How much can you actually withdraw?

What a “safe withdrawal rate” actually means

It's the share of your portfolio you take in the first year of retirement, then adjust for inflation each year after. A 4% rate on a $1M portfolio means $40,000 in year one, and $40,000 of buying power every year after that — more dollars each time, because prices rise. It is not a rule about how much you can take every year forever; it is a starting point that history suggests survives a long retirement.

Where the 4% rule came from

In 1994 William Bengen tested withdrawal rates against actual US market history and found that 4% survived every 30-year period he examined, including retirements beginning at the worst possible moments. The Trinity study, a few years later, reached broadly similar conclusions using a slightly different method. Together they gave the FIRE movement its most quoted number — and, unhelpfully, its most over-quoted one.

The experts disagree, and it's worth knowing why

Bengen himself has since revised his figure upward — to roughly 4.7% — after adding asset classes his original work didn't include, notably small-cap stocks. Morningstar, working from forward-looking return and valuation assumptions rather than historical replay, has published figures nearer 3.9%. Both are credible; they differ because one asks “what would have worked?” and the other asks “what is likely to work from here?”

That gap — call it 3.9% to 4.7% on a $1M portfolio, or $39,000 to $47,000 a year — is not a failure of the research. It is the honest width of the uncertainty, and any tool presenting a single number as the safe rate is hiding it from you. Move the inputs above and watch how much the answer depends on assumptions you chose.

Sequence risk is what actually kills plans

Two retirements can earn the identical average return over thirty years and end in completely different places, purely because of the order the returns arrived in. A bad first decade means withdrawing from a shrinking portfolio, and there is less left to recover with when markets turn. The same bad decade at the end lands on a portfolio that has already compounded, and barely registers.

This is why the worst start years in the record — 1929, the mid-1960s, 2000 — are the ones the whole field argues about, and why an average-return spreadsheet will always be more optimistic than reality deserves. The simulation above replays those sequences intact rather than smoothing them.

What this does not model

Taxes. Social Security or a pension. Healthcare costs and ACA subsidies. Long-term care. Spending that changes as you age — and it does; most people spend less in their seventies than their sixties. Any of these can move the answer more than the withdrawal rate itself, and a calculator that quietly assumes them away is telling you less than it appears to.

One more caveat worth stating plainly: historical cycles overlap, so they are not independent samples. A 95% success rate means “this plan survived 95% of the overlapping historical windows”, not “there is a 95% chance your retirement works”. The future is allowed to be worse than anything in the record.