I want to talk about a risk that I think is genuinely under-discussed, even among people who have spent years planning carefully for retirement. Most of us grow up internalising a simple mental model: earn a decent average return, withdraw a sensible percentage each year, and the math works out. I believed that model for longer than I should have. The problem is that it is wrong in one very specific, very consequential way — and the flaw only shows up at the worst possible moment, which is right after you stop working. The order in which your returns arrive turns out to matter enormously when you are drawing down a portfolio, not just accumulating one. I am still working through some of the finer calibrations of this with clients, but the core insight is solid enough that I want to write it out properly here.
the basic math problem, set out plainly ¶
Here is a toy example I use with almost every pre-retiree I work with. Imagine two retirees — call them Alex and Jordan — each starting with $1,000,000 and withdrawing $50,000 a year. Over 20 years, they both earn the exact same average annual return of 5%. The only difference is the order. Alex gets the bad years first: say, -20%, -15%, then a string of strong positive years. Jordan gets the strong years first, then the bad years at the end. Alex runs out of money somewhere around year 16. Jordan finishes year 20 with roughly $800,000 still in the portfolio. Same average return. Same withdrawal amount. Completely different outcomes. The reason is mechanical: when you withdraw $50,000 in a year your portfolio has already dropped 20%, you are selling more units to raise that cash. Those units are gone permanently. They cannot participate in the recovery. The bad early years do not just cost you the loss — they cost you the compounding on everything you were forced to sell at the bottom.
a real historical example: the 2000 to 2002 cohort ¶
This is not theoretical. Anyone who retired in January 2000 — right at the peak of the dot-com bubble — and held a standard 60/40 portfolio faced three consecutive years of negative real returns: roughly -9%, -12%, and -22% in nominal terms for the equity portion. The S&P 500 then had very strong years through the mid-2000s, but by then a retiree withdrawing at a 4% or 5% rate had already sold a significant slice of their portfolio at depressed prices. William Bengen's original 1994 research on sustainable withdrawal rates (published in the Journal of Financial Planning) identified this cohort problem implicitly — his 4% rule was partly designed to survive exactly these scenarios. But Bengen himself acknowledged the 4% figure was a floor for the worst historical sequences, not a comfortable average. I find a lot of clients have heard '4% rule' without absorbing that caveat. The 2000 retiree who held firm and spent conservatively probably survived. The one who spent at 5% or 6% likely did not, even though the 20-year average market return from 2000 to 2020 looked perfectly respectable on paper.
why accumulation is so different from decumulation ¶
During the years you are saving, sequence of returns is nearly irrelevant to your final outcome (a concept sometimes called 'sequence of returns symmetry in accumulation'). If you are putting $2,000 a month into an index fund and the market drops 30% in year three, that is actually fine — you are buying more units cheaply. The math is your friend. The moment you flip from depositing to withdrawing, the math reverses. Now a down market means you sell units cheaply to meet expenses. This asymmetry is the thing I try hardest to make visceral for clients in their late fifties. I sometimes ask them to imagine their portfolio as a lake and their spending as a pipe draining it. When the lake is rising (accumulation), a drought does not matter much — you are adding water. When you are only drawing from it (decumulation), a drought can empty it permanently before the rains return. The emotional challenge is that people who have been disciplined long-term investors rightly learned to ignore short-term volatility. That lesson, applied too rigidly in early retirement, becomes dangerous.
the mitigation strategies I actually use ¶
The good news is that sequence risk is manageable. I use a combination of approaches depending on each client's situation, and I want to be honest that there is no single clean solution — each has trade-offs. The first is a cash buffer or bucket strategy: keeping one to two years of living expenses in cash or very short-term bonds, so a bad market year does not force selling equities at a low. This is psychologically powerful as well as mathematically useful. The second is a flexible withdrawal policy. Rather than taking a fixed dollar amount each year, I work with clients to identify a floor (non-negotiable expenses covered by Social Security, pensions, or annuities) and discretionary spending that can compress in a bad market year. Guardrails-based withdrawal rules — developed by Jonathan Guyton and William Klinger — formalise this: if your portfolio drops below a certain threshold relative to your original balance, you reduce withdrawals by a set percentage. The third approach is partial annuitisation: using a portion of the portfolio to purchase an income annuity that covers the floor, removing that spending from market exposure entirely. I know annuities have a mixed reputation, and some of it is deserved (variable annuities with high fees are often not worth it), but a straightforward single-premium immediate annuity (SPIA) for base expenses is a genuinely useful tool for sequence risk specifically.
what bond allocation actually does here (and its limits) ¶
The classic advice is to hold more bonds as you age, and sequence risk is one of the reasons why. Bonds tend to lose less in equity bear markets (though 2022 was a painful exception — bonds and stocks fell together, which rattled a lot of retirees and rattled me too, frankly). A 60/40 portfolio drawdown is typically less severe than a 100% equity drawdown in a crash, which gives a retiree more time before they are forced to sell. But bonds are not a complete answer. Low bond yields can mean a 60/40 portfolio does not grow fast enough to sustain 30-year retirements. And as 2022 showed, correlation is not destiny — sometimes both asset classes fall. I have started paying more attention to research on rising equity glide paths in retirement (the work of Michael Kitces and Wade Pfau on this is worth reading), which is counterintuitive: starting with a lower equity allocation and gradually increasing it through retirement. The logic is that it limits damage in early bad years, then benefits from recovery in later years when your portfolio balance has stabilised. I do not apply this rigidly but it informs how I think about the first five to seven years post-retirement as the most critical window.
the psychological layer that the spreadsheets miss ¶
I have noticed that even clients who intellectually understand sequence risk sometimes panic and do the worst thing — selling equities at the bottom of a bear market in year two of retirement. The theoretical mitigation strategies only work if the person can stick to them. This is why I think the cash buffer is underrated: it is not just a financial tool, it is a permission slip. When a client knows they have 18 months of cash sitting separately from their investment portfolio, they can watch the portfolio drop 25% without feeling like they need to do anything. The behavioural finance literature on this (the Vanguard research team has published useful work on advisor alpha from behavioural coaching) suggests that preventing panic selling is one of the most valuable things an advisor can do. I have had clients call me during March 2020 genuinely frightened, and the single most useful thing I could say was: 'You have 22 months of cash. You do not need to touch the portfolio. Let's talk again in six weeks.' That is only possible if the structure was built before the crisis.
a practical starting point if you are within five years of retirement ¶
If I were writing a note to someone five years out from retirement (and I have written a lot of these), I would focus on three things. First, stress-test your withdrawal plan against 2000 to 2002 and 2008 to 2009 sequences specifically — not average returns. Most financial planning software can run historical Monte Carlo simulations; ask your advisor to show you the scenarios where you run out of money and what the early years looked like in those scenarios. Second, identify your floor: the minimum annual income that covers non-negotiable expenses. Then map how much of that is covered by sources completely uncorrelated to market returns (Social Security, a pension, annuity income). The larger that floor, the less exposed you are to sequence risk on the rest. Third, and this is the one people resist: build the cash buffer before you retire, not after the first bad year. It is much harder emotionally and mathematically to carve out a cash reserve after your portfolio has already dropped 20%. Build it during the last working years, when you still have income and the market may still be high. I think of it as buying insurance at the right time rather than after the accident.
Sequence-of-returns risk will not be in the news until a generation of recent retirees starts running into it, and by then it is too late to prepare. I find that the clients who take this most seriously are the ones who saw a parent or older sibling retire right before a crash and scramble. If you have not had that lived reference point, I hope this note helps. I am always happy to think through specific situations — drop me a message if something here raised a question I did not answer.