Sequence of Returns Risk: Why Order Matters

Two people retire on the same day with the same illustrative one million dollars, the same portfolio, and the same withdrawal plan. Over the next ten years they receive the identical set of annual returns: the same ten numbers, averaging the same 6.00 percent, adding up to the same total. One of them ends the decade with roughly 914,000 dollars. The other ends it with roughly 1,242,000 dollars. Nothing separates them except the order in which those returns arrived.

That difference is sequence of returns risk, and this analysis is built around demonstrating it rather than describing it. Below, one invented ten year series is run twice, forwards and backwards, through two situations: a saver who touches nothing, and a retiree who withdraws every year. The saver finishes in exactly the same place both times. The retiree does not, and the reason is arithmetic you can check with a calculator. From there the analysis works through what actually reduces the exposure, what each of those defences costs, and why the 4 percent rule analysis on this site starts from a number so far below any historical average. Put your own balance and spending into the retirement number calculator before you read the sequences.

What sequence of returns risk actually means

Sequence of returns risk is the risk that the ORDER of your investment returns, independent of their average, determines your outcome. It is not the risk that markets fall. It is not volatility, though volatility is the raw material it works with. It is the interaction between market movements and the cash you are moving in or out of the portfolio while those movements happen.

The clean way to state it: when no money moves, a portfolio’s ending value depends only on the set of returns, not their arrangement. When money moves, the ending value depends on both. Every dollar you withdraw is withdrawn at a particular moment, from a particular balance, and the size of that balance relative to the withdrawal is what decides how much of the portfolio the withdrawal represents.

That is why sequence risk is usually discussed as a retirement problem. Retirees are the people running a stream of withdrawals against an uncertain market for decades. But it applies to anyone with cash flows: a business drawing on a reserve, a family funding tuition from investments, an endowment with a spending policy. It also applies in reverse to savers making regular contributions, where the effect runs the other way and a weak early market is helpful rather than harmful.

The rest of this analysis stops describing and starts computing, because the honest version of this topic is a demonstration rather than a warning. Every number below comes from one invented series of returns, stated openly and used consistently.

The setup for the demonstration

Here is the entire dataset. Ten annual returns, invented for this analysis and not drawn from any real market or index: minus 18 percent, minus 12 percent, plus 6 percent, minus 7 percent, plus 21 percent, plus 9 percent, plus 16 percent, plus 4 percent, plus 19 percent, plus 22 percent.

Those ten numbers add to 60, so their arithmetic average is exactly 6.00 percent a year. Multiply the ten growth factors together (0.82, 0.88, 1.06, 0.93, 1.21, 1.09, 1.16, 1.04, 1.19, 1.22) and you get about 1.6432, a total return of about 64.32 percent over the decade. The compound annual growth rate behind that is about 5.09 percent, comfortably below the 6.00 percent arithmetic average, for the reasons set out in our compound growth rate breakdown.

Call the order above Sequence A, the bad years first case: three of the first four years are negative and the portfolio starts in a hole. Sequence B is the identical list read backwards, so it opens plus 22, plus 19, plus 4, plus 16, plus 9, plus 21, then minus 7, plus 6, minus 12, minus 18. Same ten numbers. Same average. Same compound growth rate. Same total return.

Treat these returns as real, meaning already adjusted for inflation, so the spending figures below can stay constant without an inflation column. The starting portfolio is an illustrative 1,000,000 dollars in both cases.

Why order is irrelevant while you are still accumulating

Start with the case that surprises people, because it is the control group for everything that follows. A saver holds the illustrative 1,000,000 dollars, adds nothing, withdraws nothing, and simply lets the decade happen.

Under Sequence A, the balance multiplies by 0.82, then 0.88, then 1.06, and so on down the list, and it finishes at about 1,643,219 dollars. Under Sequence B the multipliers arrive in the opposite order and the balance finishes at about 1,643,219 dollars. Not approximately the same. Identical, to the cent, because multiplication is commutative: rearranging factors cannot change a product.

This is worth sitting with, because it means every scary story about sequence risk is irrelevant to a person who is not touching the money. A buy and hold investor with a thirty year horizon and no withdrawals is exposed to what the returns ARE, not to when they show up. The worst decade of their life could be the first or the last and their ending balance would be unchanged.

There is one qualification. A saver making regular contributions is not order neutral either, but the effect runs in their favour when markets are weak early: contributions during a decline buy more units, which is the mechanism behind our dollar cost averaging analysis. A saver’s sequence risk is the risk of strong markets early and a crash right at the end, when the balance is largest, which is the same red zone problem approached from the other side.

Why order decides everything once you are withdrawing

Now give both portfolios a job. Each retiree takes 40,000 dollars at the start of every year, before that year’s return is applied, and takes it in real terms so the amount stays constant. On an illustrative 1,000,000 dollar portfolio that is a 4.0 percent initial withdrawal rate, which is the figure most retirement planning starts from.

Sequence A, with the bad years first, finishes the decade at about 914,226 dollars.

Sequence B, with the identical returns reversed, finishes at about 1,242,136 dollars.

The gap is about 327,910 dollars, or roughly 36 percent more money for the second retiree. Both withdrew exactly 400,000 dollars across the ten years. Both received the same ten returns. Both would report the same 5.09 percent compound growth rate if you asked them what the market did. One of them is in a materially weaker position than the other for the twenty or thirty years still to come, and no decision either of them made explains it.

The mechanism is simple once you see it. Subtracting a fixed amount breaks the chain of multiplications into segments, and the size of the balance at each subtraction determines how big a bite that fixed amount takes. Multiply first and subtract later, and the subtraction is small relative to the pot. Subtract first and multiply later, and the multiplication is applied to a pot that has already been reduced. Reordering the returns reorders which of those two situations you are in, year by year.

The bad years first decade year by year

Watching Sequence A unfold makes the damage concrete. Every line below is start of year balance, minus the 40,000 dollar withdrawal, multiplied by that year’s return factor.

Year 1 opens at 1,000,000, drops to 960,000 after the withdrawal, falls 18 percent and closes at 787,200. Year 2 opens at 787,200, drops to 747,200, falls 12 percent and closes at 657,536. Year 3 opens at 657,536, drops to 617,536, gains 6 percent and closes at 654,588. Year 4 opens at 654,588, drops to 614,588, falls 7 percent and closes at 571,567.

Four years in, an illustrative million dollars is a little over 571,000 dollars. The retiree has spent 160,000 dollars, so the market took roughly 268,000 dollars of it, and the portfolio has not had a single good year to work with. Year 5 finally delivers plus 21 percent, but it applies to 531,567 dollars after that year’s withdrawal, so it adds about 111,600 dollars rather than the 210,000 dollars the same percentage would have added to the original balance.

The decade then runs plus 9, plus 16, plus 4, plus 19 and plus 22, five consecutive positive years, and the portfolio still only claws back to about 914,226 dollars. Even a strong five year finish cannot undo the arithmetic of the opening, because the recovery is being applied to a base that the opening permanently shrank. That asymmetry is the whole of sequence risk in one paragraph.

The good years first decade year by year

Sequence B tells a completely different story with the same numbers. Year 1 opens at 1,000,000, drops to 960,000 after the withdrawal, gains 22 percent and closes at 1,171,200. Year 2 opens at 1,171,200, drops to 1,131,200, gains 19 percent and closes at 1,346,128. Year 3 closes at about 1,358,373. Year 4 gains 16 percent and closes near 1,529,313. Year 5 closes near 1,623,351.

At the halfway point this retiree has withdrawn 200,000 dollars and is up more than 62 percent on the starting balance. The Sequence A retiree, at the same point, is at about 643,196 dollars, down about 35.7 percent. Two portfolios that started identically and received identical returns are now separated by a factor of about 2.52, entirely on ordering.

Sequence B then hits its bad stretch: minus 7, then plus 6, then minus 12, then minus 18. Those are the same four unpleasant years that wrecked Sequence A, and they do real damage here too, taking the balance from about 1,915,855 dollars at the start of year 7 down to about 1,242,136 dollars at the end of year 10. That is a decline of roughly 35 percent across the final four years.

And yet this retiree ends the decade 24 percent richer than they started, with a portfolio that can support far more spending than the other one. The bad years happened. They simply happened to a portfolio that had already grown large enough to absorb them, and to withdrawals that were a trivial fraction of the balance while it did.

What the identical withdrawals actually cost

Here is the cleanest way to state the whole phenomenon, and it reconciles exactly. The no withdrawal ending value, in either order, is about 1,643,219 dollars. That is the benchmark: what the money would have become untouched.

Sequence A ended at about 914,226 dollars, so the withdrawals cost it about 728,993 dollars of ending value. Sequence B ended at about 1,242,136 dollars, so the withdrawals cost it about 401,083 dollars of ending value. Both retirees withdrew 400,000 dollars in face value. One of them paid 1.82 times face for it. The other paid almost exactly face.

The difference between those two costs, 728,993 minus 401,083, is about 327,910 dollars, which is precisely the gap between the two ending balances. Every dollar of the divergence is accounted for. Nothing is hand waved.

The reason a withdrawal can cost more than its face value is that the withdrawn dollars would otherwise have compounded. Take 40,000 dollars out at the start of year 2 of Sequence A and you forgo everything the remaining nine years would have done to it, which on that sequence is a multiple of about 2.00, or roughly 80,000 dollars of ending value from a 40,000 dollar withdrawal. Take the same 40,000 dollars out of Sequence B at a point where the remaining years multiply by only about 0.72, and it costs about 29,000 dollars of ending value.

That is the honest core of sequence risk. It is not that bad years are worse than good years. It is that withdrawals made from a depleted portfolio remove capital that the recovery would have worked hardest on.

Why selling in a down year is permanent

Percentages hide the mechanism, so switch to units. Suppose the portfolio holds fund units priced at an illustrative 100 dollars each, so 1,000,000 dollars is 10,000 units. A 40,000 dollar withdrawal at that price sells 400 units.

Now suppose the same 40,000 dollars has to be raised after prices have fallen 18 percent, so units are 82 dollars. Raising 40,000 dollars now takes about 488 units rather than 400. Put another way, the units sold were worth about 48,780 dollars before the decline, so the retiree surrendered about 8,780 dollars of pre decline value to fund a 40,000 dollar year.

The permanence is the important part. Those extra 88 units are not temporarily down. They are gone. When prices recover to 100 dollars, they are not in the account to recover. The 8,780 dollars is not an unrealized loss waiting to reverse; it is a realized reduction in the number of units the future recovery can be applied to.

This is exactly why a buy and hold saver is immune. The saver who sells nothing during the decline still owns all 10,000 units when the price comes back, so the decline was a paper event. The retiree who had to sell owns fewer units forever after. Two people, one market, and only one of them converted a temporary price move into a permanent quantity loss.

Repeat that across a run of bad years, as Sequence A does four times in its first five years, and the unit count falls fast enough that the eventual recovery has much less to work with. That is the arithmetic underneath every mitigation strategy discussed later in this analysis.

The effective withdrawal rate is not the rate you chose

Retirees choose a withdrawal rate once, at the start, and then never see it again. What they actually experience each year is the withdrawal divided by the current balance, and that number moves with the market whether they notice or not.

Under Sequence A, the 40,000 dollar withdrawal starts as 4.00 percent of the portfolio, then becomes 5.08 percent in year 2, 6.08 percent in year 3, 6.11 percent in year 4 and 7.00 percent in year 5. At its peak, this retiree is drawing 7.00 percent of their remaining capital while nominally following a 4 percent plan. They did not raise their spending. The denominator collapsed underneath it.

Under Sequence B the same 40,000 dollars falls to 3.42 percent in year 2, 2.97 percent in year 3, and reaches a low of about 2.09 percent in year 7. That retiree is nominally on the same 4 percent plan and is in practice drawing about half the rate.

This is why a fixed real withdrawal, the classic form of the 4 percent rule, is quietly the most sequence sensitive rule you can adopt. It refuses to respond to the denominator. Anything that lets spending flex with the balance, even slightly, converts part of the risk from a permanent capital problem into a temporary lifestyle problem, and the levers section below puts numbers on that trade.

Check what your own withdrawal represents as a percentage of your current balance in the retirement number calculator, and then ask what it would represent after a 20 percent decline.

Why a retiree drawdown runs deeper than the market decline

One more comparison sharpens the point. Take the first five years of Sequence A and apply them to a portfolio with no withdrawals at all. An illustrative 1,000,000 dollars becomes about 860,737 dollars, a decline of about 13.9 percent. That is what the market did over those five years.

The retiree taking 40,000 dollars a year through the identical five years is at about 643,196 dollars, a decline of about 35.7 percent. Their drawdown is roughly two and a half times deeper than the market’s, and it is not because they own anything different.

Part of that is simply the 200,000 dollars they spent. But the arithmetic does not work out to 860,737 minus 200,000 either. It works out to about 643,196, which is roughly 17,500 dollars lower still, because each withdrawn amount also forfeited whatever the remaining years would have done to it. The spending explains most of the gap and the lost compounding explains the rest.

The practical consequence sits in the recovery requirement. Getting from 643,196 dollars back to the original 1,000,000 dollars needs a gain of about 55.5 percent, and that is before funding any further spending. Getting the untouched portfolio from 860,737 back to 1,000,000 needs about 16.2 percent. Same market, same five years, and two entirely different holes to climb out of.

This is the reason retirement planning cannot be done with an average return. An average tells you where the market went. It tells you nothing about the depth of the hole a spending plan dug while it went there.

What each portfolio can safely support afterwards

Ending balances are abstractions. Translate them into spending and the divergence becomes a lifestyle difference. Apply an illustrative 4 percent rate to each ten year ending balance and Sequence A supports about 36,569 dollars a year while Sequence B supports about 49,685 dollars a year.

Look at it from the other direction and it is starker. If both retirees simply continue withdrawing their original 40,000 dollars, that is about 4.38 percent of Sequence A’s remaining capital and about 3.22 percent of Sequence B’s. One of them is now spending above the rate their plan was built on, with a full retirement still ahead. The other has a large margin.

Nothing about the Sequence A retiree’s behaviour was wrong. They chose a conventional rate, withdrew exactly what they planned, and never panicked. The market handed them the same returns as the other retiree and simply delivered them in an unlucky order. That is the uncomfortable truth this topic keeps arriving at: a large part of a retirement outcome is not a decision.

What IS a decision is how much exposure you leave to that ordering. Everything from here on is about that: what reduces the exposure, by how much, and at what cost. Run your own target balance and spending through the retirement number calculator to see which of the levers below would move your number most.

The retirement red zone

If order matters, then timing matters, and the timing that matters most is concentrated. The portfolio is at its largest right around the day you stop working, so a given percentage decline costs the most dollars there. Withdrawals are beginning at the same moment, so declines start being locked in rather than ridden out. Those two facts overlap for a window commonly described as roughly the five years before and the five years after retirement.

The label is a rule of thumb, not a measured constant. Nothing changes at the boundary, the window is longer for someone retiring early and shorter for someone with a large pension covering most of their spending, and a portfolio still faces sequence exposure long after year ten. What the label captures correctly is that risk is not spread evenly across a retirement. It is front loaded.

The practical reading is that decisions taken in that window carry unusual weight. Asset allocation, the size of a cash buffer, the initial withdrawal rate, and whether spending can flex all matter more in years minus five to plus five than they will at year twenty five. Our note on portfolio management covers the mechanics of setting an allocation deliberately rather than by drift.

It also means the pre retirement half of the window deserves as much attention as the post retirement half. A saver five years from stopping has the largest balance of their life and the shortest remaining runway to recover from a decline. That is sequence risk too, and it starts before the first withdrawal.

Lever one flexible spending

The single most effective defence is also the least popular: spend less when the portfolio is down. It works because it directly attacks the mechanism, reducing the number of units sold at depressed prices in exactly the years that matter.

Put a number on it using Sequence A. Suppose the retiree adopts a simple rule: in any year that follows a negative return, cut spending by 15 percent, from 40,000 dollars to 34,000 dollars. On this sequence that triggers in years 2, 3 and 5. Total withdrawals over the decade drop from 400,000 dollars to 382,000 dollars, a reduction of 18,000 dollars.

The ending balance rises from about 914,226 dollars to about 953,773 dollars, an improvement of about 39,547 dollars. Forgoing 18,000 dollars of spending bought more than twice that in ending capital, because the money was retained at the worst possible moment for selling and then participated in the five good years that followed.

The cost is not hidden. It is three years of living on 15 percent less, and it arrives at the same time as the market news that made you nervous. A flexible rule also has to be set in advance, in writing, because deciding mid decline whether this particular decline qualifies is exactly the decision people get wrong. A rule you will actually follow is worth more than an optimal rule you will abandon, and any figure quoted here depends entirely on the invented series behind it.

Lever two a cash and short bond buffer

The second lever attacks the same mechanism from the other side: if you do not have to sell equities in a down year, the down year cannot force a permanent unit loss. That means holding some spending in assets that are not moving with the stock market.

A common structure sizes the buffer in years of spending rather than percentages. On the illustrative plan used here, two years of spending in cash is 80,000 dollars, and a five year ladder of short bonds covering the following five years of spending is 200,000 dollars. Against a 1,000,000 dollar portfolio that is 8 percent in cash and 20 percent in bonds, leaving 72 percent in equities. Our note on building a bond ladder covers how the rungs and maturities are set up.

The cost is expected return. Money parked in cash and short bonds is money not compounding at the equity rate, and over a long retirement that drag is real. In a good sequence like B, the buffer is pure cost: those funds would have been better off invested every single year. The buffer is insurance, and insurance loses money most of the time by design.

There is also a discipline requirement. A buffer only works if it is actually spent during declines and refilled during recoveries. A buffer that sits untouched through a downturn because it feels safer to keep it is not a strategy, it is just a lower expected return.

Lever three a lower initial withdrawal rate

Starting lower is the bluntest defence and the most reliable. A smaller withdrawal is a smaller fraction of the portfolio in every state of the world, so it shrinks the effective rate ceiling that a bad sequence can push you to.

On Sequence A, dropping the withdrawal from 40,000 dollars to 35,000 dollars a year takes the ending balance from about 914,226 dollars to about 1,005,350 dollars. The retiree withdrew 50,000 dollars less across the decade and finished about 91,124 dollars ahead, again because the retained dollars compounded through the strong second half.

That change is a 3.5 percent initial rate rather than 4.0 percent. Inverted into a savings target it means the same 40,000 dollars of spending would require about 1,143,000 dollars rather than 1,000,000 dollars, roughly 14 percent more capital, which our retirement number analysis works through as a savings problem rather than a withdrawal one.

The cost is obvious and large: either you live on less, or you work longer to accumulate more, or both. There is no version of this lever that is free. What makes it attractive anyway is that it is the only lever that does not depend on you executing correctly during a downturn. It is set once, at the start, and then requires nothing of your nerve.

The peak effective rate figures show why it works. Sequence A pushed a 4.0 percent plan to 7.00 percent at its worst. A 3.5 percent plan facing the identical market would have peaked meaningfully lower, and the difference between those peaks is the difference between a stressed plan and a broken one.

Lever four working a little longer

An extra year of work does three things at once, which is why it moves the needle more than its length suggests. It adds a year of savings, it removes a year of withdrawals, and it shortens the retirement the portfolio must fund by a year.

Illustratively, if a final year of work adds 30,000 dollars of savings and defers one 40,000 dollar withdrawal, the portfolio enters retirement with roughly 70,000 dollars more in play, about 7 percent of the illustrative starting balance, before counting whatever the market did that year. On a plan already near the edge, 7 percent is not a rounding error.

There is a version of this lever that responds to sequence directly: treat the retirement date as flexible rather than fixed, and be willing to delay by a year if the market falls sharply in the months before it. That converts a portion of sequence risk into schedule risk, which for many people is the more tolerable of the two.

The cost is a year of your life, and that is not a small price. It is also not always available. Health, caring responsibilities, redundancy and the realities of hiring at older ages all make the working longer lever less reliable than a spreadsheet implies. Treat it as one option among several rather than the default answer, and be honest about whether it is genuinely under your control.

Lever five a rising equity glide path

Conventional advice is to reduce equity exposure steadily with age. A rising glide path does the opposite across the red zone: it starts retirement with a lower equity share than usual and increases it over the following decade or two.

The logic follows directly from the arithmetic above. The most dangerous moment for equity exposure is when the balance is largest and withdrawals are beginning, so holding less equity right there reduces the worst case. As the years pass and the portfolio either recovers or proves durable, the remaining horizon shortens but the sequence exposure falls, so more equity becomes tolerable again.

The cost is straightforward: in a good sequence, you own less of the thing that went up during the years it went up most. Under Sequence B, a lower equity share in the opening five years would have left the retiree meaningfully poorer for no benefit at all. This lever, like the buffer, is paid for in good states of the world.

It also demands rebalancing discipline, because a rising path means buying equities after declines, which is emotionally the hardest trade there is. Our rebalancing walkthrough covers the mechanics of setting bands and executing them without judgment calls. A glide path that exists only on paper does nothing.

The dividend funded withdrawal defence

A commonly proposed defence is to hold income producing assets and spend only the income, never the principal. The appeal is obvious against everything above: if the cash comes from dividends and interest rather than from sales, you are not converting a price decline into a permanent unit loss.

The mechanism is genuinely sound as far as it goes. Dividends are declared per share, not as a percentage of the market value, so a company paying an illustrative 1.40 dollars a share pays it whether the share trades at 40 dollars or 28 dollars. A retiree whose spending is fully covered by that stream can hold every unit through a decline and let the recovery apply to the full position.

The problem is coverage. On an illustrative 1,000,000 dollar portfolio, a dividend tilted allocation yielding 3.5 percent produces about 35,000 dollars a year, which covers 87.5 percent of a 40,000 dollar requirement, leaving 5,000 dollars to come from sales. A broad market allocation yielding closer to 1.5 percent produces about 15,000 dollars, or 37.5 percent of the need, leaving 25,000 dollars to be sold every year regardless of prices. Our dividend income sizing analysis works through what portfolio size a given yield actually requires.

Building toward full coverage means either a much larger portfolio or a much higher yield, and reaching for yield introduces its own concentration and payout sustainability problems. Partial coverage is still worth having. It is just not the same thing as immunity.

Why a dividend cut is the real risk not a price fall

The honest limitation of the income approach is not that prices fall. It is that payouts can fall too, and they tend to fall in the same conditions that cut prices.

Trace it through Sequence A. The portfolio starts year 3 at about 657,536 dollars. If the dollar payout held steady at 35,000 dollars through the decline, the retiree is fine on income and the yield on the depressed balance now reads about 5.32 percent, which looks like a windfall and is really just a smaller denominator. Every dollar of that stream is a dollar of shares not sold, and that is exactly the protection the strategy promised.

Now suppose payouts are cut by an illustrative 20 percent in the same downturn. The 35,000 dollars becomes 28,000 dollars, and the 12,000 dollar shortfall has to be raised by selling into the decline. At year 3 prices that means selling more units per dollar raised, precisely when unit preservation matters most. Alternatively, measure the income against the shrunken balance: 3.5 percent of 657,536 dollars is only about 23,014 dollars, leaving nearly 17,000 dollars to be funded from principal if the yield rather than the dollar payout is what holds.

Neither scenario is a reason to dismiss income investing. It is a reason to size the plan on the dollar payout you would still receive after a cut, not on the payout in a good year, and to treat a high starting yield as a claim requiring scrutiny rather than a solved problem. Our note on evaluating dividend stocks covers what makes a payout more or less likely to survive a downturn.

What a spending buffer looks like as a share of the portfolio

Turning the buffer lever into a picture helps, because “hold some cash” is not a plan and “hold two years of spending” is. On the illustrative plan running through this analysis, 40,000 dollars of annual spending sets the unit of measurement, and the buffer is built in multiples of it.

Two years of spending in cash is 80,000 dollars. Five further years covered by a short bond ladder is 200,000 dollars. The remaining 720,000 dollars stays in growth assets. Expressed as shares of the 1,000,000 dollar portfolio, that is 8 percent, 20 percent and 72 percent.

The point of expressing it this way is that the buffer’s job is defined in time, not in risk tolerance. Seven years of spending covered without touching equities is a specific claim about how long a decline can be waited out. Someone who wants ten years covered would hold 400,000 dollars outside equities, which is 40 percent of the portfolio and a much heavier drag on long run growth.

Choosing between those is a real trade with no correct answer, and it depends on other income, timeline and temperament rather than on arithmetic.

How sequence risk shaped the 4 percent rule

Sequence risk is the reason a sustainable withdrawal rate is not simply the expected return. If order did not matter, a portfolio expected to compound at 5.09 percent could support 5.09 percent a year indefinitely, and the invented decade above would raise no concerns at all: the average was 6.00 percent, comfortably above the 4.0 percent being withdrawn.

The Sequence A retiree withdrew below the average return for ten years and still ended the decade down. That is the whole justification for setting a starting rate well beneath any historical average. The gap between the average return and the safe withdrawal rate is not conservatism for its own sake. It is the price of surviving a bad opening.

This is also why the rule is usually stated as an INITIAL rate applied to a starting balance, then adjusted for inflation, rather than as a percentage of the current balance each year. It is trying to deliver a stable income, which is what people actually want, and stable income is precisely the feature that makes it sequence sensitive. Our 4 percent rule analysis covers what the rule claims, what it assumes, and the several places it breaks.

The corollary is worth stating plainly: a 4 percent starting rate is not a promise. It is a rate chosen so that a plan survives poor sequences, which means it will look far too cautious in most sequences and will still be under strain in the worst ones. A rule calibrated to the bad case necessarily leaves money unspent in the good case.

Why an average return cannot warn you

Every summary statistic used in investing is blind to this problem, and the blindness is structural rather than accidental. An arithmetic average adds the returns and divides, discarding order by construction. A compound growth rate uses only the beginning value, the ending value and the elapsed time, so it discards not just order but the entire path.

Both of the sequences here would report the same 6.00 percent average and the same 5.09 percent compound growth rate. Any performance table would show them as identical. Any projection built on either number would produce one line, not two, and would give no hint that the same inputs can produce a 327,910 dollar spread in outcomes once withdrawals are running.

This is not an argument against those measures. They are the right tools for comparing results on a common scale, and our compound growth rate breakdown works through exactly what they do and do not capture. It is an argument against using them alone for a withdrawal plan.

The alternative is to plan against sequences rather than averages: ask what happens if the first five years are bad, then what happens if they are good, and check that the plan is tolerable in both. That is what stress testing means in practice, and it does not require sophisticated software. Running one deliberately unpleasant opening decade by hand, as this analysis just did, teaches more than a hundred projections at an assumed constant rate.

What sequence of returns risk is not

Three misreadings show up often enough to be worth separating out.

It is not the same as volatility. Volatility is how much returns vary. Sequence risk is what happens when that variation interacts with cash flows. A volatile portfolio held by someone with no withdrawals has no sequence risk at all, and a modestly volatile portfolio with large withdrawals has plenty.

It is not an argument for avoiding equities. The Sequence A retiree still ended the decade with about 914,226 dollars because the equity exposure eventually delivered. A portfolio with too little growth exposure faces a slower but equally lethal problem, as inflation erodes what a fixed pot can buy across a thirty year retirement. The 4 percent rule analysis covers why the rule quietly depends on holding meaningful growth assets.

It is not something you can time your way out of. Nobody knows in advance whether they are living through Sequence A or Sequence B, and the difference is only visible in hindsight. Attempting to sidestep the bad sequence by exiting the market on a forecast introduces a different and better documented failure mode, which is being out of the market for the recovery years that Sequence A depended on to reach 914,226 dollars at all.

The honest posture is preparation rather than prediction: build a plan that is survivable in the bad ordering, accept that it will underspend in the good one, and treat that asymmetry as the cost of not knowing which one you got.

Where the withdrawal comes from and how rebalancing fits

Sequence risk turns a bookkeeping question into a strategic one: when you need 40,000 dollars, which holding do you sell? A portfolio with a single asset has no choice. A portfolio with several does, and using that choice well is most of what a buffer strategy actually is.

The general principle is to fund spending from whatever is least depressed. In practice that means taking the year’s cash from accumulated dividends and interest first, then from the cash and short bond slice if equities are down, and from equities when they are at or above target weight. That sequence is not clever, it is just a rule that avoids forced selling at the worst prices.

Rebalancing does the same job from the other direction. Selling the asset that has run up to fund spending simultaneously trims the overweight position and raises cash without touching the one that is down. In a year when equities are strong, the withdrawal and the rebalance are the same transaction, which is the cleanest possible outcome. Our rebalancing walkthrough sets out how to define target weights and tolerance bands so this is mechanical rather than discretionary.

The one thing to avoid is inventing a policy during a decline. Write down in advance which sleeve funds spending under which conditions, and how the buffer gets refilled after it is used. A written withdrawal policy is unglamorous and does more for a retirement plan than any allocation tweak, because it removes the single decision most likely to be made badly.

A short annual routine for the red zone years

Sequence risk rewards a small amount of regular attention rather than constant monitoring. An annual review, run at the same point each year, covers almost everything that matters.

Start by writing down the current balance and dividing this year’s planned withdrawal by it. That effective rate is the number that tells you whether the plan is under strain, and it is invisible unless you calculate it deliberately. On the Sequence A path it would have crossed 6 percent by year 3, which is a signal a full year before anything felt like a crisis.

Second, count how many years of spending sit outside equities. If the buffer covers fewer years than you decided it should, refilling it is this year’s job. If it covers more, that money is drag and should go back to work. Third, check allocation against target and rebalance if the bands are breached. Fourth, if a flexible spending rule was written down, check whether it triggered and act on it before deciding whether you feel like it.

Fifth, note whether any income the plan depends on has changed, particularly a reduced dividend or interest stream, since that changes how much has to come from sales. None of this takes more than an hour, and none of it requires forecasting anything. Test your revised figures in the retirement number calculator once a year and treat any large move in the required number as a prompt to talk to a qualified financial professional rather than to act immediately.

Common misreadings of sequence risk

The first is treating it as a reason for despair. The gap between Sequence A and Sequence B is large, but the Sequence A retiree is not ruined; they are constrained, and the constraints are manageable with the levers above if they are recognised early. Sequence risk is a planning parameter, not a verdict.

The second is over engineering. Elaborate withdrawal rules, complex bucket structures and rigid glide paths all look impressive and all depend on being executed consistently for thirty years, often by someone in their eighties or by a surviving spouse who did not build the system. A simple rule that will still be followed in year twenty five beats a sophisticated one that will not.

The third is assuming the illustration is a forecast. Every number in this analysis comes from a made up sequence of ten returns chosen to make the arithmetic legible. Real markets do not deliver tidy reversals, real spending is not constant in real terms, taxes take a share of every withdrawal, and a real retirement runs two or three times longer than the decade modelled here. The direction of the effect is robust. The magnitudes are not.

The fourth is thinking the problem ends after the red zone. It thins out, because a portfolio that survived its opening decade is usually large relative to remaining spending, but a fixed real withdrawal against a shrinking balance can still push the effective rate up late in a long retirement. The annual check above is worth keeping.

The bottom line

Order matters or it does not, entirely depending on whether money is leaving the account. A saver who touches nothing finishes at the same place whichever way the returns are shuffled, because multiplication does not care about sequence: on the invented decade used here, an illustrative 1,000,000 dollars becomes about 1,643,219 dollars either way. A retiree withdrawing 40,000 dollars a year through the identical returns finishes at about 914,226 dollars if the weak years come first and about 1,242,136 dollars if they come last. Same returns, same 6.00 percent average, same 400,000 dollars spent, a 327,910 dollar difference.

The mechanism is that a withdrawal during a decline sells units that never come back, so the recovery has less to work on. That is why the effective withdrawal rate on the bad path climbed to 7.00 percent while the plan still called itself a 4 percent plan, and why a 4 percent starting rate exists at all rather than something closer to the average return.

What you can do about it is bounded and specific: let spending flex, hold a buffer measured in years rather than percentages, start at a lower rate, keep the retirement date slightly negotiable, and think carefully about equity exposure through the years either side of stopping work. Each of those buys protection in the bad ordering by giving something up in the good one, and none of them is free. Since nobody knows which ordering they got until it is behind them, the reasonable posture is to build a plan that is tolerable in both and to check the effective rate once a year. Every figure above is an invented illustration, and the arithmetic is the only part worth trusting.

Dividora publishes independent educational analysis, and nothing above constitutes personalized investment, tax, or retirement advice. The ten annual returns used throughout, both orderings of them, and every balance, withdrawal, yield and allocation derived from them are invented illustrations built to make one arithmetic point clearly; they are not observations of any real market, index, fund or security, and no outcome shown should be read as a projection of what any portfolio will do. Real retirements run far longer than the decade modelled here, real spending changes, taxes reduce every withdrawal, and no rule of thumb about withdrawal rates or buffers is safe for everyone. Before setting or changing a withdrawal plan, model your own circumstances and discuss them with a qualified financial professional who can account for your tax position, other income, health and timeline.

Edited by Hamza Hai, MBA · Editor

Hamza Hai is the editor of Dividora. She holds an MBA and reviews the site's articles against our editorial standards, checking that every figure is labelled for what it is, that nothing is presented as verified fact without a source the reader can check, and that the writing stays useful to a non-specialist.

How we research, write and review · LinkedIn