Safe Withdrawal Rate Explained: The Math Behind the 4% Rule (2026)

The definition has three parts that people routinely mix up, so it's worth being precise:

• It's a first-year percentage, not an every-year percentage. You calculate 4% of the portfolio once, at retirement. After that, you increase the dollar amount by inflation — you do not recalculate 4% of whatever the portfolio happens to be worth.
• "Safe" means "survived history," not "guaranteed." The research asks: for every 30-year retirement window in US market history, what starting rate never ran out of money? It is a stress test against the past, including the Great Depression and 1970s stagflation — not a promise about the future.
• It assumes a specific portfolio. The original studies used 50–75% US stocks with the rest in intermediate bonds. Hold something meaningfully different and the historical results don't directly apply.

The SWR is the hinge of all retirement math: invert it and you get your target portfolio size — at 4%, you need 25× your annual spending; at 3.25%, about 31×. That inversion is exactly how the FIRE calculator derives your financial independence number, and the retirement planner projects whether your current savings rate gets you there. For the academic grounding, Bengen's original paper appeared in the Journal of Financial Planning (October 1994), and updated failure-rate tables across horizons and allocations are maintained in Early Retirement Now's SWR research series.

The 4% rule is not folk wisdom — it has a specific paper trail.

1994 — William Bengen. A financial planner with an MIT engineering background, Bengen tested every rolling 30-year retirement in US market data starting from 1926. His question: what fixed inflation-adjusted withdrawal rate survived even the worst starting years (retiring into 1929, or into the 1966–1982 stagflation trap)? The answer for a 50/50 stock/bond portfolio was just over 4%. He called it SAFEMAX. Notably, Bengen himself later revised his estimate upward — to roughly 4.5–4.7% with more diversified portfolios — and has said publicly that 4% was always the worst-case floor, not the expectation.

1998 — The Trinity Study. Three professors at Trinity University (Cooley, Hubbard, and Walz) reframed the question as a table of success rates: for each withdrawal rate, portfolio mix, and time horizon, what percentage of historical periods ended with money left over? Their headline result: a 4% inflation-adjusted withdrawal from a 50%+ stock portfolio succeeded in 95–100% of historical 30-year periods. That table is what turned "about 4%" into "the 4% rule."

What's changed since. Morningstar has published an annual "State of Retirement Income" update since 2021, re-running the analysis with forward-looking return estimates instead of pure history. Their recommended starting rate has bounced between 3.3% and 4.0% depending on bond yields and stock valuations at publication. Meanwhile, the Early Retirement Now research series extended the analysis to 40–60 year horizons for early retirees and found 4% fails too often at those lengths — supporting roughly 3.25–3.5% instead.

Horizon is the single biggest driver of a sustainable rate. The table below summarizes what the published research supports at each retirement length, assuming a diversified portfolio of at least 50% equities:

Two things stand out in that table. First, the difference between retiring at 65 and retiring at 50 is only about half a percentage point of withdrawal rate — but that half-point raises the required portfolio by roughly 15–20%. Second, past 50-year horizons, the rate stops falling: a portfolio that survives 50 years at 3.3% is overwhelmingly likely to have grown, not shrunk, and can effectively run forever.

The reason a "safe" rate is so much lower than the market's average return is sequence-of-returns risk. Stocks have averaged roughly 10% nominal (about 7% real) over the last century — yet the safe withdrawal rate is 4%, not 7%. The gap exists because you don't earn the average; you earn a specific sequence, and withdrawals make the order matter.

Consider two retirees with identical 30-year average returns. Retiree A hits a 40% crash in years 1–2; Retiree B hits the same crash in years 28–29. Retiree A is selling shares at depressed prices to fund withdrawals early on, permanently shrinking the base that later recovery compounds from. Retiree B rode the full compounding first. Same average, wildly different outcomes — Retiree A can run out of money while Retiree B dies with a larger portfolio than they started with.

This is also why "the market averages 10%, so I'll withdraw 7%" fails as reasoning. Withdrawing during drawdowns converts temporary paper losses into permanent realized ones. The safe withdrawal rate literature is, at its core, a body of research about surviving bad sequences.

The classic SWR framework assumes you withdraw a rigid, inflation-adjusted amount no matter what markets do. Almost nobody actually behaves that way — and the research shows flexibility is worth real money. The main published alternatives:

One practical note for US retirees: once you reach age 73 (rising to 75 for those born in 1960 or later under SECURE 2.0), Required Minimum Distributions from traditional retirement accounts are not optional. The IRS Uniform Lifetime Table dictates a minimum withdrawal percentage that starts near 3.8% and rises with age — which means late-retirement withdrawal strategy partially converges on the RMD-style approach whether you choose it or not.

To make the mechanics concrete, here is the arithmetic for a hypothetical retiree — not a recommendation, just the formula applied end to end.

Setup: $900,000 portfolio, retiring at 60, planning to 95 (a 35-year horizon). From the research table above, a 35-year horizon supports roughly a 3.6–4.0% starting rate; this example uses 3.8%.

• Year 1 withdrawal: $900,000 × 0.038 = $34,200 ($2,850/month).
• Year 2, with 3% inflation: $34,200 × 1.03 = $35,226 — calculated from last year's withdrawal, not from the new portfolio balance.
• Year 3, inflation 2.4%: $35,226 × 1.024 = $36,071.

Note what did not appear in that math: the portfolio's returns in years 1–3. Under the classic rule, market performance changes your risk, not your paycheck. If a deep bear market strikes early, the guardrails research suggests trimming withdrawals about 10% until the portfolio recovers — in this example, dropping from $35,226 to roughly $31,700 for the affected years.

Finally, remember the withdrawal is pre-tax if it comes from a traditional 401(k) or IRA. A $34,200 withdrawal might net $29,000–$31,000 after federal tax depending on the bracket — which is why the account type holding your money changes the portfolio size a given lifestyle requires.

To see these withdrawal rates in practice, our $500k retirement drawdown guide runs year-by-year scenarios at every common rate. If you're still building toward retirement, the how much to retire guide converts these rates into target portfolio sizes by age. And for the FIRE community's perspective on aggressive withdrawal rates, see the FIRE roadmap.