Longevity Financial Products
People are living longer than the pension and insurance systems built in the last century assumed, and that changes the arithmetic of paying for a very long old age. This page explains how the financial products built around that problem actually work — it is general information, never advice about what any individual should do.
"Longevity risk" covers two different things: one person outliving their own money, and a whole population outliving the mortality table an insurer, pension scheme or government used. The first is poolable, which is what life annuities, defined benefit pensions and collective defined contribution schemes do; the second cannot be pooled away and is instead transferred along a chain that ends in a small number of global reinsurers. Almost every product marketed with the word "longevity" is either an old mechanic under a new name or a savings product that pools no mortality at all — the test is whether any payment is contingent on the person being alive. CureMed is not authorised to give financial advice anywhere, and nothing here is a recommendation.
- Individual longevity risk is idiosyncratic and therefore poolable; aggregate longevity risk is systematic and can only be transferred, never eliminated — someone always holds it.
- Period life expectancy freezes one year's mortality rates forever; cohort life expectancy adds projected future improvement. The gap is around 8 to 10 years at birth for England and Wales cohorts born 1950–2000, but only about 0.7–1.0 years at age 65 on OECD projections for 2065 — transplanting the at-birth gap onto age 65 fabricates a statistic.
- A life annuity can pay more than a bond of the same duration because of the mortality credit: redistribution from those who die early to those who survive. That single mechanic explains every one of its tradeoffs, including the loss of the capital at death.
- Moving from defined benefit to defined contribution pensions did not reduce longevity risk; it disaggregated it and moved it onto individuals, who cannot pool it. DC assets are 63% of pension assets in the seven largest markets (Thinking Ahead Institute, 2026).
- The US healthspan–lifespan gap is the largest of 183 WHO member states at 12.4 years (JAMA Network Open, 2024), and longevity and care costs compound rather than offset: the same event that lengthens the drawdown horizon also raises the expected care bill.
Two different risks share the name "longevity risk"
The phrase "longevity risk" is used for two structurally different problems, and conflating them is the source of most confusion about what the products in this section do.
The Center for Retirement Research at Boston College defines the individual version narrowly as running out of assets because of a longer than expected lifespan, noting that a broader definition is simply the risk of running out of assets while alive (Arapakis & Wettstein, 2023 — a report funded by Jackson National Life Insurance Company, disclosed here because the funder sells annuities). The aggregate version is what an insurer, pension scheme or government faces when an entire population outlives the mortality table it priced against.
The difference is not academic. One of these risks disappears when you pool enough people; the other does not, no matter how many people you add. That is the structural key to everything below.
Why the distinction determines what a product can do
- Individual longevity risk is idiosyncratic: any one person's date of death is unpredictable, but the average across a large group is not. This is what makes mortality pooling work at all.
- Aggregate longevity risk is systematic: if the whole population improves faster than the table assumed, every member of the pool is affected in the same direction at the same time. Adding more members does not diversify it.
- Because aggregate risk cannot be pooled away, it can only be moved. Every institutional instrument in this section — buy-ins, buyouts, longevity swaps, reinsurance — relocates it rather than removing it.
- A defined contribution pot pools nothing at all. It is a stock of money, and the individual holds both risks alone unless they buy or join something that pools.
Retirement ages are increasingly linked to life expectancy by statute, which converts a demographic measurement into a benefit rule. The OECD reports that the eight countries with the highest future normal retirement age all link retirement age to life expectancy — Denmark, Estonia, Italy, the Netherlands, Sweden, Finland, Portugal and the Slovak Republic — with Greece also operating a link and Norway expected to introduce one at two-thirds of life-expectancy gains. Denmark shows how these links behave in practice: parliament confirmed the increase to 70 from 2040 in May 2025, while the OECD notes Denmark may soften the current one-to-one link, in which case the projected future age would be lower than 74.
Cohort vs period life expectancy — the number almost everyone reads wrong
Almost every headline life-expectancy number is a period figure, and a period figure is not a forecast of anyone's life. Understanding why is the single most useful thing this section can teach, because the mistake propagates into every downstream calculation about how long money has to last.
Period life expectancy takes the age-specific mortality rates observed in one year (or a short group of years) and asks a hypothetical question: how long would someone live if those exact rates applied, unchanged, for the rest of their life? The ONS puts it plainly — period life expectancies use mortality rates from a single year or group of years and assume those rates apply throughout the remainder of a person's life. It is a summary of this year's mortality experience, wearing the grammar of a prediction.
Cohort life expectancy takes a specific birth cohort and combines observed mortality rates with projected future rates for the years that cohort will actually live through. Because age-specific mortality rates have generally fallen over time, and a period table freezes that improvement at zero, cohort figures are higher. The ONS regards cohort life expectancy as a more appropriate measure of how long a person of a given age would be expected to live on average.
Verified magnitudes — every figure labelled period or cohort, with age, country and year
| Measure | Basis | Figure | Source |
|---|---|---|---|
| Gap between cohort and period life expectancy at birth, England & Wales, cohorts born 1950–2000 | cohort minus period, at birth | Around 8 to 10 years for both sexes | ONS, Period and cohort life expectancy explained (2023) |
| UK life expectancy at birth, 2021–2023 | period | 78.8 yrs male, 82.8 yrs female | ONS, National life tables: UK, 2021 to 2023 (2024) |
| UK life expectancy at birth for those born 2023, 2022-based projections | cohort | 86.7 yrs boys, 90.0 yrs girls | ONS, Past and projected period and cohort life tables, 2022-based (2025) |
| UK life expectancy at age 65, 2021–2023 | period | 18.5 yrs male, 21.0 yrs female | ONS, National life tables: UK, 2021 to 2023 (2024) |
| OECD-average life expectancy at age 65, 2024 | period | 18.5 yrs men, 21.6 yrs women (implied age at death 83.5 and 86.6) | OECD, Pensions at a Glance 2025 |
| Cohort-versus-period uplift at age 65 in 2065, OECD average | cohort minus period, at 65 | +1.0 yrs women, +0.7 yrs men | OECD, Pensions at a Glance 2025 |
Three further properties of these numbers
- A life expectancy is a mean, not a plan. Roughly half of any cohort outlives it, so a calculation built to the average is, by construction, insufficient for about half the people it describes. That is a mechanical property of averages, not a claim about what anyone should do.
- People are roughly right about the middle of the distribution and systematically wrong about its tail. Asked directly how long they expect to live, older individuals are approximately correct on average; asked about survival probabilities, the same people are pessimistic before roughly age 75 and optimistic after it, and those under 80 underestimate the probability of living to a very old age (Arapakis & Wettstein, 2023). The CRR notes it is precisely that tail risk which longevity products insure.
- One documented mechanism for the error is anchoring on the wrong number: media report life expectancy at birth, which is the wrong anchor for someone already at 65. The CRR also notes that parental age of death has been a poor predictor of an individual's mortality.
- Projections are revised, in both directions. The UK's Continuous Mortality Investigation raised cohort life expectancy at 65 by about three months for males and two weeks for females moving from CMI_2023 to CMI_2024, and by about eight weeks for males and six weeks for females moving to CMI_2025 — the latter attributed to record-low all-age mortality in England and Wales in 2025, down 2% on 2024. An earlier release had shown a fall.
- The improvement trend itself has slowed. Across 38 OECD countries the trend in period life expectancy at 65 rose about 1.6 years per decade for men and 1.4 for women from the mid-1990s to 2012; since about 2012 that pace has almost halved, to 0.9 and 0.8 years per decade, with the OECD identifying an estimated structural break in the series after 2012.
Annuities: the mortality credit, the product map, and the puzzle
An insurer collects premiums from many people of similar age and pays income only while each is alive. Those who die early leave balances behind, and those balances fund payments to survivors. The extra return survivors receive above what a bond of the same duration would pay is the mortality credit, sometimes called a survivorship credit.
This is why a life annuity can pay more than a bond portfolio of similar risk: the excess is not investment return, it is redistribution from the deceased to the living. That one mechanic also explains every tradeoff the product carries — the capital is gone at death, which is the price of the credit, and every feature that guarantees money back at death reduces the credit that made the product distinctive in the first place.
The theoretical literature is old and well established. Yaari (1965) showed that within his model, full annuitisation is optimal for a utility-maximising consumer with no bequest motive facing actuarially fair prices — a statement about a mathematical model's output under its own assumptions, not about any real person, whose circumstances the model does not describe. Davidoff, Brown and Diamond (2005) showed the result survives the relaxation of many of Yaari's restrictions, with substantial rather than necessarily full annuitisation remaining optimal under much wider conditions.
Product map — mechanics and tradeoffs, not a ranking
| Product | Core mechanic | Longevity risk pooled? | Principal tradeoffs |
|---|---|---|---|
| SPIA (single-premium immediate annuity) | Lump sum exchanged for income beginning immediately, paid for life | Yes | Irreversible; no bequest of that capital; a level version erodes in real terms; insurer credit risk; income depends on yields prevailing at purchase |
| Deferred income annuity (DIA), marketed as a "longevity annuity" | Premium paid now, income begins at a set future date, often 75–85, then paid for life | Yes — the most concentrated form | Nothing paid if death precedes the start date, absent a return-of-premium rider; access to the money is given up for the whole deferral period; very long insurer credit exposure |
| QLAC | A DIA inside a US qualified account, excluded from the balance used to calculate required minimum distributions until payments start | Yes | All DIA tradeoffs, plus statutory dollar limits and a deadline of age 85 for payments to begin |
| Escalating / inflation-linked annuity | Lifetime income rising annually by a fixed percentage or an index | Yes | A materially lower starting income; genuinely inflation-linked versions are, in the NBER authors' words, often scarce and expensive |
| Fixed-rate deferred annuity | Accumulates at a declared rate | No — not until annuitised | Surrender charges; rate resets. This is a savings vehicle wearing the word "annuity", not longevity insurance |
| Fixed indexed annuity (FIA) | Credited interest linked to an index, subject to caps, participation rates and spreads, with a floor of 0% | No — not until annuitised | Upside limited by parameters the insurer sets and can change on renewal; dividends typically excluded; surrender charges |
| RILA (registered index-linked, "buffer" or "structured" annuity) | Index-linked return with a buffer (insurer absorbs the first X% of loss) or floor, plus a cap on gains; an interim value adjustment applies to mid-term withdrawals | No — not until annuitised | Principal is genuinely at risk beyond the buffer; parameter complexity; a 10% buffer on a 15% index fall still leaves a 5% loss |
| Variable annuity with living-benefit rider | Sub-account investing with optional contractual guarantee floors | Partly — via the guarantee | Layered fees (mortality and expense charges plus rider charges); complex guarantee terms; insurer credit risk |
| Traditional long-term care insurance | Reimburses or indemnifies qualifying care costs | Insures morbidity, not longevity directly | Guaranteed renewable, so in-force premiums can be raised for a whole class; carrier exits; strict underwriting; use-it-or-lose-it |
| Hybrid life or annuity contract with an LTC rider | Death benefit or annuity value made accessible for qualifying care | Partly | Higher upfront cost; a capped benefit pool rather than open-ended cover; opportunity cost of the capital committed |
| Individual life insurance | Pays on death | Insures mortality — the opposite risk | Underwriting classification; in most jurisdictions not covered by genetic non-discrimination law |
| DC pot | An individual account balance | No | Investment, inflation, sequence-of-returns and longevity risk all sit with the individual; no decumulation mechanism is built in |
A deferred income annuity is the closest thing in the retail market to pure longevity insurance rather than income smoothing. It is priced against two compounding factors: the deferral period, during which the premium is invested, and — far more powerfully — mortality during deferral, because a substantial share of the pool will not reach the start date and their premiums fund those who do. The insurance analogy is exact: a DIA is to lifespan what a high-deductible policy is to a claim. It leaves the expected, self-fundable years uncovered and covers only the improbable-but-expensive tail, which means the buyer accepts a high probability of receiving nothing.
In the United States, SECURE 2.0 Act §202 created statutory room for this inside qualified accounts by eliminating the prior 25%-of-account-balance cap and raising the dollar limit from $125,000 to $200,000, indexed for inflation in $10,000 increments, with payments required to begin no later than age 85 and the limit applying per individual rather than per household. The current indexed figure changes annually and is published by the IRS; this page does not restate it, because the value circulating in commercial commentary could not be confirmed against a primary IRS notice.
That gap between what theory predicts and what people buy is known as the annuity puzzle, and NBER Working Paper 35145 (Hershfield, Shu, Brown, Hurwitz, Milevsky, Mitchell and Toland, April 2026) sorts the explanations into rational optimisation, market frictions, and behavioural factors. The rational and structural explanations include bequest motives (annuitised wealth cannot be left to heirs), pre-existing annuitisation (Social Security, defined benefit and state pensions are already life annuities, so marginal demand is lower than a model with no baseline predicts), pricing loads and adverse selection, liquidity needs against medical shocks, and incomplete markets — the paper notes inflation-linked annuities are often scarce and expensive, and that high prices can materially suppress demand.
The behavioural explanations include mortality salience (the decision requires contemplating one's own death, which experimental work finds aversive), framing and loss aversion, pessimism about tail survival probabilities, and low longevity literacy. The authors' own conclusion is that both classes matter: growing evidence indicates that cognitive limitations and behavioural biases also play a meaningful role.
What the shift from defined benefit to defined contribution actually moved
A defined benefit promise is structurally a life annuity issued by an employer: it pools longevity risk across the membership and hands the residual systematic risk to the sponsor. A defined contribution pot is a stock of money that pools nothing. The shift from one to the other did not reduce longevity risk in aggregate — it disaggregated it and moved it onto individuals, who cannot pool it.
Where each risk sits
| Defined benefit (DB) | Defined contribution (DC) | |
|---|---|---|
| What is promised | A benefit — an income formula, usually for life | A contribution — an account balance |
| Who bears investment risk | Sponsor | Member |
| Who bears inflation risk | Sponsor, to the extent indexation is promised | Member |
| Who bears longevity risk | Sponsor or insurer — the promise runs until death | Member, individually and uninsured by default |
| Decumulation | Built in | Not built in |
The Thinking Ahead Institute frames the consequence directly: DC systems were built to help people save, not to turn savings into a reliable income for life. That decumulation gap is now the recognised policy problem rather than a fringe concern, and responses are under way — in the UK a Value for Money framework, default retirement pathways and small-pot consolidation in the Pension Schemes Bill; in Australia a consultation on a retirement measurement and reporting framework for superannuation. All of these are proposals and reforms in progress, not outcomes.
One structural answer re-pools mortality inside a DC shell. Collective defined contribution schemes take DC contributions but pay a target — explicitly not guaranteed — income for life from a collective pool, so longevity risk is shared among members rather than borne individually or by a sponsor. The target can be adjusted, which is precisely what distinguishes CDC from DB.
Collective DC — status and tradeoffs
- The Royal Mail Collective Pension Plan launched on 7 October 2024 and is the only CDC scheme authorised by The Pensions Regulator, covering over 100,000 members.
- Draft regulations laid on 23 October 2025 would permit multi-employer (non-associated) CDC schemes from 31 July 2026, and The Pensions Regulator has consulted on extending its CDC code of practice to them.
- The draft regulations allow designs in which accrual rates and contributions need not be identical for all members, intended to manage cross-subsidy between employers and between generations.
- The tradeoff is explicit rather than hidden: income can be cut, and intergenerational cross-subsidy is the central design and regulatory concern. CDC is a genuine structural innovation — mortality pooling without a sponsor covenant — rather than a rebranding.
Because a DC pot has no built-in decumulation, a large research literature grew up around the question of how fast a portfolio can be drawn down. The best known contribution is William P. Bengen's "Determining Withdrawal Rates Using Historical Data" (Journal of Financial Planning, October 1994), which took historical US stock and intermediate-term government bond returns from 1926 onward, assumed a 30-year horizon and a tax-advantaged account, withdrew a fixed percentage of the initial balance in year one and then increased that dollar amount by inflation each year without re-linking it to the portfolio value, and tested every historical starting year to find the highest initial rate that never exhausted the portfolio. That worst-case figure — the SAFEMAX — was approximately 4%.
What the finding is: a historical worst-case backtest on one country's twentieth-century data, under one withdrawal mechanic and one horizon. What it is not: a rule, a recommendation, a guarantee, a forecast, or a statement about any individual. Bengen produced a descriptive statistic about a historical dataset; the industry converted it into a prescription, and that conversion — not the research — is the problem.
What the subsequent literature documented
- Calibrating bond returns to January 2013 real five-year TIPS yields while holding the historical equity premium raised the 30-year failure rate to 57% (Finke, Pfau & Blanchett, 2013).
- Those authors' own conclusion was that in most other countries sustainable initial withdrawal rates fell below 4 percent, and that there is nothing inherently safe about the 4 percent rule.
- The 30-year horizon is itself a longevity assumption. A 65-year-old cohort contains members who will live well past 95; the rule's horizon and the cohort's actual survival distribution are different objects.
- The mechanic is behaviourally implausible — it requires spending a fixed inflation-adjusted amount regardless of portfolio value, which real retirees do not do.
- Sequence-of-returns risk is the actual failure mechanism: poor early returns combined with fixed withdrawals permanently impair the capital base, so two retirees with identical average returns in a different order get different outcomes.
- The original model contained no taxes, no fees and no advice cost, and used US-specific asset classes and inflation history.
How institutions hedge longevity — and where the risk ends up
Aggregate longevity risk cannot be pooled away, so institutions that hold it trade it. The dominant instrument is the longevity swap: a pension scheme or insurer agrees to pay a fixed leg — a pre-agreed schedule of payments based on expected mortality, plus a fee — and to receive a floating leg equal to the actual benefit payments due to its actual members. If members live longer than assumed, the floating leg exceeds the fixed leg and the counterparty pays the difference. Assets stay with the scheme; only the longevity risk moves.
Indemnity versus index — and why basis risk decides the outcome
- Indemnity swaps reference the scheme's own members. They are a near-perfect hedge with no basis risk, but they are bespoke, illiquid, and require the counterparty to underwrite that specific population.
- Index-based swaps reference a published population mortality index. They are cheaper, more standardised and potentially tradeable, but they carry basis risk: the reference population's mortality may diverge from the scheme's.
- Basis risk is concrete, not theoretical. A national index and a specific pension scheme differ by socioeconomic composition, sex mix, age structure, occupation and geography. A national index moving 1% does not mean a given scheme's liability moves 1%.
The most instructive episode in this market is a failure. In November 2004 the European Investment Bank announced a 25-year, £540m longevity bond, structured and managed by BNP Paribas, with the longevity risk reinsured to PartnerRe; coupons would fall as a reference cohort died. It never launched, and was withdrawn in 2005.
Why the first longevity bond failed (Blake et al., 2006; ECB Financial Stability Review, 2006)
- Basis risk was too large. The index was a single cohort — males aged 65 in England and Wales — while real schemes also hold men in their 70s and 80s, and women. A hedge for one cohort of one sex is a poor hedge for an actual liability.
- Capital intensity: it required a large principal outlay to hedge a comparatively small amount of risk.
- Constrained reinsurance capacity: PartnerRe was not prepared to take cover above the £540m issue size, raising the question of whether capacity existed at meaningful scale.
- Unfamiliarity: the instrument required substantial market education on structure, hedge effectiveness, basis risk and liquidity that had not happened.
- The durable lesson is still the honest answer to why no liquid longevity market exists: the risk is long-dated, one-directional — almost everyone wants to hedge the same way, so natural sellers are scarce — hard to index without basis risk, and lacking the offsetting counterparty that makes interest-rate and credit markets liquid. What developed instead was a bilateral, reinsurance-based market.
Care costs, the healthspan gap, and how insurers now underwrite
Length of life and years of good health have not moved together, and that divergence is the most financially consequential fact in this section. It also breaks a common assumption: longevity risk and care-cost risk are not independent risks that partially offset. They compound, because the same underlying event — living longer than expected — simultaneously lengthens the drawdown horizon and raises the probability of reaching the ages of highest care need.
What happened to the private long-term care insurance market
- The ASPE brief documented the position in its own words as of 2016: less than 8 percent of Americans have purchased long-term care insurance, in part due to high and rising premiums and exit of insurers from the market, with stagnation or even decline in sales.
- The mechanism is a case of mispriced long-tail insurance. The standard actuarial account is that policies written in the 1980s to 2000s were priced on assumptions that proved wrong in four directions at once, all adverse: lapse rates far lower than assumed, interest rates far lower than assumed, longevity higher than assumed, and claim incidence and duration higher than assumed, particularly for cognitive impairment. This is the conventional explanation rather than a finding verified here against a named NAIC or Society of Actuaries experience study.
- Because most traditional policies were sold as guaranteed renewable — the insurer cannot cancel but can re-price an entire class, subject to state regulatory approval — the correction was applied to existing policyholders through rate increases, often decades after purchase.
- New sales shifted substantially toward hybrid or linked-benefit products: life insurance or annuities with long-term care or chronic-illness riders that pay a death benefit if care is never needed. Mechanically this removes the use-it-or-lose-it objection and, for the insurer, replaces open-ended morbidity exposure with a capped pool of benefit dollars — a genuine structural difference rather than marketing.
- Informal care from a spouse or partner is the largest source of care provision and is contingent on that person being alive and well. ASPE data show the unmarried at 65 use more formal services — 1.2 years on average versus 0.9 — so in a couple, longevity risk is joint rather than individual.
On the other side of the contract, how insurers classify risk is changing. Traditional life underwriting required a paramedical exam, blood and urine samples and attending physician statements. Accelerated underwriting replaces some or all of that with algorithmic evaluation of existing data — prescription-history databases, motor vehicle records, medical claims, credit-derived attributes and public records — routing only flagged cases to full underwriting. The benefits are real (faster decisions, lower cost, no needles) and so are the regulatory concerns: proxy discrimination, where a variable with no causal health link correlates with a protected characteristic; model opacity; and data accuracy with no clear correction route for the applicant.
The NAIC adopted its Model Bulletin on the Use of Artificial Intelligence Systems by Insurers in December 2023, building on AI Principles adopted in 2020, and its Accelerated Underwriting (A) Working Group issued regulatory guidance in June 2024 covering regulatory considerations, review strategies and information requests, explicitly grounded in those principles. Accelerated-underwriting considerations are being incorporated into the Market Regulation Handbook. State adoption counts move, so any figure for how many states have adopted the bulletin should be checked against the NAIC's own tracker rather than a secondary source.
Genetic underwriting law differs by jurisdiction — and the most widely believed fact about it is wrong
- United States: the Genetic Information Nondiscrimination Act of 2008 restricts the use of genetic information in health insurance and employment. It does not apply to life, disability or long-term care insurance. Life insurers may access medical records containing genetic test results as part of an application. Some states impose additional restrictions, so protection is state-specific and has to be checked per state.
- United Kingdom: the Code on Genetic Testing and Insurance is a voluntary agreement between the government and the Association of British Insurers, not a statute. Insurers never require or pressure any applicant to take a predictive or diagnostic genetic test; the only predictive test they may ask about is Huntington's disease, and only for life cover above £500,000. The government's three-year review published 5 March 2026 records that 93% of life insurance policies in 2024 fell below that threshold, that no insurer has applied to add a further condition in over 20 years, and that the Code remains largely fit for purpose.
- Australia: the government announced a total ban on the use of adverse genetic test results in life insurance underwriting in September 2024, issued draft legislation in September 2025, and the Treasury Laws Amendment (Genetic Testing Protections in Life Insurance and Other Measures) Bill 2025 has passed Parliament, amending the Insurance Contracts Act 1984 (Cth). Treasury states the ban applies to life insurance contract decisions made on or after six months after royal assent — reported as 8 October 2026 — with a review at five years. Individuals may still voluntarily disclose favourable results.
- Where a regime permits voluntary disclosure of favourable results but prohibits the use of adverse ones, the average risk of the non-disclosing pool rises. That is the same adverse-selection mechanic described in the annuity chapter, running in the opposite direction.
New mechanic, or new label? — and what changed in 2025–2026
Several categories marketed with longevity language are existing insurance with new packaging. That is not automatically a criticism — a clearer name can be an improvement — but an evidence-graded site should be able to say which is which, because the label often implies a mechanic the contract does not have.
What the marketing says, and what the contract does
| Marketing description | What it actually is | Genuinely new mechanic? |
|---|---|---|
| "Longevity annuity" | A deferred income annuity — a product category that has existed for over a century | No. New name, old mechanic — though the name describes the function better than the old one did |
| "Longevity insurance" | Usually a deferred income annuity with a late start date | No. Same mechanic again |
| QLAC | A deferred income annuity plus a specific US tax treatment (exclusion from the RMD balance) | Partly. The product is a DIA; the tax wrapper is genuinely new law under SECURE 2.0 §202 |
| "Living benefits" riders on life insurance | An accelerated death benefit or chronic-illness rider — access to a death benefit while alive | No. A repackaged death benefit with a trigger condition |
| Hybrid life/LTC or annuity/LTC | A life or annuity contract with a care rider drawing on a capped pool | Partly. The combination changes the insurer's risk profile from open-ended to capped morbidity exposure; the care benefit itself is not new |
| RILA / "structured" / "buffer" annuity | A deferred annuity whose credited return is an option payoff structure | Partly. The option-based crediting is real financial engineering — but it is not longevity insurance at all unless annuitised, because no payment is contingent on survival |
| "Longevity funds" / target-date decumulation funds | Investment funds with a glidepath | No. No mortality pooling, no survival contingency, no insurance |
| Collective DC | Target-benefit lifetime income from a collective pool with no sponsor covenant | Yes. A structurally different risk allocation from both DB and DC |
| Longevity swaps and indemnity risk transfer | Bilateral transfer of systematic longevity risk | Yes as an institutional market — though the underlying reinsurance mechanic is old |
| "Biological age" underwriting | Risk classification using epigenetic or biomarker data | Yes as a method — and the regulatory framework has not caught up with it |
2025–2026: what actually changed
- 7 October 2024
The first authorised CDC scheme launches
The Royal Mail Collective Pension Plan begins operating, covering over 100,000 members — still the only CDC scheme authorised by The Pensions Regulator.
The Pensions Regulator
- 2 December 2024
The healthspan–lifespan gap is quantified across 183 countries
JAMA Network Open publishes the first global comparison: the gap widened to 9.6 years, with the United States largest at 12.4 years.
Garmany & Terzic, JAMA Network Open
- May 2025
Denmark confirms a statutory retirement age of 70 from 2040
The OECD notes Denmark may soften the current one-to-one link between retirement age and life expectancy, in which case its projected future normal retirement age would be lower than 74.
OECD, Pensions at a Glance 2025
- 30 June 2025
CMI_2024 raises projected cohort life expectancy at 65
About three months higher for males and two weeks for females than CMI_2023 — a reminder that these are model outputs revised annually, and previously revised downward.
Institute and Faculty of Actuaries / CMI
- 1 September 2025
Regional analysis projects the healthspan gap widening to 2100
Communications Medicine reports gap size varying across six WHO regions and 183 member states, with life expectancy, GDP and non-communicable disease burden correlating most consistently with it.
Garmany & Terzic, Communications Medicine
- 23 October 2025
UK draft regulations would open CDC to multiple employers
Laid to permit multi-employer, non-associated CDC schemes from 31 July 2026, with The Pensions Regulator consulting on an extended code of practice.
DWP / The Pensions Regulator
- November 2025
OECD publishes Pensions at a Glance 2025
Old-age dependency ratio 33 in 2025 rising to a projected 52 by 2050; average future normal retirement age 66.4 for men and 65.9 for women; public pension spending at 8.1% of GDP on average.
OECD
- November 2025
LCP revises its decade projection for UK pension risk transfer downward
Buy-in volumes of £49bn in 2023 and £48bn in 2024; central decade projection revised down to £350bn–£550bn from £400bn–£600bn.
LCP, Pension Risk Transfer Report
- 10 December 2025
A record year for longevity swaps
Approximately £31bn transacted across the UK and the Netherlands, with Lloyds Banking Group schemes accounting for roughly £9.9bn.
Artemis
- February 2026
DC assets reach 63% of pension assets in the seven largest markets
Total assets across the 22 largest markets reached USD 68,274bn, up 9.6%; DC assets have grown 9.4% a year over ten years against 3.2% for DB.
Thinking Ahead Institute, Global Pension Assets Study 2026
- 19 February 2026
UK healthy life expectancy at birth falls to its lowest on record
60.7 years for males and 60.9 for females in 2022–2024 — down 1.8 and 2.5 years respectively on 2019–2021 — while lifespan has not fallen in step.
ONS
- 5 March 2026
UK publishes the three-year review of the genetic testing insurance Code
The Code is judged largely fit for purpose; 93% of 2024 life policies fell below the £500,000 threshold at which Huntington's disease results may be asked about; the review records evidence that some people delay predictive testing over insurance concerns.
UK Government / ABI
- 10 March 2026
CMI_2025 raises cohort life expectancy at 65 again
About eight weeks higher for males and six for females than CMI_2024, attributed to record-low all-age mortality in England and Wales in 2025, down 2% on 2024.
Institute and Faculty of Actuaries / CMI
- 23 March 2026
US retail annuity sales hit a fourth consecutive record — but not for lifetime income
$464.1bn total in 2025, of which SPIAs were $14.4bn and DIAs $4.8bn (down 3%): the products that actually pool longevity are about 4.1% of the market.
LIMRA
- 1 April 2026
Australia legislates a ban on adverse genetic test results in life underwriting
The Treasury Laws Amendment (Genetic Testing Protections in Life Insurance and Other Measures) Bill 2025 passes Parliament, applying to decisions made six months after royal assent, with a review at five years.
Australian Treasury
- April 2026
The annuity puzzle is restated with current numbers
NBER Working Paper 35145 reports that about half of consumers say they would favour an annuity while roughly 12% buy one, and that at least 60% of SPIA buyers elect refund features that reduce the mortality pooling the product exists to provide.
NBER Working Paper 35145
- 31 July 2026 (planned)
Multi-employer CDC schemes become permissible in the UK
The date set in the draft regulations laid in October 2025, subject to the regulations being made.
DWP
Longevity investing: the public companies actually working on aging
"Longevity investing" as a public-markets category is younger and thinner than the venture and private-biotech version of the same story. Most of the capital going into aging biology directly — Altos Labs, Retro Biosciences, NewLimit — is privately held and not investable by a retail investor at all. What trades publicly is a smaller, more mixed set: a handful of companies built explicitly around aging biology, and a larger set of pharmaceutical and diagnostics companies where a longevity-relevant programme is one line item in a much bigger business.
The distinction matters for reading any of this correctly. A single-asset aging-biology company is effectively a bet on one or two trial readouts; a diversified pharma company's stock will barely move on its geroscience programme regardless of how it reads out. Neither pattern is described here as attractive or unattractive — only as what it structurally is.
Publicly traded companies with disclosed aging-biology programmes
| Company | Ticker | Aging-related programme | Status |
|---|---|---|---|
| BioAge Labs | NASDAQ: BIOA | NLRP3 inflammasome inhibitor (BGE-102); APJ-agonist metabolic programme | Preclinical/early clinical; Novartis collaboration (Dec 2024) worth up to $20M upfront and $530M in milestones |
| Geron Corporation | NASDAQ: GERN | Imetelstat (Rytelo), telomerase inhibitor | FDA-approved (2024) for a blood disorder (myelodysplastic syndromes) — not for aging; included because telomerase biology is central to aging research |
| Longeveron | NASDAQ: LGVN | Laromestrocel, mesenchymal stem cell therapy for frailty and Alzheimer's | Phase 2/3, not approved |
| Unity Biotechnology | formerly NASDAQ: UBX | Senolytic small molecules targeting senescent cells | Ceased operations in 2025 after its lead programmes did not advance — included as a cautionary data point on how this sector's failures actually happen |
Two structural notes worth understanding before treating any of this as a sector overview. First, the most direct "pure aging" trial failure in this list is instructive: Unity Biotechnology spent years and substantial capital on senolytic drugs specifically for age-related disease and shut down in 2025 when the science did not translate to approvable products — a reminder that this is ordinary-biotech-failure-rate science wearing a longevity label, not a category exempt from the base rate.
Second, most of what actually moves the needle on human aging research is not funded through public markets at all. The TAME trial — the most-discussed proposed trial of whether a drug (metformin) can delay multiple age-related diseases at once — has sat effectively unfunded for years because it tests an off-patent generic with no exclusivity for any sponsor to recoup a roughly $75 million trial cost against. The US government's newer VITAL-H trial ($38M, testing rapamycin, dapagliflozin and semaglutide together) exists partly because that commercial gap is structural, not incidental. An investor looking at this space through public equities is looking at a narrow slice of where the actual science is happening.
Frequently asked questions
What happens if you outlive your retirement funds?
Mechanically, the income stops at whatever the exhausted source was providing, and what remains is whatever pays for life regardless of how long life lasts — a state pension, a defined benefit pension, or any contract whose payments are contingent on survival. This is what "individual longevity risk" names: running out of assets because of a longer than expected lifespan (Center for Retirement Research, 2023). The reason it is a risk rather than a certainty is that a life expectancy is a mean: roughly half of any cohort outlives it, so a plan built to the average is, by construction, insufficient for about half the people it describes. This page explains the mechanism; what any individual should do about it is a question for a financial adviser licensed in their own country.
What role do annuities play in planning for a very long life?
The function of a life annuity is mortality pooling. An insurer collects premiums from many people of similar age and pays income only while each is alive; those who die early leave balances that fund payments to survivors, and the excess return survivors receive over a bond of the same duration is called the mortality credit. That is why a life annuity can pay more than a bond portfolio of similar risk — the difference is redistribution, not investment return — and it also explains the tradeoffs: the capital is generally gone at death, the contract is usually irreversible, a level nominal payment erodes in real terms, and the promise depends on the insurer's solvency. Observed behaviour does not match the theory: about half of consumers say they would favour an annuity to avoid running out of money and roughly 12% buy one (NBER Working Paper 35145, 2026). Nothing here is a suggestion that any reader should or should not buy one.
Can I use life expectancy figures to work out how long my money needs to last?
Only with care, and only if you know which figure you are holding. A period life expectancy takes one year's mortality rates and assumes they apply unchanged forever — it summarises this year's mortality experience rather than forecasting a life. A cohort life expectancy adds projected future improvement for the years a birth cohort will actually live through, and the ONS regards it as the more appropriate measure of how long a person of a given age would be expected to live. The gap between them is around 8 to 10 years at birth for England and Wales cohorts born 1950–2000, but only about 0.7 to 1.0 years at age 65 on OECD projections for 2065, because a 65-year-old has far less future improvement ahead. Adding an at-birth gap to an age-65 period figure produces a fabricated number. And either way, both are averages of a wide distribution, not a horizon for any individual.
What is the mechanical difference between a traditional pension and personal retirement savings?
A defined benefit pension promises a benefit — an income formula, usually for life — and the sponsor or its insurer carries the investment, inflation and longevity risk, with decumulation built in. A defined contribution pot promises only a contribution: an account balance, with investment, inflation, sequence-of-returns and longevity risk all sitting with the member, and no decumulation mechanism included. A DB promise is structurally a life annuity issued by an employer; a DC pot pools nothing. The shift from one to the other did not reduce longevity risk in aggregate — it disaggregated it and moved it onto individuals who cannot pool it. DC assets are now 63% of pension assets across the seven largest markets, growing 9.4% a year over ten years against 3.2% for DB (Thinking Ahead Institute, 2026).
How does inflation affect long-term retirement income?
It affects each structure differently, and the difference is mechanical rather than a matter of judgement. A level nominal annuity pays the same cash amount for life, so its purchasing power falls throughout retirement at whatever rate prices rise. An escalating annuity raises the payment annually by a fixed percentage or an index, and pays for that with a materially lower starting income. Genuinely inflation-linked annuities exist but the NBER authors describe them as often scarce and expensive, noting that high prices can materially suppress demand. In a defined contribution pot, inflation risk sits with the individual alongside investment and longevity risk; in a defined benefit scheme it sits with the sponsor to the extent indexation was promised. None of this indicates which structure any particular person should hold.
Is the 4% rule safe?
"The 4% rule" is a description of a 1994 backtest, not a rule. William Bengen took historical US stock and intermediate-term government bond returns from 1926, assumed a 30-year horizon and a tax-advantaged account, withdrew a fixed percentage of the initial balance and then raised that dollar amount by inflation each year without re-linking it to the portfolio, and found the highest initial rate that survived every historical starting year — about 4%. Later work found it is not robust: calibrating bond returns to January 2013 real five-year TIPS yields while holding the historical equity premium raised the 30-year failure rate to 57%, and the authors concluded there is nothing inherently safe about the 4 percent rule (Finke, Pfau & Blanchett, 2013). Structurally, the 30-year horizon is itself a longevity assumption, the mechanic is behaviourally implausible, and the model contained no taxes, fees or advice costs. A withdrawal rate is a modelling assumption; assessing one for a real person requires a licensed adviser in that person's jurisdiction.
What is the role of life insurance in retirement planning?
Individual life insurance pays on death, so it insures mortality — the opposite risk to longevity. Riders marketed as "living benefits" are accelerated death benefits or chronic-illness riders: they allow access to a death benefit while alive under a trigger condition, which is a repackaged death benefit rather than a new mechanic. Hybrid contracts pairing life insurance or an annuity with a long-term care rider are a partial exception: the care benefit itself is not new, but capping the benefit pool genuinely changes the insurer's exposure from open-ended morbidity to a defined amount, and it removes the use-it-or-lose-it objection. One legal point is widely misunderstood: in the United States, GINA restricts genetic information in health insurance and employment but does not apply to life, disability or long-term care insurance.
How should health expenses be thought about when planning for a long life?
As a step function rather than a constant, and as a risk that compounds with longevity rather than offsetting it. Healthy retirement years carry baseline living costs, often with declining discretionary spend; years with functional limitation add care costs; high-dependency years are substantially more expensive with a very long tail. ASPE's projections for Americans turning 65 in 2015–2019 put the share developing a disability requiring long-term services and supports at 52%, with about 14% needing care beyond five years, average expected lifetime costs from 65 of $138,000 rising to $266,000 among those who ever use paid care — all in 2015 dollars, which understate today's nominal costs. Crucially, the same event that lengthens a drawdown horizon — living longer than expected — also raises the probability of reaching the ages of highest care need, so the two risks compound.
Are "longevity funds" and target-date funds a form of longevity insurance?
No. They are investment funds with a glidepath: no mortality is pooled, no payment is contingent on survival, and no insurance is involved. The general test is a single question — is any payment contingent on the person being alive? By that test, defined benefit pensions, collective DC schemes, immediate and deferred income annuities and QLACs pool longevity risk; fixed-rate deferred annuities, fixed indexed annuities and RILAs do not, unless and until they are annuitised, despite carrying the word "annuity". Applying that test to any product's literature separates longevity insurance from savings and investment products that merely happen to be used in retirement.
Does CureMed recommend any of these products, and are they available everywhere?
No, and no. CureMed is not a financial adviser, broker, insurer or intermediary, is not authorised or regulated to give financial advice in any jurisdiction, and receives no commission, referral fee or other payment from any provider of any product described here. Nothing on this page is a recommendation, a ranking, or a judgement about whether anything is suitable for anyone. On availability: the products and rules described are drawn from the markets where they are documented, principally the United States and the United Kingdom plus OECD-wide statistics. Product features, tax treatment, consumer protections and regulation differ substantially between countries, between states or regions within countries, and over time — a product described here may not exist, may work differently, or may carry different protections where you live. Any decision needs a financial adviser licensed in your own jurisdiction.
Are there publicly traded companies I can invest in that work on longevity?
A small number, and they're a narrow slice of where the actual science happens. BioAge Labs (NASDAQ: BIOA) is the closest thing to a pure-play public aging-biology company, with a Novartis collaboration signed December 2024. Geron (NASDAQ: GERN) and Longeveron (NASDAQ: LGVN) have aging-adjacent programmes within broader pipelines. Most of the best-funded aging research — Altos Labs, Retro Biosciences, NewLimit — is privately held and not investable through public markets at all. Unity Biotechnology, a senolytic-drug company, ceased operations in 2025, a useful reminder that this sector carries ordinary biotech failure risk. None of this is investment guidance — see the section above for what's actually verifiable about each company, and consult a licensed financial adviser before acting on any of it.
Sources
Every figure and claim on this page traces to one of these. Where a source is a company announcement rather than peer-reviewed research or a regulator, it is labelled as such.
- 01Pensions at a Glance 2025: OECD and G20 Indicators — OECD Publishing, Paris, 2025 · DOI 10.1787/e40274c1-en
- 02Period and cohort life expectancy explained (methodology) — Office for National Statistics, 2023
- 03National life tables: UK, 2021 to 2023 — Office for National Statistics, 2024
- 04Past and projected period and cohort life tables, 2022-based, UK: 1981 to 2072 — Office for National Statistics, 2025
- 05Health state life expectancies, UK: between 2011 to 2013 and 2021 to 2023 — Office for National Statistics, 2024
- 06Health state life expectancies, UK: between 2011 to 2013 and 2022 to 2024 — Office for National Statistics, 2026
- 07CMI Model shows further rise in cohort life expectancy (CMI_2025), 10 March 2026 — Institute and Faculty of Actuaries / Continuous Mortality Investigation, 2026
- 08CMI Model shows rise in cohort life expectancy (CMI_2024), 30 June 2025 — Institute and Faculty of Actuaries / Continuous Mortality Investigation, 2025
- 09The Annuity Puzzle Revisited: Barriers, Behavior, and Policy Paths to Lifetime Income — National Bureau of Economic Research, 2026 · NBER Working Paper 35145
- 10Longevity Risk: An Essay (Special Report) — funded by Jackson National Life Insurance Company — Center for Retirement Research at Boston College, 2023
- 11Annuities and Individual Welfare — American Economic Review (Davidoff, Brown & Diamond), 2005
- 12Bequest Motives and the Annuity Puzzle — Review of Economic Dynamics (Lockwood), 2012
- 13The 4 Percent Rule is Not Safe in a Low-Yield World — Finke, Pfau & Blanchett (SSRN), 2013 · DOI 10.2139/ssrn.2201323
- 14Long-Term Services and Supports for Older Americans: Risks and Financing (revised February 2016) — ASPE, US Department of Health and Human Services (Favreault & Dey), 2016
- 15Global Healthspan-Lifespan Gaps Among 183 World Health Organization Member States — JAMA Network Open 7(12):e2450241 (Garmany & Terzic), 2024 · PMID 39661386
- 16Healthspan-lifespan gap differs in magnitude and disease contribution across world regions — Communications Medicine 5(1):381 (Garmany & Terzic), 2025 · PMID 40890374
- 17Global Pension Assets Study 2026 — Thinking Ahead Institute (WTW) — industry research body, 2026
- 18Final U.S. Retail Annuity Sales Set New Sales High, Totaling $464.1 Billion in 2025 (23 March 2026) — LIMRA — industry research association, 2026
- 19Pension Risk Transfer Report, November 2025 — LCP — consultancy, 2025
- 20Lloyds Bank pensions transfer £4.8bn longevity risk. 2025 a record year for volumes (10 December 2025) — Artemis — trade press, 2025
- 21Living with Mortality: Longevity Bonds and Other Mortality-Linked Securities — British Actuarial Journal (Blake, Cairns, Dowd & MacMinn), 2006
- 22Financial Stability Review, December 2006, Box 14: Hedging Longevity Risk — European Central Bank, 2006
- 23Code on Genetic Testing and Insurance: 3-year review 2025 (published 5 March 2026) — UK Government and the Association of British Insurers, 2026
- 24Code on Genetic Testing and Insurance (October 2018) — UK Government and the Association of British Insurers, 2018
- 25Legislation passed to ban use of adverse genetic testing in life insurance (1 April 2026) — Australian Treasury, 2026
- 26Accelerated Underwriting — topic page and regulatory guidance (June 2024) — National Association of Insurance Commissioners, 2024
- 27Model Bulletin on the Use of Artificial Intelligence Systems by Insurers, and state adoption map — National Association of Insurance Commissioners, 2023
- 28Epigenetic Testing — The Way Ahead for Life & Health Underwriting? (February 2024) — Gen Re — reinsurer publication, 2024
- 29Consultation: extending the CDC code of practice to multi-employer schemes — The Pensions Regulator, 2025
- 30BioAge secures Novartis deal worth up to $550M biobucks — Fierce Biotech, 2024
- 31Unity Biotechnology — Wikipedia, citing company and press filings, 2025
- 32The TAME Trial for Metformin Remains Only Partially Funded — Fight Aging!, 2024
- 3313 Publicly Traded Companies Developing Longevity Therapeutics — BiopharmaTrend, 2026
Articles on this topic
- Longevity financial products for guaranteed lifetime retirement income.
Financial products that genuinely guarantee income for life, ranked on the strength and cost of the guarantee: defined-benefit and state pensions, lifetime income annuities, deferred income annuities, guaranteed-withdrawal riders, and collective schemes — with the products that use 'guaranteed' loosely and do not belong on the list.
- What are the best longevity financial products available?
The best longevity financial products ranked on how efficiently each covers the risk of outliving your money: deferred income annuities, immediate lifetime annuities, guaranteed-withdrawal riders, collective schemes, long-term-care riders, and reverse mortgages — with which product fits which situation.
- How to choose longevity financial products for retirement?
A ranked five-step method for choosing longevity financial products in retirement: size the income gap, decide which guarantee you need, weigh health and heirs, decide timing, then compare finalists on identical terms — with the questions that matter most and the mistakes the order prevents.
- Which longevity financial products ensure income for life?
Financial products that ensure income for life ranked on completeness of cover: pensions already held, lifetime income annuities, deferred income annuities, guaranteed-withdrawal riders, and collective schemes — with the exact mechanic (mortality pooling) that lets a payment continue no matter how long someone lives.
- Are longevity financial products worth it for retirees?
Whether longevity financial products are worth it for retirees, ranked by situation: those with no guaranteed-income floor, those with a thin floor, those with a strong floor already, those most worried about inflation, those most worried about capital access, and those in poor health — with the honest trade-off in each case.
- How do longevity financial products protect against outliving savings?
The mechanism by which longevity financial products protect against outliving savings — mortality pooling and the mortality credit — explained, with structures ranked on how completely they remove the risk: pensions, lifetime annuities, deferred annuities, withdrawal riders, and collective schemes, versus drawdown, which does not remove it at all.
- Longevity financial products to reduce risk of outliving savings.
Strategies to reduce the risk of outliving savings ranked by how much risk each removes relative to cost: laddered deferred annuities, partial annuitisation, guaranteed-withdrawal riders on part of a portfolio, a floor-and-upside split, delaying state pension claims, and full annuitisation — with a worked comparison.
- Tax-efficient longevity financial products for retirement planning.
The structural features of longevity financial products that tend to matter for tax efficiency, ranked: tax-qualified deferred annuities inside retirement accounts (e.g. the US QLAC), purchasing inside a pension wrapper, the exclusion ratio on non-qualified annuities, tax-deferred growth before payout, and death-benefit and estate treatment — with why specific numbers require local advice.
- Longevity financial products for stable income in old age.
Products ranked on how stable their income is in old age — resistant to market swings and, where indexed, to inflation: indexed pensions, indexed lifetime annuities, level lifetime annuities, guaranteed-withdrawal riders, and collective schemes with adjustable payments — with the difference between nominal stability and real stability.
- Longevity financial products designed for late-life healthcare costs.
Financial products for late-life healthcare and long-term-care costs ranked on whether they actually cover that risk: dedicated long-term-care insurance, long-term-care riders on life or annuity policies, hybrid life/LTC products, deferred annuities timed to the highest-cost years, health savings vehicles, and general longevity annuities that do not cover care at all.
- What is the best longevity insurance for retirees?
Longevity insurance for retirees ranked: deferred income annuities, lifetime annuities, defined-benefit pensions, guaranteed-withdrawal riders and collective schemes graded on how well they cover a very long life — with what each gives up.
- Best longevity insurance plans for guaranteed lifetime income.
Guaranteed lifetime income plans ranked on the strength and scope of the guarantee: state and defined-benefit pensions, lifetime annuities, deferred annuities, guaranteed-withdrawal riders and collective schemes — what each guarantees and what it leaves out.
Explore this section
- Public health policy
Where pension sustainability, retirement-age reform and population ageing are set as policy rather than priced as products.
- Longevity technology
Biological-age and biomarker testing — the same measurements now appearing at the edge of insurance underwriting.
- Longevity clinics
What extending healthspan is actually being sold as, and the costs that sit outside any insurance contract.
The Longevity Brief
One evidence-graded email a week: what is new in longevity research, what is hype, and the one change actually worth making.
Free · one email a week · unsubscribe anytime.