The 4% rule was built and tested for a 30-year retirement. Most people who quote it don't know that, and if you're retiring early, the length of your retirement is the whole problem. So I ran every withdrawal rate against every retirement since 1928 to see where the rule holds and where it breaks.
The "4% rule" is the most-quoted number in retirement planning: withdraw 4% of your portfolio the first year, adjust that dollar amount for inflation each year after, and you probably won't run out. It comes from William Bengen's 1994 study and the 1998 Trinity study, and both asked one specific question: would this have survived a 30-year retirement? For someone retiring at 65, that's a reasonable horizon. For someone retiring at 45, it isn't. I pointed my calculator's engine at the same kind of historical data Bengen used, but ran it across 30-, 40-, and 50-year windows. The short version: over 30 years the rule mostly holds; stretch it to an early-retirement horizon and it starts to fail.
Bengen's finding was narrow and honest: across every 30-year retirement he could test, a starting withdrawal rate of about 4%, adjusted for inflation, never ran dry. The Trinity study reached a similar conclusion using a stock-and-bond mix. Neither claimed 4% was safe forever, and neither tested a 50-year retirement, because in the 1990s almost nobody was planning one. The rule's reputation as a universal constant came later, from repetition rather than research. The honest way to use it is to remember the assumptions baked in: a 30-year horizon, US market history, and no taxes or fees.
Run a constant inflation-adjusted 4% withdrawal against every 30-year window since 1928 and it survives 65 of 68, about 96%. That's the result that earned the rule its reputation, and for a traditional retirement it's defensible. Drop to 3.5% and every single 30-year window survived. Here's the full grid, every rate against every horizon:
| Withdrawal rate | 30-year windows | 40-year | 50-year |
|---|---|---|---|
| 3.0% | 68/68 | 58/58 | 48/48 |
| 3.5% | 68/68 | 58/58 | 47/48 |
| 4.0% | 65/68 | 54/58 | 41/48 |
| 4.5% | 61/68 | 47/58 | 35/48 |
| 5.0% | 55/68 | 44/58 | 31/48 |
All-equity inflation-adjusted S&P 500 returns, full windows only, constant real-dollar withdrawals, no fees or taxes. Each cell is the number of historical starting years whose retirement never ran dry. Drawn from the same engine as the 50-year backtest.
Read down the 4% column and the rule comes apart as the horizon grows. 65 of 68 over 30 years (96%) becomes 54 of 58 over 40 years (93%), and just 41 of 48 over 50 years, one failure in seven. A one-in-seven chance of running out of money isn't a rule, it's a bet you'd lose more often than you'd ever accept. And these failures aren't ancient history: alongside a 1929 start, the cluster that breaks 4% is the mid-1960s (1965, 1966, 1968, 1969), people who retired straight into a decade of flat real returns. That's sequence-of-returns risk, and a long retirement is exactly what's most exposed to it: a bad first decade has forty more years to compound into a shortfall.
The good news is the repair is cheap, and the same backtest prices both options.
Lower the rate a little. The highest rate that never failed, what Bengen called the SAFEMAX, is 3.75% over 30 years but only 3.45% over 50. Going from 4% to 3.5% takes the 50-year record from 41/48 to 47/48. That's the single most effective change you can make, and it's the subject of the safe withdrawal rate page.
Add a cash buffer. Holding two to three years of spending in cash, and refilling it only after good market years, takes the worst of that sequence risk off the table without permanently lowering your rate. A three-year buffer on its own makes the 4% rule perfect over 30 years, 68 of 68. The cash buffer study tests every buffer size from zero to five years and finds the same narrow sweet spot.
Combine them and you get the backtest's headline result: a 3.5% rate paired with a two-to-three-year buffer survived all 48 fifty-year windows since 1928, the worst markets in the record included.
For a 30-year retirement, broadly yes — 96% historically, and a slightly lower rate or a small buffer closes the rest. For an early retirement of 40 to 50 years, no, not at 4%: one in seven of those long windows ran dry. The fix isn't to throw the rule out, it's to right-size it for your actual horizon. Use 3.5% or less, keep a two-to-three-year cash buffer, and the same history that fails 4% over 50 years comes through intact. The calculator does this for your own numbers — it picks a conservative rate, models the buffer, and shows the realistic range of outcomes across every market sequence since 1928.
See what rate your retirement can actually supportIf you want to see it in a worked example, the scenario pages run the full method end to end: retire at 45 with $1.5M, retire at 55 with $1M, or start from the can I retire? hub. Working with a round number? See whether $1 million is enough to retire.
Not financial advice. Consult a fee-only fiduciary CFP for personalized guidance. Tax figures use 2026 brackets.