The Risk of Things Going Right
Every other retirement risk is a story about something going wrong. A market falls at a bad moment. Inflation runs hot for a decade. A long-term care event arrives. Longevity risk is different: it happens when things go right. The retiree stays healthy, avoids the illness that took their parents, and keeps living into a period their plan never paid for.
That is what makes it hard to plan around. Nobody wants to model the possibility of dying at 78, and most people quietly assume they will live about as long as average, which sounds prudent and is not. A plan built to an average fails about half the time, by construction, because about half of any group outlives its own average.
This article puts real numbers on the problem. Every survival figure comes from the Social Security Administration's own actuarial life table. It shows why life expectancy is the wrong planning horizon, why a couple must plan for the second death rather than the first, what it costs to fund a retirement to 85, 90, 95, and 100, and why sharing that risk with other people costs less than carrying it alone.
Life Expectancy Is a Midpoint, Not a Deadline
Life expectancy is where the average person in a large group dies. By definition, a large share of the group lives past it. Setting a plan's end date at life expectancy is like buying a car with enough fuel to reach the average destination. About half of the drivers run out of gas.
What Longevity Risk Actually Is
Longevity risk is the chance of living longer than the money lasts. It has three features that set it apart from the other risks in a retirement plan.
It cannot be diversified away by one person. Owning more stocks spreads market risk. Holding bonds cushions a bear market. Nothing an individual owns makes their own lifespan more predictable. The only way to diversify longevity risk is across people, which is what a pension, an annuity, and Social Security do.
It makes every other risk worse. An extra ten years is ten more years of inflation, ten more years of withdrawals from a portfolio that may have taken an early hit from sequence risk, and the decade in which long-term care is most likely to be needed. A long life does not create those risks, but it gives each of them more time to compound.
It lands hardest on the person least able to fix it. Running short at 92 is not a problem that can be solved by going back to work, retaking a claiming decision, or waiting for the market to recover. It usually falls on a widow or widower living on one Social Security check, with no remaining earning years and a smaller household income than the couple planned around.
What the Life Table Actually Says
The Social Security Administration publishes a period life table every year with its Trustees Report. The current edition, used in the 2026 Trustees Report, is based on death rates observed in 2023. It gives, for each age, the probability that a person of that age dies within the year. Chaining those probabilities together gives the chance that someone who is 65 today is still alive at any later age.
The table puts life expectancy at 65 at 18.1 years for a man and 20.7 years for a woman, which points to death at about 83 and 86. Here is the full distribution behind those averages, for a man, a woman, and a married couple who are both 65:
| Alive at age | 65-year-old man | 65-year-old woman | Couple: at least one | Couple: both |
|---|---|---|---|---|
| 75 | 79.4% | 86.1% | 97.1% | 68.4% |
| 80 | 64.2% | 73.9% | 90.7% | 47.5% |
| 85 | 44.9% | 56.6% | 76.1% | 25.4% |
| 90 | 24.1% | 34.9% | 50.6% | 8.4% |
| 95 | 8.0% | 14.4% | 21.3% | 1.2% |
| 100 | 1.3% | 3.2% | 4.4% | 0.0% |
Source: SSA 2023 period life table (2026 Trustees Report). Couple figures assume the two lives are independent.
Three rows deserve slow reading.
The 85 row. A plan that ends at 85 is two years past the man's life expectancy and roughly at the woman's, so it sounds cautious. Yet for a couple, there is a 76.1% chance at least one spouse is still alive at 85. The plan that sounds cautious makes its last payment to three couples out of four with someone still living.
The 90 row. For a couple both 65, the odds that someone reaches 90 are 50.6%, a coin flip. The odds that both reach 90 are only 8.4%. Almost all of the longevity risk in a married household sits with whichever spouse lives longer.
The 95 row. One couple in five, 21.3%, still has a surviving spouse at 95. That is not a remote tail. It is better odds than most people give themselves of reaching 95, and it is before any adjustment for good health.
These Odds Are a Floor, Not a Ceiling
The SSA period table applies 2023 death rates to every future year, so it ignores the steady improvement in mortality that has held for most of the last century. It also describes the whole population, including people in poor health and people without the savings that pay for good medical care. Studies of annuity buyers and of higher-income retirees consistently find they outlive the population average. Someone reading a retirement planning article is, on average, more likely to reach 95 than this table suggests.
Why Life Expectancy Is the Wrong Planning Horizon
Take the 65-year-old man. His life expectancy is 18.1 years, so a plan that treats life expectancy as the horizon funds him through 83. On his 84th birthday the plan has nothing left, and 49.0% of 65-year-old men are still alive to celebrate it.
That is the core problem with averages. A plan built to life expectancy is a plan with a 50% chance of failure, built in from the first day. It does not matter how conservative the return assumptions are or how carefully the withdrawal rate was chosen. If the horizon is the average, the plan is designed to leave half its users without money.
The better question is not "how long will I live?" but "how long do I need to be prepared to live?" That turns the planning horizon into a probability choice, the same way a Monte Carlo analysis turns a return assumption into a success rate. Here is the age by which the chance of still being alive drops to one in two, one in four, and one in ten:
| Chance of still being alive | Man, 65 | Woman, 65 | Couple, both 65 (last survivor) |
|---|---|---|---|
| 1 in 2 | 83 | 86 | 90 |
| 1 in 4 | 89 | 92 | 94 |
| 1 in 10 | 94 | 96 | 97 |
Read across the bottom row. A couple who wants no more than a one-in-ten chance of the last survivor outliving the plan needs to plan to 97. A plan built to 95 accepts roughly a one-in-five chance. A plan built to 90, a horizon many people consider generous, accepts a coin flip.
None of these is the right answer for everyone. A single man in poor health has a genuinely shorter horizon than a healthy couple whose parents lived into their nineties. But the default should come from this table, not from the headline life expectancy, and moving away from it should be a deliberate decision with a reason attached.
Plan to 95 for Couples, and Say Why
A 95 horizon is a reasonable default for a healthy couple both 65. It covers everything except the roughly one-in-five chance of a survivor living past it, and the table above understates that chance. For a single woman in good health, or a couple with a family history of long lives, 97 or 100 is easy to defend. Whatever horizon a plan uses, state it and the odds it accepts.
The Planning Horizon Is the Second Death
Most longevity planning goes wrong for couples, and it goes wrong the same way: the couple thinks about their lifespans one at a time. He expects to reach about 83, she expects to reach about 86, and they plan to the later of the two. The table says that is still not enough.
The reason is simple arithmetic. For the household's money to be needed, only one of them has to be alive. Two independent chances of a long life add up to a much better chance that someone lives a long time. The man has a 24.1% chance of reaching 90 and the woman a 34.9% chance. The chance that at least one of them does is 50.6%, not 34.9%.
The last survivor also tends to be in the weakest financial position. When one spouse dies, the household loses the smaller of the two Social Security checks, and any pension with a reduced survivor option drops. Expenses do not fall by half: housing, insurance, utilities, and the car cost about the same for one person as for two. The survivor usually files taxes as single, with narrower brackets on nearly the same income, which is the trap described in the widow's penalty. Any debt still owed may land on the survivor too, as covered in debt after the death of a spouse.
So longevity risk for a couple is a very specific scenario: one person, usually a woman, in her late eighties or nineties, living on one check, with the portfolio doing more of the work than it ever did while both spouses were alive. That is the person the plan has to protect.
What a Longer Horizon Costs
Extending the plan from 85 to 95 sounds as though it should cost half again as much, since the plan now funds thirty-one years instead of twenty-one. It does not, because money needed decades away is cheaper to set aside today than money needed next year.
Take a household with a $40,000 a year income gap in today's dollars, meaning the spending that Social Security and any pension do not cover. Assume the portfolio earns 2% a year above inflation and each year's withdrawal is taken at the start of the year. Here is what it takes at 65 to fund that gap with certainty through each age:
| Funded through age | Years of income | Capital needed at 65 | Extra over funding to 85 |
|---|---|---|---|
| 85 | 21 | $694,057 | — |
| 90 | 26 | $820,938 | $126,881 |
| 95 | 31 | $935,858 | $241,801 |
| 100 | 36 | $1,039,945 | $345,888 |
Planning to 95 instead of 85 adds ten years of income for about 35% more capital. Planning to 100 adds fifteen years for about half again as much. Those later years are expensive, but not as expensive as they look, and they are exactly the years the survival table says a couple has real odds of needing.
The table also shows the limit of funding longevity alone. A retiree who self-funds must pick one horizon and pay for every year up to it, whether or not they live to use them. Fund to 95 and every dollar set aside for ages 86 through 95 is needed only in the scenario where someone lives that long. In the scenario where both spouses die at 84, that money was never needed for income. In the scenario where one of them lives to 98, it was not enough.
Pooling: Why Sharing the Risk Costs Less
There is a second way to fund a lifetime income, and it is the only way to remove longevity risk rather than just push the horizon out. A large group of people pool their money, and the pool pays each member an income for as long as that member lives. People who die early leave their unspent share behind. That share pays the incomes of the people who live long. Actuaries call the transfer mortality credits, and it is the engine inside every pension, every lifetime annuity, and Social Security itself.
Because the pool only has to pay the expected number of years across its members, not the worst case for each one, it can pay the same income for less capital. Here is the actuarially fair cost of a $40,000 a year lifetime income at the same 2% real rate, priced from the same SSA table:
| Funding approach | Capital needed at 65 | How long it pays |
|---|---|---|
| Self-funded through 95 | $935,858 | Through 95, then nothing |
| Pooled, a 65-year-old man | $607,629 | For as long as he lives |
| Pooled, a 65-year-old woman | $677,568 | For as long as she lives |
| Pooled, a couple both 65 (joint and survivor) | $785,073 | Until the second death, even past 100 |
The last row is the one that matters for a married household. For $785,073, less than the $935,858 it costs to self-fund through 95, a pool can promise the same income until the second spouse dies, whether that is at 80 or 105. The couple has swapped the 21.3% chance of outliving a 95 horizon for a zero chance of outliving the income.
That is the whole case for guaranteed lifetime income, and the case against it deserves equal space. The pooled figures are an actuarially fair price. They assume the general-population mortality in the SSA table and no insurer expenses or profit. A real annuity is priced on annuitant mortality, which is longer, plus the insurer's costs, and most commercial annuities are not indexed to inflation. Pooling also means giving up control and legacy: the money belongs to the pool, and a retiree who dies early leaves nothing from that portion to heirs. The mechanism is real. The discount a buyer actually sees is smaller than this table shows, and it has to be weighed against what is given up. See annuities in retirement for the types and trade-offs, and annuity laddering for buying guaranteed income in tranches rather than all at once.
Why the Pool Beats the Bond Ladder
A bond ladder that pays $40,000 a year to 95 has to hold the full $40,000 for every year, because it has no way to know whether the owner will be alive to collect it. A pool holds only the survivors' share. At 95, when only about one couple in five in a joint pool still has a survivor being paid, the pool needs about a fifth of the money the ladder does for that year. That difference, summed across every year, is the mortality credit.
Where Longevity Risk Meets the Other Risks
Longevity risk rarely arrives alone, and planning for it in isolation misses the ways it amplifies the rest of the plan.
Longevity and withdrawal rates. The 4% rule was tested over thirty-year periods. For a couple both 65, a thirty-year horizon ends at 95, which the table above says a survivor passes about one time in five. A couple retiring at 60 with a thirty-year plan is planning to 90, which roughly half of couples outlive. A longer horizon needs a lower starting withdrawal rate or a guaranteed floor underneath it.
Longevity and inflation. The extra years are the most expensive years. At 3% inflation, a budget costs 2.4 times as much at 95 as it did at 65. Every year past the planned horizon is a year at the highest prices of the whole retirement, funded with whatever is left.
Longevity and success rates. A Monte Carlo success rate usually answers "does the money last to the horizon?", not "does the money last as long as the client?" A plan with a 90% success rate to 90 still fails in every scenario where the survivor reaches 91 and the money does not. The horizon is an input to the success rate, and it deserves as much scrutiny as the return assumptions.
Longevity and healthcare. The probability of needing long-term care rises sharply in the late eighties and nineties. A plan that stretches the portfolio to 95 but ignores care costs has solved the income problem only to face a larger one in the same decade.
How to Hedge Longevity Risk
Roughly in order of cost-effectiveness:
Delay Social Security, especially the higher earner. Social Security is an inflation-indexed lifetime annuity backed by the federal government, and delaying it is the cheapest way most households can buy more of it. Each year of delay past full retirement age raises the benefit by 8%, for life. For a married couple, the higher earner's benefit becomes the survivor benefit, so delaying it raises the income of whichever spouse lives longer, which is exactly the person longevity risk falls on. See survivor benefits and the higher-earner delay and when to claim.
Choose the survivor option on a pension carefully. A single-life pension pays more while both spouses are alive and nothing to the survivor after the pensioner dies. For a couple, that concentrates the risk on the very person most likely to face it. The trade-off, including when life insurance can replace the survivor option, is covered in pension maximization.
Build a guaranteed income floor. Cover essential spending with income that lasts for life, from Social Security, pensions, and if needed a lifetime annuity. The portfolio then funds discretionary spending, and a bad market or a long life cuts into travel rather than groceries. Building a paycheck in retirement walks through the structure.
Consider deferred longevity income. A deferred income annuity bought at 65 and starting at 80 or 85 is one of the most efficient longevity hedges available, because the insurer pays out only to those who reach the start age. A small premium buys a large late-life income, and the portfolio only has to be managed to the start date rather than to an open-ended horizon. Inside an IRA, a qualifying longevity annuity contract can also be excluded from required minimum distribution calculations, subject to IRS limits.
Keep growth assets for the long tail. Money that will not be spent for twenty years is not short-term money. A retiree planning to 95 needs part of the portfolio invested for growth, even at 65, because the last decade of a long retirement is funded by assets that have to outrun inflation for a very long time. The bucket strategy is built around this idea.
Keep home equity in reserve. A paid-off house is a last-resort longevity reserve. It can be sold, downsized, or tapped through a reverse mortgage in the scenario where the portfolio runs short in the late nineties. That is a reason to plan it into the long tail rather than into the early years.
Common Mistakes With Longevity Planning
Using life expectancy as the horizon. The most common error and the most expensive one. It builds a 50% failure rate into the plan on the first day.
Planning each spouse's life separately. The question is not how long each spouse will live but how long the longer-lived one will. For a couple both 65, the odds of a survivor at 90 are 50.6%, far higher than either spouse's own odds.
Anchoring on parents' lifespans. Family history matters, but a parent who died at 78 in 1995 died under the mortality rates of 1995. Those rates have improved since. Family history is a reason to adjust the table, not to replace it.
Treating the table as exact. The SSA period table is conservative for most readers of a planning article, for the reasons in the callout above. Use it as a floor and lean longer for healthy clients with means.
Buying longevity insurance with the whole portfolio, or with none of it. Annuitizing everything gives up liquidity, flexibility, and legacy. Annuitizing nothing makes the portfolio carry the full tail risk alone. For most households the answer is a floor: enough guaranteed income to cover essentials, with the rest invested.
Ignoring the survivor's income drop. A plan that works for two people at $8,000 a month can fail for one person at $5,500 a month, even though the survivor spends less. Model the household after the first death, not just the household at retirement.
Why This Matters for Advisors
Longevity is the conversation clients most want to skip. They are comfortable discussing a market crash and uncomfortable discussing their own death, and the result is that most plans quietly default to a horizon nobody actually chose.
The conversation that works starts with the couple table. Show a married couple the 50.6% chance that one of them is alive at 90 and the 21.3% chance at 95. Most have never seen the joint figure, and it lands differently from any individual life expectancy they have heard. Then point to the "both" column: 8.4% at 90. The question stops being "will we live that long?" and becomes "is whoever is left going to be all right?", which is the question that actually drives the plan.
It also reframes product conversations honestly. A lifetime annuity is not a return play. It is a way to buy the pool's help with the long tail, and the pooled-versus-self-funded table shows the mechanism without a sales pitch. Presenting the trade-offs in the same breath, including the loss of liquidity and legacy, is what makes the recommendation defensible.
And the horizon belongs in the file. A planning horizon is an assumption with a failure probability attached. If a plan uses 90, the record should show that it accepts roughly a coin-flip chance of a surviving spouse outliving it, and that the client understood and chose that.
How RetirementForge Helps
The Longevity Credit Visualizer shows clients how mortality pooling works: a cohort shrinking year by year, the payment split into interest, principal, and mortality credit, and a bond ladder paying the same income side by side until it runs dry. It turns the pooling section of this article into a picture a client can follow in a meeting.
Pair it with the Monte Carlo Simulation to test a plan at more than one horizon rather than assuming one, the Social Security Optimizer to see how much lifetime, inflation-indexed income a delay buys the surviving spouse, and the Retirement Income Gap Analyzer to size the guaranteed floor. Every assumption is documented and every client session is captured in an immutable audit trail. Get started free and show a client the age their plan actually needs to reach.
This article is for educational purposes only and does not constitute financial, tax, or legal advice. Survival probabilities are calculated from the Social Security Administration's 2023 period life table as published with the 2026 Trustees Report, assume independent lives for couples, and describe the general population rather than any individual. Funding and pooling costs are illustrative calculations at a stated real return and are not annuity quotes; actual annuity pricing reflects insurer mortality assumptions, expenses, and interest rates. Consult a qualified financial advisor for guidance specific to your situation.