Here is the strangest fact in retirement math: once you start withdrawing, your average return stops deciding your outcome. The order of the returns takes over.
Meet two retirees. Each starts with $1,000,000 and withdraws $60,000 a year. Each earns the exact same 25 annual returns — three bad years (−25%, −15%, −10%) and twenty-two good ones (+12% each). Same numbers, same 8.56% arithmetic average. The only difference: Retiree A hits the three bad years first, right at retirement. Retiree B hits them last, in years 23–25.
| End of year | Retiree A (bad years first) | Retiree B (bad years last) |
|---|---|---|
| Start | $1,000,000 | $1,000,000 |
| Year 3 | $413,850 | $1,202,464 |
| Year 10 | $309,550 | $2,052,924 |
| Year 15 | $164,362 | $3,236,783 |
| Year 18 | $28,452 | $4,344,983 |
| Year 19 | Depleted | $4,806,381 |
| Year 25 | $0 | $3,598,251 |
Assumptions: $1,000,000 starting balance; $60,000 withdrawn at the end of each year; annual returns of −25%, −15%, −10% and twenty-two years of +12%, applied in opposite order for A and B. Hypothetical returns chosen to illustrate the mechanics of sequence risk; no fund, index, or Capital Wealth model performed this way, and taxes, fees, and inflation are ignored.
Retiree A is broke six years early. Retiree B dies with $3.6 million. And here is the kicker: with no withdrawals, both portfolios end at an identical $6.94 million — the order of returns literally doesn’t matter to a portfolio nobody is touching. Sequence risk isn’t a market phenomenon. It’s a withdrawal phenomenon: every dollar you sell in a down year is a dollar that can never recover. Selling at a loss to eat converts a temporary decline into a permanent one, year after year, until the compounding runs backward.
The math has a geography: the roughly five years on either side of retirement day dominate the outcome. Before that window, a crash hurts but you’re still buying, not selling, and there’s time to recover. After it, the portfolio has (in the good scenarios) grown a cushion. Inside that ten-year window you have the largest balance of your life, the withdrawals have just begun, and there’s no paycheck to wait it out. A 30% drawdown at 45 is a buying opportunity. The same drawdown at 65 — while you’re drawing — is the Retiree A table above.
This is also the hidden assumption inside the famous 4% rule. Bengen’s 1994 work didn’t find that 4% “usually works.” It found that 4% survived the worst historical starting years — retiring into 1929, or into the 1966–1982 inflation grinder. The rule is a sequence-risk artifact: the safe withdrawal rate is set not by average markets but by the unlucky orders. Retire into a good decade and 4% leaves millions unspent; retire into a bad one and 4% is the line between the two rows of that table.
And if this all sounds theoretical, 2022 ran the experiment live. Stocks fell about 18% (S&P 500) and the U.S. investment-grade bond index fell about 13% — the classic 60/40 portfolio had one of its worst years on record, with both engines failing at once, while inflation pushed the withdrawals up. Anyone who retired around 2021 and kept drawing a fixed lifestyle from a falling portfolio was living the left-hand column in real time.
Sequence risk can’t be predicted — nobody knows which decade they’re retiring into. But it can be engineered around, and the tools have real trade-offs worth stating plainly:
A cash buffer / income floor. Hold 1–3 years of spending in cash or short-term bonds, or cover the non-negotiable bills with guaranteed lifetime income (a pension, Social Security, or an annuity), so a bear market never forces you to sell stocks to eat. Cost: cash drags on returns in the good decades, and annuity guarantees are paid for in fees and flexibility — the floor discussion is worth having with the contract on the table.
Flexible “guardrail” withdrawals. Instead of a fixed inflation-adjusted draw, cut spending modestly when the portfolio falls below a guardrail and give yourself raises above it. This measurably extends portfolio life — but it requires a budget that can flex, which is a lifestyle design question, not a spreadsheet setting.
Risk-capacity allocation. Size the equity exposure to how much loss the plan can absorb in the fragile decade — not to your age or your nerve. Less equity near retirement blunts sequence risk but cedes long-run growth; the right answer depends on how much of your spending is already guaranteed.
Which brings us to the CalSTRS and CalPERS advantage, because it changes the whole problem: your pension is the floor. A teacher or deputy whose formula benefit already covers the mortgage and the groceries has, in effect, pre-purchased the expensive fix. The portfolio’s job is no longer “don’t run out” — it’s discretionary spending, inflation topping, and legacy. That retiree can often hold more equity through the fragile decade than a birthday-based glide path would ever allow, precisely because a bad first decade can’t break the household. Risk capacity, not age, sets the allocation — and a guaranteed check is capacity.
Map your own fragile decade: how many years until (or since) retirement day? Add up your guaranteed monthly income — pension, Social Security, any annuity — against your non-negotiable expenses. The gap between them is the only part of your spending exposed to sequence risk, and the portfolio should be built around that number. Then stress-test it: what does your plan look like if the first three years are 2022, twice? Our interactive sequence-of-returns explainer lets you watch the two-retiree math move, and the calculators let you run your own numbers.