An average return tells you almost nothing about what a retiree actually experienced. Two people can hold identical portfolios, earn an identical average return over an identical period, and end up thousands of dollars apart. The variable that separates them is the order in which the returns arrived.
This is sequence-of-returns risk, and it is one of the few concepts in personal finance where a small worked example settles the argument better than any explanation.
The arithmetic
Take two retirees. Each starts with $100,000 and withdraws $5,000 at the end of each year. Over five years each earns the same set of annual returns: minus 15%, minus 10%, plus 15%, plus 20% and plus 10%. Those average out to 4% a year for both.
The only difference is the order. Retiree A gets the two losing years first. Retiree B gets them last.
Retiree A, bad years first:
| Year | Return | Balance after return | After $5,000 withdrawal |
|---|---|---|---|
| 1 | −15% | $85,000 | $80,000 |
| 2 | −10% | $72,000 | $67,000 |
| 3 | +15% | $77,050 | $72,050 |
| 4 | +20% | $86,460 | $81,460 |
| 5 | +10% | $89,606 | $84,606 |
Retiree B, good years first:
| Year | Return | Balance after return | After $5,000 withdrawal |
|---|---|---|---|
| 1 | +10% | $110,000 | $105,000 |
| 2 | +15% | $120,750 | $115,750 |
| 3 | +20% | $138,900 | $133,900 |
| 4 | −10% | $120,510 | $115,510 |
| 5 | −15% | $98,183 | $93,183 |
Same portfolio, same average return, same withdrawals. Retiree B finishes with $93,183 and Retiree A with $84,606, a gap of $8,577, or about 10%.
Note what this example does not show: neither retiree ran out of money. Five years is a short window and these returns are mild. The point of the example is that the gap opens at all, from nothing but ordering. Extend the horizon to a thirty-year retirement, and steepen the early losses to something resembling 2000 to 2002 or 2007 to 2009, and the same mechanism is what separates a portfolio that lasts from one that does not.
Why the order matters at all
The mechanism is simple once you see it. Withdrawing a fixed dollar amount from a portfolio that has just fallen means selling a larger fraction of it. Those units are gone, and they are not there to participate in the recovery.
Retiree A's balance bottoms at $67,000 after two years. The subsequent 15%, 20% and 10% gains are real, but they compound on a base that has been cut by a third. Retiree B's early gains do the opposite: they lift the balance to $133,900 before the losses arrive, so the same percentage declines land on a bigger number and still leave more standing.
This is also why the risk is asymmetric with respect to timing. Someone still working, contributing, and withdrawing nothing is largely insulated. A bear market for a 35-year-old with three decades of contributions ahead is a period of buying at lower prices. The identical bear market for someone who retired last year is forced selling into a decline. Practitioners often describe the years immediately before and after retirement as the window where this risk is concentrated, for exactly this reason.
Where the 4% rule comes from, and what it actually claims
The best-known response to this problem is the "4% rule," which traces to research by the financial adviser William Bengen published in the Journal of Financial Planning in 1994. Bengen tested withdrawal rates against historical US market data and found that an initial withdrawal of about 4% of the portfolio, adjusted thereafter for inflation, would have survived every historical starting year in his sample, including the worst ones.
That last clause is the whole point, and it is routinely lost. The 4% figure is not an average return and it is not a prediction. It is a number derived from the worst sequence in the historical record. Bengen later revised his estimate upward when a wider set of asset classes was included. Both the original figure and the revisions rest on one country's historical returns over a specific period, which is a real limitation: a future sequence worse than anything in the sample is not excluded by the method.
Treat it as one analysis with stated assumptions, not a rule of nature.
What people do about it
None of the common responses eliminates the risk. Each trades it for something else.
Holding a cash or short-bond buffer. Keeping one to two years of planned withdrawals in cash means a downturn does not force the sale of equities at depressed prices. The cost is the return given up on that cash during rising markets, which over a long retirement is not trivial.
Varying withdrawals. Taking less in poor years and more in good ones directly attacks the mechanism, because it reduces the forced selling exactly when selling hurts most. The cost is income that is not predictable, which is harder to live with than it sounds when the variable being cut is your annual spending.
Working longer. Additional years shorten the withdrawal period, add contributions, and let the portfolio recover before drawdown begins. The cost is self-evident.
These are the approaches that come up most often in practitioner research, presented here as trade-offs rather than recommendations. Which, if any, is appropriate depends on circumstances this article knows nothing about. What is general, and worth carrying away, is the underlying point: for anyone drawing down a portfolio, the average return is not the number that determines the outcome.
