Congressional Budget OfficeSkip Navigation
Home Red Bullet Publications Red Bullet Cost Estimates Red Bullet About CBO Red Bullet Press Red Bullet Employment Red Bullet Contact Us Red Bullet Director's Blog Red Bullet   RSS
The Long-Term Budget Outlook
December 2003
PDF



CHAPTER
2
The Long-Term Outlook
for Social Security

Social Security is by far the federal government's largest income-redistribution program. The Old-Age and Survivors Insurance part of the program pays benefits to retired workers and their dependents and survivors. The Disability Insurance (DI) part makes payments to disabled workers who are younger than the normal retirement age and to their dependents. In all, about 46 million people now receive Social Security benefits.(1)

Driven largely by repeated expansions of the program during its first 40 years, spending for Social Security benefits steadily increased relative to the size of the economy, reaching 4 percent of gross domestic product in 1975 (see Figure 2-1). Since then, that spending has generally fluctuated between 4.0 and 4.5 percent of GDP. In 2003, it accounted for an estimated 4.2 percent of GDP.
 
Figure 2-1.
Spending for Social Security

(Percentage of GDP)
Source: Congressional Budget Office.

 

The Outlook for Social Security Spending

The cost of the Social Security program will rise significantly in coming decades--a change that has long been foreseen. Average benefits typically grow when the economy does (because the earnings on which those benefits are based increase). However, in the future, the total amount of Social Security benefits paid will grow faster than the economy because of changes in demographic structure. As the baby-boom generation reaches retirement age and as decreasing mortality leads to longer lives and longer retirements, a larger share of the population will draw Social Security benefits.(2) Moreover, the number of people age 65 or older will double during the next 30 years, while the number of adults under age 65 will grow by less than 15 percent--meaning that in three decades, the older population will be more than one-third the size of the younger group, compared with one-fifth today (see Figure 2-2). Consequently, the Congressional Budget Office estimates that unless changes are made to Social Security, spending for the program will rise to 4.9 percent of GDP in 2020, 5.9 percent in 2030, and 6.2 percent in 2050.
 
Figure 2-2.
The Population Age 65 or Older as a Percentage of the Population Ages 20 to 64

(Percent)
Source: Congressional Budget Office.

Discussions of Social Security frequently address the status of the program's trust funds. However, this chapter considers total Social Security outlays, which represent a claim on government resources, whether from trust funds or other sources. (Revenues, the means of providing such resources, are examined in Chapter 5.)
 

How Social Security Functions

In general, workers are eligible for retirement benefits if they are age 62 or older and have paid sufficient Social Security taxes for at least 10 years. Workers whose employment has been limited because of a physical or mental disability can become eligible for DI benefits at an earlier age with a shorter employment history.

When retired or disabled workers first claim Social Security benefits, they receive payments based on their average level of earnings over their working lifetime; those payments are subsequently adjusted to reflect annual changes in the cost of living. The formula used to translate average earnings into benefits is progressive--in other words, it replaces a larger share of preretirement earnings for people with lower average earnings than it does for people with higher earnings. Both the earnings history and the specific dollar amounts included in the formula are indexed for changes in average annual earnings for the labor force as a whole. That indexation causes initial benefits for future recipients to grow in real terms (beyond the level of inflation).(3)

For retirement benefits, a final adjustment is made based on the age at which the recipient chooses to start claiming benefits--the longer a person waits (up to age 70), the higher the benefit level. That final adjustment is intended to be "actuarially fair," so that an individual's total lifetime benefits will be equally valuable regardless of when he or she begins collecting them.(4)

For workers born before 1938, the age of eligibility for full retirement benefits--referred to as Social Security's "normal retirement age"--was 65. Under current law, that age is gradually increasing and will be 67 for people born in 1960 or later.(5) Workers will still be able to choose to begin receiving reduced benefits as early as age 62.

People who retired at age 65 in 2003 having had average earnings throughout their career were eligible for an annual benefit of about $13,800. That amount replaced roughly 40 percent of their previous annual earnings. Under current law, the replacement rate will be smaller for workers with average earnings who retire at age 65 in the future, mainly on account of the scheduled increase in the normal retirement age. Nevertheless, because initial benefits are indexed to average wages, which grow over time, the real value of those benefits will continue to rise.
 

Options for Slowing the Growth of Social Security Spending

Three broad approaches for slowing the rise in Social Security benefits have received considerable attention. First, policymakers could reduce the size of the initial payments that new Social Security beneficiaries are scheduled to receive. Second, they could increase the age at which workers become eligible for full retirement benefits. Third, policymakers could reduce the annual cost-of-living adjustments that beneficiaries receive once they become eligible for benefits. Because more than 99 percent of Social Security outlays are benefit payments, however, any attempt to reduce spending must center on their growth.

Reform proposals that incorporate individual accounts are not addressed in this report. Because those packages encompass a broad range of proposed changes and vary in scope (accounts of different sizes, voluntary versus mandatory participation, and direct or indirect offsets to Social Security benefits), their potential budgetary effects vary widely and no simple generic option can adequately characterize them.(6)

If policymakers decide to slow the growth of Social Security benefits, considerations of both fairness and economic efficiency point toward enacting new legislation long before the changes take effect. People often consider the size of their expected Social Security benefits when they decide how much to save and how long to work. Altering that size gradually would give people more time to plan for and adjust to the changes.

The rest of this chapter looks at specific options that represent ways to implement the three broad approaches described above. The estimates of savings included with the options are intended to indicate the relative magnitudes of alternative changes. Specific estimates of savings would depend on the details of individual proposals.

Constrain the Increase in Initial Benefits

The most straightforward way to reduce the growth of Social Security spending would be to slow the rate at which initial benefits rise from one cohort of recipients to the next. Each new group of eligible beneficiaries would then receive lower benefits than scheduled under current law. However, that approach would not alter the benefits of people who were already on the rolls before the change took effect.

One method of doing that, which has received considerable attention, would be to change the way initial benefits were calculated so that they grew with prices instead of wages. The benefits awarded to each succeeding cohort would still rise in nominal terms, but only by enough to keep up with inflation. Whether that option reduced benefits would depend on how benefits were measured:(7)

  • In real terms, annual benefits would be unchanged. Future retirees would have the same purchasing power as retirees today.

  • Measured over a lifetime, total benefits would still increase because longevity is expected to keep growing.

  • Compared with wages, annual benefits would fall (in other words, the replacement rate would decline).

  • Compared with the amounts that future beneficiaries are scheduled to receive under current law, benefits would fall.

The decline relative to currently scheduled benefits would increase for each future cohort of retirees. Under the specific option modeled here, initial benefits would grow with prices for people turning 62 in 2011 or later. If real wages grew at an average rate of 1.3 percent per year, as this option assumes, the projected impact on future benefits would be large. For example, workers who became eligible for benefits in 2030 would receive 23 percent less under this option than they would under the current rules. Workers who became eligible in 2050 would receive about 41 percent less.

Adopting this option would reduce outlays for Social Security in 2050 by about 34 percent from the level projected under current law. As a result, those outlays would equal 4.1 percent of GDP instead of 6.2 percent (see Figure 2-3). Thereafter, Social Security spending would continue to decline as a share of GDP.
 
Figure 2-3.
Federal Spending Under Illustrative Options for Slowing the Growth of Social Security

(Percentage of GDP)
Source: Congressional Budget Office.
Notes: See the text of this chapter for descriptions of the various options.
COLAs = cost-of-living adjustments.

For simplicity, this illustrative option would result in the same percentage change in benefit levels for all beneficiaries in a cohort. However, a comprehensive policy proposal could include other adjustments that would protect lower-income beneficiaries from the proposed changes--for example, by setting a minimum benefit level or by making the existing benefit formula more progressive.

Raise the Retirement Age

Since benefit levels are designed to be actuarially fair regardless of the age at which someone begins receiving benefits, changing the early-retirement age from 62 would have relatively little effect on total Social Security spending, although it might induce people to work longer and therefore pay more payroll taxes. In contrast, raising the normal retirement age would result in lower benefits for all retirees, no matter when they chose to start receiving them, because the number of years over which workers earned benefits would fall. For retirees, increasing the normal retirement age would be equivalent to reducing annual benefit levels.

Some Members of Congress and others have recommended accelerating the current shift to a normal retirement age of 67 and raising that age further thereafter. Proponents of such a change point out that current 65-year-olds are projected to live much longer than was the case in the early days of the Social Security system and that life expectancy will almost certainly continue to grow.

Debate about the level of Social Security benefits tends to focus on how much people will receive each month rather than on how much they will receive over their lifetime. But because of increasing longevity, a commitment to give retired workers a certain amount of monthly benefits at age 62 in, say, 2030 is likely to be more expensive over the recipients' lifetime than that same commitment made to retirees today. The swell of the baby-boom generation will cause most of the growth in the number of Social Security beneficiaries over the next 30 years. But in the longer term, the growth in beneficiaries--and in costs--will be driven by increasing longevity. Linking the normal retirement age to future increases in life expectancy is one way of dealing with that source of cost growth.

The specific option considered here (which is illustrated in Table 2-1) would speed up the transition to a normal retirement age of 67 and then raise that age further to keep pace with assumed future increases in life expectancy. For workers born in 1949, the normal retirement age would be 67. Thereafter, the retirement age would increase by two months per year until it reached 70 for people born in 1967. After that, it would rise by one month every other year. As under current law, workers would still be able to receive reduced benefits starting at age 62, but the amounts of the reductions would be larger.
             
Table 2-1.
The Increase in the Normal Retirement Age Under Current Law and an Illustrative Option

  Percentage Reduction
in Social Security Benefits
for Early Retirement

Year of Birth Year of
Turning
Age 65
Social Security's
Normal
Retirement Age
Retirement
at Age 62
Retirement
at Age 65

Under Current Law
1943 2008 66 25   7  
1960 2025 67 30   13  
Under Illustrative Optiona
1943 2008 66 25   7  
1949 2014 67 30   13  
1955 2020 68 35   20  
1961 2026 69 40   25  
1967 2032 70 45   30  
1991 2056 71 50   35  

Source: Congressional Budget Office based on information from the Social Security Administration's Office of the Actuary.
a. Under this option, the normal retirement age would reach 67 for workers born in 1949. The retirement age would increase by two months for each year thereafter until it reached 70 for people born in 1967, and then it would increase by one month for every two years.

This option would produce substantial savings in relation to projected spending levels under current law: by 2050, the savings would be about 19 percent. Outlays would be 5.0 percent of GDP instead of 6.2 percent in that year and thereafter would continue to decline slightly as a share of GDP.

This specific option would not affect the scheduled benefits of workers who qualified for Disability Insurance. Thus, as DI benefits became relatively more attractive, older workers nearing retirement would be more likely to apply for them. To avoid strengthening that incentive, policymakers could make similar adjustments to scheduled DI benefits--for example, by setting the benefits for workers who qualified for DI at the level those workers would have received if they had retired at a specific age, such as 65 or 67. (Under current law, their benefits equal the amount they would have received when retiring at the normal retirement age.)

Current projections of Social Security outlays are sensitive to projections of life expectancy. If future beneficiaries live longer than expected, government outlays will be higher than anticipated. Under a system in which the normal retirement age varied with life expectancy, beneficiaries would either have to work longer and start receiving benefits later or have to accept lower annual payments. In either case, their total lifetime benefits would no longer grow as a result of increases in longevity.

Reduce Cost-of-Living Adjustments

Each year, the Social Security Administration makes a cost-of-living adjustment (COLA) to monthly benefits, raising them by the percentage increase in the consumer price index (CPI). The CPI is not intended to be a "true" cost-of-living index, and many economists believe that it grows faster than such an index would. However, they disagree about the extent to which the CPI overstates inflation. In 1996, the Advisory Commission to Study the Consumer Price Index (known as the Boskin Commission) concluded that the CPI probably overstated the change in the cost of living by between 0.8 percentage points and 1.6 percentage points per year.(8) Since the commission's report was issued, the Bureau of Labor Statistics has made several modifications to the way in which it calculates the CPI, thereby eliminating most of the identified problems with the index. But even if all of the technical issues with the CPI have been corrected, the separate issue remains of whether the index properly measures the cost of living for Social Security beneficiaries. Their cost of living may grow faster than that of other consumers because of their different purchasing patterns.

Some policymakers suggest that Social Security law be changed to provide for a lower COLA--one equal to the annual increase in the CPI minus a specified number of percentage points. If in fact the CPI still overstates increases in the cost of living for Social Security recipients, policymakers can reduce the adjustment by an appropriate amount without making benefits any lower in real terms than they were when the recipients became eligible for them.

If the CPI accurately measures increases in the cost of living, a reduction in the COLA will result in each beneficiary's experiencing an annual decline in real benefits. If the CPI currently understates the change in the cost of living for Social Security recipients--perhaps because of differences between the purchasing patterns of beneficiaries and other consumers--then the existing decline in real benefits will be exacerbated.

The effects of such a change would differ from the impact of an across-the-board constraint on the increase in initial benefits (or an equivalent rise in the normal retirement age) in two ways. First, limiting the increase in initial benefits would have a progressively larger effect on each cohort. The impact on the baby-boom generation would be small, and current beneficiaries would not be affected. Reducing the COLA, by contrast, would affect all beneficiaries to some extent, and the benefits of all future cohorts would be reduced by roughly the same percentage.

Second, the effect of a lower COLA would accumulate over the years, so the change would generally have the largest impact on people who collected Social Security benefits the longest: older retirees, widows, and disabled beneficiaries. Even a relatively small annual cut in the COLA would greatly reduce benefits for those recipients. For example, if benefits were adjusted for the annual increase in the CPI minus 1 percentage point, by age 75 retired beneficiaries would incur a 12 percent reduction in benefits compared with what they would have received under current law. By age 85, their benefits would be 21 percent lower than under current law. And by age 95, they would incur a 28 percent cut.

If the COLA was set to equal the increase in the CPI minus 1 percentage point beginning in December 2004, Social Security outlays would be about 11 percent lower by 2050 than the amount projected under current law. Most of that reduction (in percentage terms) would be achieved by 2030. For example, outlays in 2030 would be 5.3 percent of GDP instead of 5.9 percent. Unlike in the previous two options, however, spending would continue to grow as a percentage of GDP in later years.

Alternatively, lawmakers might choose to reduce cost-of-living adjustments only for Social Security recipients whose benefits or income was above specified levels; however, doing that would lessen the savings. (Some beneficiaries with low income and few assets would receive Supplemental Security Income benefits, which would offset some or all of the reduction in their Social Security benefits. The estimate above does not account for that offset, which would slightly reduce the amount of savings.)


1.  For a fuller discussion of how the Social Security program works and how changes to it might affect the nation's ability to deal with impending demographic shifts, see Congressional Budget Office, Social Security: A Primer (September 2001).
2.  For more information about the retirement prospects of the baby-boom generation, see Congressional Budget Office, Baby Boomers' Retirement Prospects: An Overview (November 2003).
3.  For a more detailed description of that formula, see Congressional Budget Office, Social Security: A Primer, Chapter 2.
4.  The adjustment is not yet completely actuarially fair, but it will be for beneficiaries who turn 65 in 2008 or later.
5.  Specifically, the normal retirement age will rise by two months per birth year for people born from 1939 through 1943 and again by two months per year for people born from 1955 through 1960.
6.  See Congressional Budget Office, Social Security: A Primer, for a discussion of issues relating to proposals for individual accounts.
7.  See Congressional Budget Office, Measuring Changes to Social Security Benefits, Long-Range Fiscal Policy Brief No. 11 (December 1, 2003).
8.  Advisory Commission to Study the Consumer Price Index, Toward a More Accurate Measure of the Cost of Living: Final Report to the Senate Finance Committee (December 1996).

Previous Page Table of Contents Next Page