social security research paper

Social Security

Research, statistics & policy analysis, research & analysis publication types, international update.

A monthly publication covering recent developments in foreign private and public pensions, social security, and retirement.

Issue Papers

Papers that summarize a complex issue in a comprehensible, concise paper with graphics.

ORES Working Papers

Unedited, preliminary papers prepared to solicit comments and feedback from the wider research community.

Policy Briefs

Briefs that summarize MINT results for policy proposals or the history, rationale, or effects of current or proposed policies.

Research and Statistics Notes

Short notes that provide addenda to material already published, summarize preliminary findings, or review or explain data or analysis tools/resources.

Research Summaries

One-page summaries of select research papers.

Social Security Bulletin

SSA 's quarterly research journal, the Social Security Bulletin , comprehensively covers Social Security–related topics of broad interest. Articles focus on the Social Security and Supplemental Security Income programs, the income of the aged, private pension coverage, the poverty status of various groups (such as women and children), reform of Social Security, and other timely subjects. The Perspectives section presents refereed contributions by outside scholars on similar issues.

Discontinued or limited series publications.

An official website of the United States government

Official websites use .gov A .gov website belongs to an official government organization in the United States.

Secure .gov websites use HTTPS A lock ( Lock Locked padlock icon ) or https:// means you've safely connected to the .gov website. Share sensitive information only on official, secure websites.

• Publications
• Account settings
• Advanced Search
• Journal List

Social security and mortality: The role of income support policies and population health in the United States

Peter s arno, james s house, deborah viola, clyde schechter.

• Author information
• Article notes
• Copyright and License information

Corresponding author

Peter S. Arno, PhD, is a Professor, Director of Doctoral Studies, Department of Health Policy and Management, School of Health Sciences and Practice, New York Medical College, Valhalla, NY.

James S. House, PhD, is a Senior Research Scientist, Professor of Sociology and Research Professor, Survey Research Center, University of Michigan, Ann Arbor, MI.

Deborah Viola, PhD, is an Associate Professor, and Associate Director of Doctoral Studies, Department of Health Policy and Management, School of Health Sciences and Practice, New York Medical College, Valhalla, NY.

Clyde B. Schechter, MD, is an Associate Professor, Department of Family and Social Medicine, Albert Einstein College of Medicine, Bronx, NY.

Issue date 2011 May.

Social Security is the most important and effective income support program ever introduced in the United States, alleviating the burden of poverty for millions of elderly Americans. We explored the possible role of Social Security in reducing mortality among the elderly. In support of this hypothesis, we found that declines in mortality among the elderly exceeded those among younger age groups following the initial implementation of Social Security in 1940, and also in the periods following marked improvements in Social Security benefits via legislation and indexing of benefits that occurred between the mid-1960s and the early 1970s. A better understanding of the link between Social Security and health status among the elderly would add a significant and missing dimension to the public discourse over the future of Social Security, and the potential role of income support programs in reducing health-related socioeconomic disparities and improving population health.

Keywords: Social Security, income support, social epidemiology, health policy

Introduction

Low socioeconomic position, measured variously in terms of poverty, income, wealth, education, or occupation, has been repeatedly linked to a greater burden of disease and death in the United States and other countries. 1 – 4 While this has been one of the most consistent findings in social and epidemiological research for decades, it is surprising that the major social programs that were designed to ease the burden of poverty or otherwise improve economic wellbeing of the US population have only rarely been examined with respect to health and disease. 5 – 7

Social Security (specifically, Old Age and Survivor Insurance) is the single most important and effective income support program ever introduced in the United States. It has been credited with alleviating the burden of poverty for millions of Americans, particularly for the elderly. 8 Recent census data indicate that over 40 per cent of the US population aged 65 years or older is kept out of poverty by Social Security. 9 Supplemental Security Income (SSI), which provides cash or benefits to low-income elderly people, has also helped to improve the economic status of the poorest stratum of Social Security beneficiaries.

Most public attention to the Social Security system from the media and politicians focuses on the system’s long-range financial problems. Some economists and policymakers have been pushing for cuts in Social Security benefits as a way of addressing our long-term federal budget deficits. The National Commission on Fiscal Responsibility and Reform established by President Obama to address the nation’s long-term fiscal crisis, has, for example, recommended steps to alter Social Security by lowering benefits and raising payroll taxes. 10 Legislation introduced in early 2010 by Representative Paul Ryan (Republican, Wisconsin), titled the ‘Roadmap for America’s Future Act of 2010’, would reduce future Social Security benefits and raise the age before benefits could be obtained. 11 Far less attention has been paid to the program’s unsurpassed record of reducing poverty and providing a safety net among the elderly. Still less attention considers the possible effect on the public’s health.

Our study explores the potential impact of Social Security on mortality among the elderly in the United States. We hope it will call attention to the likely relationship between major socioeconomic programs – such as Social Security – and the public’s health. Our goal is to broaden the perspectives of those in health, economics, social policy, and research regarding ways of improving population health and reducing health-related disparities.

Social Security and the Economic Wellbeing of the Elderly

The economic wellbeing of the elderly has improved dramatically since World War II. 12 – 14 Poverty fell from 35.2 per cent for persons aged 65 and older in 1959 (the first year official statistics were tabulated) to 9.7 per cent in 2008. 15 This compares with children, for example, whose poverty rates fell far less – from 27.3 to 19 per cent over the same time period. As early as the 1940s, the average monthly Social Security benefit accounted for 16 to 17 per cent of median family income for those aged 65 and older. (The percentage of income generated by Social Security benefits is estimated here by dividing the median Social Security benefit by the median family income in current dollars for each year.) This proportion rose to 28 per cent in 1950, when Congress mandated the first in a series of benefit increases. With the help of additional mandated increases, which improved the real income of the elderly substantially during the 1960s and early 1970s, and the indexing of benefits to inflation, begun in 1975, Social Security benefits now account for 28 to 38 per cent of median income among the elderly. 16

An early attempt to quantify the economic status of the elderly was conducted by Marjorie Shearon of the Social Security Board’s Bureau of Research and Statistics immediately before the introduction of monthly Social Security benefits and published in 1938. 17 Forty years later, the Social Security Administration published a comparative analysis, tracking the economic progress of the population aged 65 years and older. 18 It found, for example, that in 1937, 47.5 per cent of the non-institutionalized elderly had no income whatsoever (presumably relying entirely on friends, relatives, and charity). By 1978, the group reporting no income dropped dramatically to only 1.5 per cent. The single most important factor accounting for this change was Social Security benefits, followed by a distant second, the growth of private pensions.

At present, more detailed evidence describes the role of Social Security in alleviating the burdens of poverty and providing a safety net for retirees. Social Security remains a key source of income for most of its recipients. In 2008, more than half (63.9 per cent) of the beneficiaries aged 65 years or older received 50 per cent or more of their total income from their monthly Social Security checks. For a third of elderly beneficiaries (34.2 per cent), it provided 90 per cent or more of their income. 19 , 20 In addition, the reliance of the elderly on private, defined benefit retirement plans has declined dramatically. 21

Statistics on poverty are no less compelling. Nearly 44 per cent of the elderly are kept out of poverty by Social Security. 9 This impact on poverty is more substantial than all other government programs combined. According to one study, nine of every 10 elderly people lifted from poverty by government benefit programs, including state and local cash assistance programs, are lifted out by Social Security. 22

Finally, it can be argued that Social Security improves the social and economic status of the elderly in at least two dimensions: (1) level of current income, and (2) freedom to exit from unpleasant or onerous work and/or increase the amount of leisure time. The introduction of Social Security accelerated retirement at younger ages. In 1942, for example, approximately 48 per cent of men and 9 per cent of women aged 65 years or older participated in the labor force. 23 Although the figures for women have not changed as much over the years (13.3 per cent as of 2008), the percentage of men working past 65 has declined significantly to 21.5 per cent in 2008. 22

Social Security and Health of the Elderly

Analysts have observed that socioeconomic disparities in health and mortality are large in infancy and early childhood, diminish through very early adulthood, widen dramatically through middle and early old age, and then diminish in old age. 24 There are multiple potential causes of diminishing socioeconomic disparities in health and mortality in the oldest age group, but few have been carefully evaluated. Some evidence suggests that differential rates of mortality by socioeconomic status at younger ages are not the primary explanation. 25 Particularly in the United States, one plausible cause of reduced health disparities is that the population aged 65 years or older is the only one that has a substantial social welfare safety net – both income support via Social Security since 1940 and later universal health insurance via Medicare (the Federal program of health insurance for the elderly) since 1965.

Despite the important role of income support provided by Social Security and the well-known relationship between socioeconomic status and health, surprisingly little research has tried to demonstrate a link between Social Security benefits and health consequences. The major exception is the work of Jere Behrman, Robin Sickles, and Paul Taubman. 26 They found that an additional $1000 of Social Security benefits lowered the mortality hazard by 10–20 per cent. Given their study sample and design (heads of households from the Retirement History Survey who were 58–63 in 1969 with 10 years of mortality follow-up), they concluded that this reduction in mortality was a function of the inverse relationship between income and mortality. As the level of Social Security benefits is related to one’s career earnings history, the additional $1000 of benefits reflects higher lifetime earnings. They also found that SSI has an even larger impact on mortality than Social Security benefits, which they attributed to the fact that SSI goes exclusively to the poor and near-poor elderly.

In contrast, Snyder and Evans examined the impact of differing levels of Social Security benefits on mortality. 27 As a result of the changes made in the 1970s to lower costs of the program, for beneficiaries with identical earning histories, those born before 1 January 1917 received a higher benefit compared to those born after this date. Their study showed that the beneficiaries with lower payments were more likely to work and had lower mortality rates after age 65 years. Their results seem to suggest that source of income was as important as amount of income. They conclude that the time spent working decreased social isolation, a co-factor in mortality, suggesting that work had a positive health impact. One critique of their study is that their results measured the effect of increases for those beneficiaries who were wealthier and healthier and did not reflect the impact on the poorer elderly. Handwerker pointed out that mortality differences observed by Snyder and Evans occurred even before this cohort began collecting Social Security benefits. 28

More recently, Herd et al studied the changes in SSI benefits and changes in disability for beneficiaries aged 65 years and older. 6 They conclude that higher SSI benefit levels reduced disability in this group and their results remained robust regardless of the disability measure used (self-reported health condition or census Activities of Daily living (ADL) measure) and whether or not beneficiaries also received Medicaid (state programs to pay for medical care used by the indigent) payments for medical services. In addition, a preliminary analysis of the old-age assistance program from 1930 to 1955 by Balan-Cohen demonstrates an inverse relationship between income support for the poor elderly and mortality. 5 These findings affirm that income policies have potential for improving health outcomes among the elderly.

A few studies in other nations have considered the effect of state pension programs on health. One, in South Africa, looked at the impact of expanded pension payments to older people. 29 In households where incomes were pooled with other earners, self-reported health (including psychological health) of adults was improved, children living in the households were taller, and overall nutrition and sanitation were improved. Another study looked at the effect of the 1996 pension crisis on Russian households, where almost one-third of pensioners went without benefits for a prolonged period of time. 30 What was the health impact? The authors found that poverty rates doubled, nutrition standards decreased, and pensioners who received no payments were 5 per cent more likely to die in the 2 years immediately following the crisis. In a more recent study looking at the impact of pension reform in China, the authors found that as ‘pension wealth declines, households reduce their expenditures on education and health for children’. 31 Finally, in an analysis of 18 OECD countries, Norström and Palme found that old-age pensions had a significant inverse impact on all-cause mortality among the elderly. 32 In other words, the existence of public pension income affects household consumption and may have an indirect effect on health outcomes. On balance, the few available data about the role of Social Security and public pensions on the health of the elderly are consistent with our hypothesis that Social Security in the United States has resulted in improvements in the health of the elderly.

The purpose of our analysis is to evaluate more directly the hypothesis that Social Security has improved the health and longevity of the elderly population and hence contributed to overall improvements in population health over the last seven decades. We model ‘exposure’ to Social Security for those aged 65 years plus, specifically to Old-Age and Survivors Insurance benefits, capturing program effects among age groups, and over time within age groups. We model overall and non-infectious disease mortality as functions of age and time period using piecewise linear regression. According to Andersson et al , ‘Linear segmented regression analysis is a partly controlled design where the trend before the event of interest is used as a control period and can be controlled by the use of a comparison area or group that was not affected by the policy studied’. 33 This approach allows us to consider pre- and post-policy temporal trends. The limitations in our approach include the challenge of ‘separating the effect of time from that of the policy, accounting for heterogeneity in the policy effect’. 34 Mortality decline is assumed to be linear over time. 35 Wald tests were used to determine the significance of differences in declines of mortality rates. Our residual plots show no signs of autocorrelation.

We test our hypothesis in two specific ways. If the data support the hypothesis that Social Security has a positive impact on health of the elderly, we would observe:

A steeper decline in mortality in the elderly compared with younger age groups during the time period that followed the initial implementation of Social Security in 1940.

A steeper decline in mortality in the elderly compared with younger age groups in the time period that followed the legislated improvements in Social Security benefits and then after indexing of Social Security benefits that occurred between the mid-1960s and the early 1970s.

Data sources

Age group-specific mortality rates from all causes and from infectious diseases for the years 1900 through 1996 were provided by Gregory Armstrong at the Centers for Disease Control and Prevention (CDC). These data were used in a published analysis of trends in infectious disease mortality in the United States during the twentieth century. 36 Total population counts for each age grouping in each year were also provided. The age groups represented in the original data are infants (less than 1 year old), children (ages 1 to 4 years), 10-year age groupings from age 5 through 84 years, and older than age 84 years. As infants, children, and young adults are not relevant for testing our hypotheses, the age groups younger than 45 years were excluded. Thus, for our analyses, we constructed 10-year age-groupings from age 45 to age 74 years (45–54 years, 55–64 years, and 65–74 years) and a grouping for everyone older than 74 years. The oldest group was calculated by combining the original CDC data for 75 to 84 year olds and those 85 years and older, using a population-weighted average for the two subgroups.

Mortality data were grouped by calendar years into five ‘eras’: 1900–1939, 1940–1954, 1955–1969, 1970–1984, and 1985–1996. The eras were defined by our study hypotheses that highlight two crucial periods around which we should observe shifts in mortality patterns: 1940–1954, the period following the introduction of monthly Social Security benefits (1940) and 1970–1984, which reflected a period of substantial growth in the amount of Social Security benefits received. Similarly, our analyses and results will focus on these same two crucial periods. The years 1900–1939 are the ‘control’ period before the Social Security program was initiated; 1955–1969 is an era in which Social Security existed but did not grow substantially; and finally we distinguish 1985–1996 because of the new impact of HIV disease on infectious disease mortality.

Non-infectious disease mortality rates were calculated as the all-cause mortality rate minus the infectious disease mortality rate.

Statistical analyses

To test Hypotheses 1 and 2, we first graphed age-specific mortality rates over time for each age group and conducted piecewise linear regression analyses to test the statistical significance of observed differences in rates of change for the two crucial periods of interest (1940–1954 and 1970–1984) compared with the other eras for the two age groups under 65 years, compared with the two age groups 65 years and older.

We then modeled the logarithm of the all-cause and non-infectious disease mortality rates assuming that separate rates of decline might be found for each age group and all time trends were piecewise, log-linear. When the natural logarithm of mortality rate is regressed against time, as in these models, the slope of the regression line is an estimate of the average annual relative change in the untransformed rate variable. For example, a regression coefficient associated with a year that equals –0.02 can be interpreted as an average annual decline of 2 per cent in the untransformed mortality rate. Wald tests were used to compare the rates of change for each of the four outcomes (all-cause mortality, non-infectious disease mortality, and their respective log transformations) for each age group with that of the next youngest age group in each era and to compare the rates of change for each era with that of the previous era in each age group. We include results for both the log-transformed and untransformed mortality data, as there is some contention as to which provides the best evidence, especially from a policy perspective. 37

Figure 1 shows age-specific mortality over the study period for four age groups, 45–54 years, 55–64 years, 65–74 years, and 75–84 years. These data suggest that the rate of mortality decline was approximately steady before 1940 for all four age groups, but just after that year there was marked deceleration of mortality among both age groups 65 years and older compared with the two age groups younger than 65 years. This steeper decline in mortality in the over-65 year groups compared with younger people in the period that followed the initial implementation of Social Security in 1940 is consistent with our hypothesis. As elderly mortality declines could have been simply due to economic growth, we conducted additional sensitivity analyses that included controls for gross domestic product growth over this time period; our results did not substantially change from what is presented below.

All-cause mortality by age group, the United States, 1900–1997 deaths per 100 000 population.

The greater rate of decline since 1940 does not appear to have been uniform over the entire post-1940 period. Rather, in both of the two age groups 65 years and older, there was a sharp deceleration of mortality in both the 1940–1954 and 1970–1984 eras, with sharpest rate of decline in the first half of both of these intervals. These steeper declines in mortality in the over-65 age group compared to the younger age groups specifically in the periods that followed the initial implementation of Social Security (1940) and the marked improvements in benefits, indexing of Social Security and increased enrollment (beginning in the late 1960s), are also consistent with our hypothesis. 35

Linear regression analyses of these data confirmed that, although mortality rates for all age groups fell throughout the entire study period, the 65 years and older groups experienced a statistically significant greater decline in mortality after 1940 compared with the earlier period (see Figure 2 ). This is in stark contrast to the groups younger than 65 years, whose mortality rate declines before and after 1940 remained virtually the same. Furthermore, the rates of mortality decline for these younger age groups were modest and not statistically different across the time periods.

Results of linear regression analysis comparing average annual mortality decline for pre- and post-1940 eras for each of four age groupings: all-cause mortality by age group, the United States, 1900–1997.

Table 1 illustrates the rates of decline in mortality for each age group and era for each of the four outcomes (all-cause mortality, non-infectious disease mortality, and their respective log transformations). The first of the three columns in each era represents either the average annual absolute decline in mortality rates or the log average annual decline in mortality rates for each age group. The log average annual decline can be interpreted as an annual percent decline, for example for the age group 65–74 years in the era 1940–1954, there was an average annual decline of 1.49 per cent in the mortality rate. The second column is the P -value associated with a Wald test of the statistical significance of the difference between declines in mortality rates in that era and the preceding one (for example, 1900–1939), thus representing an era-to-era comparison within each age group. The third column is the P -value associated with a Wald test of the statistical significance of the difference between declines in mortality rates of that age group with that of the next youngest age group within a particular era (for example, 55–64 years). As expected, larger reductions in average annual absolute mortality rates are consistently observed in the groups aged 65 years and older compared with those younger than 65 years with the one exception of non-infectious disease mortality for those ≥75 in 1930–1939. This latter group (≥75) clearly had a significant drop in infectious disease mortality during this time period as well.

Decline in mortality rates for people aged 45 years and older, 1900–1984

For untransformed mortality data (All Cause and Non-infectious Disease), values represent average annual absolute decline in mortality; for log transformations, values can be interpreted as an annual per cent decline in mortality.

Era-to-era comparison within age groups.

Age grouping to age grouping comparisons within era.

The annual percent declines within eras, however, are inconsistent across younger and older age groups. We would expect to find the era-to-era comparisons to more clearly demonstrate differences in mortality declines across age groups than within era comparisons. For example, the dramatic decline in cardiovascular disease mortality in the 1970s and 1980s was most pronounced in the younger age groups and may account for larger mortality declines in these age groups during this era. 38 , 39 Similarly, there are undeniably a host of unmeasured factors affecting mortality among the age cohorts born in the last half of the nineteenth century particularly affecting mortality in the 1900–1930 era.

The era-to-era comparisons clearly indicate that average absolute mortality declines (and the annual percent declines) were experienced within all age groups. The magnitude of these declines is, however, greater for the 1970–1984 period compared to the 1955–1969 period, as well as the 1940–1954 period compared to the 1900–1939 period. These patterns of differences between eras largely held for both absolute declines in mortality rates, as well as annual percent declines and support our hypotheses.

Summary and Discussion

Taken together, we believe these data largely support our hypothesis that improvements in socioeconomic status via the Social Security program have had a beneficial impact on the health of the elderly in the United States. Our most convincing test showed that non-infectious disease mortality declined throughout the first 80 years of the twentieth century. The deceleration of mortality observed in the first crucial period beginning 1940 was substantially more pronounced for those aged 65 years and older than for the younger groups. As non-infectious disease mortality accounted for more than 90 per cent of the decline in all-cause mortality, it is clear that the difference in deceleration of the non-infectious disease mortality between the older and younger age groups across the relevant eras is large enough to admit explanation by a factor specifically affecting the elderly, that is, Social Security.

Unadjusted comparisons of age groups and parallelisms among time series cannot exclude alternative explanations because the data are also consistent with other interpretations of the changing nature of and access to medical care. Specifically the period from 1940 to 1955 was also one in which the use of antibiotics in the treatment of infectious disease emerged and grew in clinical medicine, and the period 1971 to 1985 saw the growth of Medicare and improvement in the treatment and prevention of life-threatening chronic diseases, especially heart disease. 35 , 40 Limited availability of aggregate data makes it difficult to estimate empirically the relative importance of these changes in medical care versus the introduction and then indexing of Social Security cash benefits. It should be noted that we find significant effects in mortality decline even before the introduction of Medicare.

With our data we were however able to estimate rates of declines in various causes of death to approximate the impact of antibiotics, the effects of which should be seen almost entirely in changing rates of infectious disease among all age groups. Consistent with the patterns we originally hypothesized and tested, non-infectious disease mortality rates declined throughout the 1900s in both the young and elderly age strata, but the rates of decline for the elderly were in almost all instances higher than those for the younger age groups. To determine whether the income effect of Social Security truly exists as an independent phenomenon when other confounding effects like antibiotic use or medical technology are adequately controlled awaits further research.

Caveats exist with respect to our analyses beyond antibiotic use. We could not distinguish between income levels of beneficiaries within each era. It is likely that the beneficial impact of Social Security on health could be correlated with the proportion of total income the benefit represents. We also cannot distinguish between the significance of access to medical care versus income effects on other co-factors of mortality, including psychosocial, economic, and environmental conditions of life that may be mitigated via improved nutrition and housing, reductions of psychosocial (especially financial) stress, and increased resources for adapting to stress and threats to health. 5 , 24 Similarly, both the decline in smoking rates and the civil rights movement have been shown to have had salutary impacts on population health beginning in the 1960s. 35 , 41 These distinctions could well strengthen our results in key population subgroups and gain significance as the political debate on universal health coverage continues and as researchers investigate the correlation between other social and economic factors on improving population health.

Overall, we believe that the data we have examined are consistent with the hypothesis that Social Security improved the health of the elderly by improving living conditions and increasing access to medical care. Social Security’s role in alleviating the burden of poverty for millions of Americans and its potential link to improved health status among the elderly should add a cautionary note to policymakers as they are poised to radically alter the future of Social Security for generations to come.

Acknowledgments

Support for this article was provided, in part, by a RWJF Investigator Award in Health Policy Research (Arno) from the Robert Wood Johnson Foundation, Princeton, New Jersey and by grant P60-MD0005-03 from the National Institute on Minority Health and Health Disparities (Arno, Schechter). Special thanks to Nancy Sohler for her preliminary assistance with the analysis and to Robert G. Hughes, David Mechanic, Al Tarlov, Lynn Rogut, and Hal Strelnick for their support and encouragement.

• 1. Adler NE, Rehkopf DH. U.S. disparities in health: Descriptions, causes, and mechanisms. Annual Review of Public Health. 2008;29:235–252. doi: 10.1146/annurev.publhealth.29.020907.090852. [ DOI ] [ PubMed ] [ Google Scholar ]
• 2. Marmot M, Wilkinson RG. Social Determinants of Health. 2. Oxford: UK Oxford University Press; 2006. [ Google Scholar ]
• 3. Lantz PM, House JS, Lepkowski LM, Williams DR, Mero RP, Chen J. Socioeconomic factors, health behaviors, and mortality. Journal of the American Medical Association. 1998;279:1703–1708. doi: 10.1001/jama.279.21.1703. [ DOI ] [ PubMed ] [ Google Scholar ]
• 4. Pappas G, Queen S, Hadden W, Fisher G. The increasing disparity in mortality between socioeconomic groups in the United States, 1960 and 1986. New England Journal of Medicine. 1993;329:103–109. doi: 10.1056/NEJM199307083290207. [ DOI ] [ PubMed ] [ Google Scholar ]
• 5. Balan-Cohen A. The effect of income on elderly mortality: Evidence from the Old Age Assistance Program in the United States. American Economic Association Conference Papers; 2009. [accessed 1 January 2009.]. http://www.aeaweb.org/assa/2009/author_papers.php?author_ID=6557 . [ Google Scholar ]
• 6. Herd P, Schoeni RF, House JS. Upstream solutions: Does the supplemental security income program reduce disability in the elderly? The Milbank Quarterly. 2008;86:5–45. doi: 10.1111/j.1468-0009.2007.00512.x. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ]
• 7. Connor J, Rodgers A, Priest P. Randomized studies of income supplementation: A lost opportunity to assess health outcomes. Journal of Epidemiology and Community Health. 1999;53:725–730. doi: 10.1136/jech.53.11.725. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ]
• 8. Englehardt GV, Gruber J. Working Paper 10466. Cambridge MA: National Bureau of Economic Research; 2004. Social Security and the Evolution of Elderly Poverty. [ Google Scholar ]
• 9. Romig K. Social Security Reform: Possible Effects on the Elderly Poor and Mitigation Options. Congressional Research Service Report to Congress; 2008. [accessed 6 January 2011.]. www.globalaging.org/elderrights/us/2008/socreform.pdf . [ Google Scholar ]
• 10. National Commission on Fiscal Responsibility and Reform. [accessed 18 December 2010.];The Moment of Truth: Report of the National Commission on Fiscal Responsibility and Reform. 2010 December 1; http://www.fiscalcommission.gov/news/moment-truth-report-national-commission-fiscal-responsibility-and-reform .
• 11. [accessed 12 July 2010.];H.R.4529: Roadmap for America’s Future Act of 2010. http://www.govtrack.us/congress/bill.xpd?bill=h111-4529 .
• 12. Wright JD, Devine JA. Poverty among the elderly. Journal of Long-Term Home Health Care. 1995;13(1):5–16. [ PubMed ] [ Google Scholar ]
• 13. Radner DB. Incomes of the elderly and nonelderly, 1967–92. Social Security Bulletin. 1995;58(4):82–101. [ PubMed ] [ Google Scholar ]
• 14. Ross CM, Danziger S, Smolensky E. Interpreting changes in the economic status of the elderly, 1949–1979. Contemporary Policy Issues. 1987;5(2):98–112. doi: 10.1111/j.1465-7287.1987.tb00259.x. [ DOI ] [ PubMed ] [ Google Scholar ]
• 15. DeNavas-Walt C, Proctor BD, Smith JC. Income, Poverty, and Health Insurance Coverage in the United States: 2008. Washington DC: Government Printing Office; 2009. U.S. Census Bureau, Current Population Reports; pp. 60–236. [ Google Scholar ]
• 16. Kollman G. Summary of Major Changes in the Social Security Cash Benefits Program: 1935–1996. Congressional Research Service Report to Congress; 1996. [accessed 5 January 2011.]. http://www.ssa.gov/history/pdf/crs9436.pdf . [ Google Scholar ]
• 17. Shearon M. Economic status of the aged. Social Security Bulletin. 1938;1(1):5–17. [ PubMed ] [ Google Scholar ]
• 18. Upp M. A look at the economic status of the aged then and now. Social Security Bulletin. 1982;45(3):16–22. [ PubMed ] [ Google Scholar ]
• 19. Social Security Administration. Annual statistical supplement, 2008 to the Social Security Bulletin. Office of Policy, Office of Research, Evaluation and Statistics; 2008. SSA Publication No. 13–11700. [ Google Scholar ]
• 20. Social Security Administration. Office of retirement and disability policy. [accessed 10 July 2010.];Income of the Aged Chartbook 2008. 2008 http://www.socialsecurity.gov/policy/docs/chartbooks/income_aged/2008/iac08.html#definitions .
• 21. Social Security Administration. [accessed 22 December 2010.];Defined Contribution Pension Plans and the Supplemental Security Income Program. 2006 Policy Brief No. 2006-01 http://www.ssa.gov/policy/docs/policybriefs/pb2006-01.pdf .
• 22. Purcell P. Older Workers: Employment and Retirement Trends. Congressional Research Service Report to Congress; 2009. [accessed 6 January 2011.]. http://digitalcommons.ilr.cornell.edu/key_workplace/655 . [ Google Scholar ]
• 23. U.S. Census Bureau. Current Population Reports, Series P-50, No. 2. 1945. Current Population Survey. [ Google Scholar ]
• 24. Robert SA, House JS. Socioeconomic status and health over the life course. In: Abeles RP, Gift HC, Ory MG, editors. Aging and Quality of Life. New York: Springer-Verlag; 1994. pp. 253–274. [ Google Scholar ]
• 25. Beckett M. Converging health inequalities in later life: An artifact of mortality selection? Journal of Health and Social Behavior. 2000;41(1):106–119. [ PubMed ] [ Google Scholar ]
• 26. Behrman JC, Sickles R, Taubman P. Causes, Correlates and Consequences of Death among Older Adults: Some Methodological Approaches and Substantive Analyses. Boston, MA: Kluwer Academic Publishers; 1998. pp. 109–129. [ Google Scholar ]
• 27. Snyder S, Evans W. The effect of income on mortality: Evidence from the social security notch. Review of Economics and Statistics. 2006;88(3):482–495. [ Google Scholar ]
• 28. Handwerker EW. Unpublished doctoral dissertation. University of California; Berkeley: 2007. Empirical essays in the economics of aging. [ Google Scholar ]
• 29. Case A. Does money protect health status? Evidence from South African pensions. In: Wise DA, editor. Perspectives on the Economics of Aging. Chicago, IL: University of Chicago Press; 2004. pp. 287–312. [ Google Scholar ]
• 30. Jensen RT, Richter K. The health implications of social security failure: Evidence from the Russian pension crisis. Journal of Public Economics. 2003;88:209–236. [ Google Scholar ]
• 31. Feng J, He L, Sato H. Public Pension and Household Saving: Evidence from China. Bank of Finland, Institute for Economies in Transition; 2009. BOTIT Discussion Papers, No 2009/02. [ Google Scholar ]
• 32. Norström T, Palme J. Public pension institutions and old-age mortality in a comparative perspective. International Journal of Social Welfare. 2010;19:s121–s130. [ Google Scholar ]
• 33. Andersson K, Petzold MG, Sonesson C, Lonnroth K, Carlsten A. Do policy changes in the pharmaceutical reimbursement schedule affect drug expenditures? Interrupted time series analysis of cost, volume and cost per volume trends in Sweden 1986–2002. Health Policy. 2006;79(2–3):231–243. doi: 10.1016/j.healthpol.2006.01.007. [ DOI ] [ PubMed ] [ Google Scholar ]
• 34. French B, Heagerty PJ. Analysis of longitudinal data to evaluate a policy change. Statistics in Medicine. 2008;27(24):5005–5025. doi: 10.1002/sim.3340. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ]
• 35. Cutler D, Meara E. Changes in the age distribution of mortality over the 20th century. In: Wise DA, editor. Perspectives on the Economics of Aging. Chicago, IL: University of Chicago Press; 2004. pp. 333–366. [ Google Scholar ]
• 36. Armstrong GL, Conn LA, Pinner RW. Trends in infectious disease mortality in the United States during the 20th century. Journal of the American Medical Association. 1999;281:61–66. doi: 10.1001/jama.281.1.61. [ DOI ] [ PubMed ] [ Google Scholar ]
• 37. Manning WG, Mullahy J. Estimating log models: To transform or not to transform? Journal of Health Economics. 2001;20:461–494. doi: 10.1016/s0167-6296(01)00086-8. [ DOI ] [ PubMed ] [ Google Scholar ]
• 38. CDC. Achievements in public health, 1900–1999: Decline in deaths from heart disease and stroke: United States, 1900–1999. MMWR. 1999;48:649–656. [ PubMed ] [ Google Scholar ]
• 39. National Center for Health Statistics. Chartbook for the conference on the decline in coronary heart disease mortality. U.S. Department of Health, Education, and Welfare; 1978. [accessed 12 December 2010.]. www.cdc.gov/nchs/data/misc/corltrtacc.pdf . [ Google Scholar ]
• 40. Card D, Dobkin C, Maestas N. Does Medicare save lives? The Quarterly Journal of Economics. 2009;124(2):597–636. doi: 10.1162/qjec.2009.124.2.597. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ]
• 41. Kaplan GA, Ranjit N, Burgard SA. Lifting gates, lengthening lives: Did civil rights policies improve the health of African American women in the 1960s and 1970s? In: Schoeni RF, House JS, Kaplan GA, Pollack H, editors. Making Americans Healthier. New York: Russell Sage Foundation; 2008. pp. 145–169. [ Google Scholar ]
• View on publisher site
• PDF (460.5 KB)
• Collections

Similar articles

Cited by other articles, links to ncbi databases.

• Download .nbib .nbib
• Format: AMA APA MLA NLM

Add to Collections

How Much Lifetime Social Security Benefits Are Americans Leaving On the Table?

Americans are notoriously bad savers. Large numbers are reaching old age too poor to finance retirements that could last longer than they worked. This study uses the 2018 American Community Survey to impute retirement ages for 2019 Survey of Consumer Finance (SCF) respondents. Next, we run the SCF respondents through the Fiscal Analyzer (TFA) to measure the size and distribution of forgone lifetime Social Security benefits. TFA is a life-cycle, consumption-smoothing research tool that incorporates Social Security and all other major federal and state tax and benefit policies. The program can optimize lifetime Social Security choices. We find that virtually all American workers age 45 to 62 should wait beyond age 65 to collect. More than 90 percent should wait till age 70. Only 10.2 percent appear to do so. The median loss for this age group in the present value of household lifetime discretionary spending is $182,370. Optimizing would produce a 10.4 percent increase in typical workers’ lifetime spending. For one in four, the lifetime spending gain exceeds 17 percent. For one in ten, the gain exceeds 26 percent. Among the poorest fifth of 45 to 62 year-olds, the median lifetime spending increase is 15.9 percent, with one in four gaining more than 27.4 percent.

The authors thank the Federal Reserve Bank of Atlanta, the Goodman Institute, the Alfred P. Sloan Foundation, Boston University, and Economic Security Planning, Inc. for research support. The views expressed herein are those of the authors and do not necessarily reflect the views of Economic Security Planning, Inc., the Federal Reserve Bank of Atlanta, the National Bureau of Economic Research, or Opendoor Technologies.

MARC RIS BibTeΧ

Download Citation Data

Published Versions

How Much Lifetime Social Security Benefits Are Americans Leaving on the Table? , David Altig, Laurence J. Kotlikoff, Victor Yifan Ye. in Tax Policy and the Economy, Volume 37 , Moffitt. 2023

Working Groups

Conferences, mentioned in the news, more from nber.

In addition to working papers , the NBER disseminates affiliates’ latest findings through a range of free periodicals — the NBER Reporter , the NBER Digest , the Bulletin on Retirement and Disability , the Bulletin on Health , and the Bulletin on Entrepreneurship  — as well as online conference reports , video lectures , and interviews .

• Feldstein Lecture
• Presenter: Cecilia E. Rouse

• Methods Lectures
• Presenter: Susan Athey

• Panel Discussion
• Presenters: Karen Dynan , Karen Glenn, Stephen Goss, Fatih Guvenen & James Pearce

• Search Menu
• Sign in through your institution
• The Journals of Gerontology, Series B (1995-present)
• Journal of Gerontology (1946-1994)
• Advance Articles
• Editor's Choice
• Virtual Collection
• Supplements
• Special Issues
• Calls for Papers
• Author Guidelines
• Psychological Sciences Submission Site
• Social Sciences Submission Site
• Why Submit to the GSA Portfolio?
• Advertising and Corporate Services
• Advertising
• Reprints and ePrints
• Sponsored Supplements
• Journals Career Network
• About The Journals of Gerontology, Series B
• About The Gerontological Society of America
• Editorial Board - Psychological Sciences
• Editorial Board - Social Sciences
• Self-Archiving Policy
• Dispatch Dates
• GSA Journals
• Journals on Oxford Academic
• Books on Oxford Academic

Article Contents

Data and methods, analysis and measures, conclusions, conflict of interest, acknowledgments, author contributions.

• < Previous

The Growing Socioeconomic Gap in Lifetime Social Security Retirement Benefits: Current and Future Retirees

• Article contents
• Figures & tables
• Supplementary Data

Christopher R Tamborini, Gayle L Reznik, Howard M Iams, Kenneth A Couch, The Growing Socioeconomic Gap in Lifetime Social Security Retirement Benefits: Current and Future Retirees, The Journals of Gerontology: Series B , Volume 77, Issue 4, April 2022, Pages 803–814, https://doi.org/10.1093/geronb/gbab201

• Permissions Icon Permissions

Increasing socioeconomic disparities, including in life expectancy, have important implications for the U.S. Social Security program. This study examined inter- and intracohort trends in Social Security retirement benefits, paying special attention to how lifetime benefit trajectories by socioeconomic circumstance shift across cohorts encompassing current and future retirees.

Using a dynamic microsimulation model based on representative survey data linked to administrative records, we developed a set of cohort-specific projections that estimate monthly and lifetime Social Security retirement benefits for retirees spanning the early baby boom (1945–1954) to Generation X (1965–1974) cohorts.

We found a widening socioeconomic gap in projected monthly and lifetime benefits for men and women, especially on a lifetime basis. This divergence is associated with stagnation of benefit levels among lower socioeconomic status groups coupled with upward shifts among higher strata groups. Distributional changes are linked with increasing differential mortality, but other factors also likely play a role such as rising education premiums, growing earnings inequality, and changes in women’s work and relationship histories.

Widening mortality differentials can lead to distributional changes in the U.S. Social Security program. Microsimulation methodology lends insights into how the socioeconomic gap in monthly and lifetime benefit distributions may change among future older Americans in the context of differential mortality and other demographic changes. Moving forward in time, these complex patterns could offset some of the progressivity built into the system.

The distribution of benefits from national pension programs and its impact on the economic status of older individuals and families is a longstanding scholarly and policy interest. In the United States, Social Security (Old-Age and Survivors Insurance [OASI]) is the primary source of retirement income for the majority of the older population ( Carr, 2019 ), but the program’s distributional effects are heterogeneous, particularly on a lifetime basis. One aspect of rising concern centers on changes in the distribution of benefits by socioeconomic status (SES) across generations given demographic and socioeconomic changes over past decades ( National Academy of the Sciences [NAS], 2015 ).

Among the most important trends is rising life expectancy, which lengthens the period over which individuals receive benefits. However, uneven longevity gains across socioeconomic strata, marked by greater increases among more affluent groups ( Goldman & Orszag, 2014 ; Olshansky et al., 2012 ; Waldron, 2007 ), can contribute to a widening SES gap in Social Security benefits on a lifetime basis ( NAS, 2015 ). Given expected rises in benefit differentials by SES, the progressive features of the program’s benefit formula favoring low earners may decline in the future ( Bagchi, 2019 ; Burtless, 2019 ; Goda et al., 2011 ). Longevity trends may also affect how groups fare under proposed reforms, such as raising the retirement age ( Bosworth et al., 2016 ; Congressional Budget Office, 2015 ; Reznik et al., 2019 ).

In this article, we examine the distribution of Social Security retirement benefits among current and future U.S. retirees, focusing on how the lifetime benefit trajectories of different socioeconomic groups may shift across future retiree cohorts. Taking a lifetime approach recognizes the long reach of Social Security in later life and makes it possible to study the full distributional consequences of the program, which provides monthly payments to entitled workers from initial receipt until death. A lifetime perspective also provides a way to account for changes in the mortality differentials by SES across cohorts. Despite growing interest in exploring links between longevity trends and program impacts, most studies to date rely on snapshot measures of benefits due to the difficulty in obtaining appropriate long-term data. Among studies that do examine lifetime benefits, most focus on older cohorts whose members have already died. A central problem with estimating lifetime benefits for current and future retirees is that it necessitates the projection of trends into the future.

To overcome these limitations, this study uses a dynamic microsimulation model, Modeling Income in the Near Term (MINT). Based on nationally representative data matched to over 50 years of Social Security Administration (SSA) administrative earnings and benefit records, MINT provides a unique opportunity to investigate the expected distribution of Social Security retirement benefits among current and future retirees by SES.

Our main contribution is to use a robust microsimulation model to analyze whether the increasing socioeconomic gap in lifetime Social Security benefits, discussed in recent literature ( NAS, 2015 ), is expected to continue in more recent cohorts experiencing different socioeconomic and demographic conditions throughout their life course. To address this possibility, we develop cohort-specific projections for three different generations born between 1945 and 1974. A novel aspect of the study is the inclusion of persons born from 1965 to 1974 (known as Generation X), a group of individuals experiencing rising inequality during their work lives, among other changes. We also refine the literature by including measures of SES that rank individuals on their shared lifetime earnings, which is important when estimating outcomes for women, and also by examining trends by education, household income, and race.

The U.S. Social Security (OASI) program is the most important income source for older Americans. The program provides monthly benefits to qualified workers and family members (e.g., spouse) under certain conditions from receipt until death. Program entitlement and benefit amounts are largely a function of a retired worker’s earnings history; more specifically, the 35 highest-paid years (up to the taxable maximum; there is an annual limit on earnings subject to Social Security taxes [$142,800 in 2021]) in covered employment. Retirement benefits are payable at the earliest eligibility age of 62, except for widow(er)s who can claim at age 60. Actuarial reductions apply to those claiming before their full retirement age (FRA; e.g., 67 born in 1960 and later), while delays in claiming increase benefits, up to age 70. An important program feature is the progressive benefit formula, which provides higher monthly benefits per lifetime contributions to lower earners ( SSA, 2021 ).

The spouse and survivors of entitled workers may receive auxiliary benefits, which equal a percentage of the retired worker’s benefit amount (up to 50% for spouses and up to 100% for survivors, depending on when benefits are claimed). Whether an individual receives a retired worker benefit or an auxiliary benefit depends on marriage (of at least 10 years if divorced) and the relative lifetime earnings between spouses. Dually entitled beneficiaries receive an amount equal to the auxiliary benefit when it is higher than their worker benefit alone.

There are various ways to evaluate Social Security’s distributional impact. Conceptually, the program’s twin goals of adequacy and equity are important reference points ( Iams et al., 2010 ). Benefit adequacy focuses on how well the program provides basic protection against the loss of income following retirement. Equity stresses the link between a worker’s earnings history and benefit levels, on both a monthly and lifetime basis. In combination, an important discussion centers on how well retirement benefits provide the most vulnerable with adequate resources, while also dealing with groups equitably based on their lifetime contributions ( Herd et al., 2018 ). Analyzing lifetime benefits makes it possible to study the full distributional consequences of the program and accounts for factors such as differential mortality by SES and changes in benefits over later life, such as switching to a dually entitled survivor beneficiary upon the death of a spouse ( Sandell & Iams, 1997 ).

Sources of Cohort Changes in Lifetime Social Security Retirement Benefits

A myriad of socioeconomic and demographic trends in American society, along with program design, can combine in complex ways that alter the distributional impact of the Social Security program. In this section, we clarify several key factors.

A major trend affecting lifetime benefit distributions across cohorts is rising life expectancy, which lengthens the time span during which one receives monthly benefits. Overall life expectancy at age 65 has risen from 17.2 years in 1990 to 19.5 years in 2018 ( National Center for Health Statistics, 1994 , table 6-3; Xu et al., 2020 ; figure 1). However, life expectancy gains have evolved unevenly over recent decades, with individuals in lower socioeconomic strata experiencing less improvement than higher strata groups ( Montez et al., 2011 ; Olshansky et al., 2012 ; Waldron, 2007 ). One source of this divergence is the increased influence of education (and income) on outcomes that can influence long-term health such as employment, wages, family life, and health behaviors like smoking ( Case & Deaton, 2021 ).

Rising mortality differentials by SES means that individuals in higher strata would likely receive Social Security benefits for increasingly longer time spans than their lower strata counterparts, resulting in wider gaps in lifetime benefits. This pattern is evident in a NAS (2015) study, which shows greater gains in the present value of lifetime Social Security retirement benefits for men in the highest income quartile relative to the lowest quartile between the 1930 and 1960 birth cohorts (see also Bosworth et al., 2016 ).

In addition to rising inequality in longevity, a number of other factors may drive distributional changes in the program. Female retirees will have increasingly longer work histories and greater lifetime earnings ( Goldin & Mitchell, 2017 ). Studies show that changing marital patterns (e.g., rising divorce and never marrying), which vary by SES, are associated with a declining share of women potentially eligible for auxiliary benefits ( Iams & Tamborini, 2012 ; Tamborini et al., 2009 ). Additionally, future retirees will be better educated than past retirees and more racially and ethnically diverse ( Johnson et al., 2019 ). The former trend can affect benefits directly via the relationship between education and lifetime earnings ( Tamborini et al., 2015 ) or indirectly via the association between lifetime earnings and life expectancy ( Bosley et al., 2018 ). Another trend involves rising earnings inequality and polarization between high- and low-skilled jobs ( Autor & Dorn, 2013 ).

Program changes can also influence benefit distributions. The 1983 Social Security Amendments raised the FRA and increased the delayed retirement credits ( Song & Manchester, 2009 ). Although actuarially fair, these adjustments can result in greater monthly and lifetime benefits for higher earners who tend to claim benefits later and have greater life expectancy than lower earners ( Bronshtein et al., 2019 ).

Previous Empirical Work on Lifetime Social Security Benefits

Much of what is known about lifetime Social Security benefits come from a small but influential body of literature examining the program’s redistributive impact ( Gustman & Steinmeier, 2001 ; Gustman et al., 2013 ; Leimer, 2007 ; Smith et al., 2003 ; Tan & Koedel, 2019 ). Although studies vary in their scope and methodology, they generally seek to examine how much individuals receive from Social Security over a lifetime relative to their lifetime contributions (payroll taxes). Analytic approaches include historical data, microsimulation projections, or stylized worker simulations. Using measures such as “money’s worth” or “rates of return” ( Biggs et al., 2009 ; Clingman et al., 2020 ; Goda et al., 2011 ), comparisons often are made between high/low SES groups within and across generations. For example, Smith et al. (2003) use the MINT model to estimate the lifetime benefit/tax ratio across income groups for birth cohorts from 1931 to 1960, finding continued progressivity in the system. Gustman et al. (2013) also show progressivity over time, but find substantial differences by individual or household measures. Tan and Koedel (2019) suggest more limited progressivity in future years due to increasing differential mortality (see also Bagchi, 2019 ).

A related strand of work focuses on the distribution of lifetime Social Security retirement benefits over generations in the context of increasing SES disparities in life expectancy. In this vein, the NAS (2015) microsimulation analysis finds wider gaps in lifetime benefits between high- and low-income workers from the 1930 to 1960 birth cohorts. Using different methods, Bosworth et al. (2016) corroborate this finding by comparing the 1920 and 1940 cohorts.

The present study adds to the literature by assessing trends in the distribution of lifetime (and monthly) Social Security retirement benefits by SES across different cohorts of older Americans. Using a robust microsimulation model, we develop cohort-specific projections that encompass more recent cohorts than prior studies and use more comprehensive measures of SES, including “shared” lifetime earnings.

While research is typically about something that has happened, scholars and policymakers also are interested in questions that look over extended horizons. One area of particular interest is the future conditions of older Americans ( NAS, 2015 ). In this article, we draw on microsimulation methodology, which uses representative samples of actual individuals and families to project future trends ( Mitton et al., 2000 ; Reznik et al., 2019 ).

We use SSA’s MINT model, a microsimulation model for projecting future U.S. retirement incomes. Developed over the past couple of decades by SSA with the assistance of the Brookings Institution, the RAND Corporation, and the Urban Institute, MINT has been used to examine numerous aging topics, including cohort differences in the composition of retirement income ( Butrica et al., 2012 ), potential linkages between rising obesity and future claiming patterns ( Knoll et al., 2018 ), and the impact of Social Security policy options on older individuals’ financial status ( Couch et al., 2017 ; Reznik et al., 2019 ). We utilize the most recent version here (MINT8) developed by the Urban Institute. We provide a brief overview of the model below and a fuller description in Supplementary Technical Appendix (see also Smith & Favreault, 2019 ).

MINT8 begins with a nationally representative sample of individuals and households in the U.S. Census Bureau’s 2004 and 2008 panels of the Survey of Income and Program Participation (SIPP). These data are then linked to administrative records that contain longitudinal information on respondents’ earnings, benefits, and death through 2015. Approximately 90% of SIPP respondents are matched successfully. (For nonmatched respondents, MINT uses a hot-deck statistical match to impute earnings and benefits based on the donor’s record. Match variables include age, gender, death, education, race/ethnicity, disability status, defined-contribution pension contributions, monthly earnings, immigration characteristics, and class of worker.)

When making projections, MINT observes much of the lives of respondents using the survey data and the linked administrative data up to 2015. The model then simulates the remaining life course for each person year by year until death (up to 2099), while accounting for demographic and socioeconomic changes in the population. To do so, MINT uses simultaneous regression equations, statistical matching, and projection algorithms to project a wide range of economic (e.g., Social Security benefits, pension coverage, work history), health (e.g., disability), and demographic (e.g., death, marital history) outcomes. In some instances, MINT draws from other high-quality data to calibrate projections, including the Health and Retirement Study (HRS), Panel Study of Income Dynamics (PSID), and Survey of Consumer Finances (SCF). For example, MINT uses the HRS to develop projections of health status and the retirement decision. MINT uses SCF wealth data to calibrate the SIPP’s wealth data.

Two projections central to our analysis are Social Security benefits and mortality. We compute benefits using a benefit calculator that accounts for detailed program rules based on current law. The benefit formula is applied to a person’s lifetime Social Security earnings (average of highest 35 years), observed in the administrative earnings data from 1951 through 2015, followed by projections of future earnings. The model accounts for current and former spouses’ lifetime earnings (real and simulated) and marital duration because this helps to determine the individual’s benefit type and amount, particularly for women (i.e., retired worker/auxiliary). Estimates of benefits also account for a person’s claiming age, among other factors. Benefit amounts are reduced for claiming prior to a person’s FRA or increased for claiming after FRA (up to age 70).

Mortality is a key input when estimating lifetime Social Security benefits. Death is observed directly in the administrative data through 2015 (i.e., Numident file). To project mortality for nondisabled persons 55 or older and for disabled persons 67 or older, MINT8 builds on updated sex- and age-specific continuous hazard models originally developed by RAND researchers using the PSID ( Panis & Lillard, 1999 ). The hazard of dying at time t is a function of a rich set of covariates: age splines, education, race, disability status, birth year, marital status, average earnings over the last 5 years, and lifetime earnings. The baseline and updated models can be found in the work of Smith et al. (2010 , table 2-5). Death probabilities at younger ages (i.e., <55 for nondisabled and <67 for disabled) are based on models that jointly estimate earnings and disability status as part of a hot-deck statistical matching algorithm.

MINT estimates are thoroughly validated and benchmarked ( Smith et al., 2010 , chapter 8). Model projections align well with administrative income and benefit data published by the Internal Revenue Service (IRS) and SSA and with survey data ( Smith & Favreault, 2019 ; Smith et al., 2010 ). The version of MINT we use incorporates assumptions about future socioeconomic and demographic trends using the Social Security Board of Trustees’ intermediate-cost scenario, which reflect the best estimates of future conditions ( Board of Trustees, 2019 ). Key assumptions include future death rates, fertility, disability trends, price and wage growth, and net immigration (see Supplementary Technical Appendix for more details).

MINT provides significant advantages over several other approaches that can be used to examine future retirement incomes. The NAS report (2015) utilized the Future Elderly Model (FEM), based on the HRS, to project differentials in lifetime Social Security benefits between the 1930 and 1960 cohorts. While the FEM is a robust microsimulation model ( Goldman & Orszag, 2014 ; Zissimopoulos et al., 2018 ), MINT’s design focuses more on forecasting retirement incomes than the FEM, which centers more on health outcomes. Furthermore, the match rate with administrative records is considerably higher in the SIPP than the HRS, resulting in a more representative baseline sample of lifetime earnings and benefits in the MINT model. Other less formal methods such as “back-of-the-envelope” and synthetic approaches fail to account for underlying complex relationships and their covariance structure such as interactions between economic outcomes, mortality, and family structure.

Our analysis develops projections of life expectancy at age 62 and Social Security retirement benefits for individuals born from 1945 to 1974. We divide persons into three 10-year cohorts, each of which experienced different socioeconomic and demographic conditions over their life course. The early baby boom cohort (1945–1954) was at the forefront of demographic and socioeconomic change in the American population, including women’s incorporation into the labor market. The late baby boom (1955–1964) experienced dramatically higher increases in women’s labor force participation and changes in marital patterns. Members of the Generation X cohort (1965–1974) came of age during a period of increasing economic inequality, education-based stratification, and less marriage. (We use 1974 as the stopping point for Generation X, but the entire cohort is often thought to include individuals born up through 1979.) We elected 10-year rather than 5-year birth cohorts to increase sample sizes of subgroups and thus the reliability of projections.

Our analysis sample includes individuals who survive (or are expected to survive) to age 62 and are OASI-eligible based on their own record or spouse’s record. Age 62 is chosen because it is Social Security’s early eligibility age. The main analysis sample excludes individuals who ever received a disability benefit (through the Social Security Disability Insurance [SSDI] or the Supplemental Security Insurance [SSI] programs, about 15% of the final sample). A sensitivity test including disabled beneficiaries appears in a later section. The final sample size is 43,680 (14,351 for Early Baby Boom, 15,477 for Late Baby Boom, and 13,852 for Generation X).

Outcome Variables

We examine three outcomes. Of primary interest is lifetime Social Security retirement benefits, which refers to the total sum of current-law scheduled benefits (retired worker and/or auxiliary benefits) from age 62 to death (adjusted to 2020 dollars). Unlike monthly benefits, a lifetime measure accounts for benefit changes that may occur over one’s life, such as switching from being a retired worker to a dually entitled survivor beneficiary upon the death of a spouse ( Sandell & Iams, 1997 ). It also reflects how long one lives or is expected to live and thus can account for differential mortality. As a sensitivity test, we examine lifetime benefits as the present value at age 62, discussed later on.

We also examine monthly Social Security retirement benefits using a person’s initial benefit amount (adjusted to 2020 dollars). Although monthly benefits do not account for differential mortality or changes in benefits over a person’s life course (e.g., retired worker to dually entitled survivor), they capture income that is central to a person’s economic status at a given point in time.

The final outcome is life expectancy at age 62. Estimates reflect mean years until death (i.e., age of death minus 62) for persons within a cohort (i.e., cohort-based mortality).

Socioeconomic Measures

We construct several cohort-specific measures of lifetime socioeconomic circumstance. A key indicator is lifetime earnings, which reflects a person’s wage-indexed earnings from birth until death. Importantly, because MINT matches up current and prior spouses, we are able to compute “shared” lifetime earnings by dividing a person’s total couple earnings during years of marriage by half. In years when a person was not married, we include only the person’s own earnings. We use shared rather than individual lifetime earnings in our main analysis because it is a better measure of individuals’ SES during adulthood. Some married women may have low individual lifetime earnings due to taking time out of the labor force, but nonetheless experience relatively high economic status. Shared lifetime earnings account for this possibility.

We also analyze trends by educational attainment, using three categories: less than high school, high school graduate, and bachelor’s degree or higher. Furthermore, to provide additional insights for demographic groups associated with SES differences, we examine the projections by race (non-Hispanic White and Black). A broad literature shows substantive Black/White differentials in life expectancy ( Case & Deaton, 2021 ), long-term earnings ( Cheng et al., 2019 ), and educational attainment ( Coile & Duggan, 2019 ). These differences, and how they evolve in the future, are likely to have important consequences for monthly and lifetime benefit distributions across racial groups. We do not include Hispanics or Asians in this specific analysis as there were too few in the original data set (i.e., early baby boom cohort) to make reliable subgroup estimates. Furthermore, the original baseline mortality model built into MINT does not include immigrant status and Hispanic/non-Hispanic as predictors [Panis & Lillard, 1999].

In addition, we conduct several sensitivity tests. For example, we use a person’s own lifetime earnings to rank their socioeconomic position rather than their “shared” lifetime earnings. We also examine trends by household income at age 62 to reflect income differentials when entering the traditional Social Security claiming period before many had withdrawn from the labor force. The income measure includes Social Security benefits, labor income, all government cash benefits (e.g., SSI), pension income, income from coresidents, and annuitized income from financial assets (e.g., 401k). We discuss these and additional robustness checks further in a later section.

Our projections are cohort- and gender-specific and weighted to reflect the national population. MINT’s projections entail uncertainty like any microsimulation model. Given that the projections reflect multiple statistical equations that draw from multiple SIPP panels appended to administrative records among other data sources, sampling variance cannot be derived under normal assumptions. For this reason, small differences across groups should be interpreted cautiously.

Table 1 presents MINT projections of life expectancy at age 62 and Social Security monthly and lifetime retirement benefits for men and women by cohort. We report the percentage difference for each cohort relative to the early baby boom.

Projected Life Expectancy and Social Security Benefits by Birth Cohort and Gender

Notes: Gap between cohort and EBB (%) refers to the percent difference relative to the Early Baby Boom cohort. The sample is restricted to OASI-eligible individuals who survived to age 62, excluding disabled beneficiaries. Benefits are in 2020 real U.S. dollars ($). MINT = Modeling Income in the Near Term; OASI = Old-Age and Survivors Insurance. Source : Authors’ calculations using MINT, Version 8, updated to the intermediate assumptions of the 2019 Trustees Report .

MINT projects life expectancy gains across the cohorts. Greater gains are expected for men than women, which is consistent with other research ( Sasson, 2016 ). The results also show growth in Social Security retirement benefits (all dollars are indexed to 2020). In monthly terms, average initial benefits are expected to increase 18.5%, from $1,521 among early baby boomers to $1,802 in Generation X. We observe greater gains for women due to increases in their employment and career earnings over the cohorts examined. For lifetime benefits, the projections show a greater upward shift of 23%, from $460,150 among early baby boomers to $565,762 in Generation X.

Table 2 contains cohort-specific projections of the outcome measures by quintiles of shared lifetime earnings. We report the results for the lowest (bottom 20%), middle (middle 20%), and highest (top 20%) quintiles for parsimony. To highlight distributional changes across the cohorts, we calculate the gap (or difference) between the highest quintile value and the lowest/middle quintile values, expressed as a percentage of the lowest/middle quintile values.

Projected Life Expectancy and Social Security Benefits by Shared Lifetime Earnings Quintiles and Cohorts, Men and Women

Notes: “Gap between Highest and Lowest/Middle” refers to the percent difference between the highest quintile and the lowest/middle quintiles for that cohort. For example, for men in the Early Baby Boom cohort, it is equal to [($706,852 − $215,759)/$215,759] × 100 = 227.6%. The sample is restricted to OASI-eligible individuals who survived to age 62, excluding disabled beneficiaries. Benefits are in 2020 real U.S. dollars ($). Source : Authors’ calculations using MINT, Version 8, updated to the intermediate assumptions of the 2019 Trustees Report. MINT = Modeling Income in the Near Term; OASI = Old-Age and Survivors Insurance.

In terms of life expectancy at age 62, it is not surprising that persons in the lower part of the lifetime earnings distribution are projected to live shorter lives than those at the upper end. What is noteworthy is the growing inequality in life expectancy, which is consistent with work looking at earlier cohorts ( NAS, 2015 ). For men, the projected differential in life expectancy between the high and low quintiles increased from 6.1 years (24.6–18.5) in the early baby boom cohort to 7.5 years in Generation X (26.1–18.6). For women, the projected gap widens from 8.1 years (28.7–20.6) to 8.6 years (29.6–21.1). It is also noteworthy that MINT projects a decline in life expectancy among men in the lowest quintile from the late baby boom to Generation X. This is consistent with other studies showing increasing mortality and morbidity among certain subgroups, particularly non-Hispanic Whites without a college degree ( Case & Deaton, 2015 ).

Alongside diverging life expectancy, MINT projects a widening gap in monthly and lifetime Social Security retirement benefits across quintile rankings of shared lifetime earnings, with sharper increases for lifetime benefits. For illustrative purposes, Figure 1 displays the average projected lifetime benefits for the low, middle, and high quintiles by cohort and gender. Across cohorts, we observe remarkably flat lifetime benefits for those in the bottom fifth of the distribution but sharp gains for those in the top fifth of the distribution. Similar patterns can be seen in monthly benefits (not shown), albeit with more modest differentials.

Projected lifetime Social Security benefits by shared lifetime earnings quintiles and cohorts: men and women.

These intercohort patterns translate into greater differentials in projected monthly and lifetime benefits by socioeconomic strata. As given in Table 2 , the gap in men’s lifetime benefits between the low and high quintiles rises to 337% for Generation X relative to 228% in the early baby boom cohort. For monthly benefits, the gap rises to 243% for Generation X relative to 174% in the early baby boom. An increasing differential between the middle and highest quintiles also is evident but more modest. The patterns are similar for women. For instance, the projected lifetime benefit differential between women in the lowest and highest quintiles of shared lifetime earnings rises from $470,875 (210%) for early baby boomers to $706,471 (304%) for Generation X.

Table 3 examines the outcomes by education. In terms of life expectancy, the projections show increasing education differentials, particularly for men. Regarding benefits, several striking patterns emerge. First, the education gaps in projected monthly and lifetime benefits widen across the cohorts. Second, the lifetime benefit differentials again increase to a greater extent. For instance, male college graduates in the early baby boom cohort can expect to receive, on average, lifetime benefits that equal around almost twice the value as those without a high school diploma (97%). By Generation X, male college graduates can expect to receive 153% more than those without a high school diploma. For women, similar patterns emerge. Also notable, we observe widening differentials between college and high school graduates, suggesting that divergence is not only confined to high and low attainment.

Projected Life Expectancy and Social Security Benefits by Educational Attainment and Cohorts, Men and Women

Notes: “Gap between BA+ and Education Group” refers to the percent difference between college degree holders and the lower educational attainment groups for that cohort. The sample is restricted to OASI-eligible individuals who survived to age 62, excluding disabled beneficiaries. Benefits are in 2020 real U.S. dollars ($). Source : Authors’ calculations using MINT, Version 8, updated to the intermediate assumptions of the 2019 Trustees Report . HS = high school; MINT = Modeling Income in the Near Term; OASI = Old-Age and Survivors Insurance.

In ancillary analysis, we examine the outcome measures by race, specifically for non-Hispanic Black and White beneficiaries in each cohort ( Table 4 ). For both men and women, we find substantial Black–White gaps in life expectancy at age 62, with Blacks projected to live around 2 years less. As for benefits, we also observe substantial racial gaps, with Whites having greater levels in each cohort. This is not surprising given that benefits are based on lifetime earnings. At the same time, MINT projects a modest narrowing of the Black–White gap in monthly and lifetime benefits for men (e.g., lifetime differences decline from 36% in the early baby boom to 27% in Generation X). For women, however, the trend goes the opposite way, marked by wider Black–White gaps in Generation X. We revisit these trends in the Discussion section.

Projected Life Expectancy and Social Security Benefits by Race and Cohorts, Men and Women

Notes: The sample is restricted to non-Hispanic White and non-Hispanic Black OASI-eligible individuals who survived to age 62, excluding disabled beneficiaries. Benefits are in 2020 real U.S. dollars ($). Source : Authors’ calculations using MINT, Version 8, updated to the intermediate assumptions of the 2019 Trustees Report . MINT = Modeling Income in the Near Term; OASI = Old-Age and Survivors Insurance.

Sensitivity Tests

We conducted a number of sensitivity tests. We replicated all analyses using the median rather than the mean to assess the robustness of our estimates to potential bunching and outliers. We find similar patterns. Next, we employed alternative indicators of SES. First, to provide a more granular picture of socioeconomic strata, we assessed lifetime shared earnings by deciles rather than quintiles. The results show similar patterns as our main analysis but wider gaps at the top and bottom of the distribution. Second, we ranked persons by their individual rather than shared lifetime earnings by gender, also finding divergence between the top and bottom fifth across the cohorts ( Supplementary Appendix Table A1 ). Third, we examined the outcomes by household income at age 62 to account for all income sources. Stated briefly, we also see an increasing divergence in projected monthly and lifetime retirement benefits by household income ( Supplementary Appendix Table A2 ).

We also replicated all analyses including individuals who ever received an SSDI or SSI disability benefit ( Supplementary Appendix Tables A3–A5 ). The results show similar patterns overall, but as expected, including the disabled population reduces the expected life expectancy, especially at lower socioeconomic strata. This translates to a disproportionate decline in lifetime benefits among lower SES groups, leading to somewhat larger SES gaps within each cohort than our main analysis.

Finally, additional projections used the present value of Social Security benefits, starting at age 62 and ending at death and discounted using the interest rate contained in the recent Social Security Board of Trustees report ( Supplementary Appendix Tables A6–A8 ). These estimates also show widening socioeconomic gaps in benefits.

A broad social science literature focuses attention on the distribution of benefits from national pension systems and its impact on the economic status of older individuals and families. One aspect of concern in the U.S. centers on distributional changes in Social Security retirement benefits against the backdrop of increasing differential mortality along with other socioeconomic and demographic changes ( NAS, 2015 ). In this article, we utilized MINT, which is a robust microsimulation model based on nationally representative data matched to administrative records, to assess expected cohort changes in the distribution of Social Security retirement benefits by SES among current and future U.S. retirees.

Our findings show the increasing importance of SES as a differentiator in Social Security income, especially on a lifetime basis. A central discovery is a growing divergence in projected lifetime (and monthly) benefits for those born during the years known as Generation X (1965–1974) relative to baby boomers (1945–1964). This occurs because benefits are expected to increase among more advantaged socioeconomic groups in more recent cohorts, while they stagnate among persons in lower socioeconomic strata. This finding reinforces recent work that indicates growing SES differentials in lifetime Social Security benefits using older cohorts and other models ( Bosworth et al., 2016 ; NAS, 2015 ).

Even as we find a widening socioeconomic gap in projected Social Security benefits among future retirees, our findings by race are not as clear-cut. The projections show that the Black–White gap in monthly and lifetime benefits for men, while remaining substantive, narrows slightly in the most recent cohort. This pattern is likely driven by relative gains in education and employment among Black men compared to Whites over the cohorts examined ( Coile & Duggan, 2019 ), combined with a slight narrowing (or at least stalling) of the Black–White gap in life expectancy ( Case & Deaton, 2021 ). Yet for women, we observe wider Black–White benefit gaps in Generation X. This trend is likely driven by an increase in White women’s lifetime labor force participation relative to Black women over the cohorts examined.

While investigating the mechanisms behind the observed patterns is beyond the article’s scope, it is useful to revisit some of the factors discussed earlier. One key point is that the expected widening of the socioeconomic gap in lifetime benefits is attributable, in part, to an increase in mortality differentials. Another factor entails the rising premium for higher education. Wider socioeconomic gaps for both monthly and lifetime benefits also may be attributable to aspects related to rising income inequality, changes in the occupational structure, and differences in work at older ages and benefit claiming by SES relative to changes in life expectancy. Changing work–family patterns also are important, particularly for women. Altogether, a complex combination of socioeconomic and demographic trends fuel widening differentials in benefits by socioeconomic strata, although the program’s progressive benefit formula dampens the effects undoubtedly.

Methodologically, the study demonstrates the usefulness of a microsimulation approach when examining future-oriented questions about the distributional effects of Social Security on retirees. The analysis also underscores the importance of analyzing benefit distributions from lifetime and cohort perspectives. By examining lifetime benefits across cohorts, we are able to provide insights into the full distributional consequences of the program that account for differences across cohorts. The deeper increases in the socioeconomic gap in lifetime benefits may be difficult to reverse given increasing differentials in mortality coupled with rising economic inequality.

Our findings have a number of potential policy implications. As policymakers consider reform options, a rising socioeconomic gap in lifetime Social Security benefits could raise equity concerns around proposals to increase Social Security’s FRA, particularly for those lower SES beneficiaries with shorter life spans. Increasing divergence in benefits (monthly and lifetime) by socioeconomic strata may also have importance for policies that target vulnerable groups (e.g., low earners), such as the special minimum benefit ( Herd et al., 2018 ) or various other options that seek to increase the proportion of earnings used for the first bend point in the Social Security benefit formula. Finally, understanding changes in the distribution of benefits may be of increasing importance due to the shift from defined benefit to defined contribution pensions, which may raise the economic importance of Social Security income for some lower-income groups.

Before concluding, several limitations are worth underscoring. Our estimates for Generation X contain more projected data than for earlier cohorts, with the youngest individuals being in their early 40s when the administrative data stop. Consequently, their projections contain somewhat greater uncertainty than those from the baby boom cohorts. In addition, the model’s projections are sensitive to assumptions of mortality, wage, and employment trends that may change in ways not anticipated. For example, research on COVID-19 suggests a concentration of adverse effects, such as job loss and morbidity/mortality, among disadvantaged minorities who already have shorter life expectancies than other groups ( Fairlie et al., 2020 ; Lopez et al., 2021 ). Beyond immediate duress, these patterns could result in larger socioeconomic gaps in benefits than reported herein. Another limitation of this study is that we do not account for lifetime taxes paid into the system, which is also an important metric of distributional trends.

In summary, MINT projections show a growing socioeconomic gap in monthly and lifetime Social Security retirement benefits from the baby boom cohorts to Generation X. This pattern is characterized by stagnation of benefit levels among lower lifetime earners and lesser-educated groups coupled with upward shifts among higher-income and educated groups. Moving forward in time, this pattern could offset some of the progressivity built into the system.

None declared.

The views expressed in this article are those of the authors and do not represent the views of the Social Security Administration (SSA). Access to SSA data linked to Census Bureau survey data is subject to restriction. The data are accessible at a secured site and must undergo disclosure review before their release. For researchers with access to these data, our programs used in this analysis are available upon request.

C. Tamborini planned the study, developed the framework, conducted the data analysis, wrote the initial manuscript, and revised the manuscript. G. Reznik planned the study, conducted the data analysis, and helped to write the manuscript. H. Iams and K. Couch helped develop the framework, advised on analyses, and helped write the article.

Autor , D. H. , & Dorn , D . ( 2013 ). The growth of low-skill service jobs and the polarization of the US labor market . American Economic Review , 103 ( 5 ), 1553 – 1597 . doi: 10.1257/aer.103.5.1553

Google Scholar

Bagchi , S . ( 2019 ). Differential mortality and the progressivity of Social Security . Journal of Public Economics , 177 , 104044 . doi: 10.1016/j.jpubeco.2019.07.003

Biggs , A. G. , Sarney , M. , & Tamborini , C. R . ( 2009 ). A progressivity index for Social Security (Issue Paper No. 2009-1) . Social Security Administration, Office of Retirement and Disability Policy . https://www.ssa.gov/policy/docs/issuepapers/ip2009-01.pdf

Board of Trustees, Federal Old-Age and Survivors Insurance and Federal Disability Insurance Trust Funds . ( 2019 ). The 2019 annual report of the Board of Trustees of the Federal Old-Age and Survivors Insurance and Federal Disability Insurance Trust Funds . Government Publishing Office . https://www.ssa.gov/OACT/TR/2019/index.html

Bosley , T. , Morris , M. , & Glenn , K . ( 2018 ). Mortality by career-average earnings level (Actuarial Study No. 124) . Social Security Administration, Office of the Chief Actuary . https://www.ssa.gov/oact/NOTES/pdf_studies/study124.pdf

Bosworth , B. , Burtless , G. , & Zhang , K . ( 2016 ). Later retirement, inequality in old age, and the growing gap in longevity between rich and poor . Brookings Institution . https://www.brookings.edu/wp-content/uploads/2016/02/BosworthBurtlessZhang_retirementinequalitylongevity.pdf

Google Preview

Bronshtein , G. , Scott , J. , Shoven , J. B. , & Slavov , S. N . ( 2019 ). The power of working longer . Journal of Pension Economics & Finance , 18 ( 4 ), 623 – 644 . doi: 10.1017/S1474747219000088

Burtless , G . ( 2019 ). Fixing social security: Major reform or minor repairs? The ANNALS of the American Academy of Political and Social Science , 686 ( 1 ), 38 – 62 . doi: 10.1177/0002716219872161

Butrica , B. A. , Smith , K. E. , & Iams , H. M . ( 2012 ). This is not your parents’ retirement: Comparing retirement income across generations . Social Security Bulletin , 72 ( 1 ), 37 – 58 .

Carr , D. S . ( 2019 ). Golden years? Social inequality in later life . Russell Sage Foundation . doi: 10.7758/9781610448772

Case , A. , & Deaton , A . ( 2015 ). Rising morbidity and mortality in midlife among white non-Hispanic Americans in the 21st century . Proceedings of the National Academy of Sciences of the United States of America , 112 ( 49 ), 15078 – 15083 . doi: 10.1073/pnas.1518393112

Case , A. , & Deaton , A . ( 2021 ). Life expectancy in adulthood is falling for those without a BA degree, but as educational gaps have widened, racial gaps have narrowed . Proceedings of the National Academy of Sciences of the United States of America , 118 ( 11 ), e2024777118 . doi: 10.1073/pnas.2024777118

Cheng , S. , Tamborini , C. R. , Kim , C. , & Sakamoto , A . ( 2019 ). Educational variations in cohort trends in the black–white earnings gap among men: Evidence from administrative earnings data . Demography , 56 ( 6 ), 2253 – 2277 . doi: 10.1007/s13524-019-00827-w

Clingman , M. , Burkhalter , K. , & Chaplin , C . ( 2020 ). Money’s worth ratios under the OASDI program for hypothetical workers (Actuarial Note No. 2019.7) . Social Security Administration, Office of the Chief Actuary . https://www.ssa.gov/oact/NOTES/ran7/an2019-7.pdf

Coile , C. C. , & Duggan , M. G . ( 2019 ). When labor’s lost: Health, family life, incarceration, and education in a time of declining economic opportunity for low-skilled men . Journal of Economic Perspectives , 33 ( 2 ), 191 – 210 . doi: 10.1257/jep.33.2.191

Congressional Budget Office . ( 2015 ). Social Security policy options, 2015 . Congressional Budget Office . https://www.cbo.gov/sites/default/files/114th-congress-2015-2016/reports/51011-SSOptions_OneCol-2.pdf

Couch , K. A. , Reznik , G. L. , Tamborini , C. R. , & Iams , H. M . ( 2017 ). The distributional impact of social security policy options . Research on Aging , 39 ( 1 ), 135 – 165 . doi: 10.1177/0164027516656140

Fairlie , R. , Couch , K. , & Xu , H . ( 2020 ). The impacts of COVID-19 on minority employment: First evidence from April 2020 CPS microdata (NBER Working Paper No. 27246) . National Bureau of Economic Research . doi: 10.3386/w27246

Goda , G. S. , Shoven , J. B. , & Slavov , S. N . ( 2011 ). Differential mortality by income and Social Security progressivity. In D. Wise (Ed.), Explorations in the economics of aging (pp. 189 – 204 ). University of Chicago Press . doi: 10.7208/chicago/9780226903385.003.0007

Goldin , C. , & Mitchell , J . ( 2017 ). The new life cycle of women’s employment: Disappearing humps, sagging middles, expanding tops . Journal of Economic Perspectives , 31 ( 1 ), 161 – 182 . doi: 10.1257/jep.31.1.161

Goldman , D. P. , & Orszag , P. R . ( 2014 ). The growing gap in life expectancy: Using the future elderly model to estimate implications for social security and Medicare . The American Economic Review , 104 ( 5 ), 230 – 233 . doi: 10.1257/aer.104.5.230

Gustman , A. L. , & Steinmeier , T. L . ( 2001 ). How effective is redistribution under the social security benefit formula? Journal of Public Economics , 82 ( 1 ), 1 – 28 . doi: 10.1016/S0047-2727(00)00153-5

Gustman , A. L. , Steinmeier , T. L. , & Tabatabai , N . ( 2013 ). Redistribution under the Social Security benefit formula at the individual and household levels, 1992 and 2004 . Journal of Pension Economics & Finance , 12 ( 1 ), 1 – 27 . doi: 10.1017/S1474747212000108

Herd , P. , Favreault , M. , Harrington Meyer , M. , & Smeeding , T. M . ( 2018 ). A targeted minimum benefit plan: A new proposal to reduce poverty among older Social Security recipients . The Russell Sage Foundation Journal of the Social Sciences , 4 ( 2 ), 74 – 90 . doi: 10.7758/RSF.2018.4.2.04

Iams , H. M. , Reznik , G. L. , & Tamborini , C. R . ( 2010 ). Earnings sharing in the U.S. Social Security system: A microsimulation analysis of future female retirees . The Gerontologist , 50 ( 4 ), 495 – 508 . doi: 10.1093/geront/gnq032

Iams , H. M. , & Tamborini , C. R . ( 2012 ). The implications of marital history change on women’s eligibility for Social Security wife and widow benefits, 1990–2009 . Social Security Bulletin , 72 ( 2 ), 23 – 38 .

Johnson , R. W. , Mudrazija , S. , & Wang , C. X . ( 2019 ). Older Latinos’ financial security: Resources, needs, and future prospects. In W. Vega , J. Angel , L. Gutiérrez Robledo , & K. S. Markides (Eds.), Contextualizing health and aging in the Americas (pp. 225 – 248 ). Springer . doi: 10.1007/978-3-030-00584-9_11

Knoll , M. A. Z. , Shoffner , D. , & O’Leary , S . ( 2018 ). The potential effects of obesity on social security claiming behavior and retirement benefits . The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences , 73 ( 4 ), 723 – 732 . doi: 10.1093/geronb/gbw016

Leimer , D . ( 2007 ). Cohort-specific measures of lifetime Social Security taxes and benefits (ORES Working Paper No. 110). Social Security Administration, Office of Policy . https://www.ssa.gov/policy/docs/workingpapers/wp110.html

Lopez , L. 3rd , Hart , L. H. 3rd , & Katz , M. H . ( 2021 ). Racial and ethnic health disparities related to COVID-19 . Journal of the American Medical Association , 325 ( 8 ), 719 – 720 . doi: 10.1001/jama.2020.26443

Mitton , L. , Sutherland , H. , & Weeks , M. (Eds.). ( 2000 ). Microsimulation modelling for policy analysis: Challenges and innovations . Cambridge University Press .

Montez , J. K. , Hummer , R. A. , Hayward , M. D. , Woo , H. , & Rogers , R. G . ( 2011 ). Trends in the educational gradient of U.S. adult mortality from 1986 to 2006 by race, gender, and age group . Research on Aging , 33 ( 2 ), 145 – 171 . doi: 10.1177/0164027510392388

National Academies of Sciences, Engineering, and Medicine . ( 2015 ). The growing gap in life expectancy by income: Implications for federal programs and policy responses . The National Academies Press . doi: 10.17226/19015

National Center for Health Statistics . ( 1994 ). Vital statistics of the United States, 1990, vol. II, sec. 6 life tables (DHHS Publication No. (PHS) 94-1104) . U.S. Department of Health and Human Services, Public Health Service, Centers for Disease Control and Prevention . https://www.cdc.gov/nchs/data/lifetables/life90_2acc.pdf

Olshansky , S. J. , Antonucci , T. , Berkman , L. , Binstock , R. H. , Boersch-Supan , A. , Cacioppo , J. T. , Carnes , B. A. , Carstensen , L. L. , Fried , L. P. , Goldman , D. P. , Jackson , J. , Kohli , M. , Rother , J. , Zheng , Y. , & Rowe , J . ( 2012 ). Differences in life expectancy due to race and educational differences are widening, and many may not catch up . Health Affairs (Project Hope) , 31 ( 8 ), 1803 – 1813 . doi: 10.1377/hlthaff.2011.0746

Panis , C. , & Lillard , L. A . ( 1999 ). Near term model development. Final Report (SSA Contract No. 600-96-27335) . RAND. https://www.ssa.gov/policy/docs/contractreports/NearTermModelDev.pdf

Reznik , G. L. , Couch , K. A. , Tamborini , C. R. , & Iams , H. M . ( 2019 ). Longevity‐related options for Social Security: A microsimulation approach to retirement age and mortality adjustments . Journal of Policy Analysis and Management , 38 ( 1 ), 210 – 238 . doi: 10.1002/pam.22103

Sandell , S. H. , & Iams , H. M . ( 1997 ). Reducing women’s poverty by shifting Social Security benefits from retired couples to widows . Journal of Policy Analysis and Management , 16 ( 2 ), 279 – 297 . doi: 10.1002/(SICI)1520-6688(199721)16:2<279::AID-PAM5>3.0.CO;2-F

Sasson , I . ( 2016 ). Trends in life expectancy and lifespan variation by educational attainment: United States, 1990–2010 . Demography , 53 ( 2 ), 269 – 293 . doi: 10.1007/s13524-015-0453-7

Smith , K. E. , & Favreault , M . ( 2019 ). Modeling income in the near term 8 and 2014 primer . The Urban Institute . https://www.urban.org/research/publication/modeling-income-near-term-8-and-2014-primer/view/full_report

Smith K. E. , Favreault , M. M. , Butrica , B. A. , & Issa , P . ( 2010 ). Modeling income in the near term version 6 . The Urban Institute . https://www.urban.org/research/publication/modeling-income-near-term-version-6/view/full_report

Smith , K. , Toder , E. , & Iams , H . ( 2003 ). Lifetime distributional effects of Social Security retirement benefits . Social Security Bulletin , 65 ( 1 ), 33 – 61 .

Social Security Administration . ( 2021 ). Online social security handbook . http://www.ssa.gov/OP_Home/handbook/handbook.html

Song , J. , & Manchester , J . ( 2009 ). Revisiting the 1983 Social Security reforms, 25 years later . Research on Aging , 31 ( 2 ), 233 – 260 . doi: 10.1177/0164027508328311

Tamborini , C. R. , Iams , H. M. , & Whitman , K . ( 2009 ). Marital history, race, and Social Security spouse and widow benefit eligibility in the United States . Research on Aging , 31 ( 5 ), 577 – 605 . doi: 10.1177/0164027509337196

Tamborini , C. R. , Kim , C. , & Sakamoto , A . ( 2015 ). Education and lifetime earnings in the United States . Demography , 52 ( 4 ), 1383 – 1407 . doi: 10.1007/s13524-015-0407-0

Tan , L. , & Koedel , C . ( 2019 ). The effects of differential income replacement and mortality on US Social Security redistribution . Southern Economic Journal , 86 ( 2 ), 613 – 637 . doi: 10.1002/soej.12395

Waldron , H . ( 2007 ). Trends in mortality differentials and life expectancy for male social security-covered workers, by socioeconomic status . Social Security Bulletin , 67 ( 3 ), 1 – 28 .

Xu , J. , Murphy , S. L. , Kochanek , K. D. , & Arias , E . ( 2020 ). Mortality in the United States, 2018 (NCHS Data Brief No. 355) . U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Health Statistics . https://www.cdc.gov/nchs/data/databriefs/db355-h.pdf

Zissimopoulos , J. M. , Tysinger , B. C. , St. Clair , P. A. , & Crimmins , E. M . ( 2018 ). The impact of changes in population health and mortality on future prevalence of Alzheimer’s disease and other dementias in the United States . The Journals of Gerontology, Series B: Psychological Sciences and Social Sciences , 73 ( Suppl. 1 ), S38 – S47 . doi: 10.1093/geronb/gbx147

• life expectancy
• social security
• socioeconomic factors

Supplementary data

Email alerts, citing articles via, looking for your next opportunity.

• X (formerly Twitter)
• Recommend to Your Librarian

Affiliations

• Online ISSN 1758-5368
• Copyright © 2024 The Gerontological Society of America
• About Oxford Academic
• Publish journals with us
• University press partners
• What we publish
• New features  
• Open access
• Institutional account management
• Rights and permissions
• Get help with access
• Accessibility
• Media enquiries
• Oxford University Press
• Oxford Languages
• University of Oxford

Oxford University Press is a department of the University of Oxford. It furthers the University's objective of excellence in research, scholarship, and education by publishing worldwide

• Copyright © 2024 Oxford University Press
• Cookie settings
• Cookie policy
• Privacy policy
• Legal notice

This Feature Is Available To Subscribers Only

Sign In or Create an Account

This PDF is available to Subscribers Only

For full access to this pdf, sign in to an existing account, or purchase an annual subscription.

Strengthening the Social Security safety net

Key takeaways.

• Almost one in every five Americans receives income support from Social Security.
• Social Security is financed on a pay-as-you-go basis, which means that today’s workers’ payroll taxes are used to pay benefits to today’s beneficiaries.
• Changes in population demographics, including longevity improvements, have resulted in a sharp decline in the number of workers per beneficiary , and this trend  is projected to continue.
• As a result, Social Security is facing significant financial challenges and policymakers must take action to ensure its long-term sustainability.

Social Security provides income security to more than 65 million Americans who are either retired or disabled or have experienced the death of a working spouse or parent. The program has reduced poverty among the elderly and given a measure of financial security to those who have paid into the program during the course of their careers. However, Social Security is facing significant financial challenges and policymakers must take action to ensure its long-term sustainability.

This policy brief will address how Social Security works and its importance for retirement.  It will describe how Congress changed the program in 1983 — the last time significant reforms were made — and outline the program’s current financial status as well as the differences between the risks facing Social Security now relative to 40 years ago. The policy brief will conclude with prescriptions and challenges for reforming the program.

Social Security:  The basics

Social Security was created in 1935, began collecting taxes in 1937, and paid its first benefits in 1940. Retirement benefits are calculated based on a worker's highest 35 years of earnings, where past earnings are first indexed upward for the growth of the national average wage. The earnings that determine Social Security contributions and benefits are subject to a taxable maximum, set at $160,200 in 2023. The maximum taxable salary is increased each year along with economy-wide wage growth. Retirement benefits are lowered for those who claim benefits before the full retirement age (67 for those born in 1960 and later), while those who delay claiming benefits receive an increase in their monthly payment.  Since benefits are paid until death, people who choose to claim early receive a lower monthly benefit amount for a longer expected time. 

Social Security was designed to provide a guaranteed income in retirement to protect seniors against the risk of outliving their savings. The program's benefit formula is designed to be progressive and replaces a higher percentage of pre-retirement earnings for lower lifetime earners than for higher lifetime earners.  While Social Security is not an explicit anti-poverty program — in that benefits are not based upon need — the progressive tilt of the Social Security benefit formula has reduced elderly poverty over time. Each year during retirement, the benefits are automatically increased by the rate of inflation.

Social Security is financed primarily on a pay-as-you-go basis, which means that payroll taxes collected from today’s workers are used to fund today's benefits. But Social Security also has two trust funds, the Old Age and Survivors Insurance (OASI) Trust Fund and the Disability Insurance (DI) Trust Fund. The trust funds hold the surplus funds generated by payroll taxes when the program is in surplus and draw down assets to pay benefits when the program runs deficits, as it has since 2010.

Payroll taxes are paid by employees and employers. The OASDI payroll tax rate is 6.2 percent for employees and 6.2 percent for employers for earnings below the taxable maximum ($160,200 in 2023). Self-employed individuals pay both the employee and employer portions of the payroll tax. Social Security also receives income from the taxation of Social Security benefits and interest earned on the trust funds’ assets.

The importance of Social Security in retirement

In 2015, Social Security represented 30 percent of income on average for individuals age 65 and over. Forty percent of all seniors in 2015 received 50 percent or more of their income from Social Security. Meanwhile, 14 percent of all seniors received 90 percent or more of their incomes from the program. [1] Social Security is a critical source of income for those who are widowed and low-income retirees, and Social Security represents a larger share of retirement income for women than for men.

Inequality in life expectancy has significant implications for the distributional consequences of Social Security. Lower-income individuals tend to have lower life expectancies than higher-income individuals, meaning that they may receive fewer years of Social Security benefits. This means that the increasing gap in life expectancy by income works against the progressivity in the benefit formula. [2]   

Figure 1 below shows life expectancy at birth for people at the 10th, 50th, and 90th percentile of the income distribution.  As shown in the figure, life expectancies are higher for those with higher income and, while life expectancies have increased for each group, the gains for those with more income have been larger than for those with less income.

Figure 1 :  Life expectancy by income percentile, 2001-2014

Source: Equality of Opportunity Project

The 1983 amendments and the financial status of Social Security

The 1983 congressional amendments were significant reforms to Social Security, designed to address the program's short- and long-term financial sustainability.

Due to severe inflation and lower-than-expected wages in the early 1980s, the Social Security trust funds were at immediate risk of not being able to pay scheduled benefits in 1983. Insolvency would have prompted short-term reductions to Social Security benefits, though the program was projected to return to solvency by 1990 as the large baby-boom generation reached its peak earning and taxpaying years.

The 1983 reforms addressed short-term solvency by accelerating an increase in the combined employee and employer payroll tax rate from 10.8 to 12.4 percent, previously scheduled to take place by 1990, and by delaying the 1984 Cost of Living Adjustment by six months, a step that amounted to a one-time reduction in current benefits of about 1.75 percent.

Long-term solvency was improved by scheduling a two-year increase in the full retirement age, from 65 to 67, to take place between 2000 and 2022, and making 50 percent of Social Security benefits subject to income taxes for beneficiaries with incomes above $25,000 in nominal terms, a provision that generates increasing revenue over time as nominal income increases raise the share of retirees whose benefits are subject to income taxes.

The 1983 amendments increased revenues for the program, which helped to address the program's long-term funding shortfall, which was a significant concern at the time.  However, by 1984 the Social Security trustees once again projected an actuarial deficit over a 75-year window, with a trust fund depletion date in the mid-2030s projected by the trustees as early as 1993.

The 2023 Trustees Report projected that the OASDI Trust Fund will be depleted in 2034 and that the program faces a long-run actuarial deficit equal to 3.61 percent of wages subject to payroll taxes. This means that the deficit over the 75-year window could be closed either with an immediate and permanent payroll tax rate increase of 3.61 percent; an immediate across-the-board benefit reduction of 21.3 percent; or some combination of current and future tax increases and benefit cuts. The Congressional Budget Office’s Long-Term Projections for Social Security show an even larger deficit of 4.9 percent of taxable payroll. The policy changes would need to be even more drastic in order to balance the program over a longer time horizon, as the shortfall is projected to grow larger over time.

After the trust funds are depleted, Social Security will still be able to pay a portion of promised benefits, with the level of payable benefits decreasing over time as revenues decline. The payable portion of benefits is projected to be 80 percent in 2034, declining to 74 percent by 2097, as shown in Figure 2 below.

Figure 2 :  Old-Age, Survivors, and Disability Insurance (OASDI) Income, Cost, and Expenditures as Percentages of Taxable Payroll

Source:  2023 Social Security Trustees Report, Figure II.D2.

In the 40 years since this last major reform to the Social Security program, people are living longer and fertility rates are declining, resulting in more retirees per worker.  These demographic shifts combined with the pay-as-you-go nature of Social Security — where today’s workers’ payroll taxes are used to fund today's beneficiaries’ benefits — result in a trend toward increasing annual deficits.

Figure 3 :  Fertility, Mortality, and the Number of Workers per Beneficiary

Source:  2023 Social Security Trustees Report, Figures V.A1, V.A5, II.D3.  Total Fertility Rate refers to the average number of children that would be born to a woman if she were to experience, at each age of her life, the birth rate observed in, or assumed for, the selected year, and if she were to survive the entire childbearing period.  Life Expectancy at Age 65 represents average remaining number of years expected prior to death for a person age 65, born on January 1, using the mortality rates for that year over the course of his or her remaining life.  Values beginning in 2022 are projected assuming Social Security’s intermediate cost scenario.

The effect of demographics on Social Security costs can be approximated in the following way. The average Social Security benefit in 2023 is projected to be equal to about 39 percent of the average worker’s wage in that year. If there are 2.7 workers for each beneficiary, as the Social Security actuaries estimate for 2023, the cost per worker is 14.4 percent of their wages (where 14.4 percent represents 39 percent divided by 2.7). If there were five workers per beneficiary, as there were in 1960, that same benefit could be funded at a cost of just 7.8 percent of worker’s wages. Likewise, if the worker to beneficiary ratio falls to 2.3, as is projected by the year 2035, the cost per worker rises to 16.9 percent of wages.

It is important to note that while the retirement of the baby-boom generation accelerated these trends, the underlying changes in mortality and fertility are projected to continue. Thus, Social Security is not expected to return to financial balance even after the baby-boom generation has passed on.

Principles for reforming Social Security

There are several principles that policymakers should consider when evaluating potential Social Security reforms. These include ensuring the program's long-term financial sustainability, maintaining the program's progressivity, protecting vulnerable populations, and avoiding sudden changes that could have significant impacts on beneficiaries.

However, these principles may come into conflict with each other. For instance, phasing in reforms slowly could reduce the impact of such changes on Social Security participants, but would make it more difficult to restore the program to long-run financial health. Likewise, a uniform benefit change that maintained the program’s current level of progressivity could reduce Social Security’s effectiveness at protecting the lowest-income beneficiaries.

Moreover, policymakers may consider more radical changes that alter the broader structure of Social Security. For instance, some countries, such as Australia and New Zealand, fund their retirement programs from general tax revenues rather than from a dedicated payroll tax. Likewise, those countries, as well as the United Kingdom and Canada, focus their retirement expenditures more closely on lower-income retirees while limiting benefits to higher-income seniors.

Barriers to reform and the cost of waiting

There are several barriers to reforming Social Security, including political polarization, the difficulty of making changes to a popular program, and the complexity of the program's finances and benefits.

Delaying Social Security reform could have significant costs. As the population ages, the number of beneficiaries will continue to grow, putting additional strain on the program's finances. Waiting too long to address Social Security's long-term funding shortfall could result in sudden benefit cuts or tax increases, which could have significant impacts on beneficiaries.

In this context, it is worth contrasting the financial situation that will face policymakers when the trust funds are projected to run dry in 2034 versus the conditions that prevailed in 1983, the last time Social Security faced insolvency.

While the 1983 funding crisis was driven by short-term economic changes that drew down the modest trust fund balance then held by the plan, Social Security’s projected insolvency in 2034 is largely the result of demographic changes that have been long in the making. At the same time, the 1983 funding shortfall was substantially more modest than what faces policymakers in 2034.

Had Social Security’s trust fund run out in 1983, Social Security beneficiaries would have faced benefit reductions of about 4 percent on average over a period of six years. By 1990, payroll taxes revenues from the large baby-boom generation would have returned Social Security to annual surplus and the trust funds were projected to remain solvent until between 2025 and 2034.

By contrast, if the trust funds are allowed to become exhausted in the 2030s, beneficiaries would face reductions of around 20 percent in their benefits, roughly five times larger in percentage terms than policymakers faced in 1983. Moreover, those cuts would increase over time, with practically zero possibility that the program would return to solvency on its own without structural changes. Thus, while the 1983 reforms were a model of bipartisan compromise, elected officials face an even larger task in addressing Social Security’s solvency today.

Social Security is an essential program that provides critical support to millions of retirees, survivors, and disabled individuals. However, the program's long-term financial sustainability is at risk due to demographic and economic changes. Policymakers must consider a range of potential reforms to ensure that the program can continue to provide vital support to current and future generations of beneficiaries.

While reforming Social Security is challenging, policymakers must act to address the program's long-term funding shortfall and ensure that the program can continue to meet its important mission.

[1] Dushi, Irena, and Brad Trenkamp. “Improving the measurement of retirement income of the aged population.” Social Security Administration. ORES Working Paper No. 116 (January 2021).

[2] See National Academies of Sciences, Engineering, and Medicine, and Committee on Population. “The growing gap in life expectancy by income: Implications for federal programs and policy responses.” National Academies Press, 2015.

About the Authors

• Gopi Shah Goda is a SIEPR Senior Fellow .  From July 2021 to July 2022, she served as a senior economist at the White House Council of Economic Advisers.  She conducts research that informs how policy can best serve aging populations.  She studies the sustainability of public programs serving the elderly, how individuals make healthcare, saving and retirement decisions as they age, and the broader implications of the COVID-19 pandemic on health, labor supply and entitlement programs. 
• Andrew G. Biggs is a Tad and Dianne Taube Policy Fellow at SIEPR. He is an expert on Social Security reform, state and local government pensions, and public sector pay and benefits. He is a senior fellow at the American Enterprise Institute. In the mid-2000s, Biggs was the principal deputy commissioner of the Social Security Administration, where he oversaw SSA’s policy research efforts.
• SIEPR Research Assistant Bradley Strauss contributed to this policy brief.

Related Topics

• Policy Brief
• Taxes and Public Spending

More Publications

The impacts of covid-19 on minority unemployment: first evidence from april 2020 cps microdata, winner curse, reserve prices and endogenous entry: empirical insights from ebay auctions, is high student loan debt always a problem.