Graham Stephan's 3% withdrawal rate calculator

Stephan says he personally follows 3%, not 4%. Run that rate against your balance, return and retirement length, and see what the extra caution costs.

Graham Stephan Reviewed Aug 16, 2026 Rule located in Graham’s Newsletter (grahamstephan.substack.com), 2023

Interactive model

Put your numbers through the rule

The source establishes the rule. The values below belong to you, and the output is JMM's deterministic calculation.

Make the assumptions yours

Every field recalculates immediately. Changed values can be copied into a shareable URL.

Retirement balance

Withdrawal rate

Compare against

Your annual return scenario

Inflation scenario

Retirement length

years

Years the money lasts40of the 40-year horizon

First-year income: $30,000
Ending balance in today’s dollars: $10,715,216
Years funded at 4%: 40
Return after inflation: 7.5%

Ending balance at 3%

$10,715,216

Ending balance at 4%

$8,272,209

Starting balance

$1,000,000

Balances are shown in today’s dollars, so a flat line means spending power is being preserved. The return and inflation figures are your scenario, not a forecast, and one fixed return hides the sequence risk that decides real retirements.

What is left, in today’s dollars

3% withdrawal4% withdrawalYears into retirement

Inspect the calculationThe same result in a readable record view

Year 5

At 3%: $1,248,309
At 4%: $1,185,869

Year 10

At 3%: $1,604,788
At 4%: $1,452,707

Year 15

At 3%: $2,116,560
At 4%: $1,835,788

Year 20

At 3%: $2,851,275
At 4%: $2,385,750

Year 25

At 3%: $3,906,054
At 4%: $3,175,292

Year 30

At 3%: $5,420,324
At 4%: $4,308,781

Source and translation

What the source says, and what the calculator adds

Source-supported ruleStephan’s own newsletter post on the 4% rule sets out the original rule, the 7.5% inflation-adjusted growth assumption behind it and its 99% historical success rate over 30 years, then reports two lower figures from other researchers, and states his own position directly: he personally follows the 3% rule, on the basis that he can cut back when needed but can also plan for the worst. He names 3.5% as the alternative if the retirement is only 35 years.

JMM calculationJMM runs the rate as an actual drawdown rather than a division, so the reader can see the year the money ends rather than a reassuring percentage. This is the same tested formula the 8% page on this site uses, with the source’s own defaults: a 40-year horizon, and inflation starting at zero because the 7.5% return the post quotes is already stated as inflation-adjusted, so charging for inflation twice would be the model’s error rather than the rule’s.

Formula and methodYear-one income = balance × rate. Income in year n = year-one income × (1 + inflation)^{n − 1}. Balance after year n = (balance − income) × (1 + return). The model reports the first year the balance reaches zero, or the balance surviving the horizon, and runs the same path at the comparison rate. Every withdrawal rate is also a multiple: the portfolio a given income demands is income ÷ rate, which is 33.3× spending at 3% and 25× at 4%.

G Graham’s Newsletter (grahamstephan.substack.com) Jan 18, 2023The 4% RuleSection “What should you do?”, the sentence “That’s why I personally follow the 3% rule, knowing that I can easily cut back when needed but I can also ‘plan for the worst’”, with 3.5% named in the same passage for a 35-year retirement. Earlier in the same post: the original rule is described as spending 4% of the portfolio a year, with a $1 million portfolio producing $40,000, a growth assumption of “7.5% every year adjusted for inflation”, and a “99% chance of not running out of money during a typical 30-year retirement” on a 75-25 stock and bond portfolio. The post also reports a 2.7% figure attributed to Ben Felix and a Vanguard range of 2.6% to 2.8%, rising to 3.3% on lower fees and 4.0% under a dynamic spending rule.

Decision notes

What changes the answer

The rate is a statement about the portfolio you must build, not the income you may take

Every withdrawal rate is a multiplier in disguise. Four percent means 25 times annual spending; 3% means 33.3 times. On $60,000 of spending that is $1.5 million against $2 million, and the extra $500,000 is what the caution costs in working years.

Three percent is a horizon adjustment, not a disagreement about markets

The 4% rule was tested on 30-year retirements. The source’s own alternative of 3.5% for 35 years shows the logic: the longer the retirement, the lower the sustainable rate. Someone stopping work at 40 is running a 50-year plan, and the 4% evidence base simply does not cover it.

The source undercuts its own number in both directions

The same post reports researchers at 2.7% and Vanguard at 2.6% to 2.8%, and also notes that a dynamic spending rule lifts the safe rate back to 4.0%. Three percent is a midpoint between a rigid low number and a flexible high one, which is exactly what the author says he is doing when he says he can cut back.

Right number, wrong reason to admire it: the flexibility is doing the work, not the 3%.

JMM has no argument with 3% as a starting rate for a genuinely long retirement, and it is a welcome corrective to the 8% figure published elsewhere on this site. But read the sentence carefully and the rule is not really a rule. The stated justification is that he can easily cut back when needed. A retiree who will cut spending in a bad year is running a dynamic strategy, and the source’s own post says a dynamic rule supports 4.0%. So the honest description is not “3% is the safe rate”, it is “I have chosen to under-spend a flexible plan”, which is a defensible personal choice and a poor general instruction. The cost is not abstract. Moving from 4% to 3% adds a third to the portfolio you must accumulate, and for most earners that is measured in extra working years, taken from the healthiest end of retirement. If you can cut back, say so and plan at a higher rate with rules for when to cut. If you cannot, 3% is the number, and this page will show you what it demands.

Written by the JMM editor. This is a view, not a recommendation for your circumstances. Who writes this

Not the calculator’s limits. The rule’s.

It buys safety with years you may not want to spend
Thirty-three times spending instead of 25 is typically three to seven extra working years at a normal savings rate. Those years come out of your fifties or sixties, when health and time are worth most. A rule that is silent about what it costs is only half a rule.
For a conventional retiree with Social Security it is simply too low
Someone stopping at 66 with a state pension covering part of their essentials is running a 25 to 30-year horizon with a partial income floor. That is the situation the 4% evidence was built on. Applying an early-retirement rate to it produces a plan that dies with a very large unspent balance, which is a failure mode too, just a quiet one.
The flexibility that justifies it is not in the arithmetic
The rule as computed assumes inflation-indexed spending that never falls, which is the opposite of the behaviour the author cites as his reason for choosing 3%. Model the flexibility and the required rate rises; model the rule and the flexibility disappears. The two cannot both be counted.
A single flat return hides the only risk that actually breaks these plans
Two retirements with identical average returns end differently depending on which years were bad, because early losses are withdrawn against and never recover. Lowering the withdrawal rate reduces that exposure but does not describe it, and no fixed-rate rule can.

Limits

What this model does not know

The return is a single flat rate every year. Real markets deliver a sequence, and a bad first decade breaks plans a flat average would have funded.
The inflation field starts at zero because the source’s 7.5% is stated as already inflation-adjusted. If you raise inflation you must also raise the return to a nominal figure, or the model will understate the plan twice over.
Taxes, fees, Social Security, pensions and every other income source are excluded, which matters more at 3% than at 4% because the rule is aimed at people retiring before those start.
Spending is assumed to rise with inflation and never to be cut. The source’s own justification for 3% is the ability to cut, and the model cannot represent that flexibility.
Long-term care, health shocks and living past the entered horizon sit outside the model.

Questions people ask

Before you use the result

Does Graham Stephan reject the 4% rule?

Not exactly. His post explains the 4% rule and its 99% historical success rate over 30 years, then says he personally follows 3% because he can cut back when needed and wants to plan for the worst. He names 3.5% for a 35-year retirement.

Why does the inflation field default to zero here?

Because the 7.5% growth figure in the source is described as already adjusted for inflation. Applying inflation to the withdrawals as well as using a real return would charge for inflation twice. Raise both fields together if you prefer nominal figures.

What does 3% instead of 4% actually cost?

A third more capital. Four percent implies 25 times annual spending, 3% implies 33.3 times. On $60,000 of spending that is $2 million rather than $1.5 million.

Is this the same model as the 8% withdrawal page?

Yes, the same tested formula with different defaults and a different source. The comparison is deliberate: two published personal-finance figures, more than double apart, both presented to readers as prudent.

Take the answer further

The next question this page cannot answer

Keep researching Graham Stephan

Person fileGraham Stephan

Open the verified role and source record behind this current strategy.

General calculatorRun the underlying math without the source context

Use the reusable JMM tool with your own assumptions.

Another Graham Stephan rule20% rent rule calculator

What rent fits inside 20% of your gross income?