Sequence of Returns Risk: Why the Order of Returns Matters More Than the Average

An average return tells you almost nothing about what a retiree ends up with. This is not a statement about volatility drag or about the difference between arithmetic and geometric means. It is a statement about order — and it only becomes true the moment money starts leaving the portfolio.

The Demonstration

Take a $1,000,000 portfolio and a twenty-year sequence of annual returns:

−20%, −13%, −7%, then +9% for the remaining seventeen years.

The arithmetic mean is 5.65%. Now consider the exact same twenty returns in reverse order: seventeen years of +9%, then −7%, −13%, −20%.

With no withdrawals, both sequences end at exactly the same place: $2,801,191. Multiplication is commutative; order cannot matter when nothing enters or leaves.

Now withdraw $50,000 at the start of each year — a 5% initial withdrawal rate.

Same returns. Same average. Same number of years. Same withdrawals. A 5.2x difference in the ending balance.

Why Withdrawals Change Everything

The mechanism is straightforward once seen.

In the bad-first sequence, the year-one withdrawal of $50,000 comes out of a portfolio that then falls 20%. The retiree has permanently removed 5% of the portfolio at the top and then absorbed the decline on what remains. By the end of year three, the balance has fallen to $527,961 — nearly halved — from a combination of withdrawals and losses compounding on each other.

The critical part: the recovery years operate on a much smaller base. The seventeen years of 9% growth apply to roughly $528,000 instead of $1,000,000. The market fully recovered; the portfolio could not, because the shares sold during the decline are not there to participate.

In the good-first sequence, the early withdrawals come out of a growing portfolio. By the time the losses arrive in years 18–20, the balance is large enough that the same $50,000 withdrawal is a much smaller percentage, and the decline lands on a base that has already compounded for seventeen years.

This is the same mechanism that makes an emergency fund valuable during accumulation — selling assets into a decline converts a temporary loss into a permanent one — operating continuously and at scale.

Why It Does Not Apply During Accumulation

The symmetric point, less widely understood: for a portfolio with no flows, sequence is irrelevant. Both sequences above ended at $2,801,191.

For a portfolio with contributions, the sequence effect actually reverses. A saver contributing monthly benefits from poor early returns, because the contributions buy more shares at lower prices, and benefits from strong late returns applied to a larger accumulated base. A 30-year-old should, on a purely financial basis, prefer a bear market now and a bull market at 60.

The asymmetry between accumulation and decumulation is the whole point:

The Retirement Red Zone

Because the effect is concentrated in the years when withdrawals begin against the largest balance, practitioners commonly describe a red zone spanning roughly the five years before and five years after retirement.

Two things coincide there. The portfolio is at or near its peak, so a percentage decline costs the most in absolute dollars. And withdrawals are beginning or about to begin, which is what converts a paper decline into realized selling.

A 35% drawdown at age 60 with no withdrawals yet is recoverable over the following decade. The same drawdown at 66, with $60,000 a year coming out, is a materially different event — and the difference is entirely the flows.

Four Mitigations

None eliminates the risk. Each reduces it, and each has a cost worth naming.

1. A cash buffer or bond tent

Hold one to three years of spending in cash and short-term instruments, and draw from it during market declines instead of selling equities. The equity portfolio is left to recover, and the buffer is refilled from equities in years when markets are up.

A related structure is the rising equity glidepath, or bond tent, described by Michael Kitces and Wade Pfau: reduce equity exposure into retirement to its lowest point around the retirement date, then increase it again over the following decade. The logic is that equity exposure is most dangerous precisely at the red zone and less so afterward, once the sequence risk has largely passed.

Cost: holding several years of spending in cash creates a permanent drag on expected returns. The mitigation is real and it is not free.

2. Flexible withdrawals

Static withdrawal rules ignore what the portfolio is actually doing. Dynamic rules respond to it.

The best-known are Guyton-Klinger guardrails, which raise withdrawals when the portfolio outperforms and cut them when the withdrawal rate drifts above a ceiling. Simpler variants include skipping the annual inflation adjustment after a down year, or setting a floor of essential expenses with a discretionary layer above it that flexes.

The research is consistent that flexibility permits a higher initial withdrawal rate than a fixed rule, because the plan does not have to survive the worst case unchanged.

Cost: variable income. A retiree who cannot cut spending in a bad year cannot use this strategy, which makes the size of the discretionary layer the binding constraint.

3. A lower initial withdrawal rate

The most direct lever. The 4% figure from the Trinity study and Bengen's original work was derived specifically to survive the worst historical sequences in the US data — it is a worst case number, not an expected one. Lowering the initial rate to 3.5% or 3.25% widens the margin further.

Cost: a lower withdrawal rate means either spending less or working longer. This is the mitigation with the most explicit and least avoidable price tag. See safe withdrawal rate.

4. Guaranteed income floors

Social Security is inflation-adjusted lifetime income immune to sequence risk. Delaying benefits from 62 to 70 increases the monthly amount by roughly 76% before cost-of-living adjustments — the largest sequence-risk mitigation available to most households, and it requires no product purchase.

The general principle: covering essential expenses with guaranteed income means market declines only affect discretionary spending, which is exactly the spending that can flex.

Cost: delaying Social Security requires bridging the gap from the portfolio, which increases withdrawals in the early years — the very years when sequence risk is highest. The tradeoff is real, and the research generally still favors delay for the higher earner in a couple, largely because of the survivor benefit.

The Honest Caveat About the Evidence

Every one of these conclusions rests on backtests, and backtests have a limitation worth stating plainly: there is only one history.

The US data from 1926 to the present contains roughly one century of returns, and the overlapping 30-year retirement windows within it are far from independent — a 30-year period starting in 1965 shares 29 years with one starting in 1966. The effective number of independent 30-year observations is closer to three or four than to ninety.

This matters in both directions. The 4% rule survived the worst US sequence in the sample, which is reassuring and also a statement about a single draw from an unknown distribution. And the US sample is itself a survivor: research by Elroy Dimson, Paul Marsh, and Mike Staunton covering multiple countries over the twentieth century found that US equity returns were near the top of the international distribution, which suggests global backtests would produce lower safe withdrawal rates than US-only ones.

The reasonable conclusion is not that the framework is useless — it is the best available. It is that the precision implied by a figure like "3.87% is safe" is not supported by the data behind it, and that the flexibility mitigation matters partly because it does not require the historical distribution to repeat.

Frequently Asked Questions

Does a more conservative allocation solve this? Partly, and with a tradeoff. Lower equity exposure reduces the size of early drawdowns but also reduces expected returns, which raises the risk of running out of money over a 30-year retirement. The Trinity study's central finding was that both extremes fail: portfolios with very low equity exposure fail from insufficient growth, and very high ones fail from drawdowns. Something in the 50–75% equity range was most robust across historical sequences.

Does the risk apply to a portfolio that is never drawn down? No. A retiree living entirely on Social Security, a pension, or dividends without selling principal faces no sequence risk on the untouched portfolio — the order genuinely does not matter when there are no flows. It does apply to anyone selling assets to fund spending.

How does this interact with Required Minimum Distributions? RMDs force withdrawals from traditional accounts starting at age 73 regardless of market conditions, which means a mandatory sale into a down market. The distribution must be taken, but it does not have to be spent — moving the proceeds into a taxable account keeps the money invested, satisfying the requirement while avoiding the realization of the loss in economic terms.

Is early retirement more exposed? Yes, on two counts: a longer horizon means more sequences to survive, and there are more years before Social Security begins to provide a guaranteed floor. Early retirees also typically face a bridge period funded entirely from the portfolio, which is the highest-withdrawal, highest-sequence-risk stretch of the whole plan.

Does dollar-cost averaging help? It is the accumulation-phase analogue and works in the opposite direction — regular buying benefits from low prices. It does not address decumulation. The withdrawal-phase equivalent is the cash buffer and flexible withdrawal approach above.

This content is for educational and informational purposes only and does not constitute financial or investment advice. Equity Rank is not a registered investment adviser. All figures are arithmetic illustrations computed from the stated return sequences and withdrawal amounts; they are not projections and do not describe any actual portfolio. Historical results do not indicate future outcomes. Consider consulting a qualified financial professional about your own retirement plan.