Sequence-of-return risk is the silent decade

The order your returns arrive in can wreck a retirement that the average return says is perfectly funded. The danger years are the five before you stop working and the five after. If a crash lands in that window while you're drawing down, the damage is permanent. It's the single biggest threat to a plan that looks solid on paper.

Sequence-of-return risk explains why two people with the same savings, same average return, and same withdrawal rate can land in completely different places: one runs dry at 82, the other dies with millions to spare. The only variable that changed was when the bad years showed up. When losses coincide with withdrawals, you're forced to sell more shares at low prices, and compounding never gets the chance to repair the hole. That's the whole game.

Why the order of returns matters even when the average is the same

Withdrawals turn a temporary paper loss into a permanent one. While you're still accumulating, a 30% drop is almost good news: your contributions buy cheap, and time repairs everything. The math runs in reverse the day you stop earning and start spending. That same 30% drop now means you liquidate a bigger slice of the portfolio to fund the same lifestyle, and that slice is gone before the recovery arrives.

Run the numbers and the asymmetry is brutal. Picture two retirees, each starting with $2 million, each withdrawing an inflation-adjusted $80,000 a year, each averaging 7% a year over three decades. Retiree A hits a couple of ugly down years right out of the gate. Retiree B gets those same down years near the end. Same average, same withdrawals. A can run dry around year 24; B can finish with several million. The sequence did that, not the fund selection everyone obsesses over.

Why is this the silent decade for someone in their fifties?

If you're 52 to 62, you're standing in the most fragile window of your entire financial life, and it doesn't feel fragile at all. Your balance is near its peak. The statements look great. You've spent 30 years being told to stay the course, and that instinct served you brilliantly the whole time you were buying. It quietly stops serving you the moment the portfolio flips from feeding on contributions to feeding on itself.

This is where being in your second act actually helps you. The conventional 4% rule, which says withdraw 4% of your starting balance, adjust for inflation, and tune out, was reverse-engineered to survive the worst historical sequences, including 1966 and 1929. It's a floor for someone who refuses to ever look at the market again. You are not that person. You've run P&Ls, managed budgets, made capital calls under pressure. Your edge here isn't a better fund. It's the judgment to vary your behavior when the sequence turns against you, which is exactly the thing a rigid rule and a robo-advisor can't do.

I'll admit I got this one wrong for years. I treated "average return" as the master variable and waved off sequence risk as an academic footnote, the kind of thing advisors mention to sound rigorous. Then I watched a colleague retire in late 2007, insist on full withdrawals through 2008 and 2009 to protect his lifestyle, and never recover the ground. Same portfolio I'd have called bulletproof on paper. The sequence, plus his rigidity, ate it. I don't wave it off anymore.

How much cash should a retiree actually keep?

Enough to cover a couple of years of spending without touching stocks, and rarely more than that. Too little cash and a recession forces you to sell equities low; too much and inflation quietly erodes your spending power instead. There's no perfect number, but here are the targets I see work for people who actually live off their portfolio rather than a fat pension:

• One year of spending: the absolute floor for a cash and short-bond buffer. Below this you're exposed.
• Two to three years: the target most retirees should aim for, especially if you're retiring into a frothy market or you know you'd hate cutting back.
• More than three years: only makes sense if your withdrawal rate is unusually low (under 3%) and you value sleeping well over squeezing out returns.

The two-bucket test you can run this weekend

The rule I use fits on an index card. Before you retire, ask one question: can I cover two full years of spending without selling a single share of stock? That's the test. If the answer is no, you're exposed, and no amount of clever allocation fixes it.

The logic is plain. Most market drawdowns since World War II have recovered inside about two to three years. If you can fund your life from cash and short bonds while equities heal, the sequence can't force you to sell low. You convert a portfolio-killing event into a non-event. Build it as two buckets:

• Bucket 1, the moat (2 years of spending): cash, T-bills, a short-term Treasury or CD ladder. Boring on purpose. This is what you live on when stocks are down.
• Bucket 2, the engine (everything else): your equities, left alone to compound and recover. You refill Bucket 1 from here in good years, never in panic years.

The precise number matters less than you'd think. Some people run 18 months, some run three years. What counts is having a rule that tells you exactly what to do when the screen is red, so you're not improvising in the one moment improvisation costs you the most.

Where Claude is genuinely useful here (and where it isn't)

This is a planning problem, not a prediction problem, and that's the sweet spot for an AI like Claude. Don't ask it to forecast the market; nobody can, and it'll hedge appropriately if you push. Ask it to stress-test your own plan against bad sequences. That's work that used to mean a planner, a fat fee, and a two-week turnaround.

Take a composite of a pattern I see constantly: a 56-year-old former operations VP, call her Dana, sitting on roughly $1.8M, planning to retire at 62 and pull about $85,000 a year. She fed Claude her balance, her target withdrawal, and her Social Security timing, and asked it to walk her through what happens if a 2000-style or 2008-style sequence lands in her first three years, then to size the cash buffer that would let her ride it out without selling equities. Forty minutes of back-and-forth. She walked away knowing she needed closer to two and a half years in Bucket 1, not the six months of cash she'd been carrying. A real planner should still pressure-test the assumptions and the tax layer. But Dana showed up to that meeting fluent instead of nodding along, and she paid for advice instead of an education.

The honest limit: Claude doesn't know your actual returns, can't see the future, and shouldn't be your final word on anything with tax consequences. Use it to think faster and arrive prepared, not to replace the human who's legally on the hook for your plan.

What to actually do before you stop working

Forget timing the market; that's the one move guaranteed to make sequence risk worse. Do these instead. Build the two-year buffer in the five years before you retire, so you're not raising cash by selling into a down market the week you hand in your badge. Decide in advance, in writing, what you'll cut if the first two years are ugly, because a pre-committed 12% trim is painless and a panicked one isn't. And map your withdrawals so you're pulling from the right account in the right year, not just the most convenient one.

The whole risk compresses into a single sentence worth taping to your monitor: it's not whether the market falls, it's whether a fall forces you to sell. Build the buffer so it never can. Do that this decade, the silent one, and the sequence loses its teeth.