Guide · 8 min read · figures through 2024
Where the 4% rule stops working
The research behind 4%, the assumptions it uses, and what changes when retirement lasts forty-five years instead of thirty.
In 1994, financial planner William Bengen published a study in the Journal of Financial Planning. He replayed US market history and measured the starting withdrawal that lasted through each retirement period. His result was about 4% of the initial portfolio, with the dollar withdrawal increased for inflation each year. The later Trinity Study tested similar withdrawals across several allocations and horizons.
The familiar 4% figure therefore describes a particular set of assumptions. A longer early retirement changes one of the most important: the number of years withdrawals must last.
The four assumptions inside the number
- Thirty years. That is a retirement at 65, not at 45. Every year you add past thirty makes the sustainable rate lower, because there are more chances to meet a bad decade and less time to recover from it.
- US markets, over one particular century. The record it draws on contains the strongest equity market any country has ever produced. Using it is the honest thing to do — it is the only long, clean, daily-checkable record we have — but it is not a law of nature.
- A withdrawal that never flexes. The retiree in the model takes the same real amount in the fifth year of a crash as in a boom. Almost nobody behaves like this, and the ones who do are the ones the rule was protecting.
- No tax and no fees. Both come out of the same portfolio and neither is in the number.
What the record says once you change the horizon
Here is the same exercise this site runs everywhere else: a $1,000,000 portfolio at 80/20, replayed from every starting year between 1928 and 2024, with the withdrawal pushed up until the first starting year fails. That is the highest rate that would have survived every start on record — the strictest possible reading.
A 4% rate that failed no 30-year retirement failed 8 of 53 forty-five-year ones.
That is the same $1,000,000 portfolio spending $40,000 a year: it held in 65 of 68 thirty-year windows and 45 of 53 forty-five-year ones. Nothing changed but the number of years the money had to last.
Allocation changes the answer too
| Allocation | Rate that never failed over 40 years | 40-year windows survived at 4% |
|---|---|---|
| All stocks A broad US large-cap index, dividends reinvested. The most growth and the deepest drawdowns. | 3.35% | 51 of 58 |
| 80 / 20 stocks and bonds Stocks with a fifth in 10-year Treasuries — the usual Fat FIRE accumulation mix. | 3.59% | 52 of 58 |
| 50 / 50 stocks and bonds The classic balanced portfolio. Shallower falls, and a slower climb. | 3.38% | 47 of 58 |
| Cash and T-bills 3-month Treasury bills, rolled. Almost no nominal risk, and inflation takes most of it. | 1.67% | 0 of 58 |
Every figure computed on this page is pre-tax and before fees, using real returns from 1928 to 2024.
So what rate should a Fat FIRE plan use?
There is no single honest answer, but there is an honest way to arrive at one. Three moves, in this order:
- Start from your horizon, not from 4%. Retiring at 45 and planning to 90 is a 45-year problem. The chart above is the version of the rule that is actually about you.
- Separate the fixed spending from the flexible. Fund the part you cannot cut at the strict rate. The part you would happily suspend for two bad years can carry a higher one — and that is the real reason a Fat FIRE budget can be safer than a lean one at the same withdrawal rate.
- Check the plan against the specific years that break plans — 1929, 1937, 1966, 1973 — rather than against an average. That is a two-minute exercise in the simulator, and it tells you more than any rule of thumb.
The useful part of Bengen's work is the historical replay. It makes the retirement horizon, allocation and withdrawal policy explicit instead of burying them inside an average return.