Skip Navigation

Long-Term Budgetary Pressures and Policy Options March 1997

The outlook for the deficit appears relatively benign over the next decade. After declining for the past four years, the deficit is expected to creep up as a share of gross domestic product (GDP) from 1996 through 2007 under current laws and policies. Although the increase is fairly moderate, it is by no means the end of the story because a deeper and more fundamental problem is just over the budgetary horizon.

About 2010, the oldest members of the huge baby-boom generation will turn 65 years old and begin to draw benefits from the government's three biggest entitlement programs--Social Security, Medicare, and Medicaid. At the same time, the growth of revenues will be squeezed because the proportion of people working and paying taxes will shrink. As a result, deficits will start to mount rapidly.

Financing the growth in entitlements through ever-increasing deficits is not a workable option. Indeed, the shortfalls projected for future years would become so large that they could put an end to the upward trend in living standards that the nation has long enjoyed. Thus, current U.S. budget policies cannot be sustained indefinitely without risking substantial economic damage. At some point, the growth of spending will have to be curbed or taxes raised.

The conclusions reached here are derived from a model that the Congressional Budget Office (CBO) has developed for projecting the deficit over several decades and for examining its effects on interest rates and economic growth. CBO first reported the results of that model in Chapter 4 of its Economic and Budget Outlook: Fiscal Years 1996-2006, published in May 1996. This report updates that earlier work to reflect the revised near-term outlook for the deficit.

Obviously, projections of future events are subject to considerable uncertainty. To get a sense of the likely range of outcomes, CBO developed its projections by using a broad spectrum of possible assumptions and conditions. Although the exact outcomes are sensitive to changes in demographics, economic factors, and the interpretation of policy, a basic conclusion holds: the nation's current budget policies are unsustainable even under optimistic assumptions. The long-term budgetary problem will not resolve itself without action by policymakers.
 
 

The Aging of the U.S. Population

The proportion of elderly people in the U.S. population will increase substantially in coming decades (see Table 1). According to the Social Security Administration, the number of people age 65 and older will more than double between 1995 and 2030, whereas the number of people who are 20 to 64 years old will increase by only 20 percent. Consequently, over the next several decades, young people will have to support a growing number of the elderly.

Why Will the Number of Retirees Increase?

The expected increase in the number of elderly people stems from two factors: people are living longer and the baby-boom generation is aging. Thanks to increased education, healthier living, and improvements in the quality of medical care for older people, a larger proportion of the adult population is reaching the age of 65, and life expectancy at that age has increased by two years since 1970--about a 15 percent increase. In 1970, the average person at birth was expected to live about 71 years. By 1990, the average life span had increased to 75 years; by 2010, it is projected to increase to 78 years.

The aging of the baby boomers is also an important factor in the outlook. Before World War II, the number of births in the United States slid to a low point (see Figure 1). That generation is now reaching retirement age, and their small numbers are providing a respite from budgetary pressure. After World War II, however, the number of births soared: between 1956 and 1961, births jumped to more than 4.2 million a year and did not drop below 4 million until 1965. People born between 1946 and 1964 have been labeled the baby-boom generation, and they will begin to draw Social Security benefits for retired workers in 2008, when the oldest of them first reaches age 62.

The Slowing Growth in the Labor Force

The growth of the labor force will slow significantly when the baby boomers retire because the birth cohorts that follow the boomers are considerably smaller. After the mid-1960s, the number of births dropped to well under 4 million and did not reach that level again until 1989. The labor force will also grow more slowly as women's participation in the labor force, which escalated sharply in the 1970s and 1980s, begins to approach that of men's. The Social Security Administration projects that the average rate of growth of the labor force will slow from the 2 percent a year it achieved

Figure 1.
Number of Births in the United States, 1910-1995

SOURCE: Congressional Budget Office using data from the National Center for Health Statistics.

from 1960 to 1989 to 1 percent annually for the 1989-2010 period and 0.2 percent between 2010 and 2050.

Like all long-range projections, those for the labor force are highly uncertain. Nevertheless, the relatively high rate of growth of the labor force in the past 35 years is unlikely to continue. Higher rates of immigration could prevent some of the expected deceleration, but for the labor force to continue to grow through 2030 at even 1 percent a year, its average annual rate since 1990, rates of immigration would have to exceed those seen early in this century by a wide margin.

Despite those uncertainties, the overall message is clear: with more retirees and little growth in the number of workers, the ratio of workers to retirees will plummet in coming decades. In 1950, for each person age 65 or older, there were 7.3 people in their working years from 20 to 64. By 1990, that ratio had dropped to 4.8 to 1; by 2030, there may be only 2.8 people of working age for every person 65 and over. The United States is not alone in facing these problems: populations are graying in other industrialized countries as well (see Box 1).

How Will Demographics Affect the Budget?

Both the outlay and revenue sides of the federal ledger will be strained as the ratio of workers to retirees de-clines. Outlays for government programs that provide retirement and health benefits to the elderly will rise substantially as the number of people eligible to receive those benefits shoots up. At the same time, revenues will be pinched because the number of people working and paying taxes will grow more slowly. Moreover, as the growth of the labor force slows, economic growth will taper off, causing the growth of taxable nonlabor income, such as interest and dividends, to slow as well. Of particular concern are Social Security and Medicare's Hospital Insurance (HI) program. Because those entitlements are now structured to rely on payroll taxes, the growth of labor earnings is one of the keys to their financial health.

The projected discrepancy between spending and revenues will be a serious one. For example, the trustees for Social Security and Medicare's Hospital Insurance program project that outlays for those programs will grow from 6.4 percent of GDP in 1996 to 10.9 percent in 2050. At the same time, the inflows of funds (excluding interest) for those two programs are projected to fall from 6.6 percent of GDP in 1996 to 6.3 percent in 2050. Hence, although inflows exceed spending for those programs now, that surplus will disappear and a large gap between spending and inflows will open up. By 2050, outlays are projected to exceed inflows by about 75 percent.⁽¹⁾



The Continued Rapid Growth of Federal Health Expenditures

Rapidly rising expenditures per beneficiary in the Medicare and Medicaid programs will present a particularly serious challenge to the budget in coming years unless significant steps are taken to reduce their rate of growth. Federal spending for health care has been growing at a brisk pace for many years. Over the past decade, expenditures for Medicare have increased at an annual rate of about 10 percent; Medicaid spending has risen at a rate of about 15 percent (see Table 2). With such growth, Medicare and Medicaid have taken up an
 

increased share of the economy's income: from 1.3 percent of GDP in fiscal year 1975 to 3.8 percent in 1996.

Despite some good news about federal health costs in 1996 and the apparent recent success of private insurers in controlling their costs, many of the factors that have contributed to the fast growth of the government's health entitlement programs are still in place. CBO projects that outlays for Medicare and Medicaid will continue to rise by about 8 percent a year over the next decade, which is slower than in the past, but will still be above the overall growth of GDP. As a result, spending for those programs is projected to increase to 5.5 percent of GDP in fiscal year 2007.

Although some of that growth comes from an expansion in the number of enrollees, most of it stems from continuing increases in expenditures per enrollee at rates well in excess of inflation. Unlike Social Security, whose real (inflation-adjusted) spending for each beneficiary is set legislatively by a formula that depends on a person's wage history, Medicare and Medicaid are open-ended entitlements in the sense that they place no dollar limit on the benefits they provide to each participant. CBO projects that over the next de-

cade, federal spending per enrollee in Medicare and Medicaid will increase at more than twice the rate of inflation, as measured by the consumer price index for all urban consumers, and 50 percent above the growth rate of wages. Thus, even if the ratio of retirees to workers was not projected to increase, the health programs would consume a growing share of GDP.

Medicare's trustees assume that the growth in expenditures will slow significantly over the next two decades. Specifically, the trustees of the Hospital Insurance Trust Fund assume that, after 25 years, the cost per unit of service provided by the HI program would grow at the same rate as average hourly earnings; the trustees of the Supplementary Medicare Insurance program assume that, after 13 years, the growth in costs per enrollee would decline gradually so that they would be growing no faster than GDP per capita after 25 years. Given the historical experience of health costs per enrollee, those assumptions may be considered optimistic. Even with that slowing in per-enrollee costs, however, the trustees project that total Medicare spending will continue to climb, rising from 2.7 percent of GDP in 1996 to 8.1 percent in 2050.



The Long-Term Effects of an Aging Population

What would happen if no changes were made to U.S. budget policy in the face of the impending retirement of the baby boomers? CBO addressed that hypothetical question by projecting future budget revenues and expenditures under various economic and demographic conditions and by examining their impact on the federal deficit and the economy over the next several decades. The approach used by CBO is broadly similar to that taken by the General Accounting Office (GAO) and the Office of Management and Budget in considering the same question.⁽²⁾

Budget Assumptions

Developing computer models of the long-term implications of existing laws and policies requires making assumptions about the basic nature of policy in the absence of change. Those assumptions formed a base scenario; varying them produced alternative scenarios.

For the 1997-2007 period, CBO followed its 10-year baseline projections. Taxes and mandatory spending reflect current law, and discretionary outlays grow with inflation, subject to their statutory caps. But extending such detailed assumptions over the long run is hard to justify. For one thing, techniques that are suitable for preparing 10-year budget projections can produce misleading results when used to produce very long-run projections.

Thus, for the years after 2007, CBO did not attempt to extend its regular budgetary projections. Instead, it simply assumed that spending would grow according to some simple and reasonable rules for most categories of the budget. CBO also adopted the official long-term projections for Social Security, Medicare, and federal retirement programs prepared by other government organizations. Those projections were then adjusted for differences between CBO's economic assumptions and those of the other organizations (see Box 2 for more details).

To allow for different possibilities, CBO prepared two sets of simulations for discretionary spending. One assumes that discretionary programs after 2007 will grow at the rate of inflation, which would hold their real value constant in today's dollars. The other set assumes that discretionary programs will keep pace with the growth of the economy, which would allow the amount spent on the discretionary accounts to rise with both inflation and real economic growth.

Holding the long-term growth of discretionary programs to the rate of inflation--rather than letting them grow with the economy--is an optimistic assumption. It does not allow discretionary spending to grow with population, let alone with real income per capita. It implies that spending for those programs as a share of GDP would decline sharply from 7 percent of GDP in 1996 to 3 percent in 2050. But public demands for many categories of discretionary spending--education, infrastructure, and environmental protection, to name a
 

few--may well grow with a rising population and real incomes. Moreover, given the huge uncertainties of looking so far ahead, it may be overly optimistic to assume that the world will remain as peaceful as it is today and that the share of income spent on the military will continue to decline over the next half century.

Economic Assumptions

CBO developed its simulations of the economy using a standard model of economic growth. In that model, the production of goods and services in the economy, as measured by potential GDP, depends on hours of labor, capital, and total factor productivity (TFP). Gross domestic product also varies for cyclical reasons, but that variation averages out over time and is not considered further in this chapter.

CBO's model also provides for the way the nation's debt (the total amount that the government explicitly owes) interacts with the economy. As deficits rise, they crowd out capital investment, slow economic growth, and raise interest rates. In turn, the growth in tax revenues declines, and the cost of servicing the debt goes up. Those economic feedbacks between the deficit and the economy can significantly increase the size of the deficit--in essence, imposing a fiscal penalty rather than a dividend.

From 1997 to 2007, the base scenario follows the medium-term projections presented in CBO's Economic and Budget Outlook: Fiscal Years 1998-2007. For the years after 2007, CBO makes four assumptions about the economy. First, the annual growth in hours of work is assumed to slow to a crawl as the baby boomers leave the workforce or otherwise reduce their average hours of work. Consequently, the annual growth of total hours in the nonfarm economy drops from its average rate of 1.6 percent from 1979 through 1989 to only 0.1 percent between 2020 and 2030.⁽³⁾ Second, CBO assumes that growth of total factor productivity, which is the growth in output that is not attributable to growth in either capital or labor, would rise 1 percent each year. Third, the growth of capital depends on whether the projection includes economic feedbacks. In projections without economic feedbacks, capital grows at the same rate as the overall economy after 2007, and rising deficits have no effect on the formation of capital or economic growth. By contrast, in projections with economic feedbacks, burgeoning deficits crowd out capital investment and slow the growth of the capital stock. The effect of the deficit on capital investment in those projections is assumed to be partially offset by increased private saving and by borrowing from abroad. Finally, CBO assumes that inflation after 2007 would remain steady at 2.6 percent.

CBO made two major technical changes in its long-term budget model since it was unveiled last May. First, CBO altered its method for aggregating the components of investment into a measure of the capital stock. The new procedure is now consistent with CBO's method of preparing its medium-term (10-year) projections. (That revision also changed the definition of total factor productivity in the model.) Second, partly as a result of changing its measure of the capital stock, CBO also increased its estimate of the long-term growth of total factor productivity. Last May, CBO assumed that TFP would grow about 0.7 percent a year; it is now assumed to grow at 1 percent a year. The new rate is consistent with the historical rate of growth of CBO's revised measure of TFP from 1952 to 1996, but it is noticeably faster than what CBO assumes in its medium-term projections from 1997 to 2007.

The revision in the growth of TFP after 2007 significantly raises CBO's estimates of the growth rate of potential GDP in the long run. Last May, CBO projected that, without economic feedbacks, the trend in the annual growth rate of real GDP will slip from about 2.0 percent in 2005 to 1.3 percent in 2020, reflecting the slowing growth of the labor force. CBO now expects it to decline to 1.7 percent. Thus, although the labor force is still expected to grow much more slowly when the baby boomers retire, the pickup in TFP growth after 2007 offsets some of that decline. CBO's assumption about growth in real GDP in the long run is more optimistic than the Social Security Administration's. Implicitly, CBO incorporates the chance of a period of exceptionally high growth in productivity. Of course, making such long-term projections involves huge uncertainties, and analysts disagree about the appropriate assumption for growth in productivity.

Economists often use GDP to put a common scale on budget revenues and outlays over time, and CBO has followed that practice in this chapter. But for measuring real economic income per person, CBO used the concept of gross national product (GNP). Unlike GDP, gross national product does not include the net dividend and interest payments owed to foreigners who invest in the United States. As a result, it is a better measure than GDP of the income actually available to the U.S. population. In the projections without economic feedbacks, the growth of GNP matches that of GDP quite closely. However, in the projections with feedbacks, GNP and GDP diverge significantly because deficits are partly financed by additional borrowing from foreigners.

Simulations Without Economic Feedbacks

The assumptions described above are the key elements in the long-term simulations, and because of their critical importance, a wide range of alternative assumptions was also considered. But to keep the analysis relatively simple, CBO first presents the simulations without considering how deficits would adversely affect the economy--that is, without incorporating economic feedbacks.

Even without those feedbacks, the outlook for the budget deficit is gloomy in the early decades of the 21st century. Unless changes were made in budget policy, the deficit would increase to relatively high levels in the 2030s. Under either assumption about discretionary spending (that it rises either with the rate of inflation or at the same rate as the economy), the national income and product accounts' (NIPA) deficit would climb from 2 percent of GDP in 1996 to between 10 percent and 13 percent in 2035 (see Table 3). Moreover, the deficit

would continue to rise rapidly in the years thereafter, growing to between 13 percent and 18 percent of GDP in 2050. By any standard, the deficit would be large.

In fact, since the nation's founding, the U.S. deficit has exceeded 10 percent of GDP for only a few brief periods--and those occurred during major wars.

Figure 2.
Projections of Federal Debt and Real GNP per Capita, Using the Assumptions of the Base Scenario with Economic Feedbacks

SOURCE: Congressional Budget Office.
 
NOTES: Simulations I, II, and III are based on alternative assumptions about population and productivity growth (see Box 3 on page 15). Simulation II is the base scenario, which assumes that the population grows according to the midrange path of the Social Security Administration and that total factor productivity grows at 1 percent annually. Simulations I and III are defined so that two-thirds of the 750 alternative simulations fall between them. Thus, the chance of an outcome better than Scenario III is about 15 percent; correspondingly, the chance of an outcome worse than Scenario I is also about 15 percent.
 
The projections of real GNP per capita are truncated when debt held by the public exceeds 300 percent of gross national product.
 
GNP = gross national product.

In turn, the total amount that the government owed would soar to historic levels. Since 1790, the United States has let its federal debt exceed 100 percent of GDP only once for a brief period during World War II. Moreover, until the 1980s, the ratio of debt to GDP had never risen significantly during a period of peace and prosperity. But under the base scenario, the national debt would increase from 50 percent of GDP in 1996 to 122 percent in 2035 if discretionary spending grew with inflation. If it grew with the economy, the debt would surge to 158 percent of GDP. Because the debt would be forever growing faster than the economy, it would ultimately become unsustainable.

Little of the projected growth in federal debt would be used to finance productive government investment. Instead, the growth in borrowing would be used largely to increase consumption by elderly people and to pay interest on the debt (see Table 4). In CBO's simulations, outlays for Social Security would increase from 5 percent of GDP in 1996 to 6 percent in 2050; Medicare spending would rise from 2 percent of GDP in 1996 to 8 percent in 2050. Federal Medicaid spending would move upward from 1 percent of GDP in 1996 to about 3 percent in 2050, reflecting the growth in the cost of health care per enrollee and the increasing number of elderly people who need nursing home care. Revenues and other noninterest outlays would remain a relatively constant share of GDP.

Simulations with Economic Feedbacks

The long-term budget outlook becomes even bleaker when the simulations include the effect of the deficit on the economy. Under the optimistic assumption that dis-cretionary outlays would grow with inflation, the federal deficit would increase to 18 percent of GDP in 2035 (see Table 5). If they grew with the economy, the federal deficit would climb to 30 percent of GDP.

Those increases would clearly push federal debt to unsustainable levels. Eventually, they would greatly weaken the economy and end the upward trend in real GNP per capita that the United States has enjoyed over its history (see Figure 2). If discretionary outlays grew with inflation, federal debt would rise to 171 percent of GDP by 2035; if they grew with the economy, federal debt would surge to almost 2.5 times GDP. With federal debt growing so rapidly, the economy would enter a period of accelerating decline.

Economic feedbacks intensify the nation's long-term budgetary problems for two reasons. First, the cost of interest on the debt would soar as interest rates went up and the stock of federal debt kept getting larger. Because interest costs would be growing continually faster than the economy's income, they would eventually reach an unsustainable level. Indeed, the growth of debt would accelerate out of control as the government attempted to finance its interest payments by issuing more debt. With each new round of debt, the rate of interest that the government paid would move up, and the rate of economic growth would move down (see Figure 3). Since the 1980s, interest rates have exceeded the rate of economic growth, but that situation would grow much worse because interest payments on the debt would be rising faster than the economy's ability to service that debt. Eventually, the government would find itself caught in a vicious circle of issuing ever larger amounts of debt to pay for ever higher interest charges.

Figure 3.
Projections of Nominal Interest Rates and Economic Growth, Using the Assumptions of the Base Scenario

SOURCE: Congressional Budget Office.
 
NOTE: Discretionary spending is assumed to grow with the economy. Nominal growth rates are smoothed using a centered, three-year moving average. Economic growth rates are measured as percentage changes in nominal gross national product. Interest rates on government debt are based on a weighted average of rates on all maturities of debt.

The second reason that economic feedbacks intensify the nation's long-term budgetary problems concerns the baby boomers. The feedbacks weaken the economy, and as a result, less income is available to finance retirement benefits for the baby-boom generation. Under current law, benefits for new cohorts of retirees grow at the same rate as average wages in the economy, but benefits for previous cohorts grow at the rate of inflation. Thus, even though wages would grow more slowly as the economy weakened, federal spending for Social Security benefits would be affected only gradually. Consequently, federal outlays for Social Security would absorb a much larger fraction of the economy's income. (The Medicare and Medicaid programs do not pose quite the same problem because spending for them is not linked to past wages. Instead, CBO assumed that as the growth of wages slowed, the growth of health care costs would also slow.)

A detailed statistical accounting of the uncertainty in the assumptions about productivity and population does not overturn CBO's basic finding. To the contrary, that analysis shows that the chances are low that the nation could grow out of its long-term budgetary problems with favorable developments in productivity or demographics (see Box 3).

CBO's simulations show the economy responding smoothly to the rapidly rising debt. In actuality, however, those adjustments would probably be much more disorderly. Foreign investors might suddenly stop investing in U.S. securities, causing the exchange value of the dollar to plunge, interest rates to shoot up, and the economy to tumble into a severe recession. (Those developments have occurred in some countries with rapidly growing government debt.) Higher levels of debt might also ignite fears of inflation in the nation's financial markets, which would push up interest rates even further. Amid the anticipation of declining profits and rising rates, the stock market might collapse, and consumers--fearing economic catastrophe--might suddenly reduce their spending.⁽⁴⁾ Moreover, severe economic problems in this country could spill over to the rest of the world and might seriously affect the economies of U.S. trading partners, undermining international trade.

Yet those disturbing simulations are not predictions of what will inevitably happen. Policymakers would surely take action before the economy was driven into such dire straits. As Herbert Stein, former Chairman of the Council of Economic Advisers, once said, "If something cannot go on forever, it will stop." Nonetheless, the simulations illustrate what might occur if no changes were made in policy--and they demonstrate the importance of controlling the growth of federal debt before it gets out of hand.

A Measure of the Imbalance in U.S. Budget Policy

The underlying budgetary imbalances, though daunting, are not insurmountable. The projections are so severe in part because of the compounding effects of interest: the government would be borrowing to cover the shortfall between revenues and spending--and then borrowing again to pay the interest on that debt. Because escalating interest costs can significantly amplify even a relatively small imbalance between revenues and outlays, the projections do not necessarily imply that resolving the nation's budgetary problems would require huge changes in spending or revenues.

To summarize the magnitude of the budgetary imbalance, CBO used a standard measure for assessing the sustainability of a government's policies.⁽⁵⁾ That measure is based on a hypothetical experiment to determine by how much rates of taxation would have to be permanently raised today to prevent the debt from exceeding its current percentage of GDP for the foreseeable future.⁽⁶⁾ Larger imbalances require higher tax rates. Expressing the imbalance in terms of a tax in
 

crease is not the only way to describe the situation-- measuring it in terms of spending cuts is another--but it defines the problem in a convenient way.

The experiment is hypothetical because it would be impractical to control the growth of the debt with a sudden, major change in tax rates or spending. Moreover, the estimate does not account for the effects that increased marginal tax rates would have on incentives to work and save. Nevertheless, it provides a rough measure of the size of the "hole" in the budget and is similar in spirit to other summary measures of budgetary imbalances. For example, the trustees of the Social Security trust funds routinely estimate by how much payroll taxes would have to be raised to ensure a sufficient balance in the funds in 2070 to meet the following year's projected expenditures. Another approach, generational accounting, calculates the burden on future generations imposed by current budget policy.⁽⁷⁾

Using the sustainability measure, the budgetary imbalances are significant but manageable. Assuming that discretionary spending grew with the economy, CBO estimated that permanently increasing revenues (or reducing noninterest outlays) by 4.1 percent of GDP would keep the debt (as a percentage of GDP) at or below its current level for the foreseeable future (see Table 6). Since revenues are now about 20 percent of GDP, that amount corresponds to about 20 percent of current revenues.

Waiting to resolve the long-term imbalances in the budget would only increase the size of the problem. If policymakers waited five years before taking action on the budget, the size of the tax increase needed to keep the ratio of debt to GDP at or below current levels in the long run would increase by about 15 percent; if action was delayed 20 years, the long-term imbalances would climb by about 60 percent. Those results arise because federal debts mount as action is delayed, which crowds out capital and increases the interest cost of the debt. Moreover, if the problem was resolved through increases in marginal tax rates, delay could also reduce incentives to work and save. Those effects, however, are not included in the estimate.

Revisions in the Long-Term Outlook

The long-term outlook is slightly brighter than it was in May 1996, when CBO reported that the long-term budgetary imbalances were 5.4 percent of GDP; they now amount to just 4.1 percent.

Revisions in the long-term outlook can come from two sources: changes in CBO's 10-year projections of the budget, and changes in the assumptions underlying CBO's long-term simulations after 2007. The 10-year projections are important because they determine the level of spending, revenues, and the deficit in 2007. Other things being equal, the smaller the deficit in 2007, the brighter the long-term outlook.

About two-thirds of the improvement in the long-term outlook since last May stems from the improved outlook for the deficit over the next 10 years. As discussed in CBO's Economic and Budget Outlook: Fiscal Years 1998-2007, the outlook for the deficit has improved because of reductions in the projected growth of Medicare and Medicaid, enactment of new legislation (such as the Personal Responsibility and Work Opportunity Reconciliation Act of 1996, otherwise known as welfare reform), and improvements in the outlook for the economy. The revised outlook for the baseline deficit reduces the long-term imbalance in the budget by 0.8 percent of GDP.
 

Comparison with Other Forecasters

CBO is not alone in raising concerns about the long-term implications of the current set of commitments that the federal government has implicitly made with its budget policies. Several other organizations and academic analysts have voiced similar warnings that U.S. budget policy cannot be sustained indefinitely.⁽⁸⁾

Since 1992, the General Accounting Office has presented results showing that, if left unchecked, the federal budget deficit could grow to over 23 percent of GDP by 2025. GAO's model incorporates some economic interactions between the deficit and the economy, although it holds interest rates constant.

In 1995, the Bipartisan Commission on Entitlement and Tax Reform weighed in with another alarm. The commission saw growing imbalances between spending and revenues in the early decades of the 21st century unless changes were made to federal entitlement programs. Using a model without economic feedbacks, the commission projected budget deficits in excess of 15 percent of GDP by 2030. Its projections assumed that discretionary spending grew with the economy.

In February 1997, the Administration released an updated version of its long-term budget projections, which reflected the improved outlook for the budget over the next 10 years. Although that revision reduces some of the long-term imbalance in the budget, the Administration continues to see significant long-term budgetary problems under current policy. The Administration's calculations now show that, unless policies were changed, the deficit would grow to 4 percent in 2020 and 17 percent in 2050. (In February 1996, the Administration had expected that the deficit would grow to 6 percent by 2020 and 26 percent in 2050.) The Administration's projections are more optimistic than CBO's because it assumes that discretionary spending would grow only with inflation, and it develops its base projections without economic feedbacks. (See Box 4 for a discussion of the Administration's analysis of the effects of the President's proposal on the long-term outlook.)



Sustainable Budget Strategies

To avoid the adverse economic consequences described above, the ratio of debt to GDP must be brought under control. Two possible budget strategies would meet that goal: the first permanently balances the budget by 2002; the second holds the ratio of the deficit to GDP roughly at its current level. Both strategies are sustainable because they prevent the debt from ever growing faster than the economy. Other approaches are possible, but those two examples illustrate some of the implications such strategies have for the budget and the nation's economic outlook.

A budget that was permanently balanced would freeze the level of federal debt and continuously diminish the ratio of debt to GDP (see Table 7).⁽⁹⁾ As the economy grew, the ratio of debt to GDP would slowly decline from 50 percent of GDP in 1996 to 9 percent in 2050. Over that period, the deleterious effects of the debt on interest rates and economic growth would gradually disappear. A balanced budget would also put the United States back on its historical path, with debt declining as a share of GDP during periods of peace and prosperity. However, a ratio of debt to income as low as 9 percent would be unusual in modern history. Indeed, the debt ratio has not been so low since America's entry into World War I.

Even if the budget was not permanently balanced, the worst aspects of the base scenario could be avoided if budget policies were altered so that the deficit did not grow faster than GDP. One way to achieve that goal would be to stabilize the NIPA deficit at its current share of GDP, about 1.7 percent. If the deficit was fixed at that level, the debt would eventually stabilize at about 44 percent of GDP.

Setting goals for the ratio of debt to GDP is not a new idea. The 15-member nations of the European Union have already pledged to reduce their debt-to-income and deficit-to-income ratios. Goals are specified by the Maastricht Treaty, which aims to create a monetary union with a single European currency. With some exceptions, the treaty requires that a nation wishing to join the union must bring its combined debt from all levels of government to 60 percent of GDP or less and its combined deficit to 3 percent of GDP or less.

Implications for the Economy

Compared with the base scenario, the long-term economic outlook would be significantly brighter if policy-makers either balanced the budget permanently or stabilized the deficit at current percentages of GDP. By 2035, gross national product per capita would be 23 percent higher than in the base scenario, and that gap would grow substantially in the years thereafter (see Table 8). Of the two strategies, the balanced budget would provide the greater long-term economic gains, but at the cost of more near-term sacrifice.

The economic benefits of stabilizing the deficit are not as great as those of balancing the budget, but the differences are not large. Stabilization implies that by 2035, real GNP would be about 2 percent less than it would be under the balanced budget. That difference in GNP arises because some capital investment would still be crowded out under a deficit policy.

Implications for the Budget

Permanently balancing the budget or stabilizing the deficit would require significant changes in the government's policies. Those changes could be achieved, but they would involve paring entitlement benefits for elderly people, sharply reducing other spending, or increasing taxes.

Interest Costs. Both budget strategies would significantly reduce the amount required to service the debt compared with the base scenario. However, interest costs would decline more with a balanced budget than with a policy of permanent deficits.

With a balanced budget, the interest on the debt would eventually decline to insignificance as a share of GDP. In CBO's projections, that cost drops from 3.4 percent of GDP in 1996 to 0.5 percent in 2050 (see Table 7). The decline comes from fixing the debt in dollar terms after 2002 and from having interest rates on government debt fall relative to the rate of growth of the economy. By contrast, when the ratio of the deficit to GDP is held to current levels, interest costs do not decline as much. Instead, they stabilize at about 2.3 percent of GDP.

The pattern for interest payments has implications for the rest of the budget--the so-called primary budget. To maintain balance, the primary budget must show a surplus that exactly matches the interest payments on the debt.⁽¹⁰⁾ Thus, as interest payments declined over time, the surplus required in the rest of the budget would also fall. The projections show that the primary surplus required under a balanced budget would be 2.1 percent of GDP in 2005 but would drop to 0.5 percent

Figure 4.
Projections of Revenues if Tax Increases Are Used to Achieve Budget Goals

SOURCE: Congressional Budget Office.
 
NOTE: The balanced budget path assumes that the budget is balanced by 2002 and remains balanced thereafter. The path with the steady ratio of the deficit to gross domestic product assumes that the ratio is stabilized at its current level. The projections of the base scenario use the balanced budget economic assumptions. Receipts are as defined in the national income and product accounts.
 
GDP = gross domestic product.

by 2050. If the deficit was stabilized, the required surplus in the rest of the budget would stabilize at about 0.6 percent of GDP.

Required Policy Changes. Both strategies would require significant changes in the budget. If the budget was balanced (or the ratio of the deficit to GDP stabilized) through tax increases alone, those increases would be small in the early years but would grow considerably as the baby boomers began to retire (see Figure 4). To keep the budget balanced, federal revenues would have to rise from 21 percent of GDP in 1996 to about 27 percent in 2050.⁽¹¹⁾ Stabilizing the deficit would require smaller tax increases at first than would balancing the budget, but the additional interest costs would eventually require slightly larger increases. (That scenario does not describe an immediate tax increase such as the one mentioned earlier, but rather a gradual increase that is sufficient to keep the budget balanced.)

That result may seem surprising at first because it appears to be at odds with the common perception that deficit spending is an "easier" policy than a balanced budget. That view is certainly correct for the short run, when differences in fiscal policy have little effect on federal interest costs. But over periods as long as 30 years, a deficit policy eventually carries higher interest costs than a balanced budget policy--and those additional costs ultimately have to be financed. Although deficit spending expands current consumption above what would otherwise have been possible, that additional consumption is achieved only by sacrificing some future consumption.

Substantial reductions in current commitments for spending would also be required if budgetary actions focused solely on the spending side of the ledger (see Figure 5). Projections using the base scenario with balanced budget economic assumptions show noninterest

Figure 5.
Projections of Noninterest Outlays if Spending Cuts Are Used to Achieve Budget Goals

SOURCE: Congressional Budget Office.
 
NOTE: The balanced budget path assumes that the budget is balanced by 2002 and remains balanced thereafter. The path with the steady ratio of the deficit to gross domestic product assumes that the ratio is stabilized at its current level. The projections of the base scenario use the balanced budget economic assumptions. Noninterest outlays are as defined in the national income and product accounts.
 
GDP = gross domestic product.

outlays increasing from 19 percent in 1996 to 26 percent in 2050.⁽¹²⁾ To keep the budget balanced, nonin-terest spending would have to be cut sharply at first, and it would decline to 17 percent of GDP by 2002. But as interest costs fell, spending under a balanced budget could rise to slightly less than 20 percent of GDP in 2050. By contrast, to keep the ratio of the deficit to GDP steady, noninterest spending would have to be held at about 19 percent of GDP throughout the projection period.

Neither strategy could be carried out by focusing solely on cutting discretionary spending. Under either plan, the required changes in the budget would exceed total outlays for the discretionary accounts around 2030. The long-term budgetary situation cannot be stabilized solely by limiting the growth of this category of spending. Stability would require reductions in the growth of other spending categories or increases in taxes.



Balancing the Budget by 2002

The discussion so far has examined the implications of setting overall deficit targets for the budget and the economy. In developing a budget, however, the Congress must move beyond setting goals to making changes in specific laws. During the past year, both the Congress and the President advanced plans to balance the budget by 2002. Those proposals raise a number of issues. Would balancing the budget by 2002 by itself solve the long-term budgetary problem? Or would additional policy changes be needed? Although it is impossible to project the precise long-term impacts of specific legislative initiatives, CBO's long-range model can provide a rough assessment of how changes in policy might affect the budget over the next several decades.

To address those issues, CBO had to make specific assumptions about the path of cuts in the deficit and the distribution of those cuts among the various budget categories. Those assumptions affect the long-term outlook: other things being equal, the sooner the deficit is cut and the more that the cuts are focussed on fast-growing programs, the brighter the long-run outlook.

CBO's assumed path is broadly consistent with the plans advanced by the President and the Congress during the 104th Congress, although the allocation of the cuts in the deficit is somewhat different. In CBO's path, the cuts to the deficit roughly follow what CBO used in its estimate of the fiscal dividend in Chapter 4 of the Economic and Budget Outlook: Fiscal Years 1998-2007. (Balancing the budget will produce favorable changes in the economy, which can increase revenues and reduce spending; those resulting changes are the so-called fiscal dividend.) CBO assumed that the deficit would be reduced $17 billion in fiscal year 1998, a sum that would climb sharply in fiscal year 1999 and fiscal year 2000 and climb more slowly thereafter to reach $188 in fiscal year 2002, bringing the budget into balance in that year. In the long-term simulations, CBO assumed that the budget would remain in balance from 2003 to 2007; after 2007, spending and revenues were assumed to grow at the same rate as they do in the base scenario. The simulation thus addresses the question of whether balancing the budget in the near term-- but not dealing with the long-term pressures on the budget--will solve the nation's long-run problems.

In calculating the fiscal dividend, CBO did not have to make any specific assumptions about the mix of policies that would be used to balance the budget. All that was needed for that calculation was the total amount of deficit reduction. But to examine the effects of balancing the budget on the long-term outlook, CBO had to make assumptions about how deficit reduction would affect the levels of revenues and spending for particular programs from 1997 to 2007. In CBO's assumed path, reductions in the growth of Medicaid spending account for 5 percent of the deficit cuts; the rest is evenly divided between discretionary spending and Medicare.

The simulations show that balancing the budget by 2002 would substantially reduce the long-term budgetary imbalances in the United States, but it would not be enough to eliminate them (see Table 9). Although the budget would remain close to balance for another 10 years or so, the demands of the retired baby boomers on the Social Security, Medicare, and Medicaid programs during the 2020s would significantly increase annual budget deficits. By 2035, federal debt would climb to 91 percent of GDP and would grow rapidly thereafter. By 2055, it would exceed levels that the economy could reasonably support.

That situation obviously would be much better than what the base scenario depicts. CBO estimates that the long-term imbalances in the budget would be reduced from 4.1 percent of GDP to 2.3 percent. Thus, balancing the budget by 2002 with cuts to the level--but not the long-run growth rate--of spending would resolve about 45 percent of the long-term problem, given the package of cuts that CBO assumed.

Other packages would produce different estimates. For example, if the budget was balanced solely through reductions in discretionary spending, it would eliminate only about one-third of the long-term problem. That result illustrates one of CBO's fundamental conclusions: deficit-reduction packages that focus on fast-growing programs (like Medicare, Medicaid, or Social Security) are much more effective in resolving the long-term imbalances than those that do not.

Regardless of how the budget is balanced in the near term, additional budgetary action--such as cutting back on entitlements for the elderly or raising taxes-- would still be needed to put the budget on a sustainable path.

1. All numbers are taken from Board of Trustees, Federal Old-Age and Survivors and Disability Insurance Trust Funds, 1996 Annual Report (June 5, 1996).

2. General Accounting Office, Budget Policy: Prompt Action Necessary to Avert Long-Term Damage to the Economy, GAO/OCG-92-2 (June 1992), and The Deficit and the Economy: An Update of Long-Term Simulations, GAO/AIMD/OCE-95-119 (April 1995); "Analytical Perspectives," Budget of the United States Government, Fiscal Year 1998 (February 1997), pp. 23-30.

3. The trustees of the Old-Age and Survivors and Disability Insurance Trust Funds project a similar slowing in the growth of hours.

4. Some people might dramatically increase their saving in the face of economic collapse. In the extreme, if consumers offset all of the increase in the deficit with higher levels of private saving and invested their savings in the United States, the deficit would have no effect on GDP. But assuming that consumers would behave that way is unrealistic and risky. It is doubtful that such forward-looking people would invest in the United States, given the risk of a stock market collapse or an increase in inflation. Nonetheless, any prediction about saving under those extreme conditions is highly uncertain.

5. Olivier Blanchard, Jean-Claude Chouraqui, Robert P. Hagemann, and Nicola Sartor, "The Sustainability of Fiscal Policy: New Answers to an Old Question," OECD Economic Studies, no. 15 (Autumn 1990).

6. The additional revenues initially result in large budget surpluses, which reduce the level of the debt. As the baby boomers retire, the budget moves back into deficit, and debt climbs. The tax increase is sufficient to keep the level of debt at or below 50 percent of GDP from 1997 through 2070.

7. Congressional Budget Office, Who Pays and When? An Assessment of Generational Accounting (November 1995).

8. General Accounting Office, Budget Policy and The Deficit and the Economy; Bipartisan Commission on Entitlement and Tax Reform, Final Report to the President (January 1995); Budget of the United States Government.

9. Although the model technically assumes that the budget is balanced each year, similar results would be seen if the government allowed the budget to move into deficit during recessions--provided that the budget moved into surplus during expansions and was balanced on average.

10. Another way to think about the primary budget is that it shows all revenues and all spending for "programs" but not for interest on the debt. A primary surplus then means the amount of revenues in excess of outlays for programs.

11. Those estimates probably understate the actual size of the tax increase that would be needed because they do not account for the adverse impact that increasing marginal income (or payroll) tax rates would have on incentives to work and save.

12. Balanced budget economic assumptions are used here because they implicitly incorporate the fiscal dividend.

---