The standard advice is so familiar it barely registers as advice. As you approach and enter retirement, you sell down your stocks and buy bonds, gliding from a growth-oriented portfolio to a safer, income-oriented one. Rules of thumb encode it, "100 minus your age" in stocks, or its slightly bolder cousin "110 minus your age," and target-date funds automate it, typically arriving at something like 30% to 50% stocks by the retirement date and drifting lower from there. The logic seems unassailable: you have less time to recover from a loss, so you should take less risk.
The trouble is the word "risk," treated here as if it were a single quantity to be turned down like a dial. It isn't. A retiree faces at least two distinct dangers that pull in opposite directions, and the move from stocks to bonds does not reduce risk so much as trade one of those dangers for the other. Seeing that clearly changes the entire question.
Two risks that point opposite ways
The first danger is sequence-of-returns risk. When you are withdrawing money to live on, a market crash in the early years of retirement is far more damaging than the same crash later, because you are forced to sell depleted assets to cover spending, permanently shrinking the base from which any recovery must grow. The first five to ten years of retirement are the danger zone, and this risk is real and well documented.
The second danger is the mirror image: longevity and inflation risk, the possibility that you outlive your money or that inflation quietly erodes what it can buy. A retirement can easily last 25 to 35 years, and over that span inflation is relentless. At 3% a year, it cuts the purchasing power of a fixed sum by nearly half over 25 years, so a portfolio that merely preserves nominal value is steadily losing ground in real terms. This is a slow danger rather than a sudden one, but it ends in the same place: running short.
Here is the crux. These two risks respond to bonds and stocks in opposite ways. Bonds dampen sequence risk, because they are less volatile and steadier to draw from, but they worsen longevity risk, because over decades they rarely grow enough to outpace inflation. Stocks do the reverse, exposing you to sharp early drops while offering the long-run growth that protects against outliving your money. There is, therefore, no allocation that is simply "safe." Every mix reduces one risk by increasing the other, and the comforting idea that shifting to bonds lowers your risk is an illusion created by looking at only one of the two dangers.
Why "de-risk into bonds" is only half right
Once you hold both risks in view, the conventional advice reveals itself as a partial optimization. Selling stocks for bonds at retirement is a sound response to sequence risk, the danger that dominates the early years, and to that extent it is genuinely wise. Where it goes wrong is when it is applied as a permanent, one-way shift, leaving a retiree so bond-heavy that inflation and longevity grind the portfolio down over the decades that follow.
A 65-year-old holding 35% stocks under the old rule, or 30% in a typical target-date fund, may be reasonably protected against an early crash and dangerously exposed to a 30-year erosion. The advice optimizes for the risk you can see and feel, a market plunge, at the expense of the risk that arrives too slowly to frighten anyone, the gradual loss of purchasing power and the arithmetic of a portfolio that cannot keep up with a long life.
What the research actually suggests
This is why some of the most interesting retirement research points in a counterintuitive direction. Work by Wade Pfau, Michael Kitces, and others on the "rising equity glide path" suggests that rather than steadily reducing stocks throughout retirement, many retirees are better served by starting relatively low in equities at retirement and then increasing their stock allocation as they age. The shape looks strange until you map it onto the two risks. You hold more bonds early, when sequence risk is at its peak, effectively building a buffer to survive the dangerous first decade, and then you drift back toward stocks as that danger recedes and longevity risk becomes the dominant threat, letting equities do the long-run work of outpacing inflation.
The practical upshot is that "sell stocks, buy bonds" is roughly right for the first several years of retirement and potentially quite wrong for the decades after. The instruction most people absorb as a permanent destination is better understood as a temporary posture, and the shape and timing of the shift matter far more than the simple binary of stocks versus bonds ever captures.
Higher yields have changed the deal
It is worth acknowledging a genuine change that makes the bond side of the ledger more attractive than it was a few years ago. In the near-zero-rate era of 2020 to 2022, buying bonds meant accepting almost no yield, so protecting against sequence risk came at a steep cost in forgone growth. With bonds now yielding meaningfully more, they do more of both jobs at once, providing real income and sequence protection while sacrificing less long-term growth than they used to. That shift tilts the calculus modestly toward bonds and makes the "buy bonds" half of the traditional advice more defensible today than it was in the recent past. It does not, however, repeal the underlying trade-off; it merely improves the terms on one side of it.
Your guaranteed income is already a bond
A reframing that most allocation rules ignore can change the answer entirely: guaranteed income streams function, economically, like bonds. Social Security, and a pension if you have one, deliver steady, low-risk cash flows regardless of what markets do, which is exactly what a bond allocation is meant to provide. A retiree with substantial guaranteed income therefore already holds a large, invisible "bond" position outside the portfolio, and that means the investable portfolio can hold more stocks than a naive rule would suggest without the household as a whole being over-exposed.
Judged this way, allocation is properly a question about total wealth, the portfolio plus the capitalized value of guaranteed income, not about the portfolio in isolation. It also explains why two retirees with identical portfolios might correctly choose very different stock-and-bond mixes: the one with a healthy pension can afford far more equity in the portfolio than the one living on savings alone, because their true bond-like ballast is already in place elsewhere.
The part the math leaves out
Finally, there is a factor no formula fully captures, which is that the best allocation on paper is worthless if you cannot live with it. A great deal of retirement de-risking is driven not by optimization but by the simple desire to sleep at night, and that desire deserves respect rather than correction. Market volatility is genuinely frightening when the portfolio is no longer an abstraction but the thing paying your bills, and an allocation you panic out of during a crash, selling at the bottom, does far more damage than a theoretically suboptimal allocation you can calmly hold through the storm.
So peace of mind is a legitimate input, not a lapse in discipline. The right allocation is not the one that maximizes expected wealth in a spreadsheet; it is the one that both addresses your actual risks and that you can realistically hold through the worst year you will face, because behavior, not arithmetic, is what usually determines whether a retirement plan survives contact with a bear market.
Put it all together and the question "should I sell stocks and buy bonds?" turns out to be the wrong question, because it treats risk as one thing to be minimized when it is really a choice between two opposing dangers, each made worse by guarding too hard against the other. The more useful question is not "how do I reduce my risk?" but "which of these two risks am I more exposed to, and which can I better afford to bear?" The answer depends on how long you might live, how much guaranteed income you already have, what yields are currently on offer, and what you can actually tolerate holding when markets fall. Selling stocks for bonds is neither safe nor reckless in the abstract. It is a specific bet that the danger you most face is a near-term crash rather than a long, slow outliving of your money, and whether that bet fits your life is something only you, ideally with a fiduciary adviser who knows your full picture, are positioned to judge.
Primary sources
- The rising-equity-glide-path research of Wade Pfau and Michael Kitces for the finding that starting retirement with a lower equity allocation and increasing it over time can reduce sequence-of-returns risk more effectively than the declining or static glide paths most target-date funds use.
- William Bengen's foundational safe-withdrawal-rate research and its 4% benchmark, used to illustrate how inflation erodes a fixed withdrawal's real value over a multi-decade retirement.
- Vanguard and general target-date-fund glide-path data, showing typical funds arriving near 30% to 50% equities at the retirement date and declining thereafter, as a reference point for the conventional approach.
- Standard asset-allocation guidance on the roles of stocks and bonds in a retirement portfolio, and on the "100 minus your age" and "110 minus your age" rules of thumb.
- Current bond and inflation-protected-bond yields, which are materially higher than during the 2020-2022 near-zero-rate period.
- The total-wealth framework in which guaranteed income streams such as Social Security and pensions are treated as bond-like assets outside the portfolio.