The 4% rule in 2026
What these dates mean
Published is the original release. Updated records a material content change. Reviewed records the latest documented factual or editorial review; it does not mean personalized professional advice.
The two credible safe-withdrawal numbers disagree — 3.9% and 4.7%. Here is why, and which one your plan should use.
The 4% rule is the most-quoted number in retirement planning and the most misunderstood. It is a historical finding, not a law of finance. And in 2026 the two most credible people working on the question do not agree on what it should be.
That disagreement is not a scandal, and it is not a reason to ignore both. It is the most useful thing on this page. This spoke sits under the Retirement hub.
Where the 4% rule comes from
William Bengen, a financial planner, published Determining Withdrawal Rates Using Historical Data in the Journal of Financial Planning in October 1994. He took US market history from 1926 to 1992, ran a portfolio through every rolling 30-year retirement in it, and asked one question: what is the highest starting withdrawal rate that would have survived even the worst of them?
His answer was about 4%, on a 50/50 mix of US large-cap stocks and intermediate-term Treasuries, rebalanced annually, with the first year's dollar withdrawal increased by inflation every year after. He called it SAFEMAX — the maximum safe rate — and he never called it the 4% rule. The press did that.
The 1998 Trinity study (Cooley, Hubbard and Walz, AAII Journal) reached a compatible conclusion by a different route, reporting success rates across a grid of portfolios and horizons. Two independent findings pointing the same way is why the number stuck.
Notice what SAFEMAX actually is. It is not a forecast, and it is not an average outcome. It is the worst case in one country's recorded history. For most historical retirees, a 4% start left a very large pile of unspent money.
What the rule assumes, and how each assumption moves the number
The headline hides its assumptions. Each one is a lever.
| The original rule assumes | If your reality is different | Effect on the sustainable rate |
|---|---|---|
| A 30-year horizon | You retire at 55 and plan to 95 — 40 years | Lower |
| A 30-year horizon | You retire at 70 | Higher |
| A balanced stock/bond portfolio | Nearly all cash or bonds | Lower |
| A balanced stock/bond portfolio | Nearly all stocks | Lower — volatility, not return, sets the safe starting rate |
| US market history | Almost every other country's history was worse | Lower |
| Costs are negligible | You pay 1% in fund and advisor fees | Lower, close to one for one |
| Spending never flexes | You can cut in a bad year | Higher |
| Taxes are not modelled at all | Your money is in a traditional 401(k) | The rate is unchanged; your spendable income is lower |
The last row is the one nobody says out loud. Every published safe-withdrawal number describes the portfolio, not your bank account. A 4% withdrawal from a pre-tax IRA is 4% of taxable income, not 4% of spending money.
The disagreement: 3.9% and 4.7%
Two credible, current numbers exist. Most sites pick one and sound confident. Here is what each one is actually claiming.
| Morningstar: 3.9% | Bengen, revised: 4.7% | |
|---|---|---|
| The question it answers | What can a new retiree start at today and still have a 90% chance of money remaining after 30 years? | What is the highest rate that would have survived the single worst 30-year period in the US record? |
| Method | Forward-looking simulation using projected asset-class returns and inflation | Backtest of actual historical returns |
| Evidence base | Capital-market assumptions for the next 30 years | US market history since 1926 |
| Portfolio | 30-50% equity, remainder bonds and cash | Seven asset classes — US large, mid, small and micro-cap, international, intermediate Treasuries, T-bills |
| What "safe" means | 90% of simulated paths survive | 100% of actual historical paths survived, including the worst one |
| Horizon | 30 years | 30 years |
| Spending rule | Fixed, inflation-adjusted withdrawals | Fixed, inflation-adjusted withdrawals |
| What moves it | Today's valuations and bond yields | Adding asset classes; new history accumulating |
| It fails if | The next 30 years are worse than the assumptions | The future is worse than the worst thing that has ever happened |
These are not two estimates of one quantity. They are answers to two different questions, computed on two different portfolios, under two different definitions of success.
So: do not average them. 3.9% and 4.7% do not bracket a "true" rate sitting between them, and the midpoint of two unrelated statistics is not a statistic. Decide which question is your question.
- If your instinct is "I want to know the odds, given what markets look like now" — that is Morningstar's question.
- If your instinct is "I want a number that would have held through the worst thing that ever happened" — that is Bengen's.
Bengen's revision is worth understanding on its own terms, because it is easy to misread as bullishness. It is not. He raised the number by diversifying the portfolio, not by assuming better markets. Adding small-cap, micro-cap and international stocks to the original two-asset mix lifted the floor under the worst historical cohort. The pessimism is unchanged; the portfolio is better.
Morningstar's number is not a constant either
The figure that most sites quote as the forward-looking rate has moved every single year, because it is downstream of valuations and yields. Its own history is the best argument against treating it as precise.
| Report edition | Base-case starting rate (new retiree, 30 years, 90% success) |
|---|---|
| 2021 | 3.3% |
| 2022 | 3.8% |
| 2023 | 4.0% |
| 2024 | 3.7% |
| 2025 edition (the current one, for 2026) | 3.9% |
If you have seen "3.3%" quoted recently as Morningstar's view, that is the 2021 edition. It is five years stale.
Horizon and equity allocation
Morningstar publishes a full grid of safe starting rates by asset allocation and time horizon. We do not reproduce its cells here, because we cannot verify them outside the report, and a fabricated table on a money page is worse than no table. What the report states in plain language is:
- The 3.9% base case is for a 30-year horizon and a portfolio with 30-50% equity.
- Shorter horizon, higher rate. An older retiree planning for fewer than 30 years can reasonably spend well above the base case.
- More stocks does not mean a higher safe starting rate. Above roughly 50% equity, the extra volatility reduces the safe starting withdrawal, because it raises the odds of a bad first decade. This is the finding most people get backwards.
That third point only makes sense once you understand sequence of returns risk: a high-equity portfolio has a higher average outcome and a worse unlucky outcome, and a safe starting rate is a statement about the unlucky one.
Fees come straight off the top
A fee is a withdrawal you do not get to spend.
Draw 4% and pay 1% a year in fund expenses and advisor fees, and the portfolio is funding a 5% drain — it just only sees 4% of it. There is no published safe-rate figure that is safe at that cost level, because none of them were computed with a 1% leak in them.
This is the cheapest fix available to almost every retiree, and it is the one that requires no forecasting skill at all.
What actually beats picking a perfect number
Flexibility. Morningstar tested the common flexible spending methods against a fixed inflation-adjusted withdrawal and found that every one of them supported a higher starting rate than the rigid base case. Not because flexibility conjures returns, but because it stops you from selling into a crash to fund a raise you gave yourself.
That is the whole game. The rigid rule is what makes the number so low in the first place. A retiree who will skip an inflation increase after a bad year, or trim discretionary spending 10% when the portfolio is down, has already bought back more safety than the entire 3.9%-versus-4.7% argument is worth.
Two things to do with that, in order:
- See what a bad first decade actually does to you. Stress-test withdrawals shows the same average return in two different orders — the arithmetic is identical, the outcomes are not.
- Decide, in advance, what you would cut. A plan with a written flex rule is a different plan from one that hopes.
When the IRS picks your rate for you
Your withdrawal rate is your choice right up until it is not. Required minimum distributions begin at age 73 if you were born between 1951 and 1959, and at age 75 if you were born in 1960 or later — which is most people still planning. From that point the IRS sets a floor under withdrawals from pre-tax accounts, regardless of what your plan wanted.
RMDs do not force you to spend the money. You can pay the tax and reinvest the rest in a taxable account. What they force is the taxable event, which is why the conversion and account-location decisions you make in your sixties matter more than the second decimal place of your withdrawal rate.
Key takeaways
- The 4% rule is Bengen's 1994 SAFEMAX: the worst-case survivor of US history, on a 50/50 portfolio, over 30 years. It was never a forecast.
- The two credible 2026 numbers are Morningstar's 3.9% and Bengen's revised 4.7%. They answer different questions. Do not average them, and do not treat either as precise.
- Higher equity does not buy a higher safe starting rate. Above roughly 50% stocks, volatility works against you.
- Fees come off the sustainable rate almost one for one. Taxes come off your spendable income on top of that.
- Flexibility raises the sustainable rate more than any number-picking exercise does.
Next: turn a rate into a target on how much is enough, or test the round number on is $1 million enough.
Educational only — not financial or tax advice.
FAQ
Is the 4% rule still valid in 2026?
As a rough first pass, yes. As a precise number, no. The two most credible current estimates sit either side of 4%: Morningstar's forward-looking figure is 3.9%, and Bengen's own revised historical worst case is 4.7%. Both are defensible. They are not measuring the same thing, so the gap between them is not an error.
Why do Morningstar and Bengen disagree about the safe withdrawal rate?
Because they ask different questions. Morningstar runs a forward-looking simulation and reports the rate that survives 90% of paths given today's valuations and yields. Bengen backtests actual US market history and reports the rate that would have survived the single worst 30-year period on record, using a more diversified portfolio.
Should I just average 3.9% and 4.7% and use 4.3%?
No. Averaging two answers to two different questions produces an answer to no question at all. Decide which framing fits your plan — a forward-looking probability, or a historical worst case — and use that number. Then test what happens if you are wrong.
Does the 4% rule account for taxes?
No. Both the original study and the modern updates model the portfolio, not the tax return. A 4% withdrawal from a traditional 401(k) or IRA is taxable income. Your spendable amount is lower than your withdrawal, and how much lower depends on your account mix and your other income.
How do fees change the safe withdrawal rate?
Almost one for one. A fee is a withdrawal you do not get to spend. If you draw 4% and pay 1% in fund and advisor fees, the portfolio is actually funding a 5% drain. Every published safe-rate figure assumes costs are low; none of them assume they are zero for you.
Is 4% safe for early retirement?
It was never tested for it. Bengen's SAFEMAX and Morningstar's base case both assume a 30-year horizon. A 50-year retirement is a materially different problem, and the sustainable rate is lower. There is no widely accepted single figure for it.
Do required minimum distributions override my withdrawal rate?
Eventually, yes — for pre-tax accounts. RMDs begin at 73 if you were born 1951-1959 and at 75 if you were born in 1960 or later. From then on the IRS sets a floor under your withdrawal whether your plan wanted one or not. You can reinvest the money in a taxable account, but you cannot leave it in the IRA.
Sources and notes
- Determining Withdrawal Rates Using Historical Data (1994) — as documented in 'Revisiting William Bengen's SAFEMAX Portfolio Withdrawal Rate'Journal of Financial Planning / Financial Planning Association · Accessed 2026-07-11The origin of the rule, and the source for its actual assumptions: 1926-92 data, a 50/50 US large-cap and intermediate Treasury portfolio, annual rebalancing, a 30-year minimum, and inflation-adjusted constant-dollar withdrawals. Bengen's term was SAFEMAX; he did not call it the 4% rule.
- Retirement Savings: Choosing a Withdrawal Rate That Is SustainableAAII Journal (Cooley, Hubbard and Walz) · Accessed 2026-07-11The 1998 Trinity study. Used for the second-source confirmation of the original finding and for the success-rate framing.
- What's a Safe Retirement Withdrawal Rate for 2026?Morningstar (The State of Retirement Income, 2025 edition) · Accessed 2026-07-11Source for the 3.9% base case and its assumptions (30-year horizon, 90% success, 30-50% equity), for the prior editions' figures, for the finding that higher equity weightings lower the safe starting rate, and for the finding that flexible spending methods support a higher starting rate than fixed real withdrawals.
- Bill Bengen Boosts the '4% Rule' to 4.7%Advisor Perspectives · Accessed 2026-07-11Secondary reporting on Bengen's 2025 book 'A Richer Retirement', which is the primary source for the revised 4.7% SAFEMAX and the seven-asset-class portfolio behind it. We flag this as secondary because the underlying work is in the book, not in a public paper.
- Retirement plan and IRA required minimum distributions FAQsInternal Revenue Service · Accessed 2026-07-11Used for the RMD start ages that eventually override a chosen withdrawal rate in pre-tax accounts.