What a 2008-style crash at retirement actually does
Everyone who retires with a stock-heavy portfolio eventually asks the same question: what if 2008 happens again — the year I stop working? The honest answer is more interesting than the fear. A crash at retirement is genuinely the worst-timed crash you can have, but what it does to a plan is specific, measurable, and worth simulating rather than dreading.
What 2008 actually was
In calendar 2008, US stocks (S&P 500, dividends included) returned −36.6%. But the rest of the row matters just as much: 10-year Treasury bonds returned +20.1%, cash earned +1.4%, and inflation was a mild +3.8%[2]. A 60/40 portfolio lost about 13.9% — painful, not fatal. And 2009 followed with stocks up +25.9%. The 2008 experience for a diversified retiree was a deep one-year dent with a strong bond cushion and a fast partial rebound. Contrast 2022, when stocks fell 18.0%, bonds fell 17.8% at the same time, and inflation ran 8.0%[2] — shallower headline losses, but nowhere to hide and every withdrawn dollar worth less.
Why timing is the whole story
Sequence-of-returns risk is the observation that two retirees can earn the same average return over 30 years and end up in completely different places depending on the order[1]. The mechanism is withdrawals. A portfolio nobody touches recovers fully when prices recover — the same shares are still there. A retiree, though, sells shares every year to eat. Selling into a trough turns a temporary price decline into a permanent loss of shares, and those shares are the ones that would have done the recovering. A crash in year 25 lands on a portfolio that has already carried the plan most of the way; a crash in year one taxes every subsequent year of the plan.
A worked year one
Take a $1,000,000 portfolio at 60/40, spending $40,000 a year — the classic 4% starting point[4]. Replay 2008 as the first retirement year: stocks fall to $380,700, bonds grow to $480,400, and after the $40,000 withdrawal the portfolio sits near $821,000. Nothing has failed — but next year’s $40,000 is now a 4.9% withdrawal rate, and the portfolio needs roughly +22% just to get back to its starting dollars while withdrawals continue. An all-stock retiree fares far worse: down to about $594,000 after the withdrawal, a 6.7% effective rate — deep in the territory where historical plans start failing. The same crash 25 years in barely moves either plan’s success rate. That asymmetry — the early years matter enormously, the late years barely at all — is sequence risk in one sentence, and it is why the first five to ten years of retirement are called the “fragile decade”[3].
What actually defends a plan
Three defenses show up repeatedly in the research. Bonds and cash held near the retirement date (a “bond tent”) mean the early withdrawals come from assets that didn’t crash — exactly what saved diversified 2008 retirees[3]. Flexible spending rules that trim withdrawals a few percent after bad years keep share-selling out of the trough. And a lower starting withdrawal rate buys margin against any sequence. None of them show up in average-return arithmetic; all of them show up clearly in a Monte Carlo simulation, which is the point of running one.
Try it in Deorbit Plan
The Market assumptions panel’s Stress test section has a toggle — Replay the 2008 crash in one year — that forces one simulated year on every path to 2008’s actual returns. Choose the timing: the first year of retirement (the classic worst case), one random year per path, or a specific calendar year, and optionally replay the 2009 recovery the following year. The injected crash appears as a marker in the event lane under the Fan chart. Run it with and without the crash — or with the crash in year one versus year 25 — and compare success rates in Compare scenarios to see sequence risk in your own numbers.
Educational content only — not financial, tax, or investment advice.
See how this plays out with your own numbers. Try it in the simulator →