The 4.5% rule

The short answer

Withdrawing 4.5% a year means saving 22.2 times your annual spending. Across 69 historical 30-year retirements since 1928 it survived 87% of them, against 94% for the 4% rule. Stretch the retirement to 50 years and it falls to 59%.

How it held up, by how long the retirement lasted

The rule is usually quoted without a horizon attached, which hides the part that matters. The original research tested 30 years. An early retirement is a different question.

Historical success rate of the 4.5% rule by retirement length
Retirement lengthPeriods testedSurvivedSuccess rate
20 years 7979100%
30 years the classic test696087.0%
40 years 594169.5%
50 years 492959.2%

1928 to 2025 market data, 70% stocks and 30% bonds, spending raised by actual inflation each year. Longer horizons have fewer complete periods to test, which is why the count falls as the length rises.

The horizon is the whole story: the same 4.5% rule went from 87.0% over 30 years to 59.2% over 50. That is 27.8 points lost to nothing but time. Anyone retiring in their forties is planning for the second number, not the first, which is why a single rate quoted without a horizon is close to meaningless.

What the 4.5% rule demands

A rate is really a multiple in disguise. 4.5% means 22.2 times spending, and that multiple is what you spend working years buying.

Portfolio required by annual spending at 4.5%
Spend per yearAt 4.5%At 4%
$40,000$889K$1.00M
$60,000$1.33M$1.50M
$80,000$1.78M$2.00M
$100,000$2.22M$2.50M

Every row costs less than the 4% column, which is the appeal: $222,222 less capital on $80,000 a year. That is years of working life you do not spend saving.

What you are buying with those years is a 13% historical failure rate over 30 years. The failures were not randomly distributed either. They fell on people who happened to retire into 1929, 1966 or 1973, which is sequence risk and not bad luck spread thinly.

Against the other rules

Withdrawal rate rules compared over 30 years
RuleMultipleSurvived 30yr
4.5% (this page)22.2x87.0%
4% rule25.0x94.2%
3% rule33.3x100%
3.5% rule28.6x100%
5% rule20.0x73.9%

Test 4.5% on your own numbers

These figures assume a fixed allocation and spending that never flexes. Yours will do neither. Run the plan against the same 95 years with your actual inputs.

60 seconds, no signup required

Common questions

What is the 4.5% rule?

Withdraw 4.5% of your portfolio in the first year of retirement, then raise that dollar amount with inflation each year after. It implies a target of 22.2 times your annual spending, against 25 times for the better known 4% rule.

Does the 4.5% rule work?

Across 69 historical 30-year retirements since 1928, it survived 60 of them, or 87%. The 4% rule survived 94% of the same periods. Over 50 years the 4.5% figure falls to 59%, which is the number that matters for an early retirement.

How much do you need for the 4.5% rule?

22.2 times your annual spending. Spending $40,000 a year needs $888,889, and $80,000 needs $1,777,778. That is the whole cost of the rule: a lower rate is safer and demands a larger portfolio, so it is bought with working years.

Should I use 4.5% instead of 4%?

Only with something else absorbing the risk: a pension, a paid-off house, part-time income or genuine willingness to cut spending after a bad year. At 4.5% the historical failure rate over 30 years is 13%, and the failures were not spread evenly. They landed on people who retired just before a crash.

Assumptions on this page. Backtests use 1928 to 2025 market data at 70% stocks and 30% bonds held constant, with the first year's withdrawal raised by actual inflation each year after and never adjusted for how the portfolio performed. Nobody spends that way, and the inflexibility is why these figures are conservative. Success means the portfolio still had money at the end, which is a low bar: a plan that finishes with $1 counts. Taxes, fees and Social Security are excluded, so treat the draw as pre-tax and as the whole of your income. The spending-level pages count Social Security. Past returns are not a forecast. These are planning estimates, not advice; read the research behind the rule and the model's limitations.