February 5, 1998

How Big is the Federal Government —
and Would the Administration's Budget Make it Bigger?

Overall Spending, Defense Spending and Major Parts of
Domestic Spending at Lowest Levels in Many Years
as a Share of the Economy

by Robert Greenstein, Isaac Shapiro, and Sam Elkin


Although some contend that the Administration's budget represents a return to "big government," a review of the Administration's budget and of federal spending trends indicates this is not so. Government spending as a share of the economy is now at one of the lowest levels in recent decades and would decline further in the next few years both under current policy and under the Administration's budget.

Current Levels of Federal Expenditures

This analysis uses the standard measure of the size of the federal government — federal expenditures as a percentage of the Gross Domestic Product. (The Gross Domestic Product, or GDP, is the basic measure of the size of the U.S. economy.)

Some observers, while agreeing that the size of the federal government has fallen relative to the size of the economy, ascribe the reduction solely to a decline in defense spending. And indeed, defense spending has fallen substantially in recent years. In 1997, defense spending amounted to 3.4 percent of GDP, down from an average of 5.8 percent in the 1980s. This decline is a beneficial by-product of the end of the Cold War.

The data also show, however, that major categories of non-defense spending also are at historically low levels relative to the size of the economy.

Would the Clinton Budget Alter this Picture?

Many Administration proposals to expand programs are offset by reductions in other programs. For example, the budget includes more than $30 billion in entitlement reductions over five years.

The only proposals in the budget that would cause total government expenditures to rise, relative to what the expenditure levels would be under current policy, are those initiatives whose costs would be covered by revenue that the Administration expects to secure from tobacco legislation or a few other revenue sources, rather than from reductions in other government programs. The initiatives whose costs would be financed in this manner total $82 billion over five years, for an average of about $16 billion a year. These expenditures would not cause any increase in the deficit (or any reduction in budget surpluses) because they would be paid for, primarily through revenues anticipated from tobacco legislation. (For a further discussion of this matter, see the Center on Budget and Policy Priorities' analysis entitled "Federal Government Would Continue to Consume Less of Economy Under Administration's Budget.") Administration officials have said that if the tobacco legislation or other financing mechanisms in the budget do not materialize, they will trim their initiatives or propose other ways to finance them.(5)

This $16 billion a year in expenditures that would be financed primarily with tobacco revenues would increase the size of the government relative to what the government's size will be if no policy changes are made. But the increase would be modest and barely be noticeable when compared to the economy. The $16 billion a year involved amounts to between one-tenth and two-tenths of one percent of the Gross Domestic Product.

This increase is sufficiently small that total federal spending still would drop as a percentage of the economy. The Office of Management and Budget estimates that under the Administration's budget, federal expenditures would drop to 18.8 percent of GDP in 2003, still the lowest level since 1974.

In short, the historical trends described here showing that the federal government will shrink in relation to the economy under current policy also apply to the Administration's budget. The proposed budget initiatives would not alter these trends.

The Administration's initiatives should be judged on the merits of how effective they would be, how they would be financed, and how they affect larger fiscal policy goals. Suggestions that these initiatives will bring back big government, however, are unfounded.

These issues are examined in more detail below.


Measuring Federal Expenditures in Relation to GDP

Measuring expenditures as a share of GDP is a standard means that economists and fiscal policy analysts use to assess the size of government. This measure shows how much of the economy consists of federal spending and how that share has changed over time. (Appendix A discusses an alternative measure — federal expenditures adjusted for inflation and population growth.)

Expenditures for Government Programs

Total government spending is not a completely accurate representation of the size of the current programs and operations of the federal government because it includes interest payments on the debt. Such payments do not support any government programs or employees; rather, they are the result of the deficits of past years. Interest payments now consume a much larger portion of the budget — and of GDP — than they did in the decades before the 1980s because of the major increase in the debt that occurred during the 1980s.

Comparisons of the size of the programs and operations of the federal government over time thus should focus on program spending — that is, total expenditures for everything except interest payments on the debt:

Non-Defense Discretionary Spending

Non-defense discretionary programs — non-defense programs other than entitlements — encompass a wide range of government responsibilities, including education, health research, veterans' hospitals, law enforcement, and the federal government's administrative operations. Many grants to state and local governments are included in this part of the budget. Measured as a share of the economy, federal expenditures for non-defense discretionary functions and programs is relatively low compared with the levels over recent decades. (See Table 2.)


Defense spending's share of the economy has shrunk considerably since 1986, the height of the Reagan defense build-up. If the relative size of defense and non-defense discretionary spending remains unchanged after fiscal year 1999, defense spending will decline from 3.4 percent of GDP in fiscal year 1997 to 2.7 percent in fiscal year 2002, according to CBO's estimates. OMB's estimate is very similar. Defense spending equaled 6.3 percent of GDP in 1986.

As a share of GDP, these levels are smaller than in any year before World War II. At the same time, even at current levels, the United States devotes a larger share of its economy to defense expenditures than most or all other major industrialized nations.


Sources of Spending Reduction

The decline in defense spending has led some to conclude that this decline is the only factor that has contributed significantly to the contraction of the federal government as a share of GDP (and hence the only factor other than revenue increases that has contributed significantly to balancing the budget). Examination of the data shows, however, that this perception is not correct; spending as a share of GDP has fallen for other parts of the budget as well.

Those contending that virtually all of the reduction in expenditures has occurred in the defense part of the budget sometimes note that total federal non-defense spending, at 16.7 percent of GDP, is above the levels of both the 1970s and the late 1980s. Such statements, however, miss several important points.

First, those making this argument have neglected to look at trends in federal spending exclusive of interest payments on the debt. The large deficits that the nation rang up in the 1980s caused interest payments to rise as a share of GDP; this increase in interest payments masks the decline in spending in much of the non-defense budget, measured as a share of GDP.

Deficits stem from the imbalance between revenues and spending, and the factors that produced the large deficits of the 1980s included the 1981 tax cut and the defense build-up of that period. Counting the increase in interest payments that these deficits generated simply as an increase in domestic spending — and using the rise in interest payments to make it appear as though the domestic side of the federal government has been growing — produces a misleading result. It obscures the decline that has occurred in expenditures for much of the non-defense part of the federal government, taken as a share of GDP.

Federal expenditures for non-defense programs — total federal expenditures except for defense spending and interest payments on the debt — totaled 13.6 percent of GDP in 1997. This was lower than the percentage for every year from 1975 through 1983 and for all years from 1991 through 1996. It was above the percentage for the late 1980s. (See Table 3.)

Even these figures cloud part of the picture. In recent years, Medicare and Medicaid expenditures have risen substantially. But spending on non-defense programs other than Medicare and Medicaid has dropped as a share of GDP.

CBO estimates that total federal expenditures for non-defense programs other than Medicare and Medicaid will fluctuate between 10.1 percent and 10.3 percent of GDP between 1997 and 2001. These levels are lower than in any other year since 1971, with the sole exception of 1989 when spending on non-defense programs other than Medicare and Medicaid also stood at 10.3 percent of GDP. In several years in the 1970s, federal expenditures for non-defense programs other than Medicare or Medicaid equaled or exceeded 13 percent of GDP, well above the current level. These data demonstrate that defense is not the only part of the federal government that has shrunk significantly as a share of the economy.

Furthermore, CBO projects that spending for non-defense programs other than Medicare and Medicaid will drop to 9.9 percent of GDP in 2002. That would equal the same level at which it stood 31 years earlier, in 1971. (OMB estimates that under the Clinton budget, spending for non-defense programs other than Medicare and Medicaid would equal 10.1 percent of GDP. There is no year after 1971 when this part of the budget has consumed a smaller percentage of GDP.)

Even among entitlements other than Medicare and Medicaid, the story is one of some shrinkage as a share of GDP. Federal expenditures on entitlements and other mandatory spending other than Medicare and Medicaid equaled 7.4 percent of GDP in 1997 and will edge up to 7.7 percent by 2002, according to CBO. Both the 7.4 percent and the 7.7 percent figures are below the share of GDP that federal expenditures for mandatory programs other than Medicare and Medicaid constituted in every year from 1973 to 1994.(7) (See Table 3.) The OMB estimates for the Clinton budget are similar.

It should not be surprising that Medicare and Medicaid have been rising as a share of GDP. In recent decades, health care costs in both the public and private sectors have risen faster than the economy, particularly as advances in medical technology have prolonged life, but often at significant cost. In addition, employers have been insuring a smaller share of employees since at least 1987 (the first year for which Census data on this matter are available), leading more low-wage workers and their families to turn to Medicaid for coverage and helping lead policymakers to broaden Medicaid to cover more children in low-income working families. Demographic factors as well have pushed up Medicare costs; in 1996, some 12 percent of the population was elderly, up from 9.6 percent in 1970.


Federal Taxes vs. Federal Spending

In recent weeks, some Members of Congress and some commentators have cited a CBO estimate that federal taxes now equal 19.9 percent of GDP and have observed that this is the highest level since the end of World War II. Such statements should not lead, however, to a conclusion that the federal government is bigger than ever. Total federal expenditures have exceeded 19.9 percent of GDP every year since 1974. Federal taxes were below this share, and the imbalance caused our large deficits.

The recent increase in federal taxes as a share of GDP has helped reduce the deficit to close to zero, but the decline in federal spending has been a larger factor in erasing the deficit over the past decade than the rise in revenues. Federal spending as a percentage of GDP is currently 2.4 percentage points lower than its average level in the 1980s. By comparison, federal tax receipts as a percentage of GDP are 1.5 percentage points above the 1980s average.

It also bears noting that the share of the GDP consisting of federal taxes is projected to decline over the next few years. In part, this decline reflects the design of the tax cuts under the 1997 tax legislation. A number of the tax cuts in the 1997 tax law either phase in gradually over several years or do not take effect for several years. As a result, taxpayers will experience new tax cuts each year for the next few years, and revenues will edge down modestly as a share of GDP. In addition, as CBO director June O'Neill recently testified, current levels of tax revenues are elevated in large part because capital gains income is unusually high due to the stunning rise in the stock market and to such factors as the recovery of the commercial real estate market. As O'Neill observed, "the special factors responsible for 1997's revenue surge cannot grow at the current pace indefinitely..."(8) For all of these reasons, CBO projects that revenue will decline from 19.9 percent of GDP in 1997 to 19.4 percent of GDP in 2002 and 19.3 percent in 2003 and succeeding years. OMB projects that under the Clinton budget, revenue will decline to 19.6 percent of GDP in 2003, with the higher level of revenue under the Administration's budget stemming in part from the anticipated proceeds from tobacco legislation and in part from estimating differences.

It also should be noted that the fact that revenues are at a high point for recent decades when measured as a share of GDP does not mean the typical family's tax burden is at a high point when measured as a share of its income. As O'Neill has explained, tax receipts have risen as a share of GDP in the last few years "mainly because realizations of capital gains were unusually high and because a larger share of income was earned by people at the top of the income ladder, who are taxed at higher rates."(9) These developments do not materially affect the percentage of income paid in federal taxes by the typical family — i.e., the family in the middle of the income spectrum. The percentage of income the typical family pays in taxes is not at a high point for recent decades and was reduced somewhat by the 1997 tax legislation. (A forthcoming Center paper will address the issue of tax burdens in greater detail.)


Appendix A. Measuring the Size of the Federal Government

This analysis measures federal spending relative to the size of the economy. Such a measure addresses the following questions:

This measure does not examine whether federal expenditures have increased at a faster pace than inflation. It turns out that both the size of the economy and overall federal expenditures have increased faster than inflation, with the pace of economic growth exceeding the pace of federal expenditure growth.

Part of the increase in federal spending in excess of inflation reflects population growth. Excluding Medicare and Medicaid (which grew after accounting for general inflation and population increases), inflation-adjusted federal expenditures per capita will be lower in 2002 under current policy than in any year since 1982, according to the CBO projections. If Medicare and Medicaid are included in the calculation, inflation-adjusted federal expenditures per capita have been more or less flat since 1990, ticking and up and down within a relatively narrow range. In 2002, inflation-adjusted federal expenditures per capita are projected to be in the lower part of this range. The OMB estimates of expenditures under the Clinton budget yield similar results.(10)

Table 1


Table 2


Table 3

End Notes

1. The figures in this paper are based on data from two sources — the Congressional Budget Office (especially CBO's new report, The Economic and Budget Outlook: Fiscal Years 1999-2008, January 1998) and the Budget of the United States Government, Fiscal Year 1999.

2. The figure for 2002 was estimated by the Center on Budget and Policy Priorities based on CBO data. The Center estimate assumes that the relative size of total defense and non-defense discretionary expenditures will remain unchanged after 1999.

3. These calculations include "offsetting receipts" other than receipts related to Medicare. As the Congressional Budget Office notes in its latest report on the budget and the economy, "Offsetting receipts are income that the government records as negative spending." Another way of calculating expenditures for non-defense programs other than Medicare and Medicaid is to exclude all offsetting receipts. This alternative method yields virtually the same result; if all offsetting receipts are excluded, expenditures for non-defense programs other than Medicare and Medicaid are projected to be lower as a share of GDP in 2002 than in any year since 1970.

4. Mandatory spending other than Medicare and Medicaid expenditures equaled 7.7 percent of GDP in 1995 and 7.6 percent of GDP in 1996.

5. Bureau of National Affairs, Daily Tax Report, January 23, 1998, G-4. Similar Administration statements were reported in the February 3 Washington Post.

6. This analysis includes spending from the Violent Crime Reduction trust fund as part of non-defense discretionary spending. From 1998 through 2000, expenditures from this trust fund are subject to a separate cap from other discretionary spending.

7. Federal spending on entitlements and other mandatory programs other than Medicare and Medicaid equaled 7.7 percent of GDP in 1995 and 7.6 percent of GDP in 1996.

8. Statement of June E. O'Neill, Director, Congressional Budget Office, before the Senate Budget Committee, January 28, 1998.

9. O'Neill, p. 6.

10. The inflation-adjusted growth in these programs would be smaller if, instead of using the general inflation rate to make adjustments for all programs, costs in Medicare and Medicaid were adjusted to reflect changes in the inflation rate for health costs.