There is an intuitive but dangerous belief that the day you retire, you should move everything into the safest possible assets and stop taking any risk. It feels prudent. For a retirement that might last thirty years or more, it can quietly be the riskiest choice of all, because it hands your savings entirely over to inflation.
Your working life may end at retirement. Your money’s working life does not.
The long horizon that remains
Someone retiring today could easily need their savings to last three decades. Over that span, inflation does not pause politely — prices keep rising, and a pot held entirely in cash steadily loses the power to buy the life it was meant to fund. A portfolio with no growth engine can run out not because it was spent recklessly, but because it stood still while costs climbed.
The horizon at the start of retirement is still long enough that some growth is not a luxury; it is protection.
Balancing growth and safety
The answer is not to stay aggressive, but not to abandon growth either. Most retirement approaches keep a meaningful slice in growth assets to outpace inflation over the decades, alongside enough safety to cover near-term spending and steady the ride. The exact mix depends on your pot, your other income and your temperament.
The buffer of safe assets lets you leave the growth portion alone through downturns, so it can keep compounding for the many years of retirement still ahead.
Growth as a life-extender
Framed correctly, keeping some growth in retirement is not gambling — it is what allows the money to last as long as you might. A portfolio that quietly keeps pace with, or ahead of, rising prices is far more likely to still be there in your later years than one frozen in cash and slowly melting.
Retirement changes what your money is for. It does not end its need to grow.
How long the money actually has to last
The planning horizon at retirement is longer than most people assume, and the reason is that life expectancy figures are frequently misread. A published average at birth includes everyone who died young; the relevant figure is conditional life expectancy at your current age, which is considerably higher.
For someone reaching their mid-sixties in good health, a substantial probability of living into their nineties is realistic, and for a couple the relevant question is how long at least one of them survives, which is longer still. That is a thirty-year horizon at minimum and potentially more.
Thirty years is not a short-term problem. It is the same length as an entire accumulation phase, and treating the money as though it needed only to be preserved for a few years produces an allocation calibrated to the wrong question entirely.
The risk of being too safe
A portfolio moved entirely into cash and short-dated instruments at retirement has eliminated market risk and accepted a different one: that rising prices erode the purchasing power of the withdrawals over a horizon long enough for the erosion to be substantial.
The arithmetic is unforgiving. At even moderate rates of price increases, purchasing power halves over a couple of decades, which means a retiree relying entirely on safe assets faces a standard of living in their late eighties materially below the one they planned. That is a real risk, it is close to certain rather than probabilistic, and it produces no alarming statements along the way.
This is the strongest argument for retaining meaningful growth exposure well into retirement. The objective is not to maximise returns but to maintain purchasing power across a horizon where the alternative is a guaranteed slow decline.
Two horizons in one portfolio
The resolution is to stop thinking of a retirement portfolio as a single pot with one horizon. The money needed in the next few years has a short horizon and belongs in cash and short-dated instruments. The money that will not be touched for twenty years has a long horizon and can be invested accordingly.
Structured that way, the apparent conflict disappears. The near-term reserve provides the stability that prevents forced selling, addressing the sequence problem described elsewhere on this site. The long-term portion provides the growth that keeps the later decades funded.
This is sometimes formalised as a bucket approach, with a cash tier, an intermediate tier and a growth tier, replenished from each other during favourable periods. The specific implementation matters less than the underlying recognition that different parts of the money are answering different questions.
The withdrawal is not the whole balance
A useful reframing for anyone anxious about holding equities in retirement: in any given year you are only withdrawing a small percentage of the total. The overwhelming majority of the portfolio is not being touched this year, or next, and its horizon is measured in decades.
Seen that way, a decline affects a balance you were not going to spend for fifteen years, which is precisely the situation in which equity exposure is appropriate. The portion actually needed soon is the cash reserve, which was never exposed.
This does not make declines pleasant, and it does clarify what is actually at risk. The anxiety typically attaches to the headline balance figure, which is the least relevant number, while the figure that determines whether this year works — the cash reserve — is unaffected by anything markets did.
What a long horizon buys later
There is a category of expense concentrated at the end of a long retirement that most plans handle poorly: care costs, which can be substantial and arrive at the point when the portfolio has been drawn on for decades.
A portfolio that maintained growth exposure through retirement is in a materially better position to meet these than one that preserved a nominal balance while losing purchasing power. This is the practical payoff of the argument above, and it arrives at the moment of maximum need.
It is also the reason that the very end of a long retirement is not the time to be at the most conservative allocation, which is the opposite of the usual assumption. Where an estate is intended to pass to someone else, the relevant horizon extends beyond your own lifetime entirely, and the allocation should reflect whose money it is going to become.
How much growth, and how to decide
None of this argues for an aggressive portfolio in retirement. It argues against the assumption that safety is a single direction, and for an allocation determined by the same three inputs discussed throughout this site: capacity, tolerance and need.
Capacity is higher than it appears once the cash reserve is separated out, since the invested portion genuinely has a long horizon. Tolerance frequently falls in retirement, which is a legitimate constraint and should be respected rather than argued with. Need depends on how much of your spending is covered by guaranteed income sources and how much the portfolio must generate.
Someone whose essential costs are fully covered by state and defined benefit income has a very different situation from someone whose portfolio must fund everything, and the appropriate allocation differs accordingly. The general point stands for both: the money does not stop needing to grow on the day the salary stops, and a plan that assumes otherwise is solving a much shorter problem than the one that exists. None of this is financial advice.
The all-at-once transition to safety
A specific and common error is treating the retirement date as a switch: an accumulation portfolio one week and a conservative one the next. This concentrates a large reallocation into a single moment chosen for reasons entirely unrelated to markets.
If that moment happens to follow a decline, the reallocation locks in the loss permanently. If it follows a strong run, it may be reasonable. Nobody knows which in advance, which means the arrangement is a substantial bet on a date selected by an employment calendar.
The alternative is a gradual shift over several years approaching the date, which is what target-date funds implement automatically and what anyone managing their own portfolio can replicate with a schedule. It removes the single-moment risk entirely and costs nothing beyond deciding it in advance.
Reviewing a drawdown portfolio
A portfolio being drawn on needs a different review cadence from one being built. Annual is still right for the strategy, and the cash reserve needs checking more often, since its depletion determines when the next replenishment should happen.
The annual questions are short: is the reserve at its intended level, has the withdrawal rate drifted relative to the current balance, has spending changed materially, and has anything altered in the other income sources. Half an hour, once a year, on a fixed date.
What should not be part of it is reconsidering the allocation on the basis of the last twelve months, which is the review most people actually perform. The allocation should change when the horizon or the circumstances change, and a decade of retirement is long enough that both will — which is precisely why the review needs to be scheduled rather than prompted by the market.
The couple's horizon is longer than either individual's
For a two-person household, the relevant planning horizon is not either partner's life expectancy but the point at which the second one dies, which is meaningfully longer than the first. A plan built on a single life expectancy will be too short for a couple.
This has a second consequence that is easier to overlook: the surviving partner's income frequently falls when the first dies, since some pension entitlements reduce or cease, while household costs do not fall proportionally. Housing, utilities and most fixed costs are barely affected by one fewer person.
Checking the survivor provisions on every income source is a specific and finite piece of work, and it is worth doing well before it is needed. The combination of a longer horizon and a lower income at the far end of it is exactly the situation that a growth allocation maintained through retirement is best placed to address.
The shape of spending across a retirement
Retirement spending is not flat, and assuming it is produces a plan calibrated to a pattern that rarely occurs. The commonly observed shape is higher in the early active years, when travel and activity are at their peak; lower through a middle period as activity reduces; and potentially much higher at the end if care is required.
This shape has a useful implication. Higher withdrawals in the early years may be entirely sustainable if the middle period genuinely costs less, which means a rigid inflation-adjusted withdrawal understates what is affordable early and may overstate what is needed later.
It also identifies where the growth allocation earns its keep. The potential late-life costs are the largest single uncertainty in most plans, and they sit thirty years out — a horizon over which a portfolio that maintained growth exposure is in a substantially better position than one that did not. That is the practical case for everything argued in this article, stated as a specific rather than a principle. None of this is financial advice.