Ask a room full of people how much money you need to retire and you'll get a lot of confident-sounding non-answers. "A million dollars." "80% of your salary." "As much as you can save." None of those is actually a number you can aim at. A million is a round digit, 80% is a guess about someone else's life, and "as much as you can" never ends.
The 25x rule is the opposite. It takes one thing you can actually measure — what you spend in a year — and hands you back a target. It isn't magic, it has real failure cases, and if you want to retire early you probably need more than 25. But as a starting point for answering "what's my number?", nothing else is this simple and this well-supported. Here's where it comes from, how to run it on your own finances, and exactly where it breaks.
The one-line version
Multiply your annual expenses by 25. That's the portfolio that, historically, could fund those expenses for a 30-year retirement.
- Spend $40,000 a year → $1,000,000
- Spend $50,000 a year → $1,250,000
- Spend $80,000 a year → $2,000,000
Notice what's missing: your income. The rule doesn't ask what you earn, because your income isn't what your portfolio has to replace. Your spending is. Two people who earn the same salary but live on different amounts get two completely different numbers — and the one who spends less reaches the finish line faster regardless of title or paycheck.
That single inversion, expenses instead of income, is why this rule spread so far through the FIRE community. It means the target isn't set by your boss or the job market. It's set by a line item you control.
Where the 25 actually comes from
Twenty-five isn't a number someone pulled out of the air. It's the 4% safe withdrawal rate flipped upside down: 1 ÷ 0.04 = 25. To understand the multiple, you have to understand where the 4% came from, because most retellings get the details slightly wrong.
In October 1994, a financial planner named William Bengen published an article in the Journal of Financial Planning called "Determining Withdrawal Rates Using Historical Data." He was pushing back against the lazy practice of plugging average stock and bond returns into a calculator. Averages hide what actually breaks a retirement, which is the order in which returns show up: two people can earn the same average return over thirty years and end up with wildly different outcomes depending on which years landed first.
So Bengen did something more honest. Using a portfolio of 50% stocks and 50% intermediate-term Treasury bonds, rebalanced every year, he walked through every 30-year retirement window in the U.S. data starting in 1926 — the Depression, the war years, the brutal inflation of the 1970s. He withdrew a fixed amount in year one, then raised that dollar amount with actual inflation every year after, and asked what starting percentage survived every single window.
His answer in the original paper was 4%. Here's the verbatim reasoning worth keeping: at 4%, no historical case ever ran out before 33 years, and most portfolios lasted 50 years or longer. Push the first-year withdrawal to 4.25%, and the worst case could run dry in as little as 28 years. The retiree who defined the floor was the one who started in 1966, right into a decade where stocks fell while inflation shredded purchasing power.
About that "4.15%" floating around online: Bengen didn't print it in 1994. That sharper number came out of his later, finer-grained work; the paper itself deliberately handed readers the round, safer 4%. The principle underneath it he named the SAFEMAX — the highest starting withdrawal rate that held up even in the worst stretch the historical record could throw at it.
The idea got its famous stress test in 1998, when Cooley, Hubbard and Walz of Trinity University ran every 30-year window in the same data and arranged the answers across a grid of rates and portfolios — the "Trinity Study." One detail almost nobody mentions: they used long-term corporate bonds where Bengen had used intermediate-term Treasuries, so the two papers are close cousins rather than the identical test. Their inflation-adjusted grid is where 4% became the standard answer: a 50/50 stock-bond mix survived 95% of the 30-year periods, and a 75%-stock portfolio about 98% (the same 95% and 98% William Sharpe later cited in his own critique of the rule).
| Starting withdrawal rate | Implied multiple | What it roughly means |
|---|---|---|
| 3.0% | ~33x | Very long retirements, heavy safety margin |
| 3.5% | ~29x | Early retirement, 35–40 year horizon |
| 3.9% | ~26x | Morningstar's 2025 forward-looking 30-year figure |
| 4.0% | 25x | The classic rule, 30-year horizon |
| 4.5% | ~22x | Bengen's later, more diversified estimate |
It's 25 times spending — and that word is doing a lot of work
The most common way people mangle this rule is multiplying the wrong number. Not your gross salary. Not your peak earning year. Your actual annual spending — the dollars that leave your accounts.
To put real American household budgets around the math, the Bureau of Labor Statistics runs a Consumer Expenditure Survey that tracks average annual spending by income group. These are total expenditures across all U.S. consumer units for 2024, multiplied by 25 for scale:
| Income group (2024) | Avg annual spending | x25 |
|---|---|---|
| Lowest 20% | $35,046 | $876,150 |
| Second 20% | $50,054 | $1,251,350 |
| Middle 20% | $66,900 | $1,672,500 |
| Fourth 20% | $89,972 | $2,249,300 |
| Highest 20% | $150,342 | $3,758,550 |
Don't read that table as your personal target — read it as a sense of scale, with a warning baked in. The survey's "total expenditures" line includes things like personal insurance and pension contributions, which are payments you make while working and that largely disappear once you retire. It also reflects raising kids, commuting, and a mortgage that may be paid off by your sixties. Your retirement number is based on what you'll spend after those items fall away and after healthcare and travel potentially rise — not on today's gross number copied wholesale.
This is also where the rule hands you a lever most people never pull. Because the target is spending times 25, cutting your annual costs works on two sides at once. Every $1 of yearly spending you permanently erase takes $25 off the portfolio you have to build. Knock out $10,000 of recurring cost and your target falls by a quarter-million dollars — and on top of that, the money you're no longer burning each year gets invested. You're moving the goal nearer while saving harder toward it, both ends at once.
How to actually run the number
Skip the benchmarks and do it on your own statements.
- Track real spending for a full year if you can, or at least pull three months of bank and card activity and gross it up, adding annual costs that don't show monthly — property taxes, insurance premiums, vacations, car replacement.
- Subtract work-life expenses that end in retirement (commuting, office costs, retirement-account contributions) and add things that tend to rise (healthcare, travel in the early go-go years).
- Subtract guaranteed income. Social Security, a pension, or an annuity covers part of your spending, and only the gap has to come from your portfolio. The average retired-worker Social Security benefit in December 2025 was about $2,071 a month, or roughly $24,856 a year. If you spend $66,900 and expect that average benefit, the portfolio only has to cover about $42,044 — a target near $1,051,000 instead of $1,672,500.
- Multiply the remaining gap by 25 (or by a higher multiple if you're retiring early — see below).
One timing detail matters enormously for early retirees: Social Security doesn't start when you're 45. If you leave work decades before your claiming age, every dollar of those bridge years comes from the portfolio, and the guaranteed-income discount only kicks in later. Run the early years at full spending, then layer Social Security in once it actually starts.
The 4% detail almost everyone gets wrong
The 4% applies once, to your balance in year one. After that, you keep the same dollar amount and only raise it for inflation. You do not withdraw 4% of whatever the account happens to be worth every year.
On a $1,000,000 portfolio at 3% inflation, that looks like $40,000 in year one, about $41,200 in year two, and roughly $42,436 in year three. The actual percentage of your current balance bounces around as markets move — some years it's under 4%, some years over — but your spending stays smooth. That's the whole mechanism: flexible portfolio, fixed-in-real-terms spending.
Two more fine-print items Bengen was clear about. The 4% is before investment fees and before taxes, so a fund layer of expenses and a tax bill on withdrawals both come out of it. And it assumes a stock-heavy, annually rebalanced portfolio — he favored holding between 50% and 75% in stocks, not a bond-heavy pile that quietly loses to inflation over thirty years.
When 25x is not enough
The rule was engineered around a 30-year retirement, which fits someone stopping around 65. It does not automatically fit someone stopping at 45, whose money might need to work for 50 years. Stretch the horizon and the safe starting rate drops.
Morningstar's 2025 retirement-income research, using forward-looking return assumptions rather than just the historical record, put the highest safe starting rate at 3.9% for a 30-year horizon at a 90% success rate — about 26x — and 3.5% for a 35-year horizon, about 29x. Much of the early-retirement crowd already plans on 3% to 3.5%, which means 29x to 33x expenses, precisely because a 45- or 50-year runway and an early sequence-of-returns hit leave no room to rebuild.
That phrase — sequence of returns — is the true villain. Morningstar found that retirees who hit poor returns and high inflation in their first five years, and kept spending unchanged, were far more likely to run out. The same downturn at year 25 barely matters. Holding a year or two of expenses in cash, staying willing to trim discretionary spending in a bear market, or using guardrails that ratchet spending down when the portfolio falls can buy a lot of safety without saving a giant extra pile.
Don't mistake the 4% floor for a historical average. It's set at the rate that survived the single worst starting point on record; across history, the typical retiree could have spent considerably more. Bengen himself, after adding small-cap, micro-cap, and international holdings, later raised his own SAFEMAX estimate — to roughly 4.5%, and in his book A Richer Retirement around 4.7% for a more diversified portfolio. Reasonable people land anywhere from about 3.5% to 4.7% depending on horizon, flexibility, and how much they trust the future to resemble the past. That whole span is the honest takeaway: treat the figure as a planning band you adjust, not a fixed constant handed down from above.
How to actually use it
Treat 25x as a back-of-the-envelope target and a way to measure progress, not a guarantee. If you're aiming for a conventional mid-sixties retirement with some spending flexibility, 25x is a thoroughly reasonable bullseye. If you want out at 45, aim toward 29x to 33x, keep a cash buffer, and stay willing to flex. Either way, run the math on your real spending, subtract guaranteed income only when it actually arrives, remember fees and taxes come out first, and re-check every few years rather than setting it once.
The real gift of the rule was never precision. It's that it turns "save as much as possible" — a slogan with no finish line — into a number tied to the one variable you fully control. Spend less, and the number moves twice. That's the part worth keeping, long after you've argued about whether the exact safe rate is 3.9, 4, or 4.7.
If you want to run your own figure, the FIRE number calculator does the 25x math on your spending, the 4% rule calculator models withdrawals over your own horizon, and the savings rate calculator shows how changing the gap between income and spending moves your finish date.
Related articles
- What Is FIRE? Financial Independence, Retire Early Explained — the big-picture movement this rule powers
- Lean FIRE, Fat FIRE, Coast FIRE, and Barista FIRE: The Four Types Explained — how different lifestyles pick different spending targets
- Coast FIRE Numbers by Age: When You Can Stop Saving for Retirement — what the number looks like if you only let compounding do the rest
- A Roth Conversion Ladder for Early Retirement: A Step-by-Step Guide — how early retirees access that money before 59½
- Health Insurance for Early Retirement: How to Get Covered Before 65 — budgeting one of the costs 25x has to cover
Sources
- William P. Bengen, "Determining Withdrawal Rates Using Historical Data," Journal of Financial Planning, October 1994 — original paper (PDF)
- Jason S. Scott, William F. Sharpe, and John G. Watson, "The 4% Rule — At What Price?," Stanford (2008), which recounts Bengen's work and the Trinity Study's 95%/98% success rates — paper (PDF)
- Morningstar, "What's a Safe Retirement Withdrawal Rate for 2026?" (Dec 2025), reporting 3.9% for 30 years and 3.5% for 35 years — morningstar.com
- U.S. Social Security Administration, Monthly Statistical Snapshot, Table 2 (December 2025) — ssa.gov
- U.S. Bureau of Labor Statistics, "Consumer Expenditures — 2024" (Economic News Release, Dec. 19, 2025), Table C, average annual expenditures by income quintile — bls.gov
Disclaimer
This article is for general education and information only and does not constitute financial, tax, or legal advice. The 25x rule and the 4% safe withdrawal rate are historical rules of thumb, not guarantees of future performance; actual market returns, inflation, taxes, and personal circumstances vary. Consider speaking with a qualified financial professional before making retirement decisions.
Published by Coast FIRE Journal.