The luck of when you retire — why the order of returns decides everything

Picture two people who retire on the same day with the same one million dollars. They invest identically. Over the thirty years that follow, the market delivers exactly the same set of annual returns to both: the same big up years, the same brutal crashes, the same dull middling years. At the end, the average return each earned is identical to the last decimal, and they each took out the same inflation-adjusted income every year.

Common sense says they end up in the same place. Common sense is wrong, and the gap between what feels true and what is true here is enormous. One can be comfortably wealthy, with more than they started with. The other can be destitute, having burned through the entire portfolio with a decade of retirement still to go. Same savings, same average return, same spending, same investments. Opposite outcomes.

What separates them is not skill, fees, or asset allocation. It is luck — specifically, the luck of *when* the good and bad years showed up. Get a crash in your first few years and it can be fatal; get the very same crash in year twenty and you will barely notice. The order of returns, which no investor controls and few even think about, turns out to decide almost everything. This article is about why.

The strange thing about averages

Start with a fact that sounds like it should make the problem disappear. If you have a lump sum invested and you never add to it and never take anything out, the order of your returns does not matter at all. You can shuffle thirty years of gains and losses into any sequence you like and your final balance is exactly the same every time.

The reason is just arithmetic. Growing money is multiplication, and multiplication does not care about order: a +10% then a −20% gives the same result as −20% then +10%, because 1.10 × 0.80 equals 0.80 × 1.10. String together thirty such factors and the product is fixed however you rearrange them. For a buy-and-hold lump sum, only the compound (geometric) growth rate over the whole period matters; the path is irrelevant to the destination.

Run the experiment. Take a five-year stretch with one terrible year (−50%) and four strong recoveries (+25% each) — an average of +10% a year. Put the crash first, last, or in the middle: a million-dollar lump sum with no withdrawals lands on exactly $1,220,703 every single time. But add withdrawals and the symmetry shatters.

Both retirees get the crash — A in year 1, B in year 5 — and the identical average (+10%) and identical total withdrawn ($250,000). With no withdrawals they converge on the same $1,220,703. With withdrawals they end $216,210 apart, B ending ~27% richer, purely from order.

What changes the instant you start spending

When you stop leaving the money alone and start taking income out of it, the comfortable symmetry breaks completely. Every year you sell a slice of your portfolio to fund your life, and the price you get depends on what the market is doing when you sell. The mechanism in one sentence: when the market falls and you keep selling to eat, you are selling more shares at low prices, permanently shrinking the base of capital that could have ridden the recovery back up. A withdrawal in a downturn does double damage — it removes cash you needed, and it removes it at the worst exchange rate. Those shares are gone; when the rebound comes, it lifts a smaller pile.

This is why the early years are so dangerous and the later years so forgiving. A crash in year one hits the largest portfolio you will ever have, and then you spend years selling into the hole. A crash in year twenty hits a portfolio with two decades of compounding behind it and only a handful of withdrawals left to fund — the damage is real, but the portfolio does not need to survive much longer.

There is a subtle point that makes sequence risk deeper than the usual "watch out for volatility." For the lump-sum investor, volatility hurts in exactly one way: the bumps drag the compound return slightly below the simple average. But that penalty is *order-independent*. Our two retirees have identical returns, so identical volatility and identical compound growth. By every measure a textbook would use, they are the same — and yet one is ruined. The retiree who is spending faces a second, separate penalty the lump-sum investor never sees, and it is driven entirely by the sequence.

Two paths, same returns: converging vs diverging

Watch a real retirement break

Five years and one crash cannot do the idea justice. Stretch it to a full thirty-year retirement and the gap stops being a difference in comfort and becomes the difference between solvency and ruin.

Take a realistic set of thirty annual returns for a balanced portfolio — a couple of severe down years (a 37% collapse, a 22% fall), some mild losses, a long middle of ordinary gains, a few strong years near the top. It averages 9.0% in simple terms and compounds at 7.8%. Held as an untouched lump sum, a million dollars grows to roughly $9.5 million over thirty years — the same figure however you arrange the years.

Now retire on it with the textbook 4% rule: withdraw $40,000 in year one, then raise the dollar amount 3% a year for inflation. Give one retiree the returns with the worst years stacked at the front, the other with the best years first. Same thirty returns, same 9.0% average, same 7.8% compound rate, same withdrawal rule.

The retirement that broke vs the retirement that soared

Worst-first and best-first are the extremes — real markets do not sort returns tidily, so real retirements land between these poles. But the poles define the stakes. When literally the only thing you change is the order, and the outcome swings from total ruin to a $7 million estate, you cannot blame anything else.

Why year one is fatal and year twenty is survivable

Make the timing precise. Imagine a single one-time crash of 40%. Hold everything else constant and just slide that crash through the calendar — year one, year two, on through year thirty. The result is a steep, lopsided curve. A 40% crash in year one is close to catastrophic; in year five it is bad but recoverable; by year fifteen a minor inconvenience; by year twenty-five almost irrelevant, because the portfolio only needs to last a few more years.

Where the same crash lands — damage by timing

There is an asymmetry worth naming. The early-crash retiree suffers the crash *plus* a feedback loop: as the portfolio falls, the fixed real withdrawal becomes a larger and larger percentage of what remains. A $40,000 withdrawal is 4% of a million; after a bad couple of years it might be 7%, 8%, even 10% of the shrunken balance. The withdrawal rate the retiree never chose to raise has effectively ballooned, and at those rates almost no portfolio survives — while inflation keeps pushing the dollar amount higher. The lucky retiree experiences the mirror image: withdrawals shrink toward a rounding error as the portfolio swells.

The runaway withdrawal rate

This is the engine inside the 4% rule

You have heard of the 4% rule: withdraw 4% of your starting portfolio, adjust for inflation, and reasonably expect the money to last about thirty years. What is less understood is that the rule is, at its core, an answer to the problem this article describes — it exists because of sequence risk.

The number comes from William Bengen's 1994 work, which tested withdrawal rates against the actual historical record back to the 1920s. He ran the same strategy across every thirty-year period he could and asked: what is the highest initial rate that would have survived *every* historical period, including the worst one? That worst-surviving rate — his SAFEMAX — came out to roughly 4%. (We rebuild that analysis from scratch in the 4% rule, revisited.)

Sit with what "survived the worst period" means. The 4% figure is not calibrated to the average retirement — on average you could have withdrawn 5, 6, even 7% and been fine, because most thirty-year stretches were kind. The 4% is set by the single unluckiest *starting year*, and the worst case is defined entirely by sequence. The most instructive is not the 1929 retiree but the cohort who retired around 1966: the Depression was followed by deflation and recovery, so a 4% strategy came through, but the 1966 retiree faced fifteen-plus years of going nowhere in real terms, with 1970s stagflation pairing weak markets and high inflation that forced ever-larger withdrawals from a stagnant portfolio. That is the textbook bad sequence — the binding constraint that pins the safe rate near 4%.

So when someone debates whether the safe rate is 3.5, 4 or 4.5%, what they are really arguing about is how much protection to buy against an unlucky sequence. The entire rule is a sequence-risk insurance premium, quoted as a spending cut.

Retirement-date roulette

The most unsettling implication: the single largest factor in whether your retirement succeeds may be the calendar — and the calendar is not something you chose. Two equally diligent savers who stop working three or four years apart can face wildly different odds because one walked into a bull market and the other into a crash.

Same rule, different start dates (1966 vs 1982)

You cannot eliminate this luck, but you are not blind to it. There is one weak but real signal: market valuations when you retire. Retire into expensive markets — a high cyclically-adjusted price-to-earnings ratio (CAPE) — and future returns have historically tended to be lower and sequence risk higher. It is a tendency, not a forecast, but enough to justify more caution about your withdrawal rate when starting from rich valuations — the rational core behind a "dynamic," valuation-aware safe rate.

What you can actually do about it

Sequence risk is manageable — its teeth can be pulled even though you cannot choose your returns. The single highest-leverage defence: protect the first decade. Sequence risk is overwhelmingly front-loaded, so any strategy that lets you avoid selling stocks into an early downturn does most of the work.

The most reliable form of protection is holding two to five years of spending in assets that do not crash with equities — cash, short-term bonds, a money-market fund — and drawing income from that reserve during declines instead of selling shares at depressed prices (a cash buffer or "bucket" strategy). It severs the exact mechanism that does the damage, precisely when it matters. The second pillar is flexibility: the 4% rule assumes you raise withdrawals with inflation regardless of markets, which is what turns an early crash into a death spiral. A retiree who can flex spending 10–20% in a downturn has bought an enormous margin of safety for free.

How to measure your own exposure

The first step in managing sequence risk is seeing it — and it is measurable. The right way to pressure-test a plan is not "what return do I expect on average," but "what happens if the bad years come first?" That means running your strategy against the worst historical sequences and thousands of simulated orderings, and looking at the share of paths in which your money survives. A plan that succeeds on average but fails in a quarter of bad-sequence scenarios is not safe; it is a bet.

Two numbers are worth watching: your effective withdrawal rate (annual spending ÷ current portfolio), tracked yearly so a decline that pushes it well above your starting rate is an early warning to flex; and your safe withdrawal rate itself, which depends on your horizon, asset mix, flexibility, and the valuations you retire into.

This is exactly what the safe withdrawal rate calculator is built to run — testing your withdrawals across the historical record, including the unlucky 1966-style starts that set the 4% rule, and showing how much a cash buffer, a lower initial rate, or flexible spending shifts your odds. (To see how long a given pot lasts under a fixed draw, pair it with the how-long-will-it-last calculator.) If you take one action after reading this: stop planning around your average return, and start planning around your worst sequence.

The uncomfortable conclusion

We like to believe retirement security is earned — that diligent saving and a good long-run return will be rewarded with a comfortable old age. Mostly they are. But sitting on top of all that effort is a layer of pure luck — the order in which your returns arrive — and in the unlucky cases that luck can overwhelm everything you did right. Two retirees, identical in every way a spreadsheet can measure, can end three decades apart, one wealthy and one broke, because of nothing more than a reshuffling of the same set of years.

The 4% rule, for all its fame, is best understood not as a promise but as a confession: it is the number that survives the cruelest order history has thrown at us, set that low precisely because order can be so cruel. You cannot control when you retire relative to the market, and you cannot control the sequence you will face. What you *can* control is how exposed you let yourself be to it. Hold a buffer so you never have to sell low. Stay flexible so a bad year does not compound into a bad decade. Be more cautious when valuations are rich. Floor your essentials if you can. None of it changes your luck — all of it changes how much your luck is allowed to matter. In a problem defined by what you cannot control, that distinction is the whole game.