The longevity dividend

The longevity dividend is the argument that a modest slowing of biological aging would yield health and economic benefits exceeding those of eliminating any single disease, and that public research funding should be reallocated accordingly. It was named in 2006 by S. Jay Olshansky, Daniel Perry, Richard Miller, and Robert Butler, who proposed that a seven-year delay in the onset of age-related decline would deliver larger population health gains than curing cancer or heart disease.¹ It is the economic version of the Geroscience hypothesis and the main policy case the aging-research field makes.

§The competing-risks argument

The technical core is simple and largely undisputed. Age-related diseases compete, so a person spared one dies of another. Modelling the elimination of Cancer produces a smaller gain in life expectancy than intuition suggests, because the cohort that survives it remains at high risk of cardiovascular disease, dementia, and frailty. Attacking the shared upstream process, described in the Hallmarks of aging, shifts the whole risk profile instead of removing one branch of it.

Why disease-by-disease has a ceiling

US cause-elimination life tables for 1999-2001 found that removing cancer as a cause of death would raise life expectancy at birth by more than three years, and removing heart disease by almost four.² Those are real gains, but far smaller than the diseases' salience suggests, because the people saved remain old.

The 2006 proposal promised more than health. Promoting it at a September 2006 symposium, the Alliance for Aging Research, a Washington advocacy group whose executive director Daniel Perry was a co-author, reported Olshansky as saying the dividend would include lower health care costs, higher savings and greater worker productivity, and stated that less than 0.1% of the National Institutes of Health's $28 billion 2006 budget went to the biology of aging.³ The health-cost part of that promise is the one later research has tested most directly.

§The Goldman simulation

A 2013 study by Dana Goldman, David Cutler, Olshansky and colleagues gave the argument its best-known quantitative form.⁴ It used the Future Elderly Model, a microsimulation of older Americans' health and spending, to compare a status quo against scenarios in which cancer or heart disease incidence fell by 25% between 2010 and 2030, and a hypothetical "delayed aging" scenario.

The delayed-aging assumptions are specific. Mortality and the onset of six chronic conditions and disability were cut by 1.25% for each year of life lived above age fifty, phased in from 2010 to 2030; the slope of intrinsic mortality at ages 15 to 50 fell by 20% by 2050. Every intervention was assumed to cost nothing, and cumulative figures were discounted at 3% a year. No such intervention exists; the authors call the scenario hypothetical.

Under those assumptions, life expectancy at age 51 in 2030 rose from 35.8 years to 38.0, against about one extra year in each disease scenario. Valuing the extra quality-adjusted life years at $100,000 each gave a social benefit of about $7.1 trillion over fifty years. The disease-specific scenarios showed diminishing returns, which the authors attributed mainly to competing risks.

The same model also asks whether slowing aging saves public money, and in it the answer is no. Per capita Medicare spending was lower in the delayed-aging scenario, but more people survived to qualify for programmes, so aggregate spending rose: delayed aging added nearly $420 billion to the entitlement deficit in 2060, 70% of it from Medicare and Medicaid. The authors found that raising the Medicare eligibility age from 65 to 68 and the Social Security normal retirement age from 67 to 68 would more than offset the extra cost.

§The large valuations

The most-cited figures come from a 2021 analysis by Andrew Scott, Martin Ellison, and David Sinclair in Nature Aging.⁵ Their model is calibrated to US data, with life expectancy at birth of 78.9 years, healthy life expectancy of 68.5 years and an average value of statistical life of US$11.5 million for adults aged 25 to 65. Summed over the 2017 US population and future births, a slowdown in aging that adds one year of life expectancy is worth US$37.6 trillion and one that adds ten years US$366.8 trillion, usually rounded to $38 trillion and $367 trillion. Using a 2% interest rate, the authors convert the ten-year figure to US$7.2 trillion a year, 33.6% of 2019 GDP.

The paper compares three idealised cases. In the "Struldbrugg" case, named for Swift's immortals, life is extended while health declines at the usual rate; in the "Dorian Gray" case health improves within a fixed lifespan; in the "Peter Pan" case aging itself slows. An extra year of healthy life expectancy was always worth more than an extra year of life expectancy, and slowing aging was worth more than extending life alone. The Struldbrugg values were lower but still positive, so the model does not show that extending life in poor health has negative value.

The authors also simulated Metformin and the TAME trial, using effect sizes from an observational study of 41,204 men with diabetes whose average age was 75, and found benefits often matching or exceeding complete eradication of cancer, dementia or cardiovascular disease. They state that metformin's efficacy awaits confirmation from trials such as TAME, and that their dollar values are sensitive to calibration while the ranking of treatments is not. They describe a "virtuous circle" in which better aging raises the value of further improvement, although in their own table a second slowdown in 2050 is worth about 1-2% less than the first for small gains. Sinclair discloses commercial interests in several longevity companies.

§How the numbers are made

The valuations rest on the value of a statistical life, a figure derived from what people pay to reduce small risks of death, such as wage premiums for dangerous work. Scott and colleagues chose it partly because government agencies already use it to evaluate policies.⁵ Goldman's team used a simpler price per quality-adjusted life year. Either way, multiplying a per-person value by a population produces very large aggregates very quickly.

The resulting number is a measure of welfare, not output. It is an aggregate of what people would notionally pay for a risk reduction; it does not appear in an economy as money, it is not tax revenue, and no government could spend it as a budget. Setting a $38 trillion dividend beside national GDP compares quantities of different kinds, which is why the annualised GDP share in the Scott paper is a scale reference rather than a forecast of income.

A separate and more tractable question is the fiscal one: what happens to public budgets. The Goldman simulation suggests that health spending per person falls while total spending and pension outlays rise, and that in the one fix the authors modelled, raising the Medicare and Social Security eligibility ages, the gap more than closes. The dividend argument assumes such changes will happen, and that assumption concerns institutions rather than biology.

§Would slowing aging cut health spending?

A popular version of the argument holds that slowing aging would make most medical care unnecessary and that governments would fund it because young bodies are cheaper. The evidence supports a narrower claim.

Spending does concentrate in old age. In the United States in 2020, people aged 65 and over were 17% of the population and accounted for 37% of personal health care spending, $22,356 per person; per capita spending on those over 85 ($35,995) was more than 8.5 times that on children ($4,217).⁶ The other 63% went to people under 65, and the 2020 figures do not separate out the effects of COVID-19. Globally, 92 diseases whose adult incidence rises steeply with age accounted for 51.3% of the disease burden among adults in 2017, measured in disability-adjusted life years.⁷ On the geroscience premise, slowing aging would act mainly on that half; most injuries and infections, mental disorders and maternity care fall outside it.

Much of the age gradient reflects closeness to death. A 1999 Swiss study of people who died between 1983 and 1992 found that spending in the last two years of life depended on remaining lifetime rather than calendar age, at least beyond 65, and its authors called population aging a possible "red herring" for health spending.⁸ English hospital data later showed proximity to death acting as a proxy for morbidity.⁹ On this reading, postponing death moves the costly final period later instead of removing it.

Lifetime totals point the same way. Using US Medicare survey data, a 70-year-old with no functional limitation had a life expectancy of 14.3 years and expected cumulative health spending of about $136,000 (1998 dollars), against 11.6 years and about $145,000 for someone limited in at least one activity of daily living.¹⁰ A Dutch simulation found lifetime health spending highest for healthy-living nonsmokers, lowest for smokers and intermediate for obese people, because the healthier groups lived longer and incurred costs for unrelated diseases in the added years.¹¹ Healthier lives may cost about the same as less healthy ones, or more; the economic case for slowing aging is the value of the extra healthy years, not a smaller health budget.

§Counterarguments

Morbidity has not clearly compressed. James Fries proposed in 1980 that medical progress would push the onset of disability toward a fixed lifespan limit, compressing illness into a short terminal period.¹² Scott and colleagues read Global Burden of Disease data as showing that the proportion of life in good health has stayed broadly constant, implying more years in poor health.⁵ Compression of morbidity examines the evidence. If geroprotective drugs behave like earlier medical advances, they would extend both curves together, the Struldbrugg-like outcome the valuations price lowest.

Pensions. Adding years at the end without changing retirement norms raises pension outlays; the demographic causes of population aging are covered in Overpopulation and life extension. By 2023 one in four OECD countries linked statutory retirement ages to life expectancy, among them Denmark, Estonia, Finland, Greece, Italy, the Netherlands and Portugal.¹³ Denmark's one-to-one link projects a normal retirement age of 74 for a man entering the labour market in 2022. The OECD notes that the Netherlands moved to a slower two-thirds link before its link took effect and warns that the political sustainability of one-to-one links may be weak; discussions were under way in Denmark on moving to a slower link. Automatic indexing is the policy the dividend argument implies, and its durability is untested.

There is no drug to fund. The dividend is an argument for research rather than a description of an available intervention, and it should not be confused with the much stronger claim examined in Longevity escape velocity. No pharmaceutical has been shown to slow aging in humans, and in 2024 NPR reported that the US Food and Drug Administration did not recognise aging as a disease to treat.¹⁴ The GLP-1 receptor agonists, sometimes offered as a counter-example, treat metabolic disease rather than aging. Metformin has an observational signal and a trial designed to test it, TAME, planned to enrol 3,000 people aged 65 to 79 for six years; in 2024 NPR reported that the National Institute on Aging had set aside about $5 million against an estimated cost of $45 million to $70 million, and Nir Barzilai said that because metformin is generic, no company stands to profit from the trial; he had turned to philanthropists and foundations and had some pledges. Senolytics have produced modest or null results in the human studies reported so far; Rapamycin has strong mouse data and no completed long-term human outcome trial; unregulated practices such as sauna and cold-water immersion rest on observational cohorts and short-term physiological measures rather than outcome trials. The best-supported geroprotector remains the one described in Exercise as a geroprotector, which requires no research funding at all.

Distribution. A dividend accruing mostly to populations that already live longest would widen the existing gap in life expectancy between rich and poor. Nothing in the argument's structure addresses who receives the years, a problem developed in Access and inequality.

What the estimate is for

Supporters treat the trillion-dollar figures as evidence that aging research is systematically underfunded relative to its expected return. Critics note that the same method applied to any large mortality reduction produces similar numbers, so the figure demonstrates that mortality is valuable rather than that geroscience will deliver. Neither side's figures show a fiscal saving: the one US simulation cited here with a budget line finds higher total entitlement spending unless eligibility ages rise.

§Policy implications

The dividend argument has had two concrete effects. First, it supplied the rationale for treating aging as a legitimate indication in regulatory terms, the reasoning behind the proposed TAME trial, structured less to demonstrate efficacy than to establish that a regulator would accept an aging-related composite endpoint at all. Second, it shaped prize competitions and philanthropic funding: the endpoints chosen by XPRIZE Healthspan are functional restoration in muscle, cognition, and immunity rather than lifespan, precisely because functional years are what the economics values.

Its effect on public research budgets is harder to show. The disease-specific structure of medical research funding has proved resistant to the competing-risks argument, and the fiscal results above suggest why a finance ministry might hesitate: the welfare gain is large and diffuse, while the extra pension and health outlays land on public budgets unless eligibility ages move.

§The unresolved question

The argument turns on whether an intervention that extends life also extends health, and the field has no accepted way to measure that in advance. Aging biomarkers describes the search for a surrogate endpoint that regulators would accept; without one, a trial testing the dividend's central premise would need to run for decades on hard outcomes. Until such a trial exists, the economic case rests on an assumption about biology that the biological evidence has not yet confirmed, and the fiscal case rests on a second assumption, about retirement ages, that politics has not yet confirmed either.

§See also

• Geroscience hypothesis
• Compression of morbidity
• Healthspan
• Access and inequality
• Overpopulation and life extension
• Aging biomarkers
• XPRIZE Healthspan
• Longevity escape velocity

§References

Footnotes

1. paperOlshansky, S.J., Perry, D., Miller, R.A., Butler, R.N. "In Pursuit of the Longevity Dividend." The Scientist, 20, 28-36, 2006.↩A position piece proposing a research policy; the seven-year delay is a stipulated scenario rather than a result from any study.

2. reportArias, E., Heron, M., Tejada-Vera, B. "United States Life Tables Eliminating Certain Causes of Death, 1999-2001." National Vital Statistics Reports, 61(9), National Center for Health Statistics, 2013. PMID 24968617↩Period life tables for 1999-2001; removing a cause in a table assumes other causes are unaffected.

3. statementAlliance for Aging Research. "In Pursuit of the 'Longevity Dividend'." Press release, 12 September 2006. https://www.agingresearch.org/news/in-pursuit-of-the-longevity-dividend/↩An advocacy group's announcement of the proposal; the NIH funding share is its own figure, not an audited breakdown.

4. paperGoldman, D.P., Cutler, D., Rowe, J.W., Michaud, P.C., Sullivan, J., Peneva, D., Olshansky, S.J. "Substantial Health and Economic Returns from Delayed Aging May Warrant a New Focus for Medical Research." Health Affairs, 32(10), 1698-1705, 2013. doi:10.1377/hlthaff.2013.0052↩A microsimulation of US scenarios, not a trial; the delayed-aging case assumes a cost-free intervention that does not exist.

5. paperScott, A.J., Ellison, M., Sinclair, D.A. "The economic value of targeting aging." Nature Aging, 1, 616-623, 2021. doi:10.1038/s43587-021-00080-0↩ ↩² ↩³The trillion-dollar figures are aggregated willingness to pay for reduced mortality risk, not output, revenue or any sum a budget could contain.

6. reportCenters for Medicare & Medicaid Services. "U.S. Personal Health Care Spending by Age and Sex: 2020 Highlights." National Health Expenditure Data.↩Covers personal health care only; CMS notes the 2020 estimates do not separate COVID-19 effects from demographic factors.

7. paperChang, A.Y., Skirbekk, V.F., Tyrovolas, S., Kassebaum, N.J., Dieleman, J.L. "Measuring population ageing: an analysis of the Global Burden of Disease Study 2017." Lancet Public Health, 4(3), e159-e167, 2019. doi:10.1016/S2468-2667(19)30019-2↩Age-related diseases are defined statistically, as those whose adult incidence rises quadratically with age.

8. paperZweifel, P., Felder, S., Meiers, M. "Ageing of population and health care expenditure: a red herring?" Health Economics, 8(6), 485-496, 1999. PMID 10544314↩Data on people who died in 1983-1992; the finding concerns spending in the last eight quarters of life, beyond age 65.

9. paperHowdon, D., Rice, N. "Health care expenditures, age, proximity to death and morbidity: Implications for an ageing population." Journal of Health Economics, 57, 60-74, 2018. doi:10.1016/j.jhealeco.2017.11.001↩English inpatient hospital spending only, in two samples of about 40,000 people who died within the study window.

10. paperLubitz, J., Cai, L., Kramarow, E., Lentzner, H. "Health, life expectancy, and health care spending among the elderly." New England Journal of Medicine, 349(11), 1048-1055, 2003. doi:10.1056/NEJMsa020614↩Medicare Current Beneficiary Survey data from 1992-1998, simulated from age 70; dollar figures are in 1998 dollars.

11. papervan Baal, P.H. et al. "Lifetime medical costs of obesity: prevention no cure for increasing health expenditure." PLoS Medicine, 5(2), e29, 2008. doi:10.1371/journal.pmed.0050029↩A simulation model built mainly on Dutch data, covering health care costs and not pensions or productivity.

12. paperFries, J.F. "Aging, Natural Death, and the Compression of Morbidity." New England Journal of Medicine, 303, 130-135, 1980. doi:10.1056/NEJM198007173030304 ↩

13. reportOECD. Pensions at a Glance 2023: OECD and G20 Indicators. OECD Publishing, Paris, 2023. doi:10.1787/678055dd-en↩Projected retirement ages assume legislated life-expectancy links are applied as written and projected gains materialise.

14. newsAubrey, A. "A cheap drug may slow down aging. A study will determine if it works." NPR, 22 April 2024.↩Funding and cost figures are as reported in April 2024; the trial's later funding status is not covered.