Every December, a predictable astonishment sweeps the world: the holidays are expensive. Every year, car owners are ambushed by the annual service they have had annually for a decade. These are not emergencies — they are appointments. The budget-killer is not that such costs exist; it is that they arrive in lumps while income arrives in slices.
The sinking fund is the old, unglamorous answer: for each known irregular expense, save a small fixed amount every month into a labelled pot, so the lump is already sitting there when its date comes round. The name comes from old public finance — money set aside to 'sink' a debt — but the household version is simpler: it is pre-paying your own future, monthly.
How to build the list
Scan a full year of spending and pull out everything that is irregular but foreseeable: insurance renewals, vehicle service and repairs, holidays and travel, gifts and celebrations, school costs, annual subscriptions, home maintenance. For each, estimate the yearly total, divide by twelve, and that is its monthly price. A holiday that costs twelve hundred is, in truth, a hundred a month all year.
The moment of clarity for most people is the sum at the bottom: irregular-but-known costs often total several hundred a month. That money was always being spent — the sinking fund merely stops it being spent as a crisis, on a credit card, with interest attached.
Mechanics that make it stick
Keep sinking funds separate from both daily spending and the emergency fund — most banks now allow named sub-pots, and the label does real psychological work: money called 'Car — December service' does not get raided for a sale. Automate the transfers on payday so the funding happens before temptation wakes up.
Do not over-engineer it. Four to eight pots cover most lives; thirty pots is a hobby. And when a pot's expense arrives, spend the pot cheerfully — that is the system succeeding, not a setback. The strange pleasure of paying for a holiday from a pot that quietly filled itself is, for many people, the moment budgeting finally makes emotional sense.
What changes downstream
With lumps smoothed into monthly slices, the rest of your finances calm down. The emergency fund stops being nibbled by non-emergencies and stays intact for its real job. Credit card balances stop spiking each December and holiday season. And the monthly budget becomes honest: it finally reflects what your life actually costs, not just what it costs in an average month — a number many people have never truly seen.
That honesty compounds. Once every appointment-expense has a pot, a genuinely unexpected cost is rare enough to be met with curiosity instead of dread. Boring, labelled, automatic — the sinking fund is as close as personal finance gets to a cheat code, which is to say: it is arithmetic, applied early.
Why irregular is not the same as unpredictable
The distinction the whole method rests on is worth stating carefully, because conflating the two is what makes these expenses feel like emergencies. An unpredictable expense is one you could not have known about: a sudden illness, an accident, a job ending. An irregular expense is one you know is coming and cannot name the date of.
Almost everything that wrecks a monthly budget falls into the second category. Vehicles need servicing and eventually need repairs. Appliances have finite lives. Insurance renews annually. Gifts cluster around dates that are known years in advance. None of these is a surprise in any meaningful sense; they are simply not monthly.
The reason this matters is that the two categories deserve entirely different treatment. Unpredictable events are what the emergency fund exists for and cannot be planned in detail. Irregular events can be costed, totalled and divided by twelve, which converts them into a monthly figure like any other. Treating the second category as though it were the first is why the emergency fund keeps getting drained by things that were never emergencies.
Building the list from evidence rather than memory
The list is the whole exercise and it should be built from statements rather than from recollection, because memory reliably omits the expenses that only occur once a year. Going back through twelve or ideally twenty-four months and noting every payment that was not monthly produces a considerably more complete list than any amount of thinking about it.
The categories that appear most often are vehicle costs including tax, insurance and servicing; property maintenance and appliance replacement; annual insurance renewals of every kind; professional fees, subscriptions and memberships billed annually; holidays and travel; gifts and seasonal occasions; and periodic health, dental or optical costs.
The total tends to be a genuinely uncomfortable figure and the discomfort is the point. A household seeing this number for the first time is usually seeing an accurate account of why their finances feel tighter than the monthly arithmetic suggests they should. Those costs were always being paid; they were just being paid reactively, frequently through credit.
Estimating the ones without a fixed amount
Some entries have exact figures and some do not. An insurance renewal is knowable; the cost of the next appliance failure is not. The temptation is to omit the uncertain ones, which reintroduces exactly the problem the method was solving.
A workable approach for these is replacement cost divided by expected life. An appliance costing a certain amount and lasting roughly a decade justifies a tenth of that per year in the fund. Applied across the major items in a household, this produces a maintenance figure that is approximately right, which is enormously better than nothing.
For genuinely open-ended categories like vehicle repairs, an average of the last few years is a reasonable estimate, adjusted upward as the vehicle ages. The precision does not matter much. What matters is that a fund exists and is roughly the right size, so that the eventual bill draws down a balance rather than creating a crisis.
One account or several
There are two schools on the mechanics and both work. The multiple-account approach opens a separate pot for each category, which some banking apps make easy, and gives an unambiguous answer to whether a particular category is funded. The single-account approach holds one balance and tracks the allocation in a spreadsheet.
The multi-pot version is clearer and can be rigid: a well-funded holiday pot sitting alongside an underfunded car pot when the car breaks creates a decision that the single-pot version does not. The single-account version is more flexible and requires a record, without which it degenerates into an ordinary savings account whose purpose is forgotten.
The choice is largely temperamental. What is not optional is the separation from both the current account and the emergency fund. Sitting in the current account, the balance reads as spendable. Merged with the emergency fund, it destroys the ability to know whether the emergency fund is intact, which is the one thing that fund needs to be true of it.
The first year is the hard one
There is an unavoidable difficulty at the start: the annual costs continue arriving while the fund is still being built, which means the first year requires paying both the monthly contributions and the bills the contributions had not yet accumulated for. This is the reason many attempts at this fail in month three.
Two things make it manageable. The first is starting with the categories whose next occurrence is furthest away, which gives the maximum accumulation time before the first draw. The second is accepting a partial fund in the first year rather than a complete one, since even a partial balance reduces the shortfall.
It is also worth knowing that the difficulty is genuinely temporary. Once a full cycle has passed, the fund is drawing down and refilling in rhythm, and the monthly contribution becomes simply another fixed cost that requires no attention. The steady state is easy; the transition is not, and knowing that in advance prevents abandoning during the only part that is hard.
What this does to the rest of the system
The downstream effects are larger than the mechanism suggests, and they explain why this modest technique is worth the setup. The most immediate is that the emergency fund stops being consumed by non-emergencies, which means it is intact when something genuinely unforeseen happens.
The second is that credit usage falls, frequently substantially. A large proportion of ordinary consumer borrowing is incurred to meet exactly these irregular costs, and a household that meets them from a fund does not accumulate the balance that then compounds against them for years.
The third is harder to quantify and more valuable: the removal of a recurring low-grade anxiety about what is coming next. A household with a funded list knows the year ahead is provided for, which changes how a large unexpected bill feels when it arrives. That is the same relief the emergency fund provides, extended to the far more frequent category of things that were never actually surprises. None of this is financial advice; it is a description of a mechanism and what it tends to change.
The categories people always forget
Even a careful pass through statements tends to miss a consistent set of items, and they are worth listing because they are among the larger ones. Vehicle replacement is the clearest: a car has a finite life and will eventually need replacing entirely, which is a foreseeable cost that almost nobody provisions for and almost everybody finances.
Professional and regulatory costs are another: licences, registrations, insurance required to practise, continuing education requirements. These are predictable, annual, and frequently substantial for people in regulated professions who nonetheless treat them as an unwelcome surprise each year.
The third group is family-related and the most awkward to plan for because it feels cold to schedule: weddings you will be invited to, milestone birthdays, contributions toward family events, travel to see people. These occur reliably every year in some combination, and a household that has never provisioned for them meets each one from whatever is available.
When the fund is not enough and what that tells you
Occasionally a bill exceeds the fund it was supposed to come from, and the response should depend on why. If the estimate was simply too low, raise the monthly contribution for that category and continue. This is ordinary calibration and the first year of any system produces several of these corrections.
If the bill was genuinely exceptional — a repair far beyond anything foreseeable, a cost of a kind that had never occurred — then it belongs in the emergency fund category after all, and using that fund is correct. The line between the two is not always clean and it does not need to be, provided the decision is made deliberately rather than by default.
What is worth watching is a pattern of repeated shortfalls in one category, which usually indicates something more than an estimating error. A vehicle whose repair fund is consistently exhausted is telling you something about the vehicle. A property whose maintenance fund never lasts is telling you something about the property. The fund is functioning as a diagnostic at that point, which is a second use worth having.