Same Average Return, Two Very Different Retirements
Two retirees can earn the exact same average return over a decade and end up hundreds of thousands of dollars apart. The difference isn't luck or fees — it's the order the returns arrived in, once they started withdrawing. That's sequence-of-returns risk, and it lives almost entirely in your first few retirement years.
- The answer: while you're saving, reordering the same returns gives the same ending balance. Once you're withdrawing, a bad first decade forces you to sell more shares at low prices, leaving fewer to recover — so the identical returns in reverse order can leave you rich instead of broke.
- The danger zone: roughly the first five years of retirement. The portfolio is at its largest, you're selling into weakness, and there's the least time to bounce back.
- The defense: hold 1–3 years of spending in cash (a buffer or "bond tent"), cut discretionary spending in down years, don't sell equities in a crash, and start with a sensible — not maximal — withdrawal rate.
Where the AI summary above gets this wrong
"Stay invested and don't worry about timing — over the long run it's the average return that matters."
That's true for savers. Here's what it misses for retirees:
- Averages rule accumulation, not decumulation — when you add money and withdraw nothing, reordering returns gives the same result. Once you're pulling income out, the order dominates: the same average can end decades apart.
- A bad first decade can be unrecoverable — selling shares in a down market to fund spending locks in losses and leaves fewer shares to ride the rebound. "It averages out" assumes you still own the shares when it does.
- That's why the defenses exist — a cash buffer, flexible spending, and the "first five years" warning only make sense once you accept that order, not just average, decides the outcome.
Jordan turned 51 this spring and is roughly a decade from the retirement date he's penciled in. He isn't asking "what's the average return I should assume?" — he's asking the harder question: "what if a bad market lands in my first retirement years, while I'm taking income out?" So he ran the same stress test the engine runs: take one set of returns, hand it to a retiree in two different orders, and watch what happens. Here's what he found, and what he changed because of it.
01 Why order beats average in retirement
The order of returns barely matters while you're saving, and matters enormously once you're spending. The reason is mechanical, not mysterious: when you only add money and never withdraw, your ending balance is the same set of multiplications no matter how you shuffle them — multiplication is commutative, so a good-early decade and a bad-early decade with the same returns finish at the exact same number.
Withdrawals break that symmetry. Each year you sell shares to fund your spending. In a down year you have to sell more shares to raise the same dollars, permanently shrinking the share count that's left to participate in the eventual recovery. Do that in your first few years and the portfolio can fall into a hole it never climbs out of — even if the long-run average return is perfectly healthy.
Average return is the number people plan with, and for a retiree it is the wrong statistic. Two portfolios with identical thirty-year averages produce completely different outcomes depending on when the bad years land, because withdrawals convert a temporary decline into a permanent loss of shares.
Source: Cooley, Hubbard & Walz (1998), AAII Journal — the "Trinity study"
02 The same returns, two outcomes
Take one fixed set of annual returns and the difference is entirely in the order. Suppose a retiree starts with $1,000,000 and the market delivers, over a stretch, these numbers in some order: a couple of sharp down years mixed with several solid up years. Averaged together they're a respectable return either way. But run them bad-early versus good-early while withdrawing a fixed amount each year and the two paths diverge fast.
The bad-early retiree withdraws into a falling portfolio in the early years, selling shares cheaply just as the balance is smallest relative to the withdrawal. The good-early retiree banks gains first, so the same withdrawals are a smaller bite and there's a bigger cushion when the down years finally arrive. Same returns, same withdrawals, same average — opposite endings.
Reverse a return sequence and the average is identical to the decimal. For a saver, the ending balance is identical too. For someone withdrawing, it can differ by hundreds of thousands of dollars. The withdrawal is the whole reason the two diverge.
03 Worked example: bad-early vs good-early
Here's the test with real numbers. Both columns use the same eight annual returns — {−15%, −10%, +20%, +25%, +8%, +12%, +6%, +14%}, an 7.5% arithmetic average — and the same fixed annual withdrawal. The only difference is the order: the left column hits the two down years first; the right column saves them for last. Change the starting portfolio and the withdrawal below and watch the gap.
Shows: how the order of one identical set of returns changes the ending balance once you withdraw a fixed amount each year, over eight years. Ignores: taxes, inflation, Social Security, real market paths, dynamic spending, and fees — this isolates the sequence effect alone.
At the defaults — $1,000,000 starting, $50,000 a year withdrawn — the good-early ordering ends well ahead of the bad-early one, despite identical returns and an identical 7.5% average. Set the withdrawal to $0 and the gap collapses to zero: that's the proof that withdrawals, not the returns themselves, create the risk.
On the defaults above, the worked example shows: Same returns, same average, same withdrawal — the good-early order ends $168,863 ahead purely because of order.
04 The danger zone: the first five years
The risk concentrates in roughly the first five years of retirement, and the reason is arithmetic. Those years combine three bad ingredients at once: the portfolio is at or near its largest, so a percentage loss is the most dollars you'll ever lose; you're selling shares to fund spending, so a down market forces you to liquidate more of them; and there's the least remaining time for a recovery to undo the damage.
A retiree who sails through the first five years without a deep drawdown has largely defused sequence risk — later losses land on a portfolio that has already grown a buffer and faces fewer remaining withdrawals. The same loss in year one and year fifteen are not the same event. That asymmetry is why "just stay the course" is sound advice for a 35-year-old and incomplete advice for a 65-year-old in month three of retirement.
De-risking entirely at retirement is commonly presented as the safe response to sequence risk, and that is not what it achieves. Moving to cash removes the risk of selling into a fall and replaces it with a certainty of losing purchasing power across a thirty-year retirement — trading a possible loss for a guaranteed one. The defence is a buffer alongside growth assets, not the removal of growth assets.
Source: Cooley, Hubbard & Walz (1998), AAII Journal — the "Trinity study"
05 Buffers, bond tents, and flexible spending
The defences all do one thing: reduce how many shares you are forced to sell in a down market. None of them requires predicting anything — they change what you do when a bad year arrives, not whether you see it coming.
A cash buffer of one to three years of spending means a downturn is funded from cash rather than by selling equities at the bottom. It is the simplest and the easiest to execute under stress, which is most of its value.
A bond tent temporarily raises the bond allocation in the years around retirement — the window where sequence risk peaks — then lets it glide back down as the risk recedes. It costs some long-run return and covers exactly the period that matters.
Flexible spending is the cheapest of the three because it costs nothing to set up. Skipping an inflation increase after a poor year, or cutting discretionary spending by a set percentage, reduces withdrawals precisely when withdrawals do the most damage.
Guaranteed income does it structurally: every dollar of essential spending covered by Social Security or an annuity is a dollar you never have to raise by selling anything.
In practice these stack. A cash buffer plus flexible spending is the combination most planners reach for, because together they make it nearly certain you never sell a falling stock to buy groceries in your first retirement years — which is the entire failure mode being defended against.
| Defence | How it cushions sequence risk | The trade-off |
|---|---|---|
| Cash buffer (1–3 years of spending) | Draw from cash in a down year instead of selling shares low; refill in good years | Cash drags on returns when markets rise |
| Bond tent | Raises bonds temporarily near the retirement date, then lowers them once the danger zone has passed | Lower expected growth through the tent years |
| Flexible / guardrail spending | Cuts discretionary withdrawals in bad years and restores them in good ones, keeping shares invested | Income varies from year to year |
| Not selling equities in a drawdown | Funds spending from cash or bonds so the recovery happens on the full share count | Requires the buffer to be in place beforehand |
| Lower initial withdrawal rate | Smaller withdrawals shrink the forced-selling problem every year | Risk of underspending if returns turn out well |
Source: Cooley, Hubbard & Walz (1998), AAII Journal — the "Trinity study"
06 What it means for your withdrawal rate
Sequence risk is the reason a "safe" withdrawal rate exists at all.
Bengen's 1994 study and the Trinity study that followed did not choose 4% because it is what a portfolio earns on average — average returns would support considerably more. They chose it because it is roughly what survived the worst historical sequences, including retirements that began immediately before a bad decade.
That is a useful thing to understand, because it reframes the number. The 4% rule is not a prediction; it is a stress test result. Most retirees who followed it historically died with more money than they started with, and the rule was calibrated for the minority who did not.
It also reframes the decision. You do not have to find one magic number and live with it for thirty years. You can start at a sensible rate and then lean on a cash buffer and flexible spending, tightening only if a bad sequence actually arrives.
The rate you can support depends heavily on that flexibility. A retiree who refuses to reduce spending in a downturn needs a materially lower starting rate than one willing to skip an inflation increase after a poor year — the difference is worth roughly half a percentage point, which on a $1 million portfolio is $5,000 a year of income for the whole retirement.
The full treatment of the rate itself is in The 4% Rule Is a Starting Guardrail, Not a Law.
Source: Cooley, Hubbard & Walz (1998), AAII Journal — the "Trinity study"
07 The same crash, at four different times
Sequence risk is clearest when everything is held constant except when the bad year lands.
| A 30% fall in… | What you are doing | Damage |
|---|---|---|
| Year 10 of saving | Contributing, not withdrawing | Net positive over time — you buy more shares cheaply |
| The year before you retire | Largest balance, not yet withdrawing | Serious, but delaying a year or two is still available |
| Year 1 of retirement | Largest balance and withdrawing from it | The worst case — shares sold at the bottom never recover |
| Year 15 of retirement | Smaller balance, fewer years to fund | Uncomfortable, rarely decisive |
Only the third row is sequence risk proper, and it occupies a window of roughly five years. Every defence in this post exists to get through that window without being a forced seller.
The thing that surprised me when I ran my own numbers is how lopsided the timeline is: the first five years of retirement carry risk all out of proportion to their length. So I stopped trying to pick a magic withdrawal rate. Instead I plan to hold two to three years of spending in cash entering retirement, and I've already decided I'll cut discretionary spending in any down year rather than sell equities into a slump. Those two habits do more to protect a 30-year retirement than shaving a tenth of a percent off a withdrawal-rate estimate ever could. Order is the risk; behavior is the defense.
FAQ
What is sequence-of-returns risk?
Sequence-of-returns risk is the danger that the ORDER of investment returns — not just their average — wrecks a portfolio you are withdrawing from. Bad returns early in retirement force you to sell more shares at low prices, leaving fewer to recover, so the same set of returns in a different order can end decades apart.
Why doesn't sequence risk matter while you're still saving?
During accumulation you add money and take nothing out, so reordering the same returns produces the same ending balance — multiplication is commutative. A bad first decade while saving is even helpful, because you buy more shares cheaply. Withdrawals are what break the symmetry.
How long is the sequence-risk danger zone?
Roughly the first five years of retirement, give or take. That window has outsized influence because the portfolio is at its largest, you are selling shares to fund spending, and there is the least time to recover. Survive it without a big drawdown and the risk falls sharply.
What is a bond tent or cash buffer?
Both reduce how many shares you must sell in a down market. A cash buffer holds one to three years of spending in cash or short bonds so you draw from it instead of selling stocks low. A bond tent temporarily raises your bond allocation around the retirement date, then lowers it again once the danger zone passes.
Does cutting spending in down years actually help?
Yes. Flexible or "guardrail" spending — trimming discretionary withdrawals in down years and restoring them in good ones — leaves more shares invested through the recovery. Research on dynamic withdrawals consistently shows it improves portfolio survival versus a rigid fixed-dollar withdrawal.
Should I just pick a lower withdrawal rate to be safe?
A lower starting withdrawal rate does cushion sequence risk, but it can also mean underspending for decades if returns are fine. Most planners pair a sensible starting rate with a buffer and flexibility, so you only tighten when a bad sequence actually shows up rather than permanently spending less for a risk that may never arrive.
Sources
Regulator references
- Social Security Administration ·Period life table for the planning horizon.Last verified: 2026-09-07
- Internal Revenue Service ·RMDs force withdrawals regardless of markets.Last verified: 2026-09-07
- SEC Investor.gov ·Regulator guidance on saving and investing basics.Last verified: 2026-09-07
- Bureau of Labor Statistics ·CPI series used to deflate withdrawals.Last verified: 2026-09-07
Research
- Cooley, P. L., Hubbard, C. M. & Walz, D. T. — "Retirement Savings: Choosing a Withdrawal Rate That Is Sustainable" (the "Trinity study") · AAII Journal · 1998 · ScholarlyArticlesustainable withdrawal rates across historical return sequencesLast verified: 2026-06-21
- Pfau, Wade D. — "Reducing Retirement Risk with a Rising Equity Glide Path" · ScholarlyArticle · 2014how the order of early returns drives sustainable withdrawal outcomesLast verified: 2026-06-21
- Bengen, William P. — "Determining Withdrawal Rates Using Historical Data" · Journal of Financial Planning · 1994the original analysis behind the 4% rule and worst-case sequencesLast verified: 2026-06-21
Calculator unit tests · the assertions this page's worked example is checked against, and their last result
Changelog
- 2026-06-21 — initial publish (new format)
See how this decision plays out across your 30-year projection
Stress-test your plan against bad-early sequences, buffers, and flexible spending — month by month to age 90.
Join the Waitlist
Jordan Reeves
Founder of Talk Through Wealth. A software engineer for over a decade before turning to retirement planning, Jordan built the projection engine after watching family members get fragmented, country-by-country advice that never reconciled. He writes about retirement the way the engine computes it: month-by-month, lifetime-long, and skeptical of any rule of thumb that hasn't been run through the math.