Do My NumbersMethodology

How the money-last calculator works

The how long will my money last calculator puts two clocks side by side: the year your savings run out, and the chance that you, or one of a couple, are still alive then.

The drawdown, one year at a time

Each year the calculator takes out the gap between spending and other income, grows what is left at the return entered, and moves on. The gap rises with inflation, so the withdrawal grows each year while the balance shrinks.

W1=max⁡(0,  S−B)W_{1} = \max\left(0,\; S - B\right)
First-year withdrawal W₁: The amount taken from savings in the first year.
G1=(F0−W1)rG_{1} = \left(F_{0} - W_{1}\right) r
Growth G₁: Investment growth in the first year, on the balance left after the withdrawal.
F1=F0−W1+G1F_{1} = F_{0} - W_{1} + G_{1}
Balance after one year F₁: The balance at the end of the first year, after the withdrawal and growth.
Wt=max⁡(0,  (S−B)(1+π)t)W_{t} = \max\left(0,\; (S - B)(1 + \pi)^{t}\right)
Withdrawal needed: The spending gap t years from now, grown with inflation.
T=k+Fk/WkT = k + F_{k} / W_{k}
Years the money lasts T: How long savings cover the spending gap, including the part of the final year they cover.

The withdrawal comes out at the start of the year and growth is earned on the rest. In the final year the balance covers only part of the need, and TT counts that part, so the answer can fall between birthdays.

Survival uses the SSA table year by year, with deaths spread evenly between birthdays, and for a couple the household counts as alive while either person is (see the methodology overview for those formulas). A smoker’s death rates are multiplied by a factor, 2.0 unless changed, capped so no yearly chance of death passes 1. Past age 119 the last published rate is held.

With market ups and downs switched on, the same drawdown runs through 10,000 simulated market histories:

r1=(1+rˉ) eσLZ−σL2/2−1r_{1} = (1 + \bar{r})\,e^{\sigma_{L} Z - \sigma_{L}^{2} / 2} - 1
Return in year 1, r₁: The investment return a simulated history uses in its first year, before inflation (nominal).
r1=ws R1s+(1−ws) R1br_{1} = w_{s}\,R^{s}_{1} + (1 - w_{s})\,R^{b}_{1}
Pr⁡(D>a)=na/Nsim\Pr(D > a) = n_{a} / N_{\mathrm{sim}}
Share of histories lasting past a: The share of simulated market histories in which savings last past age a. It describes the simulation, not a forecast.

Defaults and why

  • The example household: a man of 65 and a woman of 63, $750,000 saved, spending $68,000 a year with $28,000 of Social Security. It is chosen so the money runs out after both life expectancies, which shows why the survival chance matters more than the average.
  • Return 5.5%, inflation 2.5%. Round, middle-of-the-road assumptions for a mixed portfolio, in nominal terms. They are inputs, not forecasts.
  • Simulated markets: a bell curve around your own return with the volatility of a 60/40 stock and bond mix over 1928–2025 (12.1% a year), so switching the view on changes only the ups and downs. A fixed random seed means the same inputs always give the same answer.

Data behind the projection

Each file keeps its source address, edition and retrieval date, and is tested against figures the publisher prints. All of them are listed on the data sources page.

Not in the model

  • Income tax on withdrawals, and investment fees unless you take them off the return.
  • Social Security rules such as survivor benefits, and spending that changes with age.
  • Long-term care and other large, uneven costs.
  • Improvements in death rates after 2023, and differences between your health and the population average.
  • In the simulations, market behaviour outside the chosen model.