Withdrawal
The 4% Rule, and Where It Breaks for Early Retirees
Where the 4% rule came from, what the Trinity Study actually tested, the five assumptions baked into it, and why a 40-year retirement needs a different answer than the 30-year one it was built for.
The 4% rule is the most useful sentence in personal finance and the most over-extended. It is genuinely a good starting point. It is not a law of nature, it was never claimed to be one, and the group most likely to quote it — people planning to retire decades early — is the group it was least designed for.
Here is what it actually says, what it assumed, and what to do with it when your retirement is 45 years long instead of 30.
What the research actually found
Two pieces of work created the rule.
In 1994, financial planner William Bengen asked a question nobody had answered cleanly: what is the highest withdrawal rate that would have survived every historical 30-year period, including the worst ones? He ran rolling historical windows of US stock and bond returns and found the answer sat just above 4%. He called it the “SAFEMAX” — the worst-case survivable rate, not the expected one.
In 1998, three professors at Trinity University ran a broader version across portfolio mixes and horizons, reporting success probabilities rather than a single number. That paper is where the popular framing comes from, and its headline was closer to “a 4% initial withdrawal from a stock-heavy portfolio succeeded in ~95% of historical 30-year periods” than to “4% is safe.”
Both are backtests. Both describe what would have happened in one country’s market history, over one set of overlapping windows, under one specific spending rule.
The mechanics, precisely
The rule is more specific than most people remember:
- In year one, withdraw 4% of your portfolio’s value.
- Every year after, withdraw the same dollar amount adjusted for inflation — not 4% of the new balance.
- Do this for 30 years.
- “Success” means the balance did not hit zero. Ending with $1 counts. So does ending with $8 million.
Point 2 is the one that gets lost. The rule is a fixed real income stream, deliberately blind to what the market does. That blindness is what makes it testable, and it is also exactly why it fails when it fails: it keeps withdrawing full freight into a crash.
Inverted, it gives the famous multiple: 1 ÷ 0.04 = 25. Hence 25× expenses.
The five assumptions inside the number
A 30-year horizon. This is the big one. The research tested 30 years because that was the realistic span for a 65-year-old. Retire at 45 and you may be funding 45 years — a horizon the original studies did not test and which materially changes the answer.
US market history. The 20th-century US was among the best-performing markets in the world. Studies applying the same method across other developed markets have generally found lower safe rates. Using the single best historical sample as the base case is a choice, not a neutral default.
Overlapping windows. Rolling 30-year periods drawn from about a century of data are heavily overlapping and therefore far less independent than the sample size suggests. The worst case in that record is one specific sequence of events, not the worst that can happen.
A specific portfolio, rebalanced, and no fees. Typically 50–75% equities, rebalanced annually, with no advisor fee, no fund expense, and — critically — no tax. Withdrawals from tax-deferred accounts are income. A 4% gross withdrawal can easily be a 3.4% net one, and the studies do not model that for you.
Rigid spending. Nobody actually keeps spending the same real amount while their portfolio halves. The rule’s failures assume you do. Real people cut back, and that flexibility is worth more to survival than any asset-allocation tweak.
Why the horizon changes the answer so much
Consider $50,000 a year of spending, a 5% post-retirement return and 3% inflation. Ask how much corpus each horizon requires, and what first-year withdrawal rate that implies:
| Retirement length | Corpus required | Implied first-year draw |
|---|---|---|
| 30 years | $1,138,856 | 4.39% |
| 40 years | $1,394,107 | 3.59% |
| 50 years | $1,604,702 | 3.12% |
The pattern is the whole point. The longer the retirement, the lower the withdrawal rate it can support. A 4% draw over 30 years is roughly what the research supports; the same 4% over 50 years is asking a portfolio to do something the data never tested.
Notice something else: the curve flattens. Going from 30 to 40 years costs $255,000; the next ten years cost only $211,000. Past a certain point you are essentially buying a perpetuity, and each extra decade costs less than the one before. This is why the gap between “retire at 45” and “retire at 40” is smaller than most people fear, and why the gap between “retire at 60” and “retire at 50” is larger than they expect.
What 4% is genuinely good for
Having spent 800 words on its limits, the fair thing is to say what it does well.
- As a sanity check. If your plan implies a 6% first-year draw, you do not need a Monte Carlo simulation to know something is wrong.
- As a cap on a horizon-based target. Present-value math lands your portfolio on exactly zero at life expectancy — a plan with no margin at all. Requiring that the first-year draw also sit at or below 4% adds that margin back. This is precisely how this calculator uses it: as a cushion, never as the definition.
- As a communication device. “25× your spending” gets people to start saving. No present-value formula has ever done that.
Use 4% as a ceiling on your first year, not as your target-setting rule. Size the target from your own horizon and returns, then confirm the first-year draw it implies is inside 4%. If it is not, the horizon math is telling you something and the cap should raise your number.
Adjustments worth making for a long retirement
Start lower, at 3.25–3.5%. The most common adjustment for 40+ year horizons. It costs real years of work — dropping from 4% to 3.5% raises a $1.25M target to $1.43M — but it directly buys the margin the original research never tested for.
Make spending flexible instead. Skipping the inflation raise in years after a negative return, or cutting 10% of discretionary spending during drawdowns, historically does more for survival than lowering the initial rate — and costs nothing in the years when markets behave. See withdrawal strategies for the specific rulesets.
Keep some earning capacity. Part-time income in the first five years is worth far more than the same income later, because it protects you exactly when sequence risk is at its most dangerous. This is the whole argument for Barista FIRE.
Hold a cash or bond buffer for the first two or three years. It means an early crash does not force you to sell equities at the bottom to eat.
The one-sentence version
The 4% rule is the answer to “what was the worst-case survivable withdrawal rate for a 30-year US retirement, given rigid spending and no taxes?” — a genuinely useful answer to a question that is close to, but not the same as, the one an early retiree is asking. Use it as a guardrail. Build the target from your own horizon.
You can test all of it directly: the calculator lets you set the withdrawal rate, both return assumptions, inflation and your life expectancy, and it tells you whether your number is being set by horizon math or by the withdrawal cap.