Think about the expenses that "ruin" a month: the car service, the annual insurance renewal, the festive season, the broken appliance. We label them surprises, but almost none of them are. We knew the car would need servicing and the holidays would arrive on the same date they always do. The surprise is not the cost — it is that we never set money aside for it.
A sinking fund is the antidote: a small, deliberate savings pot for a known future expense, filled a little at a time.
How it works
Take a predictable annual cost, divide it by twelve, and save that amount each month into a dedicated pot. A large yearly bill stops being a shock and becomes twelve gentle contributions you barely notice. When the bill arrives, the money is already there, waiting, and it never touches your emergency fund.
The mechanism is almost embarrassingly simple. Its power is entirely in the reframing: turning one painful lump into many painless slices.
What deserves a fund
Good candidates are costs that are large, irregular and foreseeable: annual insurance, vehicle maintenance, holidays and gifts, property upkeep, replacing an ageing laptop or phone. Anything you know is coming but not exactly when, and that hurts as a single payment, is a sinking-fund candidate.
You do not need a separate bank account for each — a simple list, or a few labelled pots within one savings account, is enough to keep them mentally distinct.
Why it changes how a budget feels
Households that use sinking funds report that money stops feeling chaotic, because the "unexpected" mostly disappears. The events still happen; they simply stop being financial ambushes. Your emergency fund, freed from covering things that were never emergencies, stays intact for genuine ones.
It is a small habit with an outsized effect on calm. Predict the predictable, and most of your money stress quietly goes away.
Why labelled money behaves differently
The effectiveness of this method rests on something economically irrational and reliably true: money assigned a purpose is treated differently from identical money without one. A balance labelled for a specific thing is spent on that thing; the same sum in an unlabelled pot gets absorbed.
Economists call this mental accounting and generally treat it as a bias, since a given sum has the same value regardless of what it is called. That is correct as a statement about the money and beside the point as a statement about behaviour, which is what actually determines outcomes.
The practical move is to use the bias deliberately rather than trying to overcome it. Naming each fund specifically — not savings but car tyres, not miscellaneous but December — changes what happens to the balance, at no cost and with no ongoing effort. It is one of the few places where a documented cognitive quirk can simply be pointed in a useful direction.
The name matters more than the structure
Following from that, the specificity of the label does real work. A fund called holiday behaves better than one called leisure, and one called February trip to see family behaves better still. The more concrete the name, the more clearly the balance belongs to something and the more obvious it is when it is being raided.
This also makes the annual review easier. A list of vaguely named funds requires remembering what each was for; a list of specific ones is self-documenting, and any that no longer correspond to something real are obvious and can be closed.
Where a bank offers named sub-accounts within a single balance, this is close to free to implement. Where it does not, a spreadsheet with one line per purpose does the same job. The mechanism is the naming rather than the technology, which means it works with whatever tools are available.
Guilt-free spending, which is the underrated benefit
The framing of this method is usually defensive: preventing costs from becoming crises. There is an equally important benefit on the other side, which is that money accumulated for a specific enjoyable purpose can be spent without any of the low-level unease that usually accompanies a large discretionary purchase.
Someone who has been putting money aside for a holiday for eleven months spends it differently from someone paying for the same holiday from general savings. The second is making a decision at the point of purchase, with all the second-guessing that involves. The first made the decision a year ago and is simply executing it.
This is a genuine improvement in the experience of spending money, and it is available at no financial cost whatsoever. For anyone who finds discretionary spending stressful — which is a large proportion of people who are otherwise good with money — it is possibly the most valuable thing this method does.
What to do with a surplus at year end
Some funds will end a cycle with money left over, because the estimate was generous or the expense did not materialise. The temptation is to sweep it into general spending, and there are better options.
Rolling it forward is the default and it is usually right, since an estimate that was generous this year may be inadequate next year and the categories are lumpy by nature. A car fund that survived a year untouched is not evidence that the car will never need work.
Where a surplus is genuinely persistent across several years, the estimate was simply too high and the monthly contribution should be reduced, with the difference redirected to whichever fund keeps running short or to long-term savings. What is worth avoiding is treating an accumulated surplus as a windfall, since that converts a functioning system back into the ad hoc arrangement it replaced.
Running this with a partner
In a household, this method works considerably better when the list is built jointly, for a reason that has nothing to do with the arithmetic. Two people almost always have different views about which irregular costs are inevitable and which are optional, and building the list surfaces that disagreement in a low-stakes context.
It also removes a common source of friction. A large annual expense arriving without provision produces a conversation about whether it should have been anticipated. The same expense arriving against a fund that both people agreed to a year earlier produces no conversation at all.
The practical arrangement is a shared set of funds for household costs, funded proportionally as described in the article on couples elsewhere on this site, alongside any individual ones each person wants. The shared list is the part that matters; the individual ones are nobody else's business.
Reviewing the list once a year
The list is not permanent and it drifts out of date faster than people expect. Circumstances change, vehicles get replaced, subscriptions end, children reach ages with different costs, and a list built three years ago will contain entries that no longer apply and omit ones that now do.
An annual review covering four questions handles this: which funds were never drawn on, which ran short, what significant irregular cost occurred that had no fund, and what has changed that will produce new costs next year. Twenty minutes, once a year.
The most useful of these is the third. Any expense that arrived without a fund is either a genuine emergency, which belongs elsewhere, or a gap in the list. Distinguishing between the two each year is how the list becomes progressively more complete, and after two or three cycles the number of genuinely unanticipated costs falls to something quite small.
Why this changes how a budget feels
The subjective effect of running this for a full cycle is out of proportion to the mechanism, and it is worth naming because it is the reason people who adopt it keep doing it. The recurring sense that something expensive is probably coming, which most households live with permanently, largely disappears.
It is replaced by something more specific: a list of known costs, each provided for, with a visible balance behind it. The unknown becomes a much smaller category, containing only genuine emergencies, which is what the emergency fund exists for and which is far less frequent than the ad hoc alternative makes it feel.
None of this requires earning more or spending less. It requires paying the same costs on a schedule rather than in response to their arrival, which changes nothing about the total and a great deal about the experience. That is an unusually good return on an hour of setup. As with everything on this site, this is educational rather than advice.
Starting with three rather than twenty
The most common way this method fails at the outset is over-engineering. Someone builds a comprehensive list of fifteen categories, sets up fifteen transfers, discovers the total exceeds what is available, and abandons the whole thing within two months.
A better start is three funds covering whichever costs have caused the most disruption in the past two years. For most households that is some combination of vehicle, property maintenance and the December cluster of gifts and travel. Three transfers, three balances, and a system that fits comfortably within what is available.
Once those have run through a full cycle and become invisible, adding a fourth is easy. The staged approach reaches the same place as the comprehensive one within a couple of years and has a far higher chance of still existing then, which is the only comparison that matters.
When the fund and the emergency fund disagree
There will be occasions where a cost arrives that could reasonably be met from either, and having a default resolves it without deliberation each time. The workable default is that anything on the list comes from its fund even if the fund is short, with the shortfall covered from the emergency fund and repaid.
This keeps the categories meaningful. A shortfall recorded as a shortfall produces the information needed to adjust the contribution next year. The same shortfall quietly absorbed by the emergency fund produces no information and leaves the estimate wrong indefinitely.
It also protects the emergency fund's integrity as a measurement. Its balance is supposed to answer one question — could we handle an income interruption — and that answer is only reliable if the balance has not been silently eroded by things that had their own provision.
The question the method is named after
The phrase in the title is worth taking seriously as a diagnostic, because it identifies precisely which costs need a fund. Any expense that has ever prompted it was, almost by definition, foreseeable and unprovided for, which is the exact category this method addresses.
Keeping a running note of every occasion the question arises, for one year, produces a better list than any amount of planning from memory. It is evidence rather than recollection, it costs nothing to maintain, and by the end of a cycle it will have identified the gaps that a statement review missed.
After two or three years of running this, the question genuinely stops occurring, which is a small and noticeable change in daily life. Costs still arrive at inconvenient moments; they simply stop being surprises, because each one draws down a balance that was accumulated for it. That is the whole claim of the method and it is a modest one that reliably delivers.