The 1693 Loan That Paid Survivors More: What Tontines Teach About Retirement
Imagine being offered an investment whose annual income could rise when another investor died. It sounds like the premise of a dark novel. Yet in 1693 the English government experimented with a financing arrangement built around exactly that principle. Investors subscribed money; payments depended on the survival of nominated lives; and as some participants died, the survivors’ claims became more valuable. The instrument belonged to a family of arrangements known as tontines.
The surprising part is not merely that the design existed. It is that its central financial question has never disappeared: how can a person finance a retirement that might last ten years or forty without either running out of money or leaving too much unused? Modern pension systems, insurers and retirement researchers still wrestle with the same uncertainty. Tontine-like structures have reappeared in technical discussions not because anyone needs to celebrate death, but because longevity is a risk that can be shared.
To understand the idea, it helps to begin with a real historical disappointment. England’s 1693 tontine was part of an ambitious government borrowing programme, but the tontine component itself attracted far less money than planned. That failure reveals as much about financial design as the famous survivor payments do.
1693: a government needed money and investors wanted a bargain
King William III’s government was fighting France in a prolonged conflict that strained public finances. The UK Parliament’s account of the Financial Revolution describes how Parliament gained a stronger role in authorizing taxes and loans, and how the cost of war encouraged new forms of public borrowing. The Bank of England would be founded in 1694, but the year before it, officials were already experimenting with long-term instruments tied to people’s lives.
A historical financing table reproduced in research on the origins of the national debt records a tontine loan dated 26 January 1693 with subscriptions of £108,100. A separate single-life annuity offer attracted £773,394. Those figures are important because popular retellings sometimes imply that a million pounds flowed into a successful tontine. In fact, the planned package relied heavily on more conventional life annuities when the survivor-sharing design failed to attract the intended amount.
The academic study by Georges Gallais-Hamonno and Christian Rietsch, examining the English tontine and life annuities of 1693, likewise describes the tontine’s weak take-up. Its abstract attributes the result to awkward design and competition from an especially attractive life-annuity alternative. That is a powerful corrective to the myth that unusual financial structures automatically appeal because they offer a spectacular reward to a few lucky survivors.
What did investors see? Historical accounts describe a proposal in which subscribers could receive relatively high early payments and then a survivor-dependent stream. The exact schedules and options in the 1693 programme were more complicated than the simplified tontines used in modern illustrations. Rather than pretending the entire historical contract can be captured by one neat percentage, we should distinguish its documented fundraising results from the basic pooling mechanism that later writers use to explain the idea.
The mechanism: income that does not have to stop at a fixed birthday
In a simple tontine, participants contribute to a common arrangement. A defined pool of payments is distributed among those still alive. As the number of surviving participants declines, the allocation to each survivor may rise. The design transfers some value associated with participants who die to those who continue living. It is an extreme form of pooling longevity risk: a person who lives unusually long receives payments for longer, partly supported by the forfeited claims of those who die earlier.
It is essential not to confuse this with a guaranteed savings account. A tontine is not inherently a promise to return each person’s original capital. Nor does every modern product labelled tontine use exactly the same rules. Some arrangements distribute investment income; others draw down pooled capital; still others use age-dependent units or mechanisms that smooth payments. The contract, legal structure, investment portfolio and survival calculations determine what people actually receive.
The OECD’s 2022 Pensions Outlook discusses tontines alongside collective defined-contribution and other non-guaranteed lifetime-income arrangements. The common features it identifies include pooling longevity risk and the ability to adjust benefits when investment returns or mortality experience differ from expectations. That is very different from a traditional insurance annuity whose provider contractually guarantees a specified payment, subject to the provider’s solvency and the contract terms.
A worked example: five people, one pool, no magical return
Suppose five retirees each contribute ₹2,00,000 to an illustrative pool, creating ₹10,00,000. Assume, purely for the example, that the pool produces ₹50,000 of distributable income in a year after all relevant investment costs, and that the agreement distributes that income equally among surviving members. There are no contributions or withdrawals from principal in this simplified illustration. We are not modelling real mortality probabilities, taxes, capital volatility or a legally available Indian product.
| Surviving participants | Distributable pool income | Income per survivor for that year |
|---|---|---|
| 5 | ₹50,000 | ₹10,000 |
| 4 | ₹50,000 | ₹12,500 |
| 3 | ₹50,000 | ₹16,666.67 (approximately) |
| 2 | ₹50,000 | ₹25,000 |
| 1 | ₹50,000 | ₹50,000 |
The arithmetic is straightforward: ₹50,000 divided by the number of survivors. The economic meaning is not. If the pool earns only ₹30,000 in a difficult year, the payment with four survivors falls to ₹7,500 each, not ₹12,500. If the contract allows principal to be spent, the pool may shrink; if costs rise, distributable income may fall. The increase in a survivor’s share is not an investment return created by another person’s death. It is a redistribution of claims within a defined arrangement.
Nor does a survivor’s larger annual payment prove they were better off than someone who died early. The person who died may have received few payments and may have surrendered a bequest. The survivor has also faced years of uncertainty and may need the income to fund an unusually long retirement. Comparing only the last surviving participant’s payment is the financial equivalent of admiring a lottery winner without examining the ticket costs and the outcomes of everyone else.
The invisible asset: protection against living longer than expected
People often fear dying before enjoying retirement savings. Financially, the opposite outcome can also be frightening: living longer than the portfolio can support. Economists call this longevity risk. A single retiree cannot know in advance whether they will need income until age 72, 88 or 101. Saving enough for the longest plausible life may leave large unused balances in many scenarios; spending aggressively risks exhaustion in the longest ones.
Pooling changes the mathematics because a group can estimate the distribution of lifetimes more reliably than any individual can predict their own. Those who die earlier do not need future lifetime payments, and their forfeited claims can help finance payments to those who remain. This is the economic logic behind mortality credits. The term is technical, but the intuition is familiar from insurance: individual uncertainty can become more manageable when risks are pooled.
That does not mean pooling removes every risk. People can live longer than expected across an entire cohort, especially as mortality rates improve. Investment returns can disappoint simultaneously for everyone. Inflation can erode the purchasing power of nominal payments. Provider insolvency, weak governance, unfair pricing and opaque fees can destroy confidence. The OECD’s research on longevity risk emphasizes the need for sound mortality assumptions and regulatory safeguards.
Why the number of participants matters more than a dramatic survivor story
Our five-person example makes the arithmetic visible, but it would be a fragile basis for real retirement income. One death changes the number of beneficiaries by 20% immediately. With hundreds or thousands of people, individual deaths tend to create smaller, more predictable adjustments, although systematic changes in longevity and market returns still affect the entire pool. The OECD notes that adequate scale is important for limiting payout volatility in non-guaranteed arrangements.
Age and health differences also matter. A pool of 65-year-olds and a pool mixing people aged 40 and 90 should not allocate equal lifetime claims without considering the expected durations of payment. Modern designs may assign age-adjusted units, use mortality tables and include rules for new entrants or closed cohorts. A poorly designed pool can quietly transfer wealth between groups in ways that participants do not understand.
This is one reason the old tontine is better treated as a design lesson than a ready-made investment product. The alluring statement that survivors get more tells us almost nothing about fairness, liquidity, expected returns or the ability to leave money to dependants. Those depend on the details that marketing slogans usually omit.
What happened to the English experiment?
The 1693 case is revealing because the government did not merely offer a novel survivor-sharing idea. It offered competing ways to lend money to the state. The historical figures—£108,100 for the tontine and £773,394 for single-life annuities in the documented financing table—suggest that investors preferred the alternative available to them. The outcome cannot be explained by one factor with certainty, but it demonstrates that investors compare complexity, perceived fairness, yield and inheritance consequences.
The surrounding institutional changes mattered too. Parliament’s authority over taxation helped make government borrowing more credible. The Bank of England, established in 1694, would create another channel for funding the state. Financial innovation was not a sequence of brilliant isolated inventions; it was a contest among contracts operating within evolving political institutions.
For modern readers, this is a useful antidote to the claim that a financial structure’s elegance guarantees its success. A contract can be mathematically interesting and commercially unattractive. Investors may rationally reject it if the alternative offers a clearer promise, better liquidity or a more understandable treatment of heirs. A designer must solve the human problem, not just the actuarial equation.
Tontine, annuity and ordinary withdrawals are different promises
An ordinary investment account allows a retiree to draw money while retaining ownership of the remaining assets, subject to market movements and withdrawal rules. A conventional life annuity typically converts a premium into a contractual stream of payments, with options that may provide for a spouse or return of purchase price depending on the product. A tontine-like pool may offer lifetime income through shared mortality experience, but payouts can vary and bequests may be limited or absent.
None is universally superior. Capilore’s earlier analysis of retirement withdrawal rates and sequence risk explores the complementary problem of drawing from an individually owned portfolio. A person who needs a guaranteed floor for food and housing may value contractual certainty more than a potentially higher variable payment. Someone with dependants may place high value on preserving a bequest. Someone with a substantial guaranteed pension may be more comfortable accepting variable income from a separate pool. Liquidity for medical emergencies, taxes, inflation protection and counterparty strength can matter more than an attractive first-year yield.
In India, the National Pension System and annuity providers operate under specific regulatory frameworks. The Pension Fund Regulatory and Development Authority issued a May 2026 circular introducing retirement-income schemes and drawdown options under NPS. That is a reminder that the menu of permitted retirement structures can change; a historical tontine is not automatically an available or approved Indian pension product. Readers should consult current PFRDA and provider documents rather than infer eligibility or guarantees from a foreign historical analogy.
Inflation can defeat even a rising nominal payment
Suppose a retiree’s payment rises from ₹10,000 to ₹10,500 in one year, a 5% nominal increase, while prices rise 4%. The approximate real increase is not 5% minus nothing; it is (1.05 ÷ 1.04) − 1, or about 0.96%. If inflation instead reaches 8%, the same nominal increase produces a real decline of roughly 2.78%. A tontine payment that rises as the group shrinks can still lose purchasing power if investment income and mortality credits do not keep up with prices.
Those examples are deliberately separate from the five-person pool table. The first table assumes a fixed ₹50,000 distributable pool income; the inflation example assumes a different hypothetical payment path. Mixing the assumptions would make the numbers look more predictive than they are. Real retirement analysis should model the actual product’s rules, return variability, longevity assumptions, fees and tax treatment.
Questions that matter more than the headline yield
Before considering any lifetime-income arrangement, ask who legally owns the assets, who manages them, and what happens if the provider fails. Find out whether payments are guaranteed or merely projected, whether they can fall, and how fees are deducted. Ask whether a surviving spouse or nominee receives anything after death, whether withdrawals are possible, and what happens if a participant needs capital urgently. Demand illustrations for long life, short life, high inflation and poor markets rather than only an average-case scenario.
Check whether the promised return is being confused with a payout rate. A contract that pays ₹60,000 a year on a ₹10,00,000 contribution has a 6% initial payout rate; that is not necessarily a 6% investment return. The payment may include the gradual return of capital, mortality credits or a guarantee priced into the contract. Comparing it with a bank deposit rate without adjusting for capital ownership and bequest rights would be misleading.
Finally, verify jurisdiction. Retirement and insurance products are regulated differently across countries. An arrangement discussed in an OECD paper may be a policy concept, not an investment that an Indian household can purchase. This article explains financial mechanisms and history, not a recommendation to buy an annuity, a tontine or any specific pension plan.
The strange instrument’s modern lesson
The old tontine is memorable because it ties one person’s income to the survival of others. But its lasting insight is less macabre than its reputation. Retirement is not just a question of earning enough interest; it is a question of matching uncertain years of life with resources that must last. Pooling can help with one part of that problem, but it introduces choices about fairness, control, inheritance, guarantees and governance.
England’s 1693 experiment also warns that cleverness is not enough. The survivor-sharing design attracted limited subscriptions while simpler life-annuity alternatives drew much more capital. Three centuries later, retirement designers still face the same test: can they offer an arrangement that people understand, trust and find useful across many possible lifetimes? The best pension innovation is not the one with the strangest mathematics. It is the one whose promises survive careful reading.
Evidence and source notes
- UK Parliament, “The Financial Revolution” — wartime finance and the institutional setting for new public debt.
- Larry Neal, “The Origins of National Debt” (historical financing table) — the 1693 tontine and life-annuity subscription figures.
- Gallais-Hamonno and Rietsch, “Tontine et rentes viagères en 1693 en Angleterre” — specialist study of the 1693 financing failure.
- OECD Pensions Outlook 2022, chapter on non-guaranteed lifetime income — current framework for tontines, pooling and payout variability.
- OECD, Mortality Assumptions and Longevity Risk — technical treatment of longevity assumptions and pension risk.
- PFRDA, May 2026 NPS retirement-income and drawdown circular — current Indian regulatory context; not evidence that historical tontines are approved Indian products.
- Bank of England archive, stock collection — historical government finance and the institution founded in 1694.
Featured image: a circa-1690 portrait of William III by Godfrey Kneller, Royal Collection RCIN 405675, public-domain reproduction via Wikimedia Commons. This is a genuine historical portrait of the king whose government pursued the 1693 financing, not an illustration of a tontine transaction.