Tontines: The Retirement Pool With a Survivor’s Bonus
Imagine receiving a larger retirement payment because several strangers in your investment pool died. Nothing was stolen, no market bet paid off, and no insurer suddenly became generous. The money simply remained inside the pool and was redistributed among the people still alive.
That is the unsettling but elegant logic of a tontine. It turns one of retirement’s hardest questions—how long will I live?—into a risk shared across a group. The idea is centuries old, frequently misunderstood, and newly relevant because longer lives are making retirement income more difficult to finance.
The risk of getting the good news
Retirement planning contains a peculiar danger. Dying early is emotionally tragic but financially inexpensive. Living far longer than expected is good news personally, yet it can exhaust a portfolio. Economists call this longevity risk: the possibility that a person outlives the assets intended to support them.
An individual cannot know whether retirement will last five years or forty. A large group is more predictable. Although nobody knows which members will die in a given year, actuaries can estimate how many are likely to remain alive. Pooling converts an unknowable personal outcome into a more manageable collective pattern.
Traditional pensions and life annuities use the same broad principle. A tontine differs mainly in where the uncertainty sits and who promises what.
A bargain among strangers
In its simplest modern form, participants contribute money to a common investment pool. The assets earn returns, fees are deducted, and payments go to surviving members. When someone dies, the portion of the pool allocated to that person is not necessarily paid to heirs. Some or all of it remains for the surviving participants.
That redistribution is often called a mortality credit. It is not an investment profit created by stocks, bonds, or clever trading. It is a transfer arising from the pool’s survival rules. Those who die earlier subsidize those who live longer.
This does not mean every tontine uses the same formula. Historical tontines often divided a fixed income stream among surviving nominees. Modern proposals may maintain individual accounts, allow limited bequests, vary payments with investment performance, or combine tontine pooling with ordinary retirement products. The common feature is that survival affects the economic claim.
The arithmetic, without the mystery
Consider a deliberately simple example. One hundred retirees each place US$10,000 into a pool, creating US$1 million. Ignore investment returns and fees for the moment. If five participants die and their combined US$50,000 stays in the arrangement, it can be reallocated among the ninety-five survivors.
That would amount to about US$526 per survivor, or roughly 5.26% of each original US$10,000 contribution. But calling this a 5.26% investment return would be misleading. The pool did not earn that money in the market. The survivors received it because the deceased members surrendered a claim under the contract.
This distinction matters whenever a retirement product advertises an attractive distribution rate. A payment can contain investment income, a return of the investor’s own capital, and mortality credits. Those three sources have very different meanings.
Before retirement, there was war
The tontine is usually associated with Lorenzo de Tonti, a seventeenth-century Neapolitan banker who proposed a financing plan to the French crown. The idea spread because governments needed money and investors wanted income.
England’s first major state tontine was launched in 1693 under William III during war with France. Subscribers provided capital to the government, while annual payments were divided among surviving nominated lives. The nominee did not always have to be the subscriber, which created an incentive to choose healthy children or young adults whose survival might keep the income claim alive for decades.
The arrangement was part public borrowing, part actuarial wager, and part administrative headache. Governments needed reliable records of births, identities, and deaths. Investors had incentives to choose unusually healthy nominees; fraud or delayed death reporting could distort payments. Long before databases and digital identity, the tontine was already a problem of verification.
Why an annuity is not quite the same thing
A life annuity normally asks an insurance company to promise a specified income for as long as the purchaser lives. The insurer assumes investment risk, longevity risk, and the danger that its pricing assumptions are wrong. Capital requirements and regulation exist because the promise depends on the insurer remaining solvent.
A tontine can avoid making the same fixed guarantee. If investment returns disappoint, or if participants live longer than expected, payments can adjust. The members bear more of the variation, while the provider administers the pool rather than promising every outcome from its own balance sheet.
That can make the structure cheaper and reduce the need for an expensive guarantee. It also means the income may fluctuate. Removing the guarantee does not remove risk; it relocates risk from the insurer to the participants.
The risks a large pool cannot swallow
Pooling handles idiosyncratic longevity risk—the uncertainty of which individual will live longest—better than it handles systematic longevity risk. If medical progress causes an entire generation to live longer than the mortality tables predicted, the whole pool is affected. There may be no short-lived subgroup large enough to offset the surprise.
Investment risk remains as well. A tontine invested in bonds can be hurt by inflation or interest-rate changes. A pool holding equities can suffer a market collapse. Weak governance can produce excessive fees, poor asset management, or opaque allocation rules.
Adverse selection also matters. People who expect to live unusually long may be more eager to join. If pricing assumes an average population but attracts a healthier one, early payment levels may prove unsustainable. Honest design requires conservative assumptions and transparent adjustment mechanisms.
Fairness is in the formula
The moral discomfort surrounding tontines often comes from the phrase “profiting from death.” Yet retirement systems already redistribute based on survival. The serious question is not whether redistribution exists, but whether the formula is fair, understandable, and resistant to manipulation.
Age is an obvious issue. A sixty-year-old and an eighty-year-old should not receive identical claims for identical deposits unless the design intentionally subsidizes one group. Health underwriting raises privacy and discrimination concerns. Allowing new members into an old pool can shift value between generations. Permitting withdrawals creates liquidity but can weaken the long-term sharing mechanism.
Inheritance creates the hardest trade-off. Every dollar promised to heirs is a dollar that cannot become a mortality credit for survivors. A product can offer both features, but it cannot maximize both. Families should be able to see exactly what is sacrificed when a bequest option is selected.
Why the old idea still matters
Modern retirement systems face a structural problem. Defined-benefit pensions have declined in many countries, while individuals increasingly manage their own retirement accounts. A personal account is excellent at recording ownership, but it does not automatically convert savings into income that lasts for an unknown lifetime.
Researchers and pension institutions have therefore revisited tontine-like arrangements, often under softer labels such as mortality-pooled funds, pooled annuity funds, or non-guaranteed lifetime income. The OECD has examined how such products can share longevity risk without requiring a full insurance guarantee.
Technology makes administration easier. Digital identity can verify participants; automated records can track deaths and beneficiaries; low-cost portfolios can reduce fees. But technology does not answer the central policy questions: who bears losses, who inherits, how payments are calculated, what happens on withdrawal, and what protection exists against misconduct.
Ask what the payout is made of
The most practical lesson is broader than tontines. Whenever an investment, pension, stable-value account, or retirement product advertises a high payout, ask what the payout is made of.
Is it interest or dividends earned on assets? Is it a return of the customer’s own principal? Is it a transfer from people who exited, died, defaulted, or surrendered benefits? Is there a contractual guarantee, and if so, whose balance sheet supports it?
A fund distributing 8% a year is not necessarily earning 8%. It may be consuming principal, relying on mortality credits, or varying benefits according to a formula. Those structures can be legitimate, but the headline percentage alone cannot reveal whether the payment is sustainable.
The asset no one can leave behind
A tontine forces a blunt choice into view. Retirement money can be designed primarily to support a long life, or primarily to preserve wealth for heirs. It cannot do both perfectly with the same dollar.
For someone concerned about leaving an estate, forfeiting capital at death may be unacceptable. For someone more worried about reaching age ninety-five with no income, mortality pooling can be valuable precisely because the unused claims of shorter-lived members support longer-lived ones.
The survivor’s bonus is therefore neither free money nor a reward for someone else’s death. It is the economic result of a contract in which a group agrees that money no longer needed by deceased members will continue doing the job for which the pool was created: paying the people who are still alive.
Important sources
CFA Institute Research Foundation, Tontines: A Practitioner’s Guide to Mortality-Pooled Investments: https://rpc.cfainstitute.org/research/foundation/2019/tontines
OECD Pensions Outlook 2020, Non-guaranteed lifetime retirement income: https://www.oecd.org/en/publications/oecd-pensions-outlook-2020_67ede41b-en/full-report/component-9.html
OECD Pensions Outlook 2022, Designing non-guaranteed lifetime retirement income: https://www.oecd.org/en/publications/oecd-pensions-outlook-2022_20c7f443-en/full-report/component-8.html
Financial History Review, Portfolio choice and longevity risk in the late seventeenth century: https://www.cambridge.org/core/journals/financial-history-review/article/abs/portfolio-choice-and-longevity-risk-in-the-late-seventeenth-century-a-reexamination-of-the-first-english-tontine/BBD3B400E5C7F60E6D82C96F1E84D5E3
Also see
Moshe A. Milevsky, King William’s Tontine: Why the Retirement Annuity of the Future Should Resemble Its Past:
https://www.cambridge.org/core/books/king-williams-tontine/742065301B20632DF9F9864CE1124530