The order of returns decides retirements that averages cannot explain
The claim
Two retirees experiencing identical average returns can face opposite outcomes depending on when the bad years arrive.
Why we rate it strong
A mathematical property of withdrawing from a volatile portfolio, demonstrable directly on the historical record.
During accumulation, the order of returns is irrelevant to the final balance if contributions are fixed — multiplication commutes. A crash early in your career is arguably good news: your ongoing contributions buy more units at lower prices.
Withdrawal reverses this completely. Selling to fund spending from a portfolio that has just fallen 30% means liquidating far more units for the same income, and those units are not there to participate in the recovery. Identical average returns arriving in a different order produce wildly different outcomes.
The historical record makes this concrete. A retiree starting in 1966 faced a decade and a half of poor real returns immediately, and a 4-5% withdrawal plan was in serious difficulty. A retiree starting in 1982 with the same plan and a similar long-run average finished with a far larger portfolio than they began with. Neither outcome had anything to do with skill.
The practical responses are unglamorous and effective: hold a cash and bond buffer sized to fund several years of spending without selling equities, and retain the ability to reduce withdrawals after a bad year. Our sequence-risk simulator runs your plan from every historical starting year so you can see the spread rather than a single average.
Where this breaks down
- The risk is concentrated in roughly the five years either side of retirement — the point of maximum portfolio value relative to remaining contributions.
- Over-correcting has its own cost. A portfolio held too conservatively through a 30-year retirement runs a serious risk of being eroded by inflation instead.
- Historical windows overlap heavily, so the apparent number of independent bad sequences is smaller than the count of start years suggests.
Sources
Follow these rather than taking our word for the summary.
William P. Bengen (1994). Determining Withdrawal Rates Using Historical Data
Journal of Financial Planning, 7(4), 171-180
Finding: Portfolio failures cluster in retirements beginning immediately before sustained real declines, not in those with low average returns.