What Is a Sinking Fund? A Monthly Set-Aside for Bills You Already Know Are Coming

Short answer: a sinking fund is money set aside in regular instalments for a specific expense that is known to be coming but does not arrive monthly: a car repair, an insurance premium, holiday gifts, a replacement phone. The sum is the annual cost divided by 12. The reason it is worth the bother is that these costs are irregular one at a time and predictable in total: the average US household spent $984 on vehicle maintenance and repairs and $1,993 on vehicle insurance in 2024, and 37% of adults told the Federal Reserve they would not cover a $400 surprise with cash. A sinking fund is not an emergency fund, and it is the wrong priority while high-interest debt is outstanding.

The idea has no trial of its own. What follows separates what is measured from what is sensible arithmetic.

What a sinking fund is, and where the name comes from

The phrase is not a personal-finance invention. In the bond market, a sinking fund provision obliges an issuer to deal with a debt gradually. The US Securities and Exchange Commission's investor glossary describes a sinking fund redemption as one that "requires the issuer to regularly redeem a fixed portion or all of the bonds in accordance with a fixed schedule." The debt is retired a piece at a time, not all at once in the final year.

The household version copies the logic and drops the legal structure. Nothing enforces it. There is no trustee and no schedule except the one the household writes down.

Why irregular expenses are more predictable than they feel

A car does not announce the month in which it will need brakes. But across a year, and across millions of households, what cars cost to keep on the road is a fairly stable figure. The Bureau of Labor Statistics' Consumer Expenditure Survey records what American households ("consumer units" in the survey's language) spend, averaged over all of them. The 2024 figures for the lumpier categories, the first, fourth and fifth rows taken from the St. Louis Fed's FRED copy of the survey series:

The monthly column, the five-category sum and the percentage below are this article's arithmetic, not BLS figures. Three cautions belong next to the table. These are averages across every household, including those with no car and those who rent, so the owned-dwelling line is diluted by renters who spend nothing on it and understates what an owner faces. Averages hide enormous variation: a household can spend nothing on repairs for two years and $2,500 in the third. And some of these lines are often already paid monthly, such as insurance on a payment plan.

Even so, five categories that tend to arrive in lumps add up to about 13% of average household spending. A budget built only from the bills that come every month is missing roughly an eighth of the year.

The Federal Reserve's annual household survey shows what happens when the lump lands on a budget that did not allow for it. In the survey covering 2024:

• 63% of adults said they would cover a hypothetical $400 emergency expense with cash or its equivalent. That leaves 37% who would borrow, sell something or not pay.
• 13% said they could not pay the $400 by any means.
• 55% had set aside three months of expenses in an emergency or rainy-day fund.
• 23% had a major, unexpected medical expense in the prior 12 months, with the median amount between $1,000 and $1,999.
• 17% did not pay all their bills in full in the month before the survey.

Some of what these figures capture is not an emergency at all: it is a known cost with an unknown date, met by a budget with no line for it.

Sinking fund vs emergency fund

The Consumer Financial Protection Bureau defines an emergency fund as "a cash reserve that's specifically set aside for unplanned expenses or financial emergencies", and gives car repairs, home repairs, medical bills and loss of income as common examples. It describes such costs as ones that are not part of routine monthly expenses, and suggests each household set its own guidelines for what counts.

The CFPB guide does not discuss predictable, non-monthly expenses as a separate category. The sinking fund idea fills that gap by reclassifying some of the CFPB's own examples. A car repair on a ten-year-old car is not really unplanned. Its date is unknown and its likelihood over a year is high.

The practical argument for keeping them apart is that earmarked money counted as a cushion overstates the cushion. A single savings balance of $3,000 looks like a reasonable buffer. If $1,200 of it is next spring's insurance premium and $600 is the holidays, the buffer is $1,200. Separate labels do nothing to the total. They make the total honest.

How much to put in: the worked table

There are only two formulas. For a recurring cost, divide the annual amount by 12. For a one-off with a deadline, divide the target by the number of months left. Costs that recur less than once a year are annualised first: a $900 phone replaced every three years is $300 a year.

The example below is illustrative: round figures chosen to show the method, not survey data and not a recommendation.

The $3,480 was always going to be spent; the only question was whether it appeared as six shocks or one flat line of $290 a month. The estimates are the weak part. The insurance figure can be read off a renewal notice. The repair figure is a guess, and the honest way to make it is to add up what was actually spent over the last two or three years and divide.

Starting late changes the monthly figure sharply. A $600 December target is $50 a month from January (12 months), $100 from June (6 months) and $200 from September (3 months).

Where the money for the contributions comes from is a separate question. Recurring charges that nobody is using are the usual first place to look, and a subscription audit is a one-off job. Annual fees belong in the table too: whether a warehouse club membership pays for itself is a different calculation, but the fee is a known annual cost either way. The larger question, how much of a paycheck to save in total, sits above all of this; sinking funds are deferred spending more than saving.

Why labelling the money seems to help

Money is fungible. A dollar in a pot called "car" is identical to a dollar in a pot called "everything". A strictly rational household would ignore the labels. The behavioural literature says households do not, and that this can be put to use.

Thaler's review describes mental accounting as the set of cognitive operations people and households use to organise, evaluate and keep track of financial activities, and argues that it matters because it violates fungibility: money in one mental account is not a perfect substitute for money in another. A sinking fund is a deliberate use of a habit people already have.

Soman and Cheema's field experiment is the closest test of earmarking itself. The workers earned 670 rupees a week and had a mean savings rate of 0.75% in the six months before the study. Each week a social worker set aside the target amount in sealed, labelled envelopes. Dividing the same sum across two envelopes rather than one raised total savings by about 72% (414 against 241 rupees), which the authors link to partitioning: opening a second envelope is a second decision. A photograph of the saver's children on the envelope also helped. The more ambitious target backfired when the money was pooled: 211 rupees saved against 269 for the low target. An envelope was opened in 85% of household-weeks under the high target, against 59% under the low one.

What these studies do not show

None of them tested the thing described in this article. The earmarking experiment involved very poor households in rural India, cash in physical envelopes, and a weekly visit from a social worker over a few months. Nothing was locked: the envelopes could be opened at any time, as a sub-account can. Thaler's paper is a synthesis, not a trial. No randomised study was found for this article showing that US households using labelled sub-accounts for irregular bills end up with less debt or more savings than those who keep one pot.

The fair summary is that the mechanism is plausible and consistent across very different settings: labels create a small psychological cost to raiding the money, and partitions create decision points. Whether that survives the move to a banking app, where transferring between pots takes two taps, is assumed and not measured. Why a monthly transfer sticks or does not is better explained by the research on how long habits take to form and break than by any budgeting system.

Sinking fund categories: how many, and which

The Soman and Cheema result argues for some partitioning, since one undifferentiated pot was the worst-performing arrangement. It does not argue for twenty. Their high-target finding is a warning: a scheme that asks for more than the household can keep up gets broken.

No study sets an optimum number. A reasonable filter is that a category deserves its own fund only if it passes three tests: the cost is larger than a normal month can absorb, it is likely within the next year or two, and the amount can be estimated from a bill or from past spending.

Three to six funds cover most households. Small or erratic items, and annual renewals, can share one fund called "irregular". Running costs that arrive monthly, such as electricity, do not need a fund; they need an accurate figure, and what appliances cost to run is a budgeting input, not a savings goal.

Where to keep sinking funds

The CFPB's test for emergency savings applies just as well here: the money should be "safe, accessible, and in a place where you're not tempted to spend it on non-emergencies." It calls a bank or credit union account "generally considered one of the safest places to put your money", and notes that cash kept at home "can be stolen, lost, or destroyed."

Money that will be spent within a year has no business being exposed to market risk. A fund for a premium due in April cannot wait for shares to recover. That points to deposit accounts, and the relevant protection is deposit insurance. The FDIC's standard is $250,000 per depositor, per insured bank, for each account ownership category. Checking accounts, savings accounts, money market deposit accounts and certificates of deposit are covered. Stocks, bonds, mutual funds, crypto assets and annuities are not.

Because the limit is counted per depositor and ownership category at each bank, labelled pots held by one person in one bank would share one limit. For most households that is far above any sinking fund balance.

Sinking fund vs savings account is therefore a slightly confused comparison. A savings account is a place. A sinking fund is a purpose. Most sinking funds are kept in a savings account, either as one account with a spreadsheet beside it or as several labelled accounts. The partitioning evidence, such as it is, favours visible separation.

What is not worth doing

Building sinking funds while carrying high-interest debt. This is the main opportunity cost and it is a matter of arithmetic. With illustrative rates, $3,000 held in savings at 4% earns about $120 in a year, while $3,000 owed on a credit card at 22% costs about $660. Holding both at once costs the difference. With no buffer at all, the next car repair goes straight back on the card, which is a fair reason to keep a small general cushion while paying debt down. It is not a reason to fund next year's holiday first.

A fund for every line of the budget. Fifteen pots with $20 each mean constantly moving money between pots to cover shortfalls, which teaches that the labels mean nothing.

False precision or chasing yield. The repair estimate will be wrong; a fund that is 30% short still turns a $900 shock into a $270 one. And the return on a few thousand dollars held for a few months is small at any realistic rate, so safety and access matter more.

What to actually do

List last year's non-monthly spending. Twelve months of bank and card statements, every charge that was not monthly and was larger than the household would absorb without noticing.

Pick three to six categories using the three tests: large, likely, estimable. Fold the rest into one general irregular fund.

Divide. Annual cost by 12, or target by months remaining. Round up.

Check the total against the budget. If the monthly sum is not affordable, the finding is not that sinking funds do not work. It is that the year's spending exceeds the year's income by that amount, and something on the list has to shrink.

Automate one transfer on payday into a separate, insured deposit account. One transfer and a simple record of how it splits is easier to maintain than six transfers.

Spend it when the bill comes. The fund is meant to hit zero.

Review once a year, when the largest renewal arrives, and reset the figures from actual spending.

Questions people ask

What is a sinking fund? A sinking fund is money saved in regular instalments for a specific, expected expense that does not occur monthly, such as an annual insurance premium, car repairs or holiday gifts. The term comes from bond finance, where issuers set money aside or redeem debt on a fixed schedule instead of paying it all at once.

What is the difference between a sinking fund and an emergency fund? An emergency fund is a cash reserve for unplanned events such as job loss or a medical bill, and ideally is never used. A sinking fund is for costs that are expected, with only the timing or exact amount uncertain, and is meant to be spent down and refilled each cycle.

What are common sinking fund categories? Insurance premiums paid annually or twice a year, car maintenance and repairs, home maintenance, gifts and holidays, out-of-pocket medical and dental costs, and device and appliance replacement. A category is worth a fund if the cost is large, likely and estimable.

How many sinking funds should I have? No study sets a number. Three to six covers most households. One field experiment in rural India found that splitting earmarked savings into two parts worked better than one pot, and that a higher target led to envelopes being opened more often, which argues against a long list of small funds.

How much should I put in a sinking fund? The expected annual cost divided by 12, or the target amount divided by the months remaining. A $1,200 annual insurance bill is $100 a month. A $900 phone replaced every three years is $300 a year, or $25 a month.

Where should I keep sinking funds? In a deposit account that is safe, accessible and separate from everyday spending. FDIC insurance covers $250,000 per depositor, per insured bank, per ownership category, and applies to checking, savings, money market deposit accounts and certificates of deposit, not to stocks, funds or crypto assets.

What are some sinking fund examples? A $600 holiday budget saved at $50 a month from January. Car insurance of $1,200 a year saved at $100 a month. A $900 annual estimate for car repairs saved at $75 a month.

Is a sinking fund worth it? For households with irregular bills and no high-interest debt, it is a low-cost way to stop predictable expenses landing as shocks. The supporting evidence is indirect: studies of mental accounting and earmarked savings in other settings. For someone paying credit card interest, clearing that balance usually saves more.

What is the difference between a sinking fund and a savings account? A savings account is a product a bank provides; a sinking fund is a purpose assigned to money. Most sinking funds sit in a savings account, either as separate labelled accounts or as one account with the split recorded elsewhere.

This article explains a budgeting method for general information and is not financial advice. Spending averages are national figures that will not match any one household, the worked examples use illustrative amounts and interest rates, and deposit insurance rules and account terms should be checked with the FDIC and the institution concerned before relying on them.