Is the 4% rule still valid?

Is the 4% rule still valid? The question assumes the rule was a promise about the future, and it never was. It was a finding about the past: the highest starting withdrawal that survived every 30-year stretch in one set of historical data. Whether it holds for you depends on how closely your situation matches the assumptions behind it. Those are worth reading, and you can test your own numbers in the how-long-will-it-last calculator as you go.

What the rule actually says

In the first year of retirement, withdraw 4% of the pot. Every year after that, withdraw the same amount again, raised by that year's inflation. The percentage is only used once. After year one you are spending a fixed real income, whatever the pot is doing.

So on a $1,000,000 pot, the first year is $40,000. If prices rise 2% in a year, the next year is $40,800, and so on. The rule answers one question: how big can that first number be before the money runs out too soon?

What Bengen tested in 1994

The rule comes from William Bengen's paper “Determining Withdrawal Rates Using Historical Data”, published in the Journal of Financial Planning in October 1994. His assumptions, as the paper states them:

• US data only, from 1926 onward, taken from the Ibbotson Associates 1992 yearbook.
• Two assets: US common stocks and intermediate-term Treasuries. His main charts use 50% of each, continually rebalanced.
• Withdrawals rising with inflation every year, exactly as described above.
• Tax-deferred accounts, so no tax is taken out. The paper deducts no fees either.
• A 30-year minimum as the test, with retirees of about 60 to 65 in mind.

On those terms, a 4% start never ran out in fewer than 33 years, and in most starting years it lasted 50 years or more, which was the longest he charted. Going to 4.25% could run out in 28 years. At 5% he called the rate risky, and at six percent or more, in his word, “gambling”.

What it never claimed

• It is not the typical outcome. 4% was the worst case. In most historical starting years the same withdrawal left a large pot behind.
• It is not a probability. Bengen reported how long each starting year lasted, not a percentage chance of success.
• It is not for any time span. It was tested against 30 years. A 45-year retirement is a different question.
• It is not about every portfolio. The result belongs to a stock-and-Treasury mix in the US. Other countries, other assets, or all cash, give different answers.
• It is not after costs. Fund and adviser fees come out of the same pot and were not in the sum.

How long 4% lasts if returns are steady

History is lumpy, but a steady return shows the bones of the arithmetic. Take $1,000,000, withdraw $40,000 in year one, and raise it by 2% a year. If returns average:

• 4% a year, the money lasts 34 years and 10 months.
• 7% a year, it never runs out inside 60 years, and after 30 years the pot has grown to about $2.8 million.
• 10% a year, it never runs out and keeps growing.

The steady return that makes the money last exactly 30 years here is 3.2% a year, about 1.2 points above inflation. For 40 years it is 4.6%, and for 50 years 5.2%. That is the first thing the rule's critics and defenders often talk past: for a 30-year span, the bar a steady return has to clear is low. For a long early retirement, it is not.

The starting rate matters a lot. At a steady 4% return, the same pot lasts:

• 3% start: 54 years 10 months
• 3.5% start: 42 years 6 months
• 4.5% start: 29 years 7 months
• 5% start: 25 years 9 months
• 6% start: 20 years 6 months

Why the order of returns decides it

If a steady 3.2% is enough, why did Bengen land on 4% rather than something higher? Because returns do not arrive steadily, and a fall early on does far more damage than the same fall late. Returns are not a straight line covers why.

Here are two hypothetical 30-year paths with exactly the same yearly returns, in a different order: three years of −15% and twenty-seven years of +9%. Both compound to about 6.3% a year.

• Bad years first: the money runs out in year 29.
• Bad years last: about $3.0 million is left after 30 years.
• A steady 6.3% every year: about $2.0 million is left.

Same average, same withdrawals, and one of the three runs dry. Bengen's worst starting years were the ones that ran into this. A 1966 start was the shortest at 33 years: it met the high inflation and falling markets of 1973 and 1974 early in retirement.

Is the 4% rule too conservative, or outdated?

Both complaints have some arithmetic behind them, and neither settles it.

• Too conservative: it is built around the worst case, so in most historical start years it left money on the table. Bengen's own charts show that.
• Outdated: the data ended in the early 1990s and covered one country, the US. Nobody knows whether the next 30 years will look like the worst of those, better, or worse.

What the arithmetic can do is show you how sensitive your own plan is. Change the return, the fee or the spending in the calculator and see how far the date moves. To go the other way, from spending to the pot it implies, use the retirement number calculator, which is where the “25 times your spending” version of the rule comes from: 25 is 1 divided by 4%.

Fees, tax and inflation shift it too

• A 1% annual fee on the example above, at a 4% return, cuts the money's life from 34 years 10 months to 29 years 1 month. A fee is a lower return by another name.
• 3% inflation instead of 2%, at a 4% return, cuts it to 29 years 2 months, because the withdrawal rises faster.
• Tax is not modelled here. Bengen assumed tax-deferred accounts, so 4% is a pre-tax withdrawal. What you keep of it depends on your country and account type.

Not financial advice. Worked examples assume a steady return, withdrawals rising with a fixed inflation rate, no tax and no fees unless stated. The return paths are illustrations, not forecasts. Historical results describe the past only.