Sequence risk is a timing problem wearing a math costume.
Two people can retire with the same $1 million, take the same $40,000 a year, and live through the same set of market returns. One dies with money left. The other runs out. The average return was identical. The order was not.
That is sequence-of-returns risk. When you are taking money out, the early years matter more than the brochure average.
Why the order suddenly matters
While you are still working and adding money, a bad year is unpleasant and often useful. Shares get cheaper. Later deposits buy more of them. Over a long career the sequence mostly washes out.
Retirement flips the cash flow. You sell instead of buy. A drop plus a withdrawal means you sell more shares to raise the same dollars. Those shares are gone. When the market recovers, they do not come back. The rebound happens on a smaller pile.
A simple version of the trap:
Start with $1 million. Withdraw $40,000.
If year one is −10%, you have about $864,000 left after the withdrawal. If year one is +10%, you have about $1,056,000. Same withdrawal. Different base. Every later return compounds on that new number. Reverse the good and bad years and the 20-year endings can be hundreds of thousands of dollars apart, even when the list of annual returns is the same list played backward.
Inflation makes it worse. The withdrawal usually rises with prices. After a bad start you are taking a larger check from a smaller account.
The dangerous window
The worst stretch is roughly the five years before retirement through the first decade after. Michael Kitces and others have shown that real returns in that first decade do most of the work in deciding whether a given withdrawal rate survives 30 years. After the portfolio has either grown a cushion or been cut down, later crashes hurt less. A 25% drop at 85, on a smaller remaining balance, is not the same event as a 25% drop at 66 while you still need three decades of checks.
This is also why the 4% rule is a historical survival test, not a promise. William Bengen’s work asked which starting withdrawal would have lasted through the worst past sequences, including retirements that began before long bear markets. The rule exists because sequence risk exists.
Accumulation versus spending, in one sentence
No withdrawals: order does not change the ending value if the returns are the same set.
Withdrawals: order can change whether the account survives.
That is the whole distinction.
What actually helps
You cannot pick your retirement-year market. You can change how much damage a bad open does.
Spending flexibility is the blunt instrument that works. Cutting withdrawals after a drop — even for a few years — keeps shares in the account for the recovery. A fixed inflation-adjusted paycheck is what turns a bear market into a permanent haircut.
Cash or short bonds covering one to three years of spending give you a place to write checks without selling stocks at the low. A three-bucket setup extends that idea: cash for the near term, bonds for the next several years, stocks for later. You refill the cash bucket from bonds or from stocks after they recover, not during the slide.
A “bond tent” does the same job with allocation instead of labeled buckets. Equity weight steps down into retirement, then can rise again once the fragile years have passed. The idea is to shrink the amount sitting in volatile assets during the window when a loss plus a withdrawal does the most harm.
Guaranteed income changes the math too. Social Security, a pension, or an annuity covers a slice of the budget so the portfolio withdrawal can fall when markets fall. Lower required selling is lower sequence risk.
Starting with a lower withdrawal rate is the unfashionable version of all of this. A 3.5% first-year draw leaves more room than 5%. The 4% figure was never a dare.
What it is not
Sequence risk is not “stocks are bad in retirement.” A portfolio that is too conservative can fail the other way: it never grows enough to keep up with 30 years of rising spending.
It is also not a reason to sit in cash from 62 onward. Cash solves the next two winters. It does not fund the 2040s.
And it is not the same as average-return risk. You can have a fine 30-year average and still fail if the ugly years arrive while the account is still large and the checks are still non-negotiable.
The plain version
If you need to sell every year, a crash at the beginning steals shares you needed for the ending. A crash at the end steals shares you were almost done using. Same market history. Different life.
Plan the first decade as if the market might be rude. Keep some spending that can flex. Keep some assets that do not need a bull market to fund next year’s groceries. Then let the growth assets do their job after the sequence has either broken you or not.