A historical result, not a promise, with five conditions attached.
By David Liu, founder of DIY9lives. Last updated November 2, 2026.
The 4% rule is the most quoted number in retirement planning and one of the least understood. It is not a law, a guarantee, or a recommendation from a regulator. It is the output of historical simulation, and it comes with conditions that are easy to lose when the number is repeated. This guide sets out what it says, where it came from, and what it assumes, alongside the retirement calculator.
Withdraw 4% of your portfolio in the first year of retirement, then increase that dollar amount each year by inflation, regardless of what the market did. A $1,000,000 portfolio therefore pays $40,000 in year one, about $41,200 in year two at 3% inflation, and so on.
The claim attached to it is that a portfolio following this pattern would have survived a 30-year retirement in the historical US data the studies used. Survived means did not run out of money before the 30 years were up.
William Bengen published the analysis in 1994 in the Journal of Financial Planning, testing withdrawal rates against historical US market returns. The Trinity study — Cooley, Hubbard and Walz, published in 1998 in the Journal of the American Association of Individual Investors — repeated the exercise across different asset mixes and horizons, and its tables are where the "4% is safe" shorthand became widespread.
Both used US market history, which is a specific and unusually favourable dataset. That is the first thing to understand about the number: it is a statement about a particular country's particular century, not a universal constant.
The 4% figure is calibrated to a 30-year horizon. Retire at 60 and plan to 95 and you need 35; the same rate has a thinner margin. Retire at 50 and the horizon is 45 years, and 4% becomes aggressive rather than conservative.
The historical success rates depend on holding equity. An all-bond portfolio at 4% has failed far more often in the same data, because inflation erodes a fixed-income portfolio that is being drawn down.
The studies assumed no fees. A 1% annual fee on a 4% withdrawal is a quarter of the income, every year, compounded against you. Fees come out of the return, and the withdrawal rate has to be lower to compensate.
The rule assumes you raise the withdrawal with inflation every year and never cut it. Real retirees cut spending in bad years, which is what makes the rule look more fragile than actual behaviour.
Withdrawals from a traditional retirement account are taxed as income, so $40,000 gross is not $40,000 of spending. The rate that matters is the after-tax one.
Two retirees can earn identical average returns over 30 years and have completely different outcomes, because the order of the returns matters when you are withdrawing. A bad first five years does lasting damage: you sell shares at low prices to fund spending, and those shares are not there for the recovery.
This is why "the average return was 7%" is not reassuring. A portfolio that falls 25% in year one while you withdraw 4% has lost about 29% of its value, and the recovery has to work on a smaller base. The same returns in a different order can turn a comfortable plan into a failed one.
The practical defences are unglamorous. Hold a cash buffer of a year or two of spending so you are not selling equity into a decline. Stay flexible about spending. And treat the withdrawal rate as a starting point to be revised annually rather than a salary.
A dynamic approach sets a range rather than a number. If the portfolio performs well, take a little more; if it falls below a threshold, take less. The common versions use a percentage of the current balance with a floor and ceiling, or an initial rate with explicit cut and raise triggers.
The trade is obvious: you give up the certainty of a fixed inflation-adjusted income and get a materially lower chance of running out. For someone with a large discretionary budget, that is an easy trade. For someone whose spending is nearly all essential, a fixed income from an annuity or a bond ladder may be worth the lower expected return.
Four percent is a reasonable planning anchor for a 30-year retirement with a stock-heavy, low-cost portfolio and some flexibility about spending. It is not a guarantee, and it is not appropriate to apply it mechanically to a 40-year horizon, a fee-heavy portfolio, or a spending plan with no room to adjust.
The more useful exercise is to model your own numbers: what you actually spend, what you will actually receive from Social Security, and what a 20% market fall in year one would do to the plan. The retirement calculator will show the balance; the hard part is deciding what you are willing to cut if the first three years are bad.
This is general information, not advice. A decision this size is worth taking to someone who can look at your whole situation, including tax, health costs and the rest of your income.
The useful version of the exercise is not "is 4% safe" but "what would I do if my plan met a bad decade". Three variables answer most of it.
Essential versus discretionary spending. Split the budget into what must be paid — housing, food, insurance, medical — and what is optional. A plan with 60% essentials has much more room to cut than one with 90%, and the withdrawal rate can be higher because the failure mode is a smaller holiday rather than a missed bill.
Other income. Social Security, a pension, or rental income reduce the amount the portfolio has to carry, and they are usually inflation-linked or at least stable. A portfolio funding $20,000 of spending rather than $40,000 can support a materially higher withdrawal percentage, because the fixed income covers the base.
Time horizon and health. A 30-year plan for a 65-year-old couple has a different shape from a 45-year plan for someone retiring at 50. Longevity runs in families, and the planning horizon should reflect it rather than the average.
Run the retirement calculator with a lower return than the historical average, add the income you expect from elsewhere, and see what rate the balance supports. If the answer is comfortable, 4% is a conservative choice. If it is tight, the honest options are to work longer, spend less, or accept more risk — and no withdrawal rule removes that choice.
Four percent is not a federal rule, a safe harbour, or a recommendation from any regulator. It is a summary of a historical simulation with assumptions attached, and it has been quoted far more often than it has been read. Treating it as a guaranteed income turns a useful planning anchor into a promise nobody can keep.
It is also not a spending rule. What you spend is a decision about your life, and the withdrawal rate is a constraint on that decision. A plan built around a fixed percentage on day one and never revisited is a plan that will either feel needlessly tight in good years or dangerously loose in bad ones.
What it is: a defensible starting point for a 30-year retirement with a stock-heavy, low-cost portfolio and some flexibility. Adjust it for your horizon, your fees, your taxes and your willingness to cut back, and revisit it every year.
Last updated: November 2, 2026. Estimates only. Not tax, legal, financial, or professional advice. See the terms.