Emergency funds: how to size and build one

An emergency fund covers two specific events: your income stops, or a cost arrives that you did not plan for and cannot postpone. A redundancy, a month of illness, a car that fails its inspection, a boiler that dies in January, an insurance excess after an accident. What these have in common is not their size. It is that the bill does not wait for payday.

Without a fund, that gap is filled by whatever credit can be arranged at short notice, and credit arranged at short notice is usually among the most expensive credit available to a household. A card balance, an overdraft, a loan taken in the same week the problem appeared. The fund is not really a savings product at all. It is self-insurance you write yourself, and what it buys is the ability to decline that borrowing.

It also buys time, which is the part people miss. Someone with four months of essentials in the bank is in a position to turn down the first job offer and hold out for a better one. Someone with nothing is under pressure to take the first offer at whatever it pays, and a low starting salary can take years to grow out of. The fund does not only cover the bill. It protects the choices you make while the bill is outstanding.

Two things are not emergency funds. A sinking fund covers costs you already know are coming: tyres wear out, the boiler has a service interval, the car tax falls due every year. Those are a schedule rather than a surprise and they belong in their own pot. General savings for a house deposit or a holiday are not it either, because spending an emergency fund on either is how it stops existing.

The generic answer is three to six months. That range is a convention that got repeated until it sounded official, rather than a figure anyone derived, which is the first reason to treat it as a starting point. The useful answer is three to six months of what, and the base is your essential outgoings, not your income and not your current total spending.

Income is the wrong base because tax and pension contributions come out of it before anything reaches your account, and most of both stops when the pay stops. Not all of it does, and this is where the tidy version misleads. Tax already owed on income you have earned still falls due, self-employed people often have payments scheduled against last year's earnings, and in the United States unemployment benefits are themselves taxable at the federal level. Total spending is the wrong base because a household under pressure cuts fast. What has to be covered is the floor: the payments that keep arriving whether or not you are working.

Add up one month of these.

• Rent or mortgage payment, plus any service charge
• Utilities, local property tax, phone and internet
• Food at the level you would actually buy under pressure, not at your current level
• Transport, including what it costs to get to interviews
• Insurance premiums that would lapse if unpaid
• Minimum payments on every debt, since missing these costs money and, in countries that run credit reporting, marks the file that lenders read later. How credit scores work covers the United States version; the mechanics and the penalty differ elsewhere
• Childcare, medicine and anything else that cannot simply stop

That total is your essentials figure, and it usually comes in well below normal monthly spending, which makes the target less frightening than the generic rule suggests. Most of it is fixed costs, and the debt-to-income calculator gives you the debt-service part of the floor in one step.

Say the figure comes to $2,750 a month. The target is then a multiple of it:

Where you sit in the range is a question about how fast your income could be replaced. What pushes the number up: a single-income household, commission or self-employed pay, a specialised role with a long hiring cycle, dependents, or a health condition that could interrupt work. What pulls it down: two stable incomes, a real notice period, income protection cover already in place, and a skill set that is hired quickly. These are the factors the range responds to rather than a scoring system, and which way any individual case lands is not something a page can work out for you.

Liquidity is how fast something turns into spendable money at a price you can predict. Both halves matter. A holding you can sell today, but only at whatever the market happens to offer, is not liquid in the sense this fund needs, because the day you need it is not a day you chose.

So the money belongs somewhere dull: an instant-access savings account, a high-yield savings account, a money market deposit account at a bank. Same-day or next-day access, a balance that does not move, no fixed term and no exit penalty. A money market fund is a different product despite the near-identical name. It is a fund rather than a deposit, so its price is not promised and no deposit guarantee stands behind it. Where the balance is a deposit it is usually covered by a government guarantee scheme up to a limit, which in the United States means FDIC insurance at banks and NCUA cover at credit unions, with other countries running their own schemes at their own limits. That limit is worth knowing before the balance gets large. Keep it in a separate account from the one your card draws on, so it is not spent by drift, but not so separate that reaching it takes a week and three signatures.

The objection is always the rate. Accounts like that usually yield less than a term deposit, and they carry a lower expected return than a bond fund or an equity fund. Expected is the load-bearing word: that is a long-run average, not something the holding hands you in the year you need it. Bond funds fall when yields rise, equities fall a long way in some years, and in some rate environments cash has out-yielded intermediate bonds outright. So price the objection. Building $11,000 over 18 months from a $2,750 buffer, an account at 4 percent has interest supply $396.29 of the target. At 1 percent it supplies $99.53, and the monthly deposit rises from $436.32 to $452.80, under 4 percent more. Three percentage points of yield across the whole build is worth less than one month's deposit.

Now price the other side. Money you cannot reach for a week is money likely to go onto a card instead. Card rates have generally run well above deposit rates, though the size of that gap moves with the rate cycle rather than sitting still, and a balance run up in one bad week can take a long time to clear. Note what this side of the comparison does not have: a fixed number. The yield you gave up is knowable in advance, the cost of being a week late is not, because it depends entirely on the bill. That asymmetry is the argument, and anyone who states it as a settled sum is overreaching.

Investing the fund carries a sharper problem than volatility on its own. Redundancies have tended to cluster in the same conditions that push asset prices down, so an invested fund can be near its low exactly when it is called on. That is a tendency in the historical record, not a mechanism that has to hold, and the relationship is far from perfect: plenty of individual job losses arrive in calm markets. It still matters here, because this particular pot only has to fail once. That correlation, rather than the average return, is the argument against investing it. Against inflation a cash account will often lose a little ground in real terms, which the real return calculator puts a number on. That small real loss is the premium on the insurance, and it is paid deliberately.

Take a card at 22 percent against a savings account paying 4. Those are the illustrative rates this page runs on, not a reading of the current market, and both move on their own schedules. On those numbers the arithmetic looks settled. Clearing the card returns exactly the rate you stop paying, with no market risk in it, and no cash account competes. Two refinements make it sharper still. A 22 percent card rate that compounds monthly costs a little over 24 percent across a year, so the headline understates it. And in the United States, savings interest is generally taxed as ordinary income while interest you avoid paying is not taxed at all, which widens the gap again; where savings sit inside a tax-free wrapper, it narrows. A strict optimiser puts every spare dollar on the debt and starts the fund afterwards.

The flaw is that the strict optimiser assumes nothing goes wrong during the payoff. With nothing set aside, the first unplanned cost goes straight back onto the card. The balance you spent four months clearing returns in an afternoon and the interest clock restarts on it. The failure being guarded against here is not misreading the rate. It is the balance being refilled faster than it is cleared, which is the whole reason the buffer exists.

So the usual sequence has three steps rather than two.

1. A starter buffer first. Roughly one month of essentials, $2,750 in the numbers above, which at $454.53 a month takes six months and less if you can find it faster. Small enough that the rate gap on it barely registers, large enough to absorb an ordinary surprise. 2. Then the expensive debt. Everything spare goes at any balance whose rate is above what the fund earns after any tax on the interest, highest rate first, while the buffer sits untouched. 3. Then finish the fund. Once the expensive debt is gone, the payment that was clearing it becomes the deposit that completes the target, and it is already a habit.

Two adjustments. Where an employer matches retirement contributions, the match is often taken before either step, because the matched portion is a large immediate return on the money paid in. It is not automatic, though, and the usual telling skips the catches: the money is locked away until retirement, the match may vest over several years and be forfeited if you leave first, and a small match against a very high card rate is a closer call than the slogan suggests. In the United States that is typically a 401(k) match, and other countries have their own arrangements, so check what your own scheme actually offers. And low-rate debt, a mortgage or a subsidised student loan, generally sits below the fund rather than above it. The debt-to-income calculator shows what the minimum payments are doing to your monthly floor while all of this runs.

Make it a standing order on payday into a named account, so the deposit happens before the money is available to be spent. The savings goal calculator at the top of this page turns any target and deadline into that figure. Three routes to the same place, all at 4 percent compounded monthly:

The three deposits sit within about 5 percent of one another. What changes is not the monthly commitment but where it lands, which is the most useful thing this arithmetic shows: pick the deposit you can actually sustain, then read off the deadline it implies, rather than fixing a deadline and discovering the deposit is impossible.

Windfalls compress the schedule far more than the rate does. A tax refund or a bonus paid straight in removes months from the term, because a short goal is almost entirely deposits and hardly any interest, as the $22.83 in the first worked example shows.

Then two habits keep it working.

• Refill it on a schedule. Spending the fund on a real emergency is the fund doing its job, not a failure. The failure is not restarting the standing order. Put the rebuild back into the budget in the same week you spend it.
• Re-price the target when your costs move. The target is a multiple of your essentials figure, and that figure moves when rent, childcare or a loan payment moves. Recheck it once a year.

And stop when it is full. Past the target, more cash buys certainty you have already bought, and that certainty has a cost in forgone return which grows with the balance. Where a surplus is better used is a question about your own goals, horizon and tax position rather than one a page can settle. The compound interest calculator shows what a balance does over a longer horizon at a rate you choose, which is a way to see the size of the trade rather than a recommendation to take it.