of the $106 billion supplemental appropriations legislation passed in June 2009, which included $80 billion for the wars in Iraq and Afghanistan, although adjustments are made, as noted above, consistent with a reduction in troop levels through 2013. Other study adjustments affect the projection of total outlays in 2019, after which outlays for this broad category of programs are assumed to grow in line with GDP through 2083.

For revenues, including Social Security and Medicare payroll taxes, the study baseline uses CBO’s projections, adjusted as noted above for extension of some expiring tax provisions and indexation of the AMT.

The combined effect of the committee’s adjustments to the CBO standard baseline is to increase estimates of future deficits and debt. The cumulative deficits forecast in CBO’s standard baseline over the 2010-2019 period, for example, increase by $2.2 trillion. The study baseline extends to 2083; see Table B-2. Over that period:

• Total outlays, which were 21 percent of GDP in 2008 and an estimated 27.6 percent in 2009, first decline from 2010 to 2012 to 22.2 percent of GDP (reflecting reduced spending associated with the 2008-2009 recession and the 2009 stimulus legislation) and then rise at an accelerating pace, reaching 60 percent of GDP in 2083.

• Social Security spending climbs from 4.3 percent of GDP today to about 6 percent in 2035, and then remains roughly stable as a percentage of GDP, reaching 6.2 percent in 2083. Although the demographic challenge posed by retirement of the baby boom generation is the largest short-term contributor to higher Social Security (and Medicare) outlays, rising life expectancy is projected to keep the share of elderly people in the population larger than it is today.

• Medicare and Medicaid more than double as a share of GDP between now and 2035, from 4.1 percent to 9.9 percent of GDP. Unlike Social Security, outlays for these programs are projected to continue growing steeply thereafter, almost doubling again—to 18.4 percent by 2083. Unlike Social Security, essentially a defined-benefit pension program in which total outlays roughly grow with the elderly population and the average wage,⁵ Medicare and Medicaid represent the government’s pledge to pay for health care at prevailing usage patterns and prices. Over the past few decades, rapid technological change, among other factors, has driven up federal spending at a rate faster than GDP growth.

• All other spending—consisting of defense and nondefense discretionary spending plus mandatory spending programs other than the “big three”—is assumed to follow the CBO baseline for the first 10 years, with the exception of the adjustments noted above. In the study baseline it is held constant as a share of GDP thereafter. Under these assumptions, this third category of spending declines gradually from 10.8 percent of GDP in 2008 and a recession-driven peak of 16.8 percent of GDP in 2009 to 8.6 percent of GDP in 2019, remaining at roughly that level thereafter.

• Revenues under current policies as defined in the study baseline rise gradually, from 17.7 percent of GDP in 2008 to 21.8 percent in 2083, somewhat less than CBO’s “alternative” baseline. That slow

rise occurs chiefly because of two factors—“real bracket creep” (because rising real incomes push taxpayers into higher brackets) and growing taxable distributions from retirement accounts as baby boomers retire. Those two factors are just big enough to outweigh other sources that are flat or gently declining.

The baseline projection of deficits and debt is derived arithmetically from the above projections, except for 2009-2019. In this period, the estimates of changes and debt depend partly on the issuance and repayment of certain debt associated with the Troubled Asset Relief Program (TARP), assistance to Fannie Mae and Freddie Mac, and other financial interventions associated with the 2008-2009 financial crisis and economic downturn. The study baseline retains CBO’s estimate of outlays for TARP and some elements of the assistance to Fannie Mae and Freddie Mac: it records their costs as outlays in the year of intervention based on the discounted present value of projected cash flows resulting from each action.

Projections over periods much shorter than 75 years are highly unreliable if viewed as forecasts. CBO and many others view Medicare and Medicaid as the key source of uncertainty in the long-run projections. Over the last few decades, annual “excess cost growth” in the two programs (growth in excess of GDP, adjusted for changes in the eligible populations) averaged 2.3 and 1.9 percent in Medicare and Medicaid, respectively. CBO assumes a slowdown in excess cost growth (to an annual average rate of 1.5 percent and 0.6 percent, respectively, for Medicare and Medicaid, over the 2019-2083 period). Furthermore, excess cost growth in the rest of the health care sector—which historically has moved in tandem with Medicare and Medicare—is projected to diverge, falling to an average of 0.5 percent (Congressional Budget Office, 2009c:28). Even with those assumptions, health care from all sources would consume an estimated 46 percent of GDP in 2080 (Congressional Budget Office, 2009c:30). If excess cost growth continued instead at about 2.5 percent annually, health care spending would account for an estimated 99 percent of GDP in 2082 (Congressional Budget Office 2007a:27).

Congressional Budget Office. (2009a). CBO’s Long-Term Projections for Social Security: 2009 Update. Washington, DC: Congressional Budget Office.

Congressional Budget Office. (2009b). The Long-Term Budget Outlook. Washington, DC: Congressional Budget Office.

Congressional Budget Office. (2009c). A Preliminary Analysis of the President’s Budget and an Update of CBO’s Budget and Economic Outlook. Washington, DC: Congressional Budget Office.