Sequence-of-Returns Risk: Why the Same Average Can End Two Ways
Two portfolios can earn the same average return and end with very different balances, because the order of those returns decides how much a withdrawal costs. This guide runs the same sequence two ways and shows how much the ordering alone is worth.
A retirement portfolio that is being drawn down faces a problem that an accumulating portfolio does not: the withdrawals happen on a fixed schedule, while the returns arrive in an unpredictable order. Because money taken out at the bottom cannot participate in the recovery that follows, two investors can experience the same set of annual returns, in different orders, and end with meaningfully different balances. This is sequence-of-returns risk, and it is arithmetic rather than opinion.
Setting up the comparison
Assume a portfolio of $500,000 at the start of a four-year period. The investor withdraws $20,000 at the end of each year, and that withdrawal is not adjusted for inflation in this simplified example. The four annual returns are −20%, +20%, −20% and +20%, so the arithmetic average is 0% either way. Only the order changes.
Case A: a bad year first
| Year | Return | Balance after return | After $20,000 withdrawal |
|---|---|---|---|
| 1 | −20% | $400,000 | $380,000 |
| 2 | +20% | $456,000 | $436,000 |
| 3 | −20% | $348,800 | $328,800 |
| 4 | +20% | $394,560 | $374,560 |
Case B: a good year first
| Year | Return | Balance after return | After $20,000 withdrawal |
|---|---|---|---|
| 1 | +20% | $600,000 | $580,000 |
| 2 | −20% | $464,000 | $444,000 |
| 3 | +20% | $532,800 | $512,800 |
| 4 | −20% | $410,240 | $390,240 |
Same four returns, same average of 0%, same withdrawals, same starting balance. Case A ends at $374,560 and Case B ends at $390,240 — a gap of $15,680, or 4.2% of the starting balance, produced purely by the order. That is the whole idea in one number.
The mechanism is easy to state: the bad year first shrinks the base on which every later withdrawal is measured, and each withdrawal then removes a larger slice of a smaller portfolio. Money withdrawn in year 1 never participates in the year 2 recovery. Had there been no withdrawals at all, the two orders would have produced exactly the same ending balance, because multiplication does not care about order. Withdrawals are what make the order matter.
Why the gap grows with the depth of the early fall
The −20% figure above is deliberately mild. Extend the case to a deeper early decline: a $500,000 portfolio withdrawing $20,000 a year, with a −30% year followed by four +10% years.
| Year | Return | Balance after return | After $20,000 withdrawal |
|---|---|---|---|
| 1 | −30% | $350,000 | $330,000 |
| 2 | +10% | $363,000 | $343,000 |
| 3 | +10% | $377,300 | $357,300 |
| 4 | +10% | $393,030 | $373,030 |
| 5 | +10% | $410,333 | $390,333 |
Now move the same −30% year to the end instead of the start, keeping the four +10% years in between:
| Year | Return | Balance after return | After $20,000 withdrawal |
|---|---|---|---|
| 1 | +10% | $550,000 | $530,000 |
| 2 | +10% | $583,000 | $563,000 |
| 3 | −30% | $394,100 | $374,100 |
| 4 | +10% | $411,510 | $391,510 |
| 5 | +10% | $430,661 | $410,661 |
The bad year late leaves $410,661; the bad year first leaves $390,333. The gap has widened to $20,328 on identical returns. The deeper the early loss, the more expensive it is, because the withdrawal made at the reduced balance locks in a share of the decline.
Why this is discussed as a risk rather than a return problem
Note what has not been assumed anywhere: no market forecasts, no claims about future returns, no assumption that any particular asset is superior. The result follows entirely from the interaction between a fixed withdrawal schedule and variable returns. That is why the standard framing treats it as a withdrawal-rate problem: a drawdown that would be uncomfortable-but-survivable for an accumulating portfolio can become permanent damage for a portfolio that is being spent.
Retirement research built on this observation typically explores the relationship between withdrawal rates, portfolio composition and the probability of running out of money under historical return sequences, and it generally reaches the conclusion that the first decade of withdrawals carries disproportionate weight. Two limitations are worth stating plainly. Such research is backward-looking: it describes what would have happened under the sequences that actually occurred in the data, and the past sequence is one draw from a distribution, not a guarantee. And the studies usually assume a fixed, inflation-adjusted withdrawal pattern, which is a simplification — real households adjust their spending, which changes the arithmetic materially.
What the arithmetic does establish
The ending balance depends on the order of returns whenever money is being withdrawn as well as on the average. That is a durable, verifiable statement about compounding with cash flows, and it is the reason the same average return can produce two different retirements. It does not establish what any particular person should hold, how much they should withdraw, or how markets will behave. It establishes only that when a portfolio is being spent, the sequence of returns is part of the arithmetic rather than a footnote to it.
The asset allocation guide covers why the mix between growth and defensive assets affects the range of sequences an investor might face, and the behaviour gap guide looks at how investor actions during a decline add to the problem. The articles page indexes the archive and the tools page has calculators for compounding with cash flows.
This guide is general educational information, and it is neither financial, tax, nor legal advice. It recommends no specific security, fund, or provider, and it describes no individual's situation. All balances, returns, withdrawal amounts and periods above are simplified illustrations with the assumptions stated; they are not forecasts, and they do not depict any real investor's outcome. Sequence-of-returns research is based on historical data and assumes patterns of contributions, withdrawals and spending that individual households may not follow, so its findings describe historical sequences rather than predicting future results. Withdrawal rates, account rules and market conditions vary by jurisdiction and change over time. Consult a licensed financial professional about your own circumstances before acting.
