Skip Navigation

Mr. Chairman, Ranking Member Spratt, and Members of the Committee, I appreciate the opportunity to appear before you to discuss the Social Security program. The Social Security Act of 1935, enacted in the midst of the Depression, is widely seen as one of the most important legislative accomplishments in U.S. history. Since its inception, Social Security has grown to become by far the largest federal program. Over the next 30 years, the aging of the baby-boom generation will pose new challenges for Social Security, the federal government, and the U.S. economy.

The Congressional Budget Office (CBO) has examined those challenges in a number of recent reports. Late last year, we published Social Security: A Primer and Uncertainty in Social Security's Long-Term Finances: A Stochastic Analysis. Last week, we released a policy brief on the long-range picture of the federal government's share of the economy. My testimony today summarizes some of the findings of those reports; it will make the following major points.

• Once the baby-boom generation retires, the portion of the nation's output that the federal government will spend on Social Security is expected to rise by 50 percent--from about 4 percent of gross domestic product (GDP) today to an estimated 6 percent in 2030.
  
  
• Addressing the growing cost of Social Security would not by itself eliminate the economic and budgetary pressure caused by the aging of the U.S. population. The rapidly escalating costs of the government's health care programs are a major source of that pressure. CBO projects that federal spending for Social Security, Medicare, and Medicaid combined will account for about 14 percent of GDP by 2030--nearly double the current share.
  
  
• Looking farther ahead, CBO projects that government spending (excluding interest on the federal debt) will rise from about 18 percent of GDP today to 28 percent in 2075 under current policies. If revenues remain within their historical range relative to GDP, the total cost of government (including interest) could double as a share of the economy: from about 19 percent of GDP today to about 40 percent in 2075. However, modest reductions in the growth of spending for federal programs could significantly slow the growth of interest costs and total outlays.
  
  
• Projections of Social Security's finances are highly uncertain, but the range of uncertainty is not adequately reflected in the low-, medium-, and high-cost scenarios used by the Social Security trustees.
  
  
• Uncertainty about long-term budgetary outcomes derives from many factors, including rates of mortality, fertility, inflation, and real wage growth. Social Security's finances are most influenced by variables, such as mortality rates, that move independently of economic growth, because outlays are affected more than receipts when people live longer. That aspect of the system's finances is important to keep in mind when designing proposals to reform Social Security. For instance, a change in the retirement age that was linked to life expectancy could automatically resolve the budgetary consequences of any future changes in mortality rates.
  
  
• The expected increase in Medicare spending highlights the divergence between economic growth and budgetary pressures even more clearly, because most of that increase results from health care costs per beneficiary growing faster than GDP per capita. No mechanism in law exists that will slow down that growth.
  
  
• Although policymakers have many goals, if they want to limit the growth of federal spending as a share of GDP, they have only two options: slow the growth of that spending or increase the growth of the economy. The nation's ability to sustain an aging population will ultimately depend on how many goods and services the economy produces and how they will be distributed, not on how much money is credited to Social Security's trust funds.

THE PRESSURES OF AN AGING POPULATION

Over the next three decades, the aging of the baby boomers (the large group born between 1946 and 1964) will put new pressure on Social Security, the federal government, and the U.S. economy. The Social Security Administration projects that the number of people age 65 or older will rise by more than 90 percent during that period (from about 36 million now to 69 million in 2030), according to its intermediate assumptions (see Figure 1). At the same time, the number of adults under age 65--who will largely be the ones paying the payroll taxes to support their elders--will grow by only about 14 percent (from 172 million to 196 million). Moreover, even after all of the baby boomers have retired, the number of elderly people is expected to keep rising at a faster rate than the number of nonelderly people as life spans continue to lengthen.
 

Figure 1. Projected Growth in the Adult Population, 2001-2075

SOURCE: Social Security Administration, The 2002 Annual Report of the Board of Trustees of the Federal Old-Age and Survivors Insurance and Disability Insurance Trust Funds (March 26, 2002), Table V.A2 (intermediate assumptions).

Perhaps even more important, as the population ages, spending on Medicare and Medicaid is likely to rise rapidly because of increases in federal costs per beneficiary as well as in the percentage of the population eligible for benefits (unless major changes are made to those programs). Medicare provides health insurance to most U.S. residents age 65 or older and to eligible disabled people, most of whom also receive Social Security benefits. Medicaid is a joint federal/state program that provides medical assistance to low-income people; in recent years, a large share of its payments have gone to provide long-term care, mainly for elderly or disabled people.
 

A LONG-RANGE PICTURE OF THE FISCAL SITUATION

How will those pressures of demographics and health care costs affect the U.S. budget and economy? To help address that question, CBO has developed a new long-range model. Using the model, we recently prepared a 125-year picture of the budget that extends from 1950 to 2075. Those projections illustrate a potential path for the budget that highlights the implications of maintaining current policies. Of course, the future path of the budget is highly uncertain and subject to wide variation. Thus, the path shown in those projections is simply a representation based on an illustrative set of key assumptions.

Although my testimony focuses on long-range projections of spending under current policies, CBO is about to unveil an expanded version of the model that will be capable of simulating the budgetary and economic effects of policy changes--including detailed proposals for Social Security, such as the introduction of private accounts. One of the more innovative features of the expanded model is its ability to perform stochastic simulation analysis, which shows the probabilities of alternative outcomes based on a statistical distribution of alternative assumptions about such factors as returns on stocks and bonds, mortality, fertility, and wage growth. The model will also include equations that reflect how people alter their work and saving in response to increases in taxes and cuts in benefits. We expect to begin releasing analyses from that model sometime this fall.

In CBO's current long-range model, Social Security spending reflects growth in the number of recipients and in wages, which determine benefits. Medicare and Medicaid spending reflects the increasing number of recipients and the age profile of enrollees as well as the rising costs of medical care. For the long-range projections described below, the growth in health care costs for each recipient of a given age is assumed to slow to a rate 1 percentage point faster than the growth rate of per capita GDP. Although seemingly high, that rate is lower than it has been in recent decades. The budget figures in these projections are expressed as a share of GDP so that the magnitude of federal revenues and spending can be observed in relation to the country's total economic activity in any given year and over time.⁽¹⁾

The projections show that it is spending for the major entitlement programs and for interest--because of the commitments involved and their sheer magnitude--that has the largest potential to constrain future Congresses. Moreover, much of the government's remaining spending consists of discretionary outlays, the levels for which are determined annually. Given the wide array of discretionary programs, that category of spending (unlike the major entitlement programs) does not easily lend itself to projections that merge economic and demographic assumptions with legislative rules for the payment of benefits. Thus, CBO's long-range projections assume that defense, nondefense discretionary, and all other spending (that is, other than for Social Security, Medicare, Medicaid, and interest) will remain fixed as a share of GDP beginning in 2012, the last year of the 10-year baseline projections that CBO published in March 2002. The projections do not incorporate the recently enacted farm bill and economic stimulus package.

Historical Trends in Spending

Spending by the federal government grew from approximately 3 percent of GDP in 1925 to about 16 percent in 1950. (Following the Depression, World War II abruptly boosted federal spending to about 42 percent of GDP, but afterward, that spending dropped and resumed a less volatile growth trend.) Since then, Social Security, Medicare, and Medicaid have together become the largest component of the federal budget (see Figure 2). In 1962, when Social Security outlays represented only 2.5 percent of GDP, and Medicare and Medicaid had not yet been created, spending for all other government activities made up about 85 percent of federal noninterest outlays. The largest share was for national defense, which accounted for half of noninterest outlays and represented 9.2 percent of GDP. By 2001, total spending for Social Security, Medicare, and Medicaid equaled 7.8 percent of GDP, about triple the 1962 share for Social Security alone. Although still constituting less than half of all federal spending, the three programs combined accounted for the largest share of total outlays. Defense spending had fallen to 3 percent of GDP, and all other noninterest spending stood at 6.3 percent. Interest costs, whose share of GDP had risen steadily from 1.2 percent in 1962 to a high of 3.3 percent in 1991, stood at 2.0 percent in 2001.
 

Figure 2. Federal Outlays, 1962-2001

SOURCE: Congressional Budget Office.

Projections to 2075

Looking ahead, CBO projects that outlays for Social Security, Medicare, and Medicaid (based on the current rules for benefits) could nearly double as a share of GDP by 2030, rising to about 14 percent. If spending for all other government activities in 2030 remained at roughly the same share of GDP as projected for 2012 (about 7 percent), Social Security, Medicare, and Medicaid would account for almost 70 percent of the government's noninterest spending. By 2050, outlays for the three programs would constitute nearly 17 percent of GDP, and by 2075, about 21 percent--exceeding the share of GDP now absorbed by all federal revenues (see Figure 3).
 

Figure 3. Federal Outlays by Category, 1950-2075

SOURCE: Congressional Budget Office.

Under the assumptions that CBO made for its long-range picture of government finances, the projected rise in spending for Social Security, Medicare, and Medicaid would drive total federal outlays well above the level seen throughout much of the post-World War II period. The government's core costs (that is, ignoring net interest on the debt) could rise from about 18 percent of GDP today to about 24 percent in 2050 and 28 percent in 2075. Left unattended, that steady escalation in spending could cause major deficits to emerge, pushing the government's debt, and its interest spending on that debt, to unprecedented levels. If revenues remain within their historical range relative to GDP, the total cost of government (including interest) could double as a share of the economy--from about 19 percent of GDP today to about 40 percent in 2075 (see Figure 4).⁽²⁾
 

Figure 4. Federal Revenues and Outlays, 1950-2075

SOURCE: Congressional Budget Office.

ISSUES TO CONSIDER IN REFORMING SOCIAL SECURITY

Several aspects of Social Security and the outlook for it as the population ages are especially important in considering changes to the program. First, throughout its long history, Social Security has had multiple goals--some related to redistributing income, others related to offsetting lost earnings. In 2000, only about two-thirds of Social Security's beneficiaries were retired workers; the rest were disabled workers, survivors of deceased workers, and workers' spouses and minor children (see Figure 5). Policymakers will need to decide whether the program's goals are still appropriate and, if so, how changes to Social Security would aid or hinder the achievement of those goals and affect various types of beneficiaries and taxpayers. Those decisions will also need to take into account the dramatic increase in the elderly population that is expected in the coming decades.

Figure 5. Distribution of Social Security Beneficiaries, by Type of Benefit Received, December 2000

SOURCE: Social Security Administration, Annual Statistical Supplement, 2001, Table 5.A1 (available at www.ssa.gov/ statistics/Supplement/2001/5a.pdf).

Second, issues about how to prepare for an aging population ultimately concern the amount of goods and services that the economy will produce and how they will be distributed, not how much money is credited to the Social Security trust funds. In that sense, the projected depletion of those funds--which is the focus of much of the popular debate about Social Security's future--is irrelevant. The challenges of adjusting to an aging population would need to be faced even if the trust funds never existed.

Third, deciding how to prepare for an aging population is likely to require weighing the interests of today's workers and Social Security beneficiaries against the interests of future workers and beneficiaries. No matter how it is packaged, any plan to increase national saving today means that the U.S. population will consume fewer goods and services now so that consumption can be greater in the future, when a larger share of the population is retired. Gone are the days when expansion of the labor force could pay for the growth of Social Security benefits. (In past decades, Social Security's payroll tax revenues grew substantially as the baby-boom generation and women of various ages entered the labor force in large numbers.) As the Congress looks at policy changes, one consideration is that future workers and Social Security beneficiaries are likely to have higher standards of living, on average, than current workers and beneficiaries do, because of future increases in productivity.
 

STRATEGIES FOR DEALING WITH AN AGING POPULATION

Spending more on elderly people may be appropriate in light of their increasing numbers, but questions can be raised about the extent to which that spending should rise. Policymakers have many goals, but if they want to limit the growth of spending on the elderly as a share of the economy, they can do so in only two ways: either by slowing the growth of that spending or by increasing the growth rate of the economy. Different options for reform would have different effects on economic growth. To the extent that those options boosted the future size of the economy and increased the nation's accumulation of assets, they could lessen the burden on future workers from government programs that serve the elderly.

My testimony focuses on three ways to prepare for an aging population that have generated a lot of public attention: paying down federal debt, creating private retirement accounts, and making changes to the benefits or revenues of the current Social Security program. Those approaches are not mutually exclusive; they could be combined in any number of ways. (In addition, many people have put forward proposals to curb the rising costs of federal health care programs. Such proposals could also help the nation deal with its impending demographic changes, but they are beyond the scope of this testimony.)

Regardless of which approach policymakers decide to take, a number of key questions should be raised about any proposed policy option:

• How would it affect economic growth over the long run?
  
  
• Would the proposed policy improve the long-term fiscal outlook faced by succeeding generations? How would it alter the taxes they pay and the benefits they receive?
  
  
• Would the policy improve the ability of Social Security and Medicare to respond to unanticipated changes in demographics (such as life expectancy) and in the economy (such as productivity growth)? Some proposals could help make those programs more adaptable to change; other proposals could reduce their flexibility.

Pay Down Debt

One strategy for preparing for the needs of an aging population is to pay down federal debt. If the government spends less than it receives in revenues (and private saving does not fall too much in response), national saving will rise, boosting the stock of private capital and expanding the productive capacity of the economy in the long run. Indeed, federal debt held by the public has fallen sharply in recent years--from about 50 percent of GDP in 1995 to about 33 percent today. That decline has freed up funds for investment in private capital.

CBO will soon update its 10-year projections for the budget, but it does not expect any significant surpluses to be available for paying down debt for at least a few years. However, if current tax and spending policies are maintained, significant budget surpluses could reemerge at some point in the next 10 years. But even paying off all of the federal debt available for redemption would not provide enough interest savings or additional economic growth to finance Social Security, Medicare, and Medicaid spending over the long run.

Create Private Accounts

A second strategy is to encourage private saving. A prominent set of proposals envisions creating private retirement accounts. Those proposals differ in many ways, but they share a common feature: the income from an account would depend on the payments made into it and the rate of return on the account's assets. Many types of accounts are possible, and their effects would vary widely.

One of the central issues is how private retirement accounts would be financed. Many proposals include a contribution from the government to help people pay for accounts. According to supporters of private accounts, diverting payroll tax revenues from the government could prevent policymakers from spending those revenues on other programs and could thus provide many of the same economic benefits as paying down debt. In addition, they argue, private accounts could allow the government to encourage asset accumulation while avoiding the problems of having the government own shares in private companies. However, because national saving consists of both private and government saving, a proposal that simply moved dollars from government saving to private saving (by financing private accounts through an increase in federal debt) would have no direct effect on national saving or capital accumulation. To raise national saving, a proposal would have to cut either government consumption, private consumption, or both.

Some people argue that private accounts would offer higher rates of return than the traditional Social Security system does, but that argument can be misleading. Social Security has a low rate of return largely because initial generations received benefits far greater than the payroll taxes they paid. That difference would have to be made up even if the Social Security system was entirely replaced by private accounts. Moreover, investing in the stock market--either through private accounts or through government purchases of stock for the Social Security trust funds--would be no panacea. Simply raising the average rate of return on assets by taking on more risk would not change the economic fundamentals. Only if the investment proposal increased national saving and enlarged the economy would it reduce future burdens.

In setting up a system of private accounts, policymakers would have to address many practical issues. How much would the system cost to administer? Would it provide insurance against downturns in the stock market? Would the system require that accounts be converted into annuities and, if so, under what conditions? How would it handle benefits for workers' families, for survivors of deceased workers, and for disabled workers? Would the system give subsidies to people with low income and intermittent work histories? How would the system be regulated and investors informed?

The answers to those questions could have implications for the economy. For example, government guarantees that people would receive a minimum level of retirement income in the event of a market downturn would probably reduce national saving below what it would be without those guarantees. And subsidies to low-income workers that were phased out as wages rose could impose implicit taxes on work and could discourage some people from working more.

Make Programmatic Changes

A third approach is to modify the current Social Security program. Changes that have been proposed include reducing benefits (for example, by raising the retirement age, calculating initial benefits using a price index rather than a wage index, or reducing annual cost-of-living adjustments) or increasing payroll taxes. The effect on the economy would depend on the particular kind of change.

Many types of benefit reductions could increase the size of the economy in the long run because they could encourage some people to save more. However, those long-term gains could take a couple of decades to materialize fully, and the effect in the near term would be uncertain. Slowing the growth of Social Security benefits could reduce the lifetime resources of some transitional generations, but it could also lead to higher wages and lower tax burdens for later generations. If benefits were to be cut, changing the law now rather than later would give workers time to adjust their plans for saving and retirement.

Raising taxes to pay for future Social Security benefits would have an uncertain effect on the size of the economy in the long run. Moreover, the effect would depend on the type of tax increase and other factors. If the revenues from a tax increase did not change the government's decisions about other spending or taxes, national saving could rise. But the extra revenues could encourage more government spending, which would limit any rise in national saving. In addition, increases in marginal tax rates on payroll or income could reduce people's incentives to work or save, also dampening any increase in national saving.

Although long-term projections of the federal budget and the economy carry huge uncertainties, one fact seems certain: the U.S. population will age significantly over the next 30 years, and unless policies are changed, spending on the elderly will rise sharply, posing new challenges for the federal government and the nation's economy.

2. For analytical purposes, these projections assume that total federal revenues will level out at 19 percent of GDP, toward the higher end of their range during the post-World War II period (18 percent was the average from 1950 through 2001).