The 4% rule: does it still work? Every retirement since 1872 tested

The 4% rule says: withdraw 4% of your retirement portfolio in the first year, raise that amount with inflation every year, and the money should last 30 years. We tested it on every 30-year retirement in US data since 1872, 125 start years in all. With 60% stocks and 40% bonds, 124 of the 125 lasted the full 30 years. The one that did not started in 1966. This guide explains where the rule comes from, what our test shows, why the first ten years of a retirement matter more than the next twenty, and what the rule does not cover. Backtested results are hypothetical.

What the 4% rule is

The 4% rule is a spending rule for a retirement portfolio. In the first year of retirement you withdraw 4% of the portfolio's value. Every year after that you withdraw the same amount, adjusted for inflation, no matter what the markets did. On a portfolio of $1,000,000 the first withdrawal is $40,000; if inflation runs at 3%, the second is $41,200, and so on.

The rule comes from William Bengen, a financial planner, who published it in 1994 in the Journal of Financial Planning. He asked a plain question: looking at US stock and bond returns since 1926, what is the highest first-year withdrawal rate that would have lasted 30 years from every starting year? His answer was a little above 4%, for a portfolio of roughly half stocks and half bonds. A 1998 study by three professors at Trinity University (Cooley, Hubbard and Walz, published in the AAII Journal) tested a range of withdrawal rates and stock–bond mixes on US data and reached a similar conclusion for a 30-year horizon. Those two papers are why "4%" became the number everyone quotes.

Two things about the rule are often missed. First, it is a worst-case number: it is the rate that survived the single worst starting year in the data, not the rate an average retiree needed. Second, it was derived from US history, so it tells you what would have worked, not what will.

What our test shows

We ran the same question on a longer data set: US stocks and 10-year Treasuries (US government bonds) from 1872 to 2025, with the portfolio rebalanced once a year back to its target mix and withdrawals raised with inflation each year. No fees are deducted unless stated. That gives 125 retirements of 30 years, starting in every year from 1872 to 1996. All figures below are backtested and hypothetical.

At 4% with a 60/40 portfolio, 124 of the 125 retirements lasted 30 years. That is 99%. The one failure started in 1966 and ran out of money in year 29. The 1966 retiree met a decade of high inflation and flat stock markets in the first ten years: withdrawals rose with inflation while the portfolio did not.

The highest rate that survived every start year depends on the mix and the horizon:

Two patterns stand out. More stocks did not buy much more safety: 60/40, 80/20 and all-stock portfolios all supported about 3.9–4.0% over 30 years. Too few stocks did cost safety: the 20/80 portfolio supported only 3.4% over 30 years and fell to 2.3–2.7% over 40 to 50 years, because bonds alone did not keep up with inflation-adjusted withdrawals. Too little risk is also a risk.

Every extra point of withdrawal costs a lot of safety. With 60/40 over 30 years: at 4.0%, one failure in 125; at 4.5%, 12 failures; at 5%, 29 failures (77% lasted); at 6%, 59 failures (53%). A 6% rule is close to a coin flip.

Fees count like withdrawals. With 1% a year in fees, the highest rate that survived every start falls from 3.9% to about 3.5% (3.45%), and 4% fails 10 times instead of once (92% lasted). Fees are the part of a retirement plan you control.

Most retirees ended with more than they started with. The median 60/40 retiree at 4% ended the 30 years with 1.4 times the starting value after inflation. The best start, 1982, ended at 5.7 times. The 1929 start ended at 0.84 times, with money left but less than it began with. 13 of the 125 ended below half of the starting value. The retirement that started in 2000 is still running; after 26 years, to 2025, it stands at about two thirds of its starting value after inflation.

Why the first years decide: sequence risk

Three retirees followed the same 4% rule with the same 60/40 portfolio. The one who started in 1982 ended with 5.7 times the money. The one who started in 1929 ended with 0.84 times. The one who started in 1966 ran out in year 29. Same rule, same mix, three very different outcomes.

What separated them was the order of returns. A retiree who withdraws a fixed, inflation-adjusted amount sells more shares when prices are low. A bad first decade, with falling or flat markets and rising prices, forces the portfolio to sell at the worst time, and there is less left to recover when markets improve. A good first decade does the opposite. This is called sequence-of-returns risk, and it is why the average return over 30 years tells you very little about whether a withdrawal plan survives.

It also explains why the worst start years cluster. Of the 125 start years, the eight that supported the lowest sustainable 60/40 rate all fall between 1962 and 1973, the run-up to the inflation of the 1970s. Those eight start years supported 3.9% to 4.3%. The best start years, 1921 and 1982, supported about 10.5%. Backtested, hypothetical (60/40, inflation-adjusted withdrawals, every start since 1872).

Two numbers, not one

That spread is the most useful thing in the data. "The safe withdrawal rate" is really two numbers: about 4% for a retiree who happens to start just before a bad decade, and about 10% for one who starts just after it. Nobody knows in advance which retiree they will be, which is why the rule is built on the worst case.

It also tells you what the rule is for. If you plan at 4% and your retirement turns out to start in a good year, you will most likely leave money behind; the median outcome was 1.4 times the starting value. If it starts in a bad year, 4% is about the most history says a 60/40 portfolio could carry for 30 years. The rule trades upside in good cases for survival in bad ones. Whether that trade suits you is a personal decision, not a data question.

What the 4% rule does not cover

It is a worst case of the US past, not a guarantee. Every number here comes from 154 years of one country's markets. The worst sequence in that history, 1966, defines the rule. A worse one is possible, and so is a better one.

It assumes rigid spending. The rule raises withdrawals with inflation every year, whatever happens. Real retirees can spend less after a bad year, which raises the survival rate, or spend more, which lowers it. Flexible spending rules exist; this guide does not test them.

It ignores taxes. Withdrawals are shown before any tax. Depending on where the money sits and where you live, the amount you can spend is lower. We give no tax advice.

It covers the portfolio only. Pensions, state benefits and other income are outside the rule. They reduce what the portfolio must supply, which is the same as choosing a lower withdrawal rate.

It assumes a fixed mix and a two-asset portfolio. The test uses US stocks and 10-year Treasuries, rebalanced yearly. Other assets, other countries and other rebalancing schedules give other numbers. Our guide to the 60/40 portfolio covers the construction side; our guide to the stock–bond correlation explains why the bond half protected less in some decades than in others.

Test your own plan

The free safe withdrawal rate calculator runs your plan against every start year since 1872: choose the withdrawal amount, the stock share, the horizon and the annual fee, and it shows how many start years survived, the worst start year, the highest rate that lasted from every start, and the range of outcomes in today's money.

Two ways to read it. If your plan survives every start year, it survived the worst sequence in US history. If it survives, say, 90% of start years, the failures tell you which decades broke it, and that is the scenario a flexible spending rule would need to handle.

The free white paper When Bonds Stop Hedging explains why the bond half of a 60/40 portfolio protected less in the 1960s and 1970s than in 2000–2020, the decades that decided the worst retirements in this test. The free Research Terminal adds the strategy-versus-buy-and-hold comparison on 30 markets and the research library; no credit card is needed.

Frequently asked questions

Hamann, H. (2026). The 4% rule: does it still work? Every retirement since 1872 tested. InvestWealth.com. https://www.investwealth.com/learn/guides/4-percent-rule

<a href="https://www.investwealth.com/learn/guides/4-percent-rule">The 4% rule: does it still work? Every retirement since 1872 tested</a> (InvestWealth.com)