Public Longevity Infrastructure

Population ageing is usually described as a single global trend. It is not. The OECD's Pensions at a Glance 2025 publishes the demographic old-age to working-age ratio — people aged 65 and over per 100 people aged 20 to 64 — drawing on UN World Population Prospects 2024, and the spread between countries is larger than the change over time in most of them.

The policy consequence is that no two governments in the table below face the same problem. Japan and Italy are managing a shift that has already happened. Korea and China face the steepest transitions ever recorded. Gulf states are building longevity infrastructure well ahead of the demographic pressure rather than in response to it — Saudi Arabia's ratio in 2024 was 4.5, roughly one-twelfth of Japan's.

The morbidity gap is the difference between how long people live and how long they live in good health. It is the number that public longevity policy exists to reduce, and every recent high-quality analysis finds it moving the wrong way.

The GBD authors also report where the extra unhealthy years appear. The widening occurred across the adult life course rather than concentrating in the final years of life — which undercuts the comfortable assumption that additional survival simply appends a short period of frailty at the very end. Globally in 2023 an average of 14.5% of life was spent in poor health, compared with 13.6% in 1990.

IHME, reporting on the same study, notes that global life expectancy rose from 64.6 years in 1990 to 73.8 in 2023 while healthy life expectancy rose from 55.9 to 63.1 — that is 9.2 extra years of life against 7.2 extra years of healthy life. The surplus is the gap.

A second, independent line of evidence reaches the same place from different data. Garmany and Terzic, using WHO Global Health Observatory data across 183 WHO member states (JAMA Network Open 2024, PMID 39661386), found the healthspan–lifespan gap had widened globally over two decades to 9.6 years, with women showing a mean gap 2.4 years wider than men, and the United States showing the largest gap at 12.4 years, underpinned by a rise in non-communicable diseases.

Their companion regional analysis (Communications Medicine 2025) reports a global median gap of 9.1 years, ranging from 6.5 years in Lesotho to 12.4 years in the United States, with a European median of 9.9 years, and finds non-communicable diseases accounting for 56% to 90% of total disease burden across regions.

Three peer-reviewed analyses, two data sources, one direction of travel. The claim that the gap is roughly a decade wide is well supported. The claim that it is closing is not supported by any of them.

The table below separates two things that press coverage routinely merges: the stated commitment, and the status of delivery against it. Where a government has published outcome data, that data is cited. Where it has not, the entry says so rather than filling the space with an announcement.

This is the best-evidenced material in the whole field and the least reported. Within a single high-income country, the deprivation gap in healthy life expectancy exceeds the healthy-life-expectancy gap between most pairs of comparable countries, and is roughly double the global healthspan–lifespan gap. Every national average on this page conceals it.

The clearest measurement comes from the Office for National Statistics. In its release of 4 July 2025 covering England and Wales in 2020–2022, the slope index of inequality — the gap across the full deprivation distribution — was 19.1 years for men and 20.2 years for women in healthy life expectancy, against 10.7 and 8.5 years in life expectancy. Deprivation costs roughly twice as many healthy years as it costs total years. A policy scorecard that tracks only life expectancy understates the inequality by about half.

The pattern is not unique to Britain. Japan's National Institute of Health and Nutrition reports a 2019 prefectural gap in healthy life expectancy of 2.33 years for men and 3.90 years for women — smaller in absolute terms, but persistent through a period in which the national figure improved. Across the EU, Eurostat's 2023 healthy-life-years data ranges from 51.2 years for men in Latvia to 71.7 in Malta: a spread of about twenty years between countries, comparable in size to the deprivation gap inside England alone.

WHO's World Report on Social Determinants of Health Equity, published on 6 May 2025 and the first such report since the 2008 Commission, provides the analytical spine. It finds that people in the highest-life-expectancy country live approximately 33 years longer on average than those in the lowest; that within countries life expectancy varies by decades depending on area and social group, with these internal inequalities often expanding rather than shrinking; and that within-country income inequality has nearly doubled over two decades and now exceeds between-country inequality.

One caution against over-claiming. Between the world's extremes the gap is still larger than any within-country gap — 33 years versus roughly 20. The defensible claim is the narrower one: inside a single high-income country, the deprivation gap in healthy life expectancy is roughly double that country's own life-expectancy gap, roughly double the global healthspan–lifespan gap, and larger than the gap between most pairs of peer countries.

There is also a global distributional fact that has to be reported alongside the GBD finding, not instead of it. WHO projects that 80% of older people worldwide will live in low- and middle-income countries by 2050, while GBD 2023 finds morbidity gaps largest in the high-SDI quintile. These are not in conflict — high-SDI countries have converted more mortality risk into survivable chronic disease — but either fact alone misleads.

The headline figure in this debate comes from a systematic review by Masters and colleagues (Journal of Epidemiology and Community Health 2017, PMID 28356325), which screened 2,957 titles and included 52 studies. Its medians are frequently quoted; its own caveats almost never are.

The strongest counter-argument is not rhetorical, it is mechanical. van Baal and colleagues (PLoS Medicine 2008, PMID 18254654) simulated a Dutch cohort from age 20 and found that until age 56 annual health expenditure was highest for obese people, and at older ages smokers incurred higher costs — but that because of differences in life expectancy, lifetime health expenditure was highest among healthy-living people and lowest for smokers, with obese individuals in between. Their conclusion: although effective obesity prevention lowers the costs of obesity-related disease, that decrease is offset by cost increases from diseases unrelated to obesity in the life-years gained, so obesity prevention may be an important and cost-effective way of improving public health, but it is not a cure for rising health expenditure.

Cohen, Neumann and Weinstein (New England Journal of Medicine 2008, PMID 18272889) argued from the Tufts Cost-Effectiveness Analysis Registry that sweeping claims about the cost-saving potential of prevention are overreaching, and that the distributions of cost-effectiveness ratios for preventive measures and for treatments are similar — prevention is not systematically cheaper than treatment. A specific percentage is often attributed to this paper; it could not be verified from the primary text, so it is not repeated here.

The measurement literature has since turned on itself. Turner and colleagues (BMJ Global Health 2023, PMID 37648275) reviewed 118 return-on-investment studies from 2018–2021 and found that analyses counting only fiscal savings and analyses incorporating monetised health benefits are fundamentally different quantities routinely reported under the same label, that the methodologies used were inconsistent and often poorly reported, and that return-on-investment metrics should be carefully interpreted before being used to inform resource-allocation decisions.

There is one setting where a large return is comparatively well established. PAHO/WHO states that every US$1 invested in the WHO Best Buys in low- and lower-middle-income countries will yield a return of at least US$7 by 2030. That figure applies to countries where baseline coverage of cheap interventions is low, and 85% of premature non-communicable disease deaths occur; it should not be extrapolated to high-income health systems.

Raising retirement ages is the most direct fiscal response to population ageing, and it is genuinely contested. What follows separates what has been legislated, what the arithmetic implies, and where researchers and policymakers disagree.

From OECD's Pensions at a Glance 2025: the average normal retirement age across OECD countries will increase from 64.7 years for men and 63.9 for women retiring in 2024, to 66.4 and 65.9 respectively for those starting their career in 2024. The normal retirement age will rise in more than half of OECD countries under current legislation. Future ages range from 62 in Colombia, Luxembourg and Slovenia to 70 or more in Denmark, Estonia, Italy, the Netherlands and Sweden. Denmark, Estonia, Finland, Greece, Italy, the Netherlands, Portugal, the Slovak Republic and Sweden have legislated links between the pension age and life expectancy — under which Denmark moves from 67 to 74, Estonia from 64.8 to 71 and Italy from 63 to 70.

OECD also supplies the neutral framing of proportionality: remaining life expectancy of men at age 65 is projected to increase on average from 18.5 to 22.7 years, so the average legislated increase in men's normal retirement ages accounts for slightly more than 40% of the average projected increase in old-age life expectancy. Roughly 40% of the gain is absorbed by working longer; roughly 60% remains as additional retirement time.

Whatever is legislated, delivery depends on a workforce that does not currently exist at the required scale. WHO's health workforce estimate is a projected shortfall of 11 million health workers by 2030, concentrated in low- and lower-middle-income countries — precisely where WHO expects 80% of the world's older people to live by 2050.

Reported as modelling rather than as fact, this literature is genuinely useful — and its most policy-relevant conclusion is not the headline number. Scott, Ellison and Sinclair find that a compression of morbidity that improves health is more valuable than further increases in life expectancy, and that targeting ageing yields larger economic gains than eradicating individual diseases. That aligns directly with the GBD 2023 evidence: the gap, not the lifespan, is where the value is.

Goldman and colleagues (Health Affairs 2013, PMID 24101058) modelled delayed ageing on US data using the Future Elderly Model, finding it could add 2.2 years of life expectancy, most of it in good health, with an economic value estimated at $7.1 trillion over fifty years, while tackling heart disease and cancer separately would yield diminishing improvements by 2060 because of competing risks. The authors state plainly that delayed ageing would greatly increase entitlement outlays, especially for Social Security, and suggest these could be offset by raising the Medicare eligibility age and the Social Security normal retirement age. The leading economic model of the longevity dividend therefore assumes retirement-age increases as part of the package — presenting the dividend without that assumption misrepresents the source.

The counterweight comes from inside the same field. Olshansky and colleagues (Nature Aging 2024, PMID 39375565), analysing 1990–2019 vital statistics from the eight longest-lived national populations plus Hong Kong and the US, report that improvements overall in life expectancy have decelerated since 1990, that survival to age 100 is unlikely to exceed 15% for females and 5% for males, and that unless the processes of biological ageing can be markedly slowed, radical human life extension is implausible in this century. Olshansky is a co-author of the delayed-ageing paper above. Both positions are consistent: the dividend is conditional on actually slowing biological ageing, and presenting either alone is less honest than presenting both.