Sinking Funds Explained: The Budget Line Nobody Sets Up

In one three-week stretch a household gets a $714 car insurance instalment, a $186 dental bill and a wedding invitation with flights attached. None of it was in the budget. All three were entirely predictable, and the last one arrived with four months of notice.

This is the month that convinces people they are bad with money. They are not. Their budget was built on twelve identical months, and no year has ever had twelve identical months in it. Sinking funds are the fix, and in the four-bucket spending plan they are the second bucket, which is the one that decides whether the whole plan is still standing in June.

What is a sinking fund, exactly?

One irregular expense, divided by twelve, set aside monthly in cash before the expense arrives.

That is the whole mechanism. The term comes from corporate finance, where a company sets aside money each year against a bond it will have to repay in a decade, and the domestic version is identical in every way that matters. You know the expense is coming. You know roughly what it costs. You pay it in instalments to yourself instead of in one lump to somebody else.

The reason it feels like new information is that the expense usually already has a name, and that name is the problem. "Car insurance" sounds like a thing that happens in April. Written down properly it is $119 a month, every month, including the eleven months when no bill arrives. A household that does not hold that $119 is not a household with lower expenses. It is a household that will borrow $714 twice a year.

Why does a budget break without them?

Because the money has to come from somewhere, and there are only three places it can come from.

It comes out of the flexible bucket, which means groceries and fuel get squeezed in the month the insurance lands, and the squeeze fails because groceries are not optional. It comes out of savings, which means the emergency fund is being spent on things that are not emergencies, and never recovers, because the next one is six weeks away. Or it goes on a card.

The third is what usually happens, and it is worth pricing. Take one household's full year of irregular expenses, $7,240, put all of it on a card at 22.4% and pay $250 a month against it with no new charges. That takes 43 months and costs $3,274 in interest.

Read that again as a ratio. The interest is 45% of the original expenses. Not because the household bought anything foolish, but because a predictable $7,240 was funded in arrears rather than in advance. The expenses were identical either way. The $3,274 is the entire cost of not having a plan for them, and it is the clearest case in household finance where the money is lost to a timing decision rather than a spending one.

The second cost is less visible and probably worse. Once the card carries the year's insurance and dentistry, the minimum payment goes up, which reduces next month's flexible bucket, which makes the next irregular expense more likely to go on the card. This is the mechanism behind the pattern described in why budgets fail in week three, and it is not about discipline in any direction.

The nine sinking funds most households need

The list goes first. Ranges rather than single figures, because a household with two old cars and a house has a genuinely different total from a renter with a bus pass.

Renters swap the third line for two smaller ones: renters insurance, which is often annual, and a moving fund, because a deposit and a van is a four-figure expense that arrives with sixty days of notice.

Nobody runs all nine. Four to six of them carry most of the money, and the useful exercise is finding out which four yours are, because it differs more than people expect. A household with a nine-year-old car and no pets has a completely different shape from one with a new lease and two dogs, and the generic advice to "save for irregular expenses" is useless precisely because it never says which.

Here is what one household's year actually looked like. Take-home $4,800 a month, one car at nine years old, a $280,000 house, no pets:

$603 a month, against take-home pay of $4,800. That is 12.6% of everything the household earns, and in its previous budget the line did not exist.

The home repairs figure is the one to argue with, and it should be argued with. One per cent of a property's value per year is a convention, not a measurement, and it is a poor fit at both ends: on a new build it is too high for the first decade, and on a sixty-year-old house with the original boiler it is optimistic. Use it as a starting number in year one and replace it with your own receipts in year two. That is the only figure in the table that is not derived from something the household actually paid.

How do you work out your own numbers?

Twelve months of statements, not three. This is the step people shorten and it is the step that decides whether the total is right, because the defining feature of these expenses is that they do not appear in a typical month. A three-month window catches roughly a quarter of them and produces a sinking-fund total that is comfortingly small and wrong.

Export a year of your checking account and every card. Sort by amount, largest first, and read down until the transactions stop being interesting, which is usually somewhere around $150. Everything above that line which happened once or twice in the year is a candidate.

Then group them by what would cause them to happen again, not by what the merchant was. Two visits to a garage and a tyre shop are one fund. A dentist, an optician and the pharmacy run after an urgent-care visit are one fund. This matters because you are not building a record of last year, you are building a forecast, and the question each group has to answer is "how much, and how likely, next year".

Three adjustments once the groups are totalled. Add anything you know is coming that did not happen last year, such as tyres that are now at four years old, or the wedding you have already been invited to. Remove anything genuinely one-off, such as the immigration fee you will not pay again. And if an expense is both large and uncertain, such as a boiler on borrowed time, price the repair rather than the replacement, and understand that the replacement is what your emergency fund is for.

Divide each total by twelve. If you are starting in September and Christmas is in three months, divide that one by three instead, fund it at the higher rate until December, and drop it to a twelfth afterwards. Pretending the annual figure is available monthly from a standing start is how the first December goes wrong.

Where should the money sit: one account or several?

One. The separation you need is informational, and a spreadsheet provides it for free.

Open one savings account, point a single standing transfer at it for the day after payday, and keep a six-row table of what each fund holds. When the insurance instalment arrives you move $714 out and reduce that row. Nine accounts give you exactly the same information, nine statements, and nine chances for a $12 monthly fee to appear somewhere you were not looking.

Some banks and credit unions offer named buckets or sub-accounts inside one savings account, and if yours does, use them. It is the same spreadsheet with the arithmetic done for you.

On the rate: the balance here averages a few thousand dollars across a year, so it is worth putting somewhere that pays. The FDIC's national average for savings accounts was 0.38% in August 2026, while online banks were paying around 3%, and the reasoning is the same as for choosing where to keep an emergency fund. One requirement is different, though, and it matters: sinking-fund money gets withdrawn several times a year on a known schedule, so an account with a withdrawal limit or a two-day transfer delay is a worse fit here than it is for an emergency fund you hope never to touch.

Sinking fund or emergency fund: which is which?

One axis separates them, and it is not the size of the expense.

Could you have seen it coming twelve months out? If yes, it is a sinking fund, however large. A $2,800 roof repair you have known about since the inspection is a sinking fund. If no, it is an emergency fund expense, however small. A $340 windscreen from a stone on the interstate is an emergency, in the only sense that matters here.

The practical consequence is that the emergency fund should be almost boring. Most of what people spend an emergency fund on is predictable, which is why it never recovers, and the three-part test for what actually counts as an emergency exists to settle the borderline cases before you are standing in a garage deciding. If you want the two held side by side with the arithmetic, the comparison is worked out here, and the answer is more interesting than it sounds: the totals come out identical, and the difference is entirely in what you do with a healthy-looking balance in July.

When this does not apply

If your essential spending already exceeds your income. Sinking funds cannot be funded out of a deficit, and the honest sequence is to close the gap first, on housing, income or debt terms. The list is still worth writing, because knowing that your real monthly cost is $603 higher than your budget says is the difference between a plan that might work and one that cannot.

If almost nothing you own can break. A renter with no car, no pets, no dependants and a phone on a monthly contract may genuinely have two funds: gifts and travel. Total them, fund them, and skip the rest. A nine-line table is not a sign of financial maturity.

If you pay insurance monthly and prefer it that way. Monthly premiums usually cost 3% to 8% more than paying annually, and the instalment is a fixed bill rather than a sinking fund. That is a defensible trade if the alternative is not funding it at all. Take the fixed-bill version, name the extra cost, and move on.

If you are mid-crisis. Someone three weeks from a shut-off notice does not need a nine-row table. They need this month solved. Come back to this when the immediate thing is handled, and start with insurance only.

Next steps

1. Export twelve months of statements and list every one-off expense over $150. This is 45 minutes and it is the only step that changes your numbers.
2. Group them into four to six funds, total each, divide by twelve. Fund the largest and most certain one this month, even if you cannot cover them all.
3. Open one savings account and set a standing transfer for the day after payday, then put the six balances in a spreadsheet you will actually open.