Sequence-of-Returns Risk: Why the First Five Years of Retirement Matter Most

Two hypothetical retirees each earn the exact same average annual return over a 25-year retirement — say, 6% a year, averaged across good years and bad. One of them retires comfortably with money to spare. The other runs out early. The difference isn't skill, luck in stock-picking, or fees. It's the order in which the good and bad years happened to arrive — a phenomenon called sequence-of-returns risk, and it's one of the more counterintuitive ideas in retirement math, because "average return" is exactly the number most people assume determines the outcome.

The same average, two different outcomes

During the accumulation years — while someone is working and adding money to a portfolio — the order of returns barely matters, because there's no ongoing withdrawal amplifying the damage of a bad year. A market decline early in a career is, if anything, a buying opportunity: new contributions purchase more shares at lower prices, and decades remain for the account to recover.

Retirement flips that dynamic. Once someone starts withdrawing a set amount each year to live on, a market decline is no longer just a paper loss — it's a paper loss combined with money simultaneously leaving the account. Withdrawing from a shrunken balance locks in losses that a portfolio in the accumulation phase would have simply ridden out.

Why withdrawals change the math

The mechanism is straightforward once you see it: a fixed-dollar (or inflation-adjusted) withdrawal represents a larger percentage of a portfolio after it has fallen in value than before. Selling a larger percentage of a smaller pool leaves fewer shares remaining to participate in the eventual recovery — so even when the market does bounce back, the portfolio has fewer units left to ride that bounce, and the math never fully catches up, even though the average return across the whole period looks identical to a scenario where the bad year came later.

A side-by-side illustration

Imagine two hypothetical retirees, each starting with $1,000,000 and withdrawing $50,000 a year, adjusted for a flat 2% inflation assumption to keep things simple. Both experience the exact same five annual returns over a five-year stretch — say, +20%, +15%, +5%, −10%, and −15% — just in a different order.

Retiree A gets the good years first: +20%, +15%, +5%, −10%, −15%. Retiree B gets the bad years first: −15%, −10%, +5%, +15%, +20%. The five-year average return is identical for both — it's the same five numbers. But because Retiree A's portfolio grew substantially before absorbing the two down years, the withdrawals in those early years came out of a much larger base, leaving more shares intact when the recovery happened. Retiree B, withdrawing from a portfolio that had already shrunk in years one and two, sold a proportionally larger chunk of a smaller balance during the worst years — and by the time the +15% and +20% years arrived, there was meaningfully less capital left to benefit from them.

Run the actual dollar trajectories and Retiree B typically ends the five-year window with a noticeably smaller balance than Retiree A, despite an identical average return and identical withdrawal amounts. Extend that same dynamic across a 25- or 30-year retirement, and an unlucky early sequence can be the difference between a portfolio that lasts and one that doesn't — again, with the same long-run average return either way.

Why this risk is concentrated early, not spread evenly

The reason this risk clusters specifically in the first several years of retirement, rather than being evenly distributed across the whole retirement, is that the portfolio is at its largest dollar value right at the start — meaning early losses are being withdrawn against the biggest base and have the most shares to permanently remove from future growth. A bad year in year 20 of a 30-year retirement, when the portfolio (and the withdrawal's percentage impact) has already been substantially drawn down regardless, simply does less structural damage than the identical bad year in year one or two.

This is also why sequence risk gets less attention in the accumulation-phase conversation and dominates the retirement-phase one: the entire asymmetry depends on withdrawals being layered on top of returns, which only happens once someone starts drawing the portfolio down.

Mitigation concepts, in general terms

None of this means retirees are helpless against the order the market happens to deal them — it means the early retirement years deserve more defensive attention than a simple "average expected return" framework would suggest. Common concepts people discuss in this context include holding a cash or short-term-bond buffer sized to cover a couple of years of expenses, so a bad market year doesn't force selling depressed shares to fund that year's withdrawal; and building flexibility into the withdrawal amount itself, so a down year can be met with a smaller withdrawal rather than a fixed one. Neither is a guarantee, and the right mix depends on a household's full financial picture — the point here is understanding why the concept matters, not a specific product or allocation recommendation.

There's also a planning-side lever worth naming: the withdrawal rate chosen at the very start of retirement implicitly bakes in an assumption about how kind the first several years of returns will turn out to be. A retiree who begins with a notably lower withdrawal rate than a textbook average-return calculation would suggest is, in effect, buying insurance against a bad early sequence — spending a bit less than the long-run average math implies is "safe," in exchange for a bigger cushion against the specific years that carry the most structural weight.

The practical takeaway isn't that average returns are meaningless — over long enough horizons, they still drive most of the outcome. It's that in the specific window where withdrawals begin, the order returns arrive in carries its own separate risk, one that a single "expected average return" number will never show you.