A Single Retirement Number Is the Wrong Output
Free retirement calculators return one number. The arithmetic is fine — three of the inputs are not. Longevity, long-term care and sequence of returns are where the estimate actually breaks.
· 4 min read
If you ask a free calculator whether you have enough to retire, it returns one number. The arithmetic behind that number is usually correct. Three of its inputs are where the estimate actually breaks — and each one fails in the same direction, which is why the single number tends to be optimistic rather than merely imprecise.
The right output is not a number. It is a range, a failure probability, and a short list of the variables you can still move.
Input one: longevity, measured from the wrong starting line
Most planning uses life expectancy at birth. The figure that matters is conditional life expectancy — how long someone who has already reached 52 tends to live — and it is materially longer, because you have already survived every risk that pulls the birth figure down. Planning to an average is also the wrong target: roughly half of people outlive it, and for a couple the relevant horizon is the second death, not the first. Under-forecasting here does not shave a year off the plan. It removes the most expensive decade of it.
Input two: the largest potential expense is usually excluded
Long-term care is left out of most consumer calculators, and it is frequently the single largest line item a retirement can face. It is not a small correction to a monthly budget — it is a different order of expense, it arrives with little notice, and in many households it lands on one spouse while the other still needs an income. A plan that models thirty years of groceries and no years of care is not conservative. It is incomplete in the one place where being wrong is unrecoverable.
Input three: when the bad years happen matters as much as whether they happen
This is sequence-of-returns risk, and it is the least intuitive of the three. Two portfolios can post the identical average return over thirty years and end in completely different places, purely because of order. Poor returns in the first few years of drawdown do lasting damage, because you are selling assets at depressed prices to fund living costs and those shares never participate in the recovery. The same poor returns twenty years later are survivable. An average-return calculator cannot see this distinction at all — which is exactly why it produces one confident number.
| Input | What a simple calculator assumes | Why it breaks |
|---|---|---|
| Longevity | Life expectancy at birth, single life | Understates the horizon; ignores that a couple plans to the second death |
| Long-term care | Excluded entirely | Omits the largest potential expense in the plan |
| Returns | A single average annual rate | Cannot represent sequence risk, which is decided by order, not average |
| Spending | One flat inflated figure | Real spending is lumpy and usually falls mid-retirement before rising with care |
| Output | “You need $X” | Hides the probability of failure and every lever you could still pull |
The three variables you control most are the three the calculator never asks you about: when you stop, what your floor is, and what the plan is if someone needs care.
What to model instead
Everyone spends their energy on the market return, which is the variable they influence least. The high-leverage inputs are the ones inside the household:
- Retirement date. Working longer does three things at once — adds contributions, removes withdrawal years, and shortens the horizon the portfolio must cover. Nothing else in the model moves three levers simultaneously.
- The spending floor. Not your budget — the number below which the plan stops being acceptable. Knowing the floor tells you how much of your spending is genuinely discretionary, and discretionary spending is what makes a bad early market survivable.
- The care plan. Insured, self-funded, or family-provided. All three are legitimate answers. Having no answer is the version that breaks the model.
A useful answer looks like this: “At this spending level, retiring at 64, the plan survives in most modelled scenarios; the ones where it fails are early-downturn paths combined with an uninsured care event; here are the two changes that fix most of those.” That is actionable. “You need $1.4 million” is not, because it tells you nothing about what to do differently on Monday.
Common questions
Is the 4% rule wrong?
It is a useful starting heuristic that was derived from a specific historical market, a 30-year horizon and a particular portfolio mix. Treated as a rough orientation it is fine. Treated as a guarantee for a 35-year retirement with a care event in it, it is being asked to do something it was never built for. The rule's real weakness is the same as the calculator's: it produces one number and hides the distribution behind it.
How much does long-term care actually cost?
It varies enormously by country, region and level of care, so any single figure quoted online is close to meaningless for your situation. The planning-relevant facts are more stable: it is expensive relative to ordinary retirement spending, most people underestimate the probability that someone in the household will need some form of it, and the duration is the variable that drives the total. Price it locally rather than nationally.
What is sequence-of-returns risk in one sentence?
It is the risk that poor investment returns arrive early in retirement rather than late — the same average return, in a worse order — which does disproportionate damage because you are withdrawing from a shrinking balance and locking in losses that never recover.
So do I have enough or not?
That question has no single-number answer, and the honest response is a conditional one: enough for what spending level, retiring in which year, surviving which market paths, with which plan for care. Once those four are specified the question becomes answerable — and usually turns out to be about the retirement date and the spending floor rather than about the portfolio.
Tags