How Much Emergency Fund Do You Actually Need?
Three to six months of what, exactly? The answer changes the target by tens of thousands of dollars.
Say your role is eliminated on a Tuesday in March. Your final paycheck lands two weeks later, and there is no severance because the company is cutting costs, not restructuring generously. Your health insurance ends on the last day of the month. Your mortgage, your car payment, your utilities and your groceries do not care about any of this. The question that matters now is not "do I have an emergency fund" — it is "how many of these Tuesdays can I absorb before something breaks?"
The standard answer is three to six months. That range is so wide it is barely advice. Six months is double three months, and for most households the difference between the two is tens of thousands of dollars sitting in a savings account instead of a brokerage account. Worse, the rule almost never specifies three to six months of what, and that ambiguity alone can swing the target by a factor of two.
Three to six months of expenses is not three to six months of income
Take a household earning $95,000 gross. After federal and state withholding, payroll taxes, and a retirement contribution, take-home might be around $71,250 a year, or roughly $5,940 a month. Now separate the spending: the non-negotiable monthly outflow — housing, utilities, insurance, minimum debt payments, groceries, transportation, childcare — comes to $4,100. The rest is discretionary.
Six months sized on gross income is $47,500. Six months sized on core expenses is $24,600. Same household, same rule of thumb, and the target differs by nearly $23,000. That is not a rounding error; it is the difference between an achievable goal and one that takes three extra years to reach.
Expenses are the correct base, because expenses are what the fund actually pays. Income-based sizing quietly builds in a buffer for taxes you will not owe on income you are not earning and for discretionary spending you would cut in week one. Use the Savings Goal Calculator with your real core-expense number, not a percentage of your salary.
There is one honest exception. If your spending is unusually inelastic — a fixed private school tuition, an expensive medical regimen, a house you cannot leave — then "core expenses" is most of your income anyway and the two methods converge. And if your income is variable rather than salaried, the relevant question shifts from "how long could I be unemployed" to "how bad can a slow quarter get," which is a different calculation entirely.
The number the median hides
Federal data on how long unemployment actually lasts is more useful here than any rule of thumb, and it points in an uncomfortable direction. In the Bureau of Labor Statistics household survey for August 2026, the median duration of unemployment was 11.4 weeks. The mean was 26.3 weeks — more than double.
A mean that is 2.3 times the median tells you the distribution has a long right tail. Most job searches end reasonably quickly. A meaningful minority run six months, nine months, longer. The same release put the number of people unemployed 27 weeks or more at 1.9 million.
Here is the part that changes how you should size the fund: an emergency fund is not there for the median outcome. The median outcome is survivable almost by definition — three months of expenses covers it, and if you land at week eleven you were never in real danger. The fund exists for the tail. Sizing to the median is like buying a seatbelt rated for the average collision. The number you want is the one that covers a bad draw, not a typical one.
What actually moves your number up or down
Rather than picking a point in the three-to-six range at random, adjust from a base of six months of core expenses using the factors that genuinely change your exposure:
| Factor | Pushes the target | Why |
|---|---|---|
| Two incomes at different employers, different industries | Down | Both incomes stopping at once is a much lower-probability event |
| Two incomes at the same employer, or in the same sector | Up | The incomes are correlated; a downturn hits both. This is the case people most often get wrong |
| Single income, dependents | Up | No second income to fall back on, and expenses are less compressible |
| Senior, specialized, or highly paid role | Up | Fewer open positions at your level; senior searches routinely run longer |
| Commission, freelance, or seasonal income | Up | You are absorbing normal volatility, not just job loss |
| Homeowner rather than renter | Up | You absorb the repairs a landlord would otherwise cover, and you cannot downsize on 30 days' notice |
| Employer-sponsored health coverage | Up | Continuation coverage after job loss typically costs far more than the payroll deduction you are used to |
| Stable public-sector or tenured position | Down | Lower probability of abrupt termination |
A dual-income household where both people work at the same company should be thinking in terms of nine to twelve months, not three. A single-earner household with a specialized role and a mortgage should also. A renter with two uncorrelated incomes, no dependents, and portable skills can defensibly sit at three.
Where to keep it, and where not to
The requirements are narrow: the money has to be available within a day or two, and it has to be worth what you think it is worth on the day you need it. That rules out anything with market risk, because emergencies correlate with recessions and recessions correlate with drawdowns — the day you get laid off is disproportionately likely to be a day your portfolio is down.
A high-yield savings or money market deposit account at an insured institution satisfies both requirements. The FDIC standard insurance amount is $250,000 per depositor, per insured bank, for each account ownership category, which is well above any reasonable emergency fund, so insurance limits are rarely the binding constraint. The Consumer Financial Protection Bureau's guidance suggests keeping the money accessible but separate from your day-to-day spending account, which is a small behavioral point that matters more than it sounds: a fund sitting in your checking balance gets spent.
Two things that are not emergency funds: an unused credit card, and a home equity line of credit. Both are borrowing capacity, both can be reduced or frozen by the lender precisely when conditions deteriorate, and both convert a cash-flow problem into a debt problem.
The opportunity cost is real. Price it honestly.
Holding cash costs something, and articles that pretend otherwise are not doing you a favor. Take $30,000 held for ten years. At 4% in a savings account it grows to $44,407. At a hypothetical 7% invested, it would reach $59,015 — a gap of $14,607. And savings interest is taxable as ordinary income each year, so at a 22% marginal rate the effective return is closer to 3.12% and the balance lands near $40,790. Call the real ten-year cost of that cash roughly $18,000 in foregone growth.
Now price the other side. Suppose you have no fund and need $20,000 during a period when the market is down 30%. Liquidating $20,000 of a portfolio at that level means selling holdings that were worth $28,571 at the prior peak — you have permanently converted a paper loss into a realized one, and you have removed those shares from the recovery. The alternative, putting $20,000 on a card at 22.99%, costs about $774 a month over three years and $7,867 in interest. Do that twice in a decade and the interest alone comes to $15,734 — close to the $18,000 of foregone growth, and that is before counting the realized loss on anything you also had to sell, or the months of carrying a payment you did not plan for.
The useful reframe: an emergency fund is not an investment that underperforms. It is insurance, and the spread between the savings rate and the expected market return is the premium. Insurance you never claim on is not wasted money. But it is also not free, which is the argument against the opposite error.
You can overfund it
Twenty-four months of expenses in cash, held by someone who is not contributing to a retirement account and has no employer match captured, is a mistake in the other direction — and it is a common one, because cash feels safe and the cost of holding it is invisible. Over a thirty-year horizon that premium compounds into a genuinely large number; run it through the Compound Interest Calculator and the shape becomes obvious. Past the point where you can survive a realistic worst case, additional cash is buying you very little additional protection at a steadily rising price.
The sequence most people should follow is unglamorous: build a small starter buffer of a month or so, capture any full employer retirement match (which is an immediate return no savings account can compete with), clear high-rate revolving debt, then finish funding the emergency reserve to your adjusted target, then invest past it. None of this is personalized advice, and none of it accounts for your tax situation or your job security. But it does replace "three to six months" with a number you can actually defend.
Related tools
- Savings Goal Calculator — work out the monthly contribution needed to hit your target by a date
- Compound Interest Calculator — see what cash held long-term costs in foregone growth
- Loan Payment Calculator — price the borrowing you would fall back on without a fund
Sources and further reading
Figures and definitions on this page are drawn from the following primary sources. If you find something out of date, tell us and we will correct it.