Sequence of Returns Risk

The same returns, reordered

Take a 30-year run of annual returns — a realistic-looking mix with an arithmetic mean of 6.63% and a compound growth rate of 6.13%. Nothing about that series changes in what follows. The only thing that changes is the order.

−15  −12   −8   22   18   14    5   −6   12    9
  16   −3   11    7   20   −9   13    6   15    4
  10   −5   19    8   12    3   17   −7   14    9

Arithmetic mean 6.63%   ·   Compound growth 6.13%

Now run a retirement through it. A €1,000,000 portfolio, withdrawing €40,000 in year one — the textbook 4% — and increasing that withdrawal by 3% each year for inflation.

One retiree finishes with nothing. The other finishes with almost twice what they started with. They earned the same returns, took the same withdrawals, and made no different decisions. The only difference is which decade the bad years landed in.

Why the order matters at all

The mechanism is straightforward once stated. A withdrawal during a downturn is funded by selling units at depressed prices. Those units are gone. When the market recovers, it recovers on a smaller base — the shares that would have participated in the rebound were spent at the bottom.

Early losses therefore compound in a way late losses cannot. A 30% fall in year one hits a full portfolio and permanently reduces everything that follows. The same fall in year 28 hits a portfolio that has already delivered most of the growth the retirement needed.

It only exists when money is moving

Run the same two sequences with no withdrawals at all and something clarifying happens:

With no cash flows the two results are identical to the euro, because multiplication does not care about order. Sequence risk is not a property of markets — it is a property of markets combined with money moving in or out.

Note the third row, which surprises people. For someone still contributing, bad years early are actively helpful — every contribution during the downturn buys more units, and those units are still held when the recovery arrives. The accumulator and the retiree face the same market with opposite exposures.

When the exposure is highest

The risk is not spread evenly across a retirement. It concentrates in a window either side of the day the withdrawals begin — commonly described as the five to ten years before and after retirement, when the portfolio is at its largest and the remaining time to recover is shortest.

• Before that window, the portfolio is small and future contributions dominate the outcome.
• Inside it, the portfolio is at maximum size and has just started to be drawn down. This is where the damage happens.
• After it, a portfolio that survived the early years has usually grown enough that later declines are absorbable.

How the risk is commonly addressed

The approaches below appear repeatedly in the literature. They are described here, not recommended — each carries its own costs, and which if any suits a given retirement is not a question that can be answered generically.

• Cash or short-bond buffers. Holding one to three years of expenses outside equities allows withdrawals to come from the buffer during a decline rather than from falling assets. The cost is the return given up on that cash in every year the buffer is not needed.
• Flexible spending rules. Guardrail systems reduce withdrawals after poor years and raise them after good ones, which directly interrupts the mechanism. The cost is a variable income.
• A lower initial withdrawal rate. Starting below 4% widens the margin, at the cost of either working longer or spending less.
• Rising equity glidepaths. Holding more bonds at the start of retirement and increasing equity exposure over time — the opposite of conventional advice — targets the risk precisely at the window where it concentrates.
• Continuing to earn. Any income that reduces or eliminates withdrawals during a downturn removes the forced selling entirely, which is why part-time work in early retirement is discussed so often.