Two portfolios can earn the exact same average annual return over twenty years and end in very different places, depending on nothing but the order the returns arrived in. This is sequence risk, and it only matters once money is being withdrawn — which is precisely the period most people are least prepared to think about it.
An accumulation portfolio does not care about order
While money is going in and nothing is coming out, the order of good and bad years is close to irrelevant to the ending balance. A down year early, followed by strong years, ends in roughly the same place as strong years first, followed by a down year late. The arithmetic of compounding without withdrawals is symmetric enough that sequence washes out.
A withdrawal portfolio cares enormously
Once withdrawals start, that symmetry breaks. A down year early in retirement forces withdrawals to be taken from a smaller balance, which permanently reduces the capital base available to participate in the recovery that follows. The portfolio never gets the chance to compound back from the low point on the full original balance, because a chunk of that balance has already been withdrawn and spent.
The reverse is also true: strong returns early in a withdrawal period build a larger cushion before any bad year arrives, and that cushion absorbs the bad year far more comfortably.
What this means in practice
Average return is the wrong number to plan around for the years immediately before and after a withdrawal period begins. What matters more in that window is the sequence — and because nobody controls market sequence, the practical response is structural rather than predictive: holding enough in short-duration, low-volatility assets to fund several years of withdrawals without selling equities into a decline, so a bad sequence in the market doesn't force a bad sequence of sales.
This is not a case for avoiding risk altogether. It is a case for understanding that the same portfolio, the same average return, and a different five-year stretch of bad luck at the start can produce meaningfully different outcomes decades later.