What is the 4% rule, and can you actually rely on it?

The 4% rule is a rule of thumb for how much a retirement portfolio can pay out each year without running dry: withdraw 4% of the portfolio's value in the first year, adjust that dollar amount for inflation every year after, and — based on the worst 30-year stretch in the historical US market record — the money holds out. Flip the fraction over and it becomes the more familiar shorthand behind a FIRE number: a portfolio worth roughly 25 times annual spending.

It is a reasonable starting point and a poor set of driving instructions. The number was built from a specific setup — a particular time horizon, a particular mix of stocks and bonds, a withdrawal that only ever moves for inflation — and any one of those assumptions can be wrong for an individual retirement without the rule itself being wrong. What follows is where the number comes from, what it quietly assumes, why the order returns arrive in matters as much as their average, and how to actually use a safe withdrawal rate instead of just quoting one.

Where the 4% rule comes from

A financial planner named William Bengen asked a specific, testable question: using real US market history, what is the highest withdrawal rate — a fixed percentage of the starting portfolio, increased each year for inflation — that never emptied a portfolio within 30 years, across every historical starting date? Tested against a portfolio split between US stocks and intermediate-term bonds, the answer came out close to 4%, even for the unluckiest retirees, whose 30 years happened to start right before a prolonged downturn. A few years later, three finance professors at Trinity University re-ran a broader version of the same test — more allocations, more withdrawal rates, more starting dates — and landed on a similar number, which is why the same idea sometimes travels under the name "the Trinity study" instead of Bengen's own.

Both are backward-looking tests, not laws of markets. They ask what would have preserved a portfolio through the worst sequence that actually happened, not what is guaranteed to happen again. Every future retirement is a new sequence the historical worst case did not include, which is worth remembering before treating 4% as a figure handed down from physics rather than one estimated from a finite set of past outcomes.

What the rule actually assumes

Behind the single number sits a specific setup, and moving any one piece of it changes the safe rate:

That last row is easy to miss and matters most. The 4% figure comes from the worst starting year in the dataset ending with something left over, not with a cushion. Most historical starting years did far better than merely surviving — many retirees following the rule would have ended up with several times their starting balance — but the rule is calibrated to the unluckiest case on record, not the typical one, which is exactly why it holds up as a floor rather than a guarantee for the case that turns out worse than any seen before.

Why more margin is usually the right call now

Two of the assumptions above argue for building in margin rather than treating 4% as precise. The first is horizon: a traditional retirement starting in the mid-sixties fits reasonably inside the 30-year window Bengen tested, but early retirement plans and FIRE targets often need 40 or 50 years of withdrawals, and a longer horizon gives a bad early sequence more time to compound its damage before growth can offset it. This is one reason some FIRE variants target 3.5% instead of 4% — the extra margin is priced directly into the multiple.

The second is starting valuations. The historical test averages across cheap markets and expensive ones, but any individual retirement starts from wherever prices happen to sit on the day withdrawals begin. Starting from richly valued markets has historically been associated with lower subsequent decade returns, which does not doom a 4% plan outright but does argue for treating the rate as an upper bound to test against, rather than a floor to build on top of. Neither point means abandon the framework — it means use it with room to spare, particularly for a longer retirement.

Sequence-of-returns risk: order matters as much as average

The single biggest reason a "safe" rate has to be lower than the average return would suggest is sequence-of-returns risk: withdrawals taken during a downturn lock in losses that would otherwise have been temporary, because the shares sold to fund spending are gone and cannot recover when the market does. The portfolio's average return over the retirement barely matters if the bad years land early, while the identical average return causes little damage if the bad years land late instead.

The cleanest way to see it is with two portfolios that experience the identical set of ten annual returns, only in reverse order of each other:

Bad years firstBad years last

Same average return, same withdrawals, opposite order: two 10-year retirements

	Bad years first	Bad years last
Year 0	1000000	1000000
Year 5	632229	1533228
Year 10	879823	1129217

Both portfolios start at $1,000,000 and withdraw $40,000 (a fixed real 4%) at the start of each year, before that year's return is applied. Both experience the identical ten annual returns — ranging from −15% to +20%, averaging 5% a year — only in reverse order. Same withdrawals, same average return, same total withdrawn.

Same ten returns, same 5% average, same $400,000 withdrawn over the decade — and a $249,394 gap in the ending balance, driven entirely by which years the losses landed in. The retiree who hit the downturn first spent the whole decade withdrawing from a smaller and smaller base at the worst possible moment; the retiree who hit it last had already built a large enough cushion that the same losses barely dented the plan. This is the mechanism the 4% rule is built to survive, and it is also why the rule's safe rate is well below the market's long-run average return rather than close to it — the gap between the two is the price of insuring against exactly this kind of bad luck showing up in year one instead of year twenty.

Flexible spending is the practical fix

A rigid, never-adjusted withdrawal is the hardest version of this problem to survive, because it keeps pulling the same real amount out through a downturn regardless of what the portfolio is doing. Flexible rules relax that in one of a few ways: skip the inflation increase in any year that follows a negative return, cap how much a withdrawal can rise or fall from one year to the next, or simply recalculate the withdrawal as a percentage of the current balance rather than the original one, so spending naturally falls after a bad year and rises after a good one. Each of these, tested against the same historical data the original rule used, supports a higher starting withdrawal rate than a rule that never bends — because the flexibility does some of the work a static number can't. The trade-off is not free: some years, spending genuinely has to come down, and a plan only works if the retiree is actually willing to make that behavioural adjustment when the numbers call for it, not just in theory.

Run your numbersRetirement decumulation→

Retirement accounts add their own wrinkle a flexible spending rule doesn't fully solve on its own. A registered retirement income fund carries a mandatory minimum withdrawal that rises with age, and by the late seventies and eighties that minimum can exceed what even a cautious rate would call for — the forced withdrawal may need to be redirected into a TFSA or a non-registered account rather than spent, to keep actual spending on the plan the flexible rule set rather than on the mandate. The decumulation planner models registered-account minimums alongside the withdrawal rate itself, which a flat 4%-of-portfolio calculation on its own does not.

Use it as a multiple, not autopilot

The most useful thing the 4% rule does is turn a spending figure into a target: multiply annual spending by 25 (or by roughly 28–29 for the more conservative 3.5%) and you have a number to save toward, not a number to withdraw against blindly once you get there. The rate is where you start the conversation with your own plan, and the sequence-of-returns math above is exactly why the conversation shouldn't stop at the starting number — a plan that only checks "am I at 25 times spending yet" ignores everything about when the retirement begins relative to the market, which the multiple alone cannot see.

The better use of the framework is as a stress test rather than a formula: work out your spending-based number with the FIRE calculator, then run the actual withdrawal path — including how it holds up if a downturn lands in the first few years rather than the middle — through a tool built to model the full sequence, not just the average. A rule of thumb earns its keep by telling you roughly where to aim; the plan that gets you there safely has to look at more than the average return the thumb was pointed at.