The 4% rule, and where it comes from.
The most cited rule in retirement planning has a specific origin (the 1998 Trinity Study), specific assumptions (US 60/40 portfolio, 30-year horizon), and specific limitations. This page walks through what the rule actually says — and what it does not.
The rule
Withdraw 4 % of your retirement portfolio in year one, then increase the withdrawal each subsequent year by inflation. Your portfolio is invested 60/40 in stocks and bonds. If history rhymes, the portfolio survives 30 years with high confidence in roughly 95 % of historical 30-year windows.
The Trinity Study
The 1998 paper by Cooley, Hubbard and Walz at Trinity University tested various withdrawal rates and asset allocations against US historical returns from 1926–1995. The 4 % rate is the highest withdrawal that survives 30 years in 95 %+ of the rolling historical windows tested. Lower rates produce higher confidence and larger residual balances; higher rates fail more often.
The study's most-quoted output is the survival-rate matrix — for a 60/40 portfolio over 30 years:
| Withdrawal rate | Survival rate (30 years) | Median ending balance |
|---|---|---|
| 3% | 100% | ~3.0× starting balance |
| 4% | 95% | ~1.7× starting balance |
| 5% | ~80% | ~0.5× starting balance |
| 6% | ~50% | ~0× (often depleted) |
| 7% | ~25% | ~0× (typically depleted) |
Why the rule may be too generous in 2026
Two structural changes since 1998 weaken the rule's empirical basis:
- Bond yields are lower. A 60/40 portfolio in 1998 earned its bond return at roughly 5 % real. In 2026 that figure is closer to 1.5–2 %. The bond sleeve contributes less.
- Equity valuations are higher. The Shiller cyclically-adjusted P/E ratio sits well above its long-run average. Future expected returns from equities are correspondingly lower than the 1926–1995 historical average.
Recent academic work by Wade Pfau, David Blanchett, and Michael Kitces suggests a more defensible safe withdrawal rate at current valuations is 3.3–3.7 % for a 30-year retirement. The 4 % figure is now best treated as the upper bound of a defensible range, not the central estimate.
What the rule does not address
- Spending lumpiness. The rule assumes constant inflation-adjusted spending. Real retirement spending typically declines from age 65–75, rises again from 75–90 (healthcare), and may spike at end of life.
- Sequence of returns. A bad first decade is much worse than a bad second decade. The rule does not adjust for the realised return path; a flexible withdrawal strategy that cuts spending after a bad year materially raises survival probability.
- State pensions. The rule applies to your portfolio. State and workplace pensions, which often inflation-adjust, reduce the load on the portfolio and effectively raise the safe rate.
- Tax. Withdrawal from tax-deferred accounts is taxed; withdrawal from Roth/PEA-equivalent is not. The same gross withdrawal supports very different net spending.
Two practical adjustments
- Use 3.5% as the planning figure. If your nest egg projection comfortably supports 3.5 % withdrawal of target spending, the plan is robust to 2026-vintage assumptions.
- Build a 2-year cash buffer. Hold 2 years of expected spending in cash equivalents. In a market drawdown, draw from cash rather than selling depressed assets. This single adjustment historically raises survival rate by 4–6 percentage points at the 4 % withdrawal rate.