Sequence-of-returns risk: the retirement risk a single number hides

What is sequence-of-returns risk?

Sequence-of-returns risk (also called sequence risk) is the risk that the order in which good and bad years arrive — while you're adding money to or taking money out of a portfolio — changes your final outcome, even when the average return is identical.

Here's the counterintuitive part. If you hold a lump sum and never add or withdraw a cent, the order of returns does not matter at all. A gain of 20% and a loss of 20% multiply out to the same place regardless of which comes first — the math is commutative. As financial planner Michael Kitces puts it, "the sequence of returns doesn't matter when there are no cash flows in and out of a portfolio, even when there is extreme volatility."

The moment cash flows enter — contributions on the way up, or withdrawals in retirement — that symmetry breaks. Now when the bad years land matters enormously, because a market drop hits a different-sized balance depending on where it falls in the sequence. That's sequence risk, and it's why "my portfolio averaged 7%" can be a dangerously incomplete sentence.

Why average returns lie in retirement

In retirement, a bad decade at the start can sink a plan that a bad decade at the end would have survived — even with the same average return over the full period. When you're withdrawing a fixed amount each year, an early crash forces you to sell more of your shrunken portfolio to cover spending, leaving less invested to recover when the market rebounds.

Early Retirement Now's Karsten Jeske ("Big ERN") states the problem directly: "you can still run out of money even if your 30-year or even 50-year average return was above your withdrawal rate. If returns were bad enough initially and you keep withdrawing through the bear market, there is a chance you will deplete your investment portfolio so severely that even the subsequent bull market and recovery will not save you — the sequence of returns matters."

Kitces's research puts a number on how much the early years dominate: the correlation between a retiree's first-decade real equity returns and their sustainable withdrawal rate is about 0.79, far stronger than the correlation with one-year returns (about 0.21) or even full 30-year returns. In plain terms, the first ten years of retirement matter far more to your outcome than any single year — or even than the long-run average you probably used to build your plan.

Worked example: same average return, three different stories

Everything below uses the same three annual returns — +20%, +5%, and −20% — just in different orders. The arithmetic average is the same either way ((20 + 5 − 20) ÷ 3 ≈ 1.67% per year), and because it's the same set of numbers, the compound (geometric) average is identical too. Every figure is a simplified illustration in round dollars — no inflation-indexing, taxes, or fees — so you can check it with a calculator. These are estimates to build intuition, not a forecast.

Why a single FI number — and a single average return — can't see this

The standard FI number, and the calculators built on it, assume one smooth return every year forever. That assumption is useful for a first estimate and blind to sequence risk by construction. Your FI number is annual spending ÷ a safe withdrawal rate; project it forward with a constant return and you get one clean FI date. Real markets never deliver the average on schedule.

We're upfront that this applies to our own free tool. The FIManager FI calculator uses a closed-form growth formula with a single real (after-inflation) return — 5% by default, editable, and shown on screen. That makes it fast, transparent, and perfect for answering "roughly how far am I?" It cannot model the order of returns, because it only has one return. That's not a bug to hide; it's the exact boundary where a point estimate stops being enough and a simulation begins.

How planners handle sequence risk: run the sequence, don't average it

The honest answer to sequence risk isn't a better single number — it's running your plan across many possible return sequences and reading the share that succeed. That's what a Monte Carlo simulation does, and it's why the output is a chance of success rather than one deterministic date.

FIManager's full plan does this in two complementary ways, both with visible assumptions:

• Randomized (parametric) sequences — draw each year's stock and bond returns from a distribution you can see (expected return, volatility, correlation, inflation), thousands of times, and count how often the plan survives.
• Real historical sequences — replay actual market history, including the option to keep runs of good and bad years in their real order (block sampling), so the 1970s or 2000s show up as the connected bad decades they actually were, not as scrambled averages.

Run enough sequences and you stop asking "what's my FI date at 7%?" and start asking "across a wide range of futures — including bad-first ones — what fraction leave me with money?" That's a claim you can act on. And because every assumption is on screen and editable, you can see why the number moved — the opposite of a black box that swings from confident to catastrophic when you nudge one input, with no explanation.

How to reduce sequence-of-returns risk

Sequence risk can't be eliminated, but it can be managed. Common, well-documented approaches — none of these are personalized advice, just the standard toolkit to discuss with your own situation in mind:

1. Plan at a more conservative withdrawal rate. The 4% rule is built on ~30-year US history; for retirements of 40–50+ years, historical simulations point to roughly 3.25–3.5%, precisely because early bad sequences failed the higher rate. A lower rate is a smaller target to sell into during a crash.
2. Hold a cash or bond buffer for early-retirement spending so you're not forced to sell equities into a downturn in the highest-risk years. (Kitces's own research is a caution here: a long bad decade "outlasts most cash reserve strategies," so a buffer is a shock absorber, not a force field.)
3. Stay flexible on spending. Trimming withdrawals during down years — rather than mechanically taking the same inflation-adjusted amount — is one of the most effective ways to blunt an early bad sequence.
4. Keep working, or coast, a little longer. Even part-time income in the first retirement years reduces withdrawals exactly when sequence risk is highest. (See Coast FIRE.)
5. Stress-test the plan, don't just point-estimate it. Before you rely on a date, run it across many sequences — including bad-first ones — and look at the chance of success and the failure cases, not one average path.

FAQ

What is sequence-of-returns risk?

It's the risk that the order of your returns — not just their average — determines your outcome once you're adding money to or withdrawing from a portfolio. Bad returns early in retirement, while you're drawing income, are far more damaging than the same bad returns later, even if the long-run average is identical.

Why do average returns "lie" in retirement?

Because the average treats every year as interchangeable, but withdrawals don't. An early crash forces you to sell more of a shrunken portfolio to fund spending, leaving less to recover. Two retirees with the same average return can end up worlds apart depending on when the bad years hit.

Does the order of returns matter if I'm not withdrawing anything?

No. For a lump sum with no contributions or withdrawals, the order is irrelevant — the returns multiply to the same result regardless of sequence. Sequence risk only appears once there are cash flows in or out.

Does sequence risk matter while I'm still saving?

Yes, but in reverse. For an accumulator still contributing, an early crash can actually help (your contributions buy in cheap), while a crash right before your FI date — when your balance is largest — does the most harm.

When is sequence risk highest?

In the years immediately around retirement — roughly the last stretch of accumulation and the first decade of withdrawals. Research finds the first ten years of retirement returns are by far the strongest predictor of whether a withdrawal rate proves sustainable.

How do you protect against sequence-of-returns risk?

Common approaches: plan at a more conservative withdrawal rate (many early retirees use 3.25–3.5% instead of 4%), hold a cash/bond buffer for early spending, stay flexible on withdrawals in down years, earn some income early in retirement, and stress-test with a chance-of-success simulation rather than a single average return.

Can a Monte Carlo simulation eliminate sequence risk?

No — nothing eliminates it. A simulation measures it: by running your plan across many randomized and real historical sequences (including bad-first ones), it reports the share that succeed, so you can see the risk instead of averaging it away.