Sequence of returns risk is one of the most underestimated dangers for anyone who starts living off a portfolio. It creates an unpleasant paradox: two portfolios with exactly the same average return can end up with very different balances, depending only on the order in which the good years and the bad years arrive. The damage is not just a temporary shortage of cash: when withdrawals and a downturn coincide, part of the loss becomes permanent.
This article explains what sequence of returns risk is, why it does the most damage once money is being taken out, and three strategies that reduce it: the bond tent, the cash buffer and dynamic withdrawals. The last section translates the same arithmetic into the situation of a trader who lives off a trading account.
What Sequence of Returns Risk Is
The term describes the danger that a run of poor returns arrives in the first years in which someone begins to draw money from a portfolio. That the order should matter at all sounds counterintuitive, because a lump sum that is left alone ends at the same value whatever the order of its returns. The sequence starts to make a difference as soon as money moves in or out, and it makes the biggest difference once money is being taken out.
Take an investment that loses 20% in the first year and gains 25% in the second. A starting capital of $100,000 drops to $80,000 and then climbs back to $100,000. Reverse the order and the result is identical: $125,000 after the first year, $100,000 after the second. Multiplication does not care about order, and a portfolio that only compounds is a chain of multiplications.
Everything changes when money is withdrawn to pay for living expenses, which is what living off investments means. If a sharp loss arrives right at the start of the decumulation phase, the years in which a portfolio is drawn down rather than built up, the withdrawals come out of a capital that has already shrunk. To raise, say, $40,000, the investor has to sell more fund units, shares or bonds precisely when their prices are low. Those units are no longer in the portfolio when the market recovers, and that is the part of the damage that never comes back.
The real problem is therefore not an insufficient average return. It is the combination of a bear market and withdrawals taken at the worst possible moment. The risk is highest in the first years of retirement, or more generally whenever someone moves from accumulating capital to living, at least in part, on what it produces. A deep bear market early in that phase can undermine the sustainability of a portfolio for many years afterward.
The Same Returns in Two Orders: A Worked Example
The example below uses a starting portfolio of $1,000,000, a fixed withdrawal of $40,000 taken at the start of each year, and ten annual returns that are the same set of numbers in two different orders. There are no taxes, fees or inflation adjustments, and the $40,000 never changes, so the only difference between the two cases is the order of the returns. Anyone can reproduce the figures in a spreadsheet.
The ten returns are +25%, +18%, +14%, +10%, +8%, +6%, +3%, -4%, -9% and -18%. Their arithmetic average is 5.3% a year and their geometric average is 4.56%, and both averages are the same whichever way the sequence is ordered. Without any withdrawal, both orders finish the decade at $1,562,366.
With the withdrawals, the two orders diverge. When the good years come first, the portfolio pays out $400,000 over ten years and still ends at $1,184,453. When the bad years come first, in the reverse order from -18% up to +25%, the same $400,000 of withdrawals leaves $837,818: a gap of $346,635, or 29% of the better outcome, produced by the same returns and the same spending.
The mechanism becomes visible if the portfolio is imagined as a fund with 10,000 units priced at $100 at the start. In the bad-first sequence the first three withdrawals cost 1,424 units, because after the first year each $40,000 has to be raised at falling prices ($82.00, then $74.62 a unit); in the good-first sequence they cost 991 units. The extra 433 units sold in the first three years are gone for good, and every later recovery applies to a smaller base. The gap keeps widening after that: by the end of the decade the bad-first sequence has sold 4,638 units against 2,419. Both orders finish with the same unit price, $156.24, so the 2,219 extra units account for the entire $346,635 difference. That is the whole of sequence of returns risk in one number.
Why the First Years Weigh So Much
The effect is not spread evenly across the horizon. In a January 2014 paper in the Journal of Financial Planning, Wade Pfau and Michael Kitces recalled a finding from Kitces's own 2008 research: over a 30-year horizon, the outcome of a withdrawal scenario is dictated almost entirely by the real returns of the portfolio in the first 15 years. A bad decade at the end matters much less, because by then the remaining horizon is short and the balance has already been secured or lost. In an October 2016 article, Kitces widened the window to include the years just before withdrawals begin. He called the last decade before retirement and the first decade after it a "retirement danger zone": in those years the portfolio is so large that its own gains and losses dwarf anything that further saving or working longer can add, so the outcome depends heavily on the sequence of returns on both sides of the retirement date.
History shows what a bad start looks like. The S&P 500 fell 49.1% from its record close of 24 March 2000 to its closing low of 9 October 2002, and 56.8% between 9 October 2007 and 9 March 2009. Someone who began withdrawing in early 2000 lived through both declines within the first decade of decumulation. The index recovered; the units sold along the way did not.
Strategy 1: The Bond Tent
One way to reduce sequence of returns risk is the "bond tent," the name Kitces used in that 2016 article for an allocation that adds a pre-retirement build-up of bonds to the rising equity glide path he had studied with Pfau, a stock allocation that increases year by year through retirement. The idea is to raise the bond share of the portfolio progressively as the start of withdrawals approaches, so that bond exposure peaks right around the beginning of decumulation, and then to let the equity share rise again once the most delicate years have passed. Drawn over time, the bond allocation goes up, peaks and comes down: the shape of a tent.
In their 2014 study, Pfau and Kitces found that a portfolio starting at 30% in stocks and finishing at 60% performed better than one that stayed at 60% throughout, and that, under the paper's baseline return assumptions, a 4% withdrawal rate over 30 years had its highest success rates, by a small margin, with glide paths starting at 20% to 40% in equities and finishing at 60% to 80%. The logic is simple. If a bear market hits in the first years, a larger bond position dampens the portfolio's volatility and provides assets that can fund withdrawals without selling stocks at depressed prices; the equity share then rises again on a preset schedule, as the bond reserve is spent down first.
This does not mean giving up equities. A portfolio that has to sustain decades of withdrawals generally still needs to outgrow inflation, and overdoing the caution creates a different risk: returns too low to last. The tent is a temporary shelter, not a permanent home.
Strategy 2: The Cash Buffer
A second strategy is a cash buffer: a reserve set aside to cover part of future spending without selling investments. The approach goes back to the financial planner Harold Evensky, whose firm has paired clients' long-term portfolios with a "cash flow reserve" since 1985; in the form it describes, the reserve covers roughly the next year or two of the withdrawals the portfolio has to fund. Such a reserve can sit in cash or in very short-term instruments.
The benefit shows up in a downturn. If equities fall quickly, the buffer pays the bills for a while instead of the portfolio, and the investments get time to recover before anything has to be sold. The cost is that cash earns less than riskier assets over the long run, sometimes nothing at all, and loses purchasing power to inflation. The size of the buffer is therefore a balance: large enough to avoid forced sales in a crisis, and small enough that most of the wealth keeps working.
Strategy 3: Dynamic Withdrawals
The third option is to give up the idea of withdrawing a rigidly fixed amount every year. Dynamic withdrawals adapt the sum taken out to the state of the portfolio: a moderate increase after good years, a temporary reduction after a sharp fall, when personal finances allow it. The rules are set in advance, so that no decision is left to the emotions of the moment: floors and ceilings on the annual withdrawal, a percentage of the portfolio's current value, or automatic cuts when the balance drops below a threshold.
Dynamic rules are easiest to judge against the fixed withdrawal they replace, and that is where the research on withdrawal rates started. William Bengen's 1994 article in the Journal of Financial Planning found that, on U.S. data going back to 1926, an initial withdrawal of 4% adjusted for inflation every year had never exhausted a portfolio of 50% stocks and 50% intermediate-term Treasuries within 30 years; that is the origin of the "4% rule." The Trinity study of 1998, by Philip Cooley, Carl Hubbard and Daniel Walz, backtested stock and bond mixes on data from 1926 to 1995 and found success rates of 95% and 98% for the same 4% inflation-adjusted withdrawal over 30 years with 50% and 75% in stocks. In a March 2006 paper, Jonathan Guyton and William Klinger turned flexibility into explicit "guardrails": if the withdrawal rate, meaning the year's withdrawal divided by the current balance, rises more than 20% above its initial level, the withdrawal is cut by 10% (a cut that no longer applies in the last 15 years of the plan); if it falls more than 20% below, it is raised by 10%. Together with the portfolio-management and withdrawal rules from Guyton's earlier work, those guardrails supported initial rates of 5.2% to 5.6% over at least 40 years, at a 99% confidence level, for portfolios with at least 65% in equities.
Recent estimates are more cautious about the rigid approach. Morningstar's State of Retirement Income report, published in December 2025, put the highest safe starting withdrawal rate for a new retiree seeking constant inflation-adjusted spending at 3.9%, over a 30-year horizon, with a 90% probability of success and 30% to 50% in equities, up from 3.7% a year earlier. The same research concluded that retirees willing to tolerate some fluctuation in their spending can start at nearly 6%. The gap between the two figures is the price of rigidity, and much of it is protection against a bad sequence.
What This Means for a Trader
A trader who lives off a trading account faces the same problem with a faster clock. Monthly withdrawals during a drawdown are the bad-years-first sequence compressed into months: every withdrawal made at the bottom shrinks the base that has to earn the recovery, and a strategy with a positive expectancy can still leave the account too small to survive its own losing streak. The article on how much capital full-time trading requires works out how large the account has to be for the withdrawals it must fund, and the one on the hidden side effect of living off trading describes the constraints and the stress that come with depending on trading income.
The three strategies translate directly. The cash buffer becomes a reserve of living expenses held outside the trading account, so that withdrawals do not have to come out of the account while it sits at the bottom of a drawdown. Dynamic withdrawals become a rule that takes out a set share of the profits earned above a floor balance, rather than a fixed sum every month. The bond tent becomes a lower risk per trade in the first months after withdrawals begin, raised again only once a cushion of profits has built up. Sizing every trade as a fraction of current equity rather than of the starting balance adds a brake that works on its own, because position size shrinks after losses; the article on risk of ruin shows how much the size of that fraction changes the odds of survival, and the one on compound interest explains why recovering from a loss always takes a larger percentage gain than the loss itself.
Sequence of returns risk cannot be eliminated, because nobody chooses the order in which the market delivers its years. What can be chosen is how much has to be sold at the worst moment. Every strategy in this article, whether it is called a tent, a buffer or a guardrail, is a way of making that number smaller.
