The Emergency Fund, Sized Honestly

Three to six months. Every article says it, and not one of them says three to six months of what. On a household taking home $5,240 a month, reading it as income instead of essential spending adds $10,410 to the target.

That is not a rounding difference. It is the difference between a goal you reach in under two years and one you are still short of in five, and which one you picked was decided by an ambiguity in a sentence nobody bothered to finish.

This article produces a number. Two of them, in fact, for two households, because a single worked example in a sizing piece reads as a coincidence.

Three to six months of what, exactly?

There are three candidates, and they are far apart.

Essential spending is the right one, and the reason is what the fund is actually for. It replaces the money you must spend, not the money you happened to be earning. In the month after a job ends, this household does not spend $5,240. It spends $3,505, because it stops eating out, cancels the gym, does not replace the coat and postpones the haircut, all within about a week and without a spreadsheet.

Sizing against take-home pay funds three to six months of your old lifestyle, including the parts you would drop on day one. That is not caution. It is a target 49% too high, and the practical effect is that people look at $31,440, decide it is impossible, and save nothing.

One qualifier, because it changes who this matters to. The gap between the two readings is the size of your slack, so it is large for a household with room and small for one without. Our second household below spends 89% of its take-home pay on essentials, and for them the two readings are only $2,094 apart at six months. The "of what" question saves a comfortable household a lot of money and a stretched household almost none, which is the reverse of how most advice is aimed.

The inputs, and where to find yours

Essential spending is the only input that takes work. Get it from statements rather than estimating, because the estimate is always low, and the error is concentrated in groceries and transport.

The test for each line is one question: in the month after your income stopped, would you still pay this? Not "should I", and not "is this important to me". Would the money leave the account.

In:

• Housing, including escrowed tax and insurance, or rent
• Utilities: electricity, gas, water, refuse
• Groceries, at the level you would actually buy, which is lower than now but not zero
• Phone and internet, because a job search needs both
• Transport you cannot avoid: fuel to interviews, insurance, the minimum maintenance to keep a car legal
• Health insurance premiums and ongoing prescriptions
• Minimum debt payments, because missing these compounds the problem
• Childcare you cannot pause, which for many people is the largest line here
• Basic household and personal care

Out: eating out, subscriptions, holidays, gifts, hobbies, retirement contributions, extra debt payments, and the sinking funds that are not urgent. Note what that last exclusion means: sinking funds are paused during an emergency, not abandoned, so a long unemployment will leave you behind on tyres and the dentist. That is the correct trade and it is worth knowing you are making it.

Here are the two households.

Household A is two salaried earners, in different industries, neither role specialised, on a health plan with a $3,000 family deductible and $1,000 on the car. Household B is one earner paid mostly on commission, in a specialised role with about four employers in the region, renting.

The replaceability multiplier

Three months is the floor. It assumes the thing an emergency fund most often has to survive is an income gap, and that a gap for one earner in an ordinary role is a matter of weeks rather than quarters.

Four adjustments move it up. Add one month for each that applies.

One more adjustment, which is not about income at all. If your health plan carries a deductible you could not pay from a current account, hold that amount as a named part of the fund rather than adding a month. Household A's combined deductibles are $4,000, which is 1.1 months of essentials, so their target becomes four months rather than three. Doing it this way rather than rounding up a month is deliberate: a deductible is a known number, and there is no reason to guess at it.

What the table does not include is anything about age, risk tolerance or how you feel about debt. Those change how quickly you should build the fund and in what order relative to investing. They do not change how many months of essentials it takes to survive a job ending, which is arithmetic.

How long does an income gap actually last?

The multiplier is only defensible if the months mean something, so here is the figure it rests on. In August 2026 the Bureau of Labor Statistics counted 1.9 million people who had been out of work for 27 weeks or more, and that group was 27.0% of everyone unemployed.

Twenty-seven weeks is 6.2 months. So roughly one unemployed person in four is still looking after the point at which a six-month fund has run out, and that is the strongest argument available for the upper half of the range.

It is also the argument against treating the number as a probability. Three quarters of people are back inside 27 weeks, which is why three months is a defensible floor for a household with a second income continuing: the realistic worst case for them is not one person unemployed for six months, it is one person unemployed for three while the other keeps earning. The fund has to cover the gap, not the whole salary.

For a single-income household, the same statistic reads completely differently. There is no second wage, so 27 weeks of no income means 27 weeks of essentials paid entirely from cash, and a one-in-four chance of exceeding that is not a risk you can reasonably run at three months. That asymmetry is the first row of the multiplier table, and it is the single largest adjustment in it.

The calculation, worked at two incomes

Look at the bottom row, because it is the point of running this twice. The standard advice is nearly right for A at its low end and wildly over at its high end. For B it is wrong in the other direction: every figure in the usual range is below what B actually needs, and the top of it is $737 short.

B earns $24,720 a year less than A and needs $5,797 more in cash. That is not a paradox, it is the whole finding. The size of an emergency fund is set by how long the income takes to replace and how little of it is discretionary, and both of those are worse for the household with less money. Which leads somewhere uncomfortable.

What if the number is out of reach?

Household B can save about $180 a month. At that rate $19,817 takes 110 months, which is a little over nine years, and telling B to aim for it as a single goal is useless advice delivered with a straight face.

So do not. Build it in three stages, and treat each stage as finished rather than partial.

For B, stage one is $2,831 and takes sixteen months at $180. That is the goal. Stage three is a number to know rather than a number to chase this year, and the honest advice for B is that the fund is the second priority behind raising the income or reducing the rent, because at 89% of take-home going on essentials, no savings rate fixes the underlying shape.

Stage one deserves the emphasis it gets here. The Federal Reserve's 2025 household survey found that 63% of adults could cover a $400 emergency expense with cash or its equivalent, and 12% said they could not pay it by any means. Almost all of the distance between those two groups is the first month of essentials, not the fourth.

Where should it sit?

Two requirements, in order: you can reach it within about a day, and it cannot fall in value. Everything else, including the interest rate, is a distant third.

That rules out more than people expect. Not a CD, because the penalty for breaking one can exceed the interest it paid. Not an index fund, because the month you need it is disproportionately likely to be a month the market is down. Not the account you spend from, for reasons that have nothing to do with finance.

A savings account at an FDIC-insured bank or an NCUA-insured credit union, in your own name, with transfers you have tested once. The rate is still worth ten minutes: the FDIC's national average for savings was 0.38% in August 2026 while online banks were advertising around 3%, and on $15,000 that gap is $393 a year. The full comparison, including what happens when you break a CD early, is in where to keep your emergency fund.

Keep it apart from your sinking funds, in a different account, even though the arithmetic of combining them is identical. The reason is not accounting, and it is the most interesting thing in this cluster: a single pot with $8,000 in it in July gets spent in July, and two pots with the same $8,000 do not.

When this does not apply

If you are carrying credit card debt above about 20%. Build stage one, take any employer match, then put everything else at the card until it is gone. A $3,505 cash cushion earning 3% while $9,000 sits at 22.4% is a guaranteed loss of about $1,700 a year, and the cushion is only worth that price for the first month of essentials, which is the one that stops the balance growing again.

If your income is irregular. You need two separate things and most advice names only one. An income buffer smooths the gap between a slow month and a normal one, and an emergency fund covers events. Build the buffer first, at three months of the salary you pay yourself, and read how to budget when your income changes every month for how the two sit together.

If you have a paid-off house and no dependants. Your essential spending may be low enough that four months of it is a small number, and the binding constraint is a medical event rather than an income gap. Size against your deductibles and your out-of-pocket maximum instead, and the months multiplier stops being the useful frame.

If you are within a year of a known large expense. A baby, a move, a roof you have been told about. That money is a sinking fund with a date on it, and holding it inside the emergency fund guarantees you will spend the emergency fund and call it something else. The three-part test settles which is which.

If you are currently unemployed. None of the above applies today. Cut to essentials, claim everything you are entitled to, and do not put a dollar into savings while the runway is being used. Sizing is for before and after.

Next steps

1. Open three months of statements and write down your essential lines using the "would the money still leave" test. Forty minutes, and it is the only figure here you cannot guess.
2. Multiply by three, then add a month for each of the four conditions that applies to you. Write the number down, including the deductibles.
3. Set stage one as this year's goal: one month of essentials. Put a standing transfer in place the day after payday and leave it alone.
4. Once stage one is funded, take your full employer retirement match before continuing to stage two. It is the one return that beats holding cash, and what to put the money in once you get there is a separate half hour.