How to raise the Social Security retirement age while protecting the poor

Key Takeaways

• Social Security's trust funds will be exhausted within a decade, and the program faces $25 trillion in long-term funding shortfalls.
• A key cause of Social Security's funding gap is rising longevity: More Americans are surviving to retirement age and living longer in retirement.
• To partially address this funding gap, many have proposed increasing Social Security's retirement age. But increased life expectancies aren't equally distributed. Wealthier Americans are living longer than ever, while life spans for lower-income Americans have stagnated.
• We propose a way to increase Social Security's retirement age that would protect lower-income Americans who aren't living longer. In the long term, this policy would address about half of Social Security's annual funding shortfalls. 

Social Security’s Old Age and Survivors Insurance (OASI) trust fund is projected to be exhausted in 2033, just eight years from now. The program as a whole faces a $26.1 trillion funding gap over the next 75 years (Social Security Trustees 2025).

Here we present a policy reform that is designed to satisfy two criteria. First, it is a partial solution based on reductions to future benefits. Second, the proposal is built around a well-understood cause of Social Security’s funding gap, which is that Americans on average are living longer. To address rising longevity, we propose gradually increasing Social Security’s full retirement age past the current age of 67.

However, our proposal also addresses the fact that rising longevity has not been distributed equally. While wealthier Americans live longer and healthier lives, life expectancies of lower-wage workers are not significantly increasing. It is possible to increase the Social Security retirement age in conjunction with changes to Social Security’s benefit formula that would not hurt the lowest-earning Americans. A stylized policy would effectively increase the full retirement age by two years for a middle-earning employee and four years for the highest earners while leaving benefits unchanged for low earners. This policy change could ultimately address roughly half of Social Security’s annual deficits.

By 2035, two years after the OASI trust fund is depleted, Social Security’s annual cash flow deficit will reach 1.06 percent of gross domestic product. While some have proposed filling the entire funding gap with additional revenues, the demands of such a policy would be steep. By comparison, the largest peacetime tax increase in U.S. history, the Tax Equity and Fiscal Responsibility Act of 1982, increased revenues by 0.98 percent of GDP over its first four years (Tempalski 2006). To meet Social Security’s funding gap by increasing the payroll tax rate would require an increase from the current low rate of 12.4 percent of wages to 15.9 percent of wages in 2035, with further gradual increases thereafter. Other options to increase revenues include increasing the maximum salary subject to Social Security taxes, which will be $176,100 in 2025, or transferring general tax revenues to Social Security.

However, resolving Social Security’s funding gap entirely through reductions in benefits would significantly reduce incomes for seniors, for whom Social Security benefits made up 19.6 percent of average incomes, net of taxes and transfers, in 2021   (Congressional Budget Office 2024). Biggs (2004) found that resolving Social Security trust fund insolvency via across-the-board benefit reductions would double the elderly poverty rate and meaningfully impact even middle-income households.

Therefore, it is probable that a package of policy changes including both revenue increases and benefit reductions will be enacted, with changes to benefits likely being phased in over time.

We outline a reform to Social Security’s full retirement age that, once fully implemented by approximately 2065, could close roughly half of Social Security’s annual funding gap via a progressive reduction in benefits. We implicitly assume the remaining half of Social Security’s funding shortfall may be filled with tax increases and that additional measures of either variety would be needed to maintain Social Security’s solvency in the shorter term.

Our proposed reform has the narrative benefit of being tied to one of the chief causes of Social Security’s funding gap — longer life expectancies — making it more understandable to the public. At the same time, our proposal is responsive to reasonable concerns that have been raised regarding increasing the Social Security retirement age.

Longer lifespans, for some

More Americans are surviving to claim Social Security retirement benefits and collecting benefits for a greater number of years. In 1940, when Social Security first paid retirement benefits, remaining life expectancy as of age 65 was about 13.7 years. Today, remaining life expectancy at age 65 is about to 20.6 years, a 50 percent increase in the duration of expected benefit receipt. By 2050, remaining life expectancy as of age 65 is projected to increase a further 1.4 years (Social Security Trustees 2025).

It therefore is unsurprising that many have proposed increasing Social Security’s “normal” or “full” retirement age. Not only would boosting the retirement age save money, it also would push Americans to work longer, which gives them more years to save and a shorter retirement that those savings must cover.

The problem with increasing the Social Security retirement age is that not all Americans are living longer. Lifespans for the rich have increased as they gave up smoking and embraced healthier lifestyles, while barely changing for the lowest-income Americans who have suffered from so-called “deaths of despair” due to chronic diseases or substance abuse. For instance, see National Academies of Sciences (2015).

Treasury Department economist Hilary Waldron, calculating annual death rates by lifetime earnings decile, found that from ages 63 through 71 individuals with lower lifetime earnings had a roughly three times greater probability of dying in a given year than did higher earners (Waldron 2013). The Government Accountability Office’s (GAO) analysis of Waldron’s work concluded that, due to differences in life spans, a retiree at the 75th percentile of the income distribution would survive 17 percent longer than the average retiree (Jeszeck, Shields et al. 2016). Quinby and Wettstein (2024) show that two disproportionately lower-income groups, Black men and less-educated whites of both genders, have shown little improvement in disability-free life expectancies, significantly complicating their prospects for working longer.

The effects of longevity differences by earnings level can be illustrated using stylized workers constructed by the Social Security Administration’s Office of the Chief Actuary (Burkhalter and Rose 2024). An individual at the 75th percentile of the lifetime earnings distribution roughly corresponds to the Social Security actuaries’ “high-wage” worker,  who earns 160 percent of the national average wage (about $115,00 annually) over their careers and who makes up about 20 percent of retirement beneficiaries. Assuming an average life expectancy as of age 65 of 20.5 years,[1] this implies that a lifetime high-wage earner would survive for about 3.5 years longer past age 65 than the average 65-year-old.[2]

Similarly, the GAO concluded that a retiree at the 25th percentile of the lifetime earnings distribution would have a 13 percent shorter duration of retirement. A worker at the 25th percentile roughly corresponds to SSA’s stylized “low-wage” worker, earning 45 percent of the national average wage (about $32,500 annually) and making up about 23 percent of retirees. The GAO’s estimates imply a life expectancy as of age 65 that is about 2.7 years below the average.

Moreover, one should expect that at further extremes — for instance, the SSA actuaries’ “very low” earner, who earns 25 percent of the national average wage and makes up 18 percent of retirees, or SSA’s “maximum wage” worker, who earns the maximum salary subject to payroll taxes for at least 35 years and comprises about 8 percent of retirees — difference in life expectancies in retirement would be even greater.

In the context of Social Security reform, increasing the retirement age for low earners in effect makes them pay for a funding problem they did not cause. To the degree that rising life spans have harmed Social Security’s funding health, lower-earning Americans largely have not contributed to that problem.

Increasing the retirement age while accounting for differential longevity

The wide dispersion of increases in longevity seemingly renders raising Social Security’s retirement age a poor solution to the program’s funding gap. However, it is possible to raise the Social Security retirement age while sparing those whose lifespans aren’t increasing. The key insight is that raising the Social Security retirement age imposes a uniform percentage benefit reduction on all retirees. However, other elements of the Social Security benefit formula can be adjusted to compensate specific groups for the benefit reductions imposed by increasing the retirement age.

For a person claiming benefits in 2024 at the full retirement age of 67, Social Security’s benefit formula replaces 90 percent of the first $926 in average indexed monthly earnings,[3] 32 percent of monthly earnings between $926 and $5,583, and 15 percent of earnings between $5,583 and the roughly $14,050 maximum taxed by Social Security (Social Security Administration 2024).

Benefits also depend upon the age of benefit claiming. For each year a person claims prior to the full retirement age, benefits are reduced by about 6.7 percent. For instance, if the full retirement age were increased by two years every retiree’s benefits would be cut by around 13.3 percent.

However, by adjusting the replacement percentages in Social Security’s benefit formula, an increase in the full retirement age can be coupled with offsetting benefit increases to hold harmless low-earning Americans.

We illustrate using the following stylized policy: The full retirement age would be increased by two years, from 67 to 69. At the same time, Social Security’s first replacement percentage would be increased from 90 percent to 106 percent, the middle factor reduced from 32 percent to 28 percent, and the top factor cut from 15 percent to 7 percent.

Using this formula, very low-wage workers  with average earnings about 25 percent of the national average would receive the same benefit they are entitled to under the current benefit formula. Middle-income workers would receive benefits 13.3 percent below the current benefit formula, equivalent to a two-year increase in the normal retirement age.[4]

Workers earning the maximum taxable wage would receive benefits 26.7 percent below the current formula. Their benefit would be reduced by approximately 13.3 percent via the two-year increase in the full retirement age and an additional 13.3 percent by reducing the benefit formula’s 15 percent replacement factor to 7 percent.  

A two-year increase in the Social Security retirement age would likely spur many individuals to delay retirement, among them including workers with lower earnings (Song and Manchester 2007).  However, low-wage workers who could not work longer would not be penalized. Their total benefit at any age at which they claimed benefits would be approximately the same as under the current benefit formula. But the very highest-earning workers would need to work substantially longer to receive what they are promised under the financially unsustainable current law formula.

Implementation

While promising, this progressive increase in the Social Security retirement age is not a magic bullet for the program’s solvency, for three reasons.

First, increasing Social Security’s retirement age would reasonably require a decade or more of prior notice. For instance, in 1983 Congress legislated an increase in Social Security’s full retirement age from 65 to 67, but that increase did not begin until the year 2000 and was not completed until 2023. By contrast, proposed pension reforms in France — which prompted riots in 2023 — would have increased the French retirement age by three years over a far shorter seven-year period (Associated Press 2022).

Second, any increase in the Social Security retirement age would affect only new beneficiaries. Benefits for current retirees would be unaffected. The full savings to Social Security would not occur for a period of roughly three decades.

Finally, even once fully implemented and phased in around 2065, this provision would reduce total retirement benefit outlays by roughly 14.9 percent, enough to fill about half of Social Security’s annual funding gap.

Therefore, while a progressive increase in the Social Security retirement age could serve as a meaningful contributor to the program’s long-term solvency, other steps would likely need to be implemented in the short and medium terms. These could include increased payroll tax revenues, transfers of revenue from the general fund of the federal government, or other reductions to Social Security benefits that could be implemented more rapidly, such as paying smaller annual Cost of Living Adjustments (COLAs).

But the example presented here shows that it is possible to raise the retirement age for the better-off Americans who are living dramatically longer, healthier lives while protecting lower-income Americans who are not.

Footnote

[1] This is the simple mean of male and female cohort life expectancy as of age 65 for 2024, as reported in the 2024 Social Security Trustees Report.

[2]This discussion draws upon Biggs, A. G. (2019). How much should the poor save for retirement? Data and simulations on retirement income adequacy among low-earning households. Prepared for presentation at the Pension Research Council Symposium.

[3] Average indexed monthly earnings are calculated by, first, indexing prior years’ nominal wages by the growth of national average wages through age 60, with earnings after age 60 left unindexed; second, summing the highest 35 years of earnings and dividing by 35 to produce average indexed annual earnings; and, finally, dividing by 12 to produce average indexed monthly earnings.

[4] That is to say, the increase in the 90% replacement factor to 106%, coupled with a reduction in the 32% factor to 28%, generates the same benefit for middle-earning workers as they would receive under the current benefit formula. The 13.3% reduction in total benefits would be solely a function of increasing the full retirement age.

References

Associated Press (2022). French leader vows to raise retirement age to 65, up from 62. Associated Press.

Biggs, A. G. (2004). The Distributional Consequences of a "No-Action" Scenario. S. S. Administration. Baltimore, Md.

Biggs, A. G. (2019). How much should the poor save for retirement? Data and simulations on retirement income adequacy among low-earning households. Prepared for presentation at the Pension Research Council Symposium.

Burkhalter, K., and K. Rose (2024). Scaled factors for hypothetical earnings examples under the 2024 Trustees Report assumptions. Actuarial Note. Social Security Administration. Baltimore, Md.

Congressional Budget Office (2024). The Distribution of Household Income in 2021.

Jeszeck, C. A., M. Shields, M. Weber, L. Hoffrey, and V. Lui (2016). Retirement Security: Shorter Life Expectancy Reduces Projected Lifetime Benefits for Lower Earners. Government Accountability Office.

National Academies of Sciences (2015). The growing gap in life expectancy by income: Implications for federal programs and policy responses.

Quinby, L. D., and G. Wettstein (2024). "Are older workers capable of working longer?" Journal of Pension Economics & Finance 23(2): 165-182.

Social Security Administration (2024). "Benefit Formula Bend Points." Retrieved August 16, 2024.

Social Security Trustees (2025). The 2025 Annual Report of the Board of Trustees of the Federal Old-Age and Survivors Insurance and Federal Disability Insurance Trust Funds. Washington, D.C., Government Printing Office.

Song, J., and J. Manchester (2007). "Have people delayed claiming retirement benefits-Responses to changes in Social Security rules." Social Security Bulletin 67: 1.

Tempalski, J. (2006). Revenue effects of major tax bills. Office of Tax Analysis Working Paper. U.S. Department of the Treasury. 81.

Waldron, H. (2013). "Mortality differentials by lifetime earnings decile: implications for evaluations of proposed Social Security law changes." Social Security Bulletin 73: 1.