Testimony on Social Security’s Finances

Phillip L. Swagel


Before the Committee on the Budget

United States Senate

Chairman Whitehouse, Ranking Member Grassley, and Members of the Committee, thank you for inviting me to testify about the Social Security program.

Social Security faces a significant financial challenge in the coming decade. Its two components, Old-Age and Survivors Insurance (OASI) and Disability Insurance (DI), are financed by revenues from payroll taxes and income taxes on benefits that are credited to separate trust funds. But those revenues are not sufficient to cover the benefits that are due under those programs.

  • In the Congressional Budget Office’s projections, the OASI trust fund is exhausted in fiscal year 2032, when its balance reaches zero. The DI trust fund is exhausted in 2052. If the two trust funds were combined, they would be exhausted in fiscal year 2033.
  • Such long-term projections are uncertain. Demographic factors are one key source of that uncertainty. Economic growth is another, particularly in projecting trust fund exhaustion. If the economy grew more quickly than projected, annual revenues would be greater and the trust funds would be exhausted later than projected. If the economy grew more slowly than projected, the opposite would occur.

CBO produces long-term projections for Social Security under two scenarios, which differ in the concepts used to estimate benefits after the projected exhaustion of the trust funds:

  • Scheduled benefits—benefit amounts are paid as scheduled under current law, regardless of whether balances in the program’s trust funds are sufficient to cover those payments.
  • Payable benefits—total benefit amounts are limited to annual revenues from payroll taxes and income taxes on benefits after the combined trust funds are exhausted.

In this testimony, I will discuss CBO’s two sets of projections of Social Security’s finances and the agency’s underlying demographic projections.

Scheduled Benefits

CBO’s baseline budget projections for Social Security are required by statute to reflect scheduled benefits.1 The agency projects that if Social Security benefits were paid as scheduled, the program’s actuarial deficit over the next 75 years would equal 1.7 percent of gross domestic product (GDP), or 5.1 percent of taxable payroll.2 In other words, trust fund balances would be sufficient to pay scheduled benefits through 2097 if payroll tax rates were increased immediately and permanently by about 5.1 percentage points (before accounting for the effects of such changes on the economy). Such an increase would boost payroll taxes from 12.4 percent to 17.5 percent, a relative rise of 41 percent. Alternatively, a reduction in scheduled benefits by an amount equivalent to 5.1 percent of taxable payroll (a relative reduction in benefits of 27 percent) would be sufficient to permit full payment of those smaller benefits through 2097. A combination of changes to taxes and benefits could also suffice.3

Such policies would not ensure Social Security’s solvency after 2097. Although policies that increased taxes or reduced benefits by a constant percentage would create annual surpluses over the next three decades, they would lead to an increase in annual deficits later.

CBO projects that if Social Security paid benefits as scheduled, outlays for the program would increase from 5.2 percent of GDP in 2023 to 7.0 percent in 2097. In those projections, most of that increase occurs over the next decade as members of the large baby-boom generation retire. That growth then slows as members of that generation die and people from later generations become eligible for Social Security, but outlays continue to rise throughout the 75-year projection period because life expectancy increases.4

Unlike outlays, revenues for Social Security remain at their current level over the 75-year period, at about 4.6 percent of GDP. Payroll tax revenues decrease, and receipts from income taxes on Social Security benefits increase, in CBO’s projections. Those changes offset each other, so the amount of tax revenues credited to the trust funds remains roughly unchanged as a percentage of GDP.

CBO’s projections of the amount by which Social Security’s spending exceeds its revenues over the long term are subject to considerable uncertainty. In addition to economic growth, demographics are a key source of that uncertainty. If demographic trends differed from CBO’s projections (as discussed later in this testimony), the effects could be especially large in the later years of the projection period, in part because those effects would compound over time.

Payable Benefits

Under current law, Social Security outlays are limited to amounts payable from annual revenues after a trust fund’s exhaustion. Therefore, the Social Security Administration would no longer be able to pay full benefits after the combined Old-Age, Survivors, and Disability Insurance (OASDI) trust funds were exhausted in 2033. It is unclear how much payments to specific beneficiaries would be reduced if total benefits were limited to the amounts payable from dedicated funding.5

CBO estimated the amount of the total annual reduction in benefits that would be necessary for the program’s outlays to match its revenues after the combined funds were exhausted. In 2034, Social Security revenues are projected to equal 75 percent of the program’s scheduled outlays, resulting in a 25 percent shortfall. Thus, CBO estimates that Social Security benefits would be reduced by 25 percent in 2034 under the payable-benefits scenario. For this analysis, CBO assumed that benefits paid to all existing and new beneficiaries would be reduced by the same percentage. In CBO’s projections, about 79 million people (roughly one-fifth of the population) receive OASDI benefits in 2033.

In CBO’s assessment, the reductions in benefit amounts under the payable-benefits scenario would have four important economic effects. The first effect would decrease economic output in the short term, and the others would increase it in the long term:

  • Retirees’ income would decrease, dampening the overall demand for goods and services and causing output to be lower in the years immediately after exhaustion than CBO currently projects.6
  • The reduction in benefits would cause some people to work more and some people to remain in the labor force longer than they would have otherwise. Both of those factors would expand the supply of labor and thus the economy’s output in the long term.
  • Some workers who have not yet retired would respond to the prospect of smaller benefit payments by boosting their saving and reducing their spending. Those changes would lessen the effect that smaller future benefit payments would have on households’ future income and spending. The resulting increases in saving would boost the stock of private capital and output in the long term.
  • Federal debt would be lower than it is in CBO’s extended baseline projections. That lower debt would increase the amount of money available for private investment in capital goods and services, boost the stock of private capital, and make output in the long term greater than it would be otherwise.

CBO estimates that lifetime spending would decrease by a larger percentage and lifetime hours of work would increase more for households with lower lifetime income than for those with higher lifetime income. If benefits were reduced equally for all beneficiaries after trust fund exhaustion, the number of Social Security beneficiaries in poverty would probably increase.7

Because payable benefits are limited by the amount of revenues credited to Social Security’s trust funds, the uncertainty of CBO’s projections of the amounts of those benefits stems mainly from uncertainty about the economic growth that generates those revenues.


CBO’s long-term projections for Social Security are based on a detailed microsimulation model that starts with data about individuals from a representative sample of the population and projects economic and demographic outcomes for that sample through time.8 The demographic and economic projections used in CBO’s long-term projections for Social Security are consistent with those used in the agency’s baseline projections and for other purposes.

In CBO’s projections, the share of people age 65 or older rises in most years of the projection period as growth of that group outpaces growth of younger age groups (see Figure 1).9 In particular, the number of people in that older age group, who are less likely to work and pay payroll taxes and who are generally eligible for Social Security benefits, grows faster than the number of people ages 25 to 54, who are more likely to work and pay payroll taxes.

Figure 1.

Population, by Age Group

Millions of People

Data source: Congressional Budget Office, using data from the Social Security Administration. See www.cbo.gov/publication/59354#data.

The population referred to in this figure is the Social Security area population, which includes all residents of the 50 states and of the District of Columbia, as well as civilian residents of U.S. territories. It also includes federal civilian employees and members of the U.S. armed forces living abroad and their dependents, U.S. citizens living abroad, and noncitizens living abroad who are eligible for Social Security benefits on the basis of their earnings while in the United States.

Population growth is determined by births, deaths, and net immigration. In CBO’s projections, fertility rates remain lower than the replacement rate (the fertility rate required for a generation to exactly replace itself in the absence of immigration), mortality rates generally continue to decline, and immigration becomes an increasingly important part of overall population growth.


CBO’s projections of fertility are based on historical trends and other factors. For the 20 years before the 2007–2009 recession, the total fertility rate for women of childbearing age—ages 14 to 49—was 2.02 children per woman, on average. (The total fertility rate represents the average number of children that a woman would have if, in each year of her life, she experienced the birth rates observed or assumed for that year and if she survived her entire childbearing period.) After peaking at 2.12 in 2007, the total fertility rate generally fell (largely because of lower fertility rates among women age 24 or younger), reaching a low of 1.64 births per woman in 2020 and then increasing to 1.67 births per woman in 2022.10

In CBO’s projections, the total fertility rate equals 1.67 births per woman in 2023 and then rises as fertility rates among women ages 30 to 49 increase. By 2030, the total fertility rate is projected to be 1.75 births per woman and remains so through 2097. That rate is below the replacement rate of 2.1 births per woman. In CBO’s projections, fertility rates rise for women of relatively older childbearing ages and fall for women of relatively younger childbearing ages. That pattern is consistent with the trends of delayed childbearing and the rising average age of mothers.

Two key sources of uncertainty affect CBO’s projections of fertility. First, if trends in fertility, such as the rising average age of mothers and delayed childbearing, differed from CBO’s expectations, then the agency’s projections of fertility rates and the age distribution of mothers would change. Second, significant uncertainty remains about the long-term effects of the coronavirus pandemic on fertility rates.


CBO projects mortality rates on the basis of historical trends; the agency then takes into account the effects of COVID-19. Until recently, mortality rates in the United States had generally declined (meaning that life expectancy had generally risen) since at least the early 20th century. For the most part, mortality rates decreased more quickly for younger people than for older people. In recent years, though, the rate of decline has slowed, and mortality rates have increased for some groups, particularly younger people. To account for the recent slowdown in the decline of mortality rates, CBO projects that mortality rates in the near term (from 2020 to 2024) will decrease at roughly the same average rate as they did from 2010 to 2019. After 2024, mortal­ity rates are projected to return to longer-term trends and decline at a faster average rate than they did between 2010 and 2019. In CBO’s projections, after 2024, mortality rates decline at the average rate that they declined between 1950 and 2019.

After projecting mortality rates on the basis of historical trends, CBO incorporated the effects of COVID-19 on mortality rates through 2043 by increasing those rates, especially for older people, who are more likely to die from that illness. Although COVID-19 was a leading cause of death in recent years, in CBO’s projections, the long-term trend of declining mortality rates continues and outweighs the projected effects of COVID-19 among the surviving and future population. As a net result, mortality rates continue to decline in CBO’s projections, and life expectancy at age 65 increases from 18.9 years in 2023 to 24.6 years in 2097.

The evolving effects of the pandemic on mortality rates are a significant source of uncertainty in CBO’s projections of those rates. Changes in the total number of deaths from COVID-19 or the age composition of those deaths could affect outcomes significantly. Mortality rates in the long term are also uncertain because factors such as the evolution of medical technology and environmental conditions may have different effects in the future than they have had in the past.

Net Immigration

For the first two decades of the projection period (from 2023 to 2043), CBO’s estimates of net immigration (the number of people who migrate to the United States in a given year minus the number who leave in that year) are based on the agency’s economic projections and assessment of recent trends in immigration. From 2044 to 2097, net immigration in a given year is projected to grow at roughly the same rate as projected overall population growth in the previous year—0.2 percent per year, on average.

In CBO’s projections, annual net immigration to the United States averages 1.1 million people per year over the 2023–2097 period. Over that period, net immigration of lawful permanent residents averages 870,000 people per year; net immigration of legal temporary residents averages 80,000 people per year; and net immigration of foreign-born people without legal status averages 190,000 people per year.11

Several key factors contribute to the uncertainty of CBO’s projections of net immigration. Changing conditions in immigrants’ countries of origin, for example, could affect outcomes significantly.

1. Congressional Budget Office, CBO Explains How It Develops the Budget Baseline (April 2023), www.cbo.gov/publication/58916.

2. The actuarial balance is the sum of the present value of projected income and the current trust fund balance minus the sum of the present value of projected outlays and a year’s worth of benefits at the end of the period. (The present value expresses a flow of current and future income or payments in terms of an equivalent lump sum received or paid today.)

3. For a discussion of policy options to address the Social Security program’s finances, see Congressional Budget Office, Options for Reducing the Deficit, 2023 to 2032—Volume I: Larger Reductions (December 2022), www.cbo.gov/publication/58164, Options for Reducing the Deficit, 2023 to 2032—Volume II: Smaller Reductions (December 2022), www.cbo.gov/publication/58163, “How Changing Social Security Could Affect Beneficiaries and the System’s Finances” (interactive tool, April 2019), www.cbo.gov/publication/54868, and Social Security Policy Options, 2015 (December 2015), www.cbo.gov/publication/51011.

4. For details about this year’s long-term projections for Social Security, see Congressional Budget Office, CBO’s 2023 Long-Term Projections for Social Security (June 2023), www.cbo.gov/publication/59184. For additional discussion of CBO’s extended baseline projections over the 2023–2053 period, see Congressional Budget Office, The 2023 Long-Term Budget Outlook (June 2023), www.cbo.gov/publication/59014.

5. Barry F. Huston, Social Security: What Would Happen If the Trust Funds Ran Out? Report RL33514, version 34 (Congressional Research Service, September 28, 2022), https://tinyurl.com/3v5t6a28.

6. Congressional Budget Office, The Long-Term Budget Outlook Under Alternative Paths for the Economy and the Budget (forthcoming).

7. Kathleen Romig, “Social Security Lifts More People Above the Poverty Line Than Any Other Program” (Center on Budget and Policy Priorities, June 2023), https://tinyurl.com/2rpvpfxd.

8. Congressional Budget Office, An Overview of CBOLT: The Congressional Budget Office Long-Term Model (April 2018), www.cbo.gov/publication/53667.

9. Congressional Budget Office, The Demographic Outlook: 2023 to 2053 (January 2023), www.cbo.gov/publication/58612.

10. Brady E. Hamilton, Joyce A. Martin, and Michelle J. K. Osterman, Births: Provisional Data for 2022, Vital Statistics Rapid Release 28 (National Center for Health Statistics, June 2023), www.cdc.gov/nchs/data/vsrr/vsrr028.pdf.

11. Lawful permanent residents are authorized to work, responsible for paying taxes, and eligible for Social Security benefits. Legal temporary residents are also eligible for Social Security benefits. Both of those types of residents must meet the eligibility requirements, such as those related to past earnings, to receive benefits. Foreign-born people without legal status are generally not eligible for Social Security benefits because they are usually not authorized to work in the United States. However, some people without legal status who are authorized to work may be eligible for Social Security benefits. Section 208(d) of the Immigration and Nationality Act allows the Department of Homeland Security to provide employment authorization, subject to certain restrictions, to asylum applicants who are not otherwise eligible to work. The Administration, by rule, allows certain recipients of Deferred Action for Childhood Arrivals to receive employment authorization. See Congressional Budget Office, The Foreign-Born Population, the U.S. Economy, and the Federal Budget (April 2023), www.cbo.gov/publication/58939.

Xinzhe Cheng, Damir Cosic, Daniel Crown, Noah Meyerson, Kurt Seibert, and Emily Stern conducted the analysis for this testimony with guidance from Molly Dahl, Elizabeth Cove Delisle, Jaeger Nelson, Sam Papenfuss, and Julie Topoleski. Aaron Betz, Kyoung Mook Lim, Charles Pineles-Mark, and Jordan Trinh (formerly of CBO) also contributed to the analysis. Leigh Angres, Lora Engdahl, and Joseph Kile provided comments. Charles Pineles-Mark fact-checked the testimony. In keeping with the Congressional Budget Office’s mandate to provide objective, impartial analysis, the testimony makes no recommendations.

Mark Doms, Jeffrey Kling, and Robert Sunshine reviewed the testimony. Caitlin Verboon edited it, and Jorge Salazar created the figure and prepared the text for publication. The testimony is available at www.cbo.gov/publication/59354.

Phillip L. Swagel