Milevsky Ch. 6: Sequence Risk and the Order of Your Returns

"In accumulation the order of returns is irrelevant. In decumulation it is nearly everything." — the asymmetry that ends retirements

Investment Background

Here is the fact that separates retirement investing from every other kind, and it is genuinely counterintuitive.

While you are saving, the order in which returns arrive does not matter. A portfolio that earns +20%, −10%, +8% ends at exactly the same place as one that earns +8%, −10%, +20%. Multiplication is commutative. This is why accumulation advice can be simple: buy the index, keep costs low, keep contributing.

The moment you begin withdrawing, that commutativity breaks. Money leaves the portfolio each year, so a decline early — when the balance is at its largest and you are selling into it — permanently removes shares that would otherwise have participated in every subsequent recovery. The same decline arriving in year 25 is close to harmless.

The magnitude is larger than most people expect. Take two retirees, each starting with $1,000,000, each withdrawing $45,000 annually adjusted for inflation, each earning exactly the same set of annual returns over thirty years — just in reverse order from one another. The one who receives the bad years first can be out of money in the early twenties of the sequence. The one who receives them last can finish with well over a million dollars. Identical average return. Identical volatility. Opposite outcomes.

Milevsky's framing is that sequence risk is not a risk you can diversify away, forecast, or out-invest. It is a structural property of withdrawing from a volatile portfolio. What you can do is change how exposed you are to it in the years when it does the damage — and that window is short.

The Wall Street Translation

Where the Risk Actually Lives

Sequence risk is overwhelmingly concentrated in the first ten years of withdrawals, and most acutely in the first five. This is sometimes called the retirement red zone.

The mechanism is simple arithmetic. A 30% decline in year 2, combined with a 4.5% withdrawal, means you are selling roughly 6.4% of the remaining portfolio to fund the same spending. Do that for two years and you have permanently liquidated a fifth of your shares at depressed prices. Those shares are not there for the recovery. The same 30% decline in year 25 hits a smaller balance over a shorter remaining horizon and is largely irrelevant.

The practical implication is that retirement risk is not spread evenly across thirty years. It is front-loaded into a decade that begins the day you stop earning. That decade deserves defenses that the following twenty years do not.

The Four Defenses That Actually Work

Defense	How it works	Cost
Cash and short-bond buffer	Spend from it in down years instead of selling equities	Small drag from holding low-return assets
Flexible spending rule	Reduce withdrawals when the portfolio falls	Requires genuine discretionary spending to cut
Rising equity glide path	Start retirement lower in equity, increase over time	Gives up some early upside
Lifetime income floor	Essentials do not depend on the portfolio at all	Liquidity, bequest, fees (see Ch. 4)

Note what is absent from that list: market timing, tactical allocation, buffered products and anything that promises to sidestep the decline. Those do not defend against sequence risk; they add cost and a new failure mode. The defenses that work all share a common feature — they reduce the need to sell equities into a decline. That is the entire mechanism.

The cash buffer is the most accessible. Two to three years of portfolio withdrawals held in cash, T-bills or short-duration bonds means that when equities fall 35%, you spend the buffer instead of liquidating shares, and refill it when markets recover. It is not free — that money earns a low return in most years — but the cost is small relative to what it prevents.

The flexible spending rule may be the most powerful per unit of sacrifice. Cutting withdrawals by 10% during a decline, and restoring them on recovery, substantially raises plan survival. The catch is that it requires having discretionary spending to cut. A household whose entire withdrawal is essential has no flexibility, which is precisely the argument for a floor.

The rising equity glide path is counterintuitive and well documented: starting retirement at, say, 45% equity and rising toward 65% over fifteen years produces better worst-case outcomes than a constant 60%, because it minimizes equity exposure exactly when sequence risk is highest.

What Does Not Change

The core portfolio stays broad, low-cost and indexed throughout. Sequence risk is not an argument for active management — an active manager's underperformance compounds against you in exactly the same way, with fees added. It is not an argument for abandoning equities, since a bond-heavy retiree simply exchanges sequence risk for inflation risk over a thirty-year horizon. It is an argument for structuring withdrawals around a low-cost index core, which is a different thing entirely.

Execution Rules

1. Build the buffer before you retire, not after. In the final two years of work, accumulate two to three years of planned withdrawals in cash and short-duration Treasuries. Building it after a decline starts means selling equities at the bottom to fund it, which is the exact behavior it exists to prevent.

2. Write the spending rule down while markets are calm. Specify it in advance: if the portfolio falls more than 20% below the planned path, cut discretionary spending by 10% until it recovers; if it rises 20% above, take a 10% raise. Rules written during a decline are written by a frightened person.

3. Glide equity upward through the red zone, not downward. Consider entering retirement near 45–50% equity and drifting toward 60–65% over ten to fifteen years. This inverts the conventional target-date glide path, and it is specifically calibrated to the fact that sequence risk peaks at the start.

4. Never sell equities in a decline to fund spending — sell from the buffer, then rebalance on recovery. Establish the ordering in writing: cash buffer first, then bonds, then equities only when equities are at or above their planned weight. This single rule converts the buffer from a comfort blanket into a mechanism.

5. Delay retirement or work part-time through a bad start rather than raising the withdrawal rate. Two years of part-time income covering half your spending in years 1–3 of a bear market does more for plan survival than any allocation change available to you. Sequence risk is a cash-flow problem, and earned income is the most direct answer to a cash-flow problem.

Retirement Application

A retiree with $1,000,000 planning $45,000 of annual withdrawals should hold roughly $100,000–$135,000 outside the growth portfolio in cash and short Treasuries — two to three years of withdrawals. The remaining $865,000–$900,000 sits in a low-cost global index allocation.

Suppose the market falls 32% in year 2. The equity sleeve drops to roughly $590,000. The retiree does not sell a share. They fund year 2 and year 3 from the buffer entirely. If the decline persists into year 4, the flexible spending rule cuts discretionary spending by 10%, extending buffer coverage further. When markets recover, they refill the buffer from equity gains and resume normal withdrawals.

Compare that to the retiree with no buffer, who sells $45,000 of equities at the bottom in year 2 and again in year 3. They have permanently disposed of roughly 15% of their shares at depressed prices and will spend the rest of retirement short of that capital.

Neither retiree predicted the crash. Neither timed anything. Neither held different funds. The difference in outcome came entirely from having two or three years of spending sitting outside the volatile portfolio and a written rule about what to sell first.

This is also the clearest illustration of how the floor from Chapters 4 and 5 interacts with the core. A household whose essential spending is covered by Social Security and, where appropriate, a bounded pooled income contract, has vastly reduced forced-selling pressure. That household can hold more equity in its index core, not less, and can hold it through the decline. The floor exists so that the index core can be left alone — which is, in the end, the only way an index core delivers what it promises.

Risk Management

• The buffer must be genuinely safe, not merely conservative. T-bills, money market funds, short Treasuries. Intermediate bond funds fell double digits in 2022 alongside equities; a buffer that falls with the thing it is protecting against is not a buffer.
• Do not over-build it. Five or more years of cash is a permanent drag on a thirty-year portfolio and creates its own inflation shortfall. Two to three years is the range where protection is high and cost is low.
• The flexible rule needs real slack. If 100% of your withdrawal is essential, you cannot cut. Sizing the essential floor so that at least 20–25% of spending is genuinely discretionary is what makes the rule usable.
• A rising glide path is not a license to be aggressive later. The point is a lower start, not a higher finish. It also requires the discipline to increase equity after a decline, which is behaviorally the hardest version of rebalancing.
• Inflation is a sequence risk too. A burst of high inflation early in retirement raises withdrawals in nominal terms at the same time returns disappoint. Inflation-linked income — Social Security, TIPS, I-bonds — defends against a sequence a cash buffer cannot.

What This Chapter Cannot Do

This chapter cannot tell you whether the next decade is a bad sequence. Nobody knows, and any strategy that requires knowing is not a strategy. The defenses here work precisely because they do not require a forecast.

It also cannot fully protect a plan that is overspending. Sequence defenses buy time and reduce forced selling; they do not rescue a 6.5% withdrawal rate. If the underlying rate is unsustainable, no buffer, glide path or rule saves it — Chapter 3's arithmetic applies first.

And it does not argue against indexing anywhere. Every defense described here is a structure built around a broad, low-cost, globally diversified index core: how much sits beside it in cash, how withdrawals are ordered, how spending flexes, and how much essential income arrives independently of it. The core itself never needed replacing. It only ever needed to be left alone long enough to work.

---

Key Takeaway: Order matters, and it matters most in the first decade of withdrawals. Hold two to three years of spending in genuinely safe assets outside the portfolio, write your spending guardrail before the decline arrives, glide equity upward rather than downward through the red zone, and cover essentials with income that does not depend on markets — so that the low-cost index core can be left alone through the years that would otherwise have broken it.