How Much Should You Really Keep in an Emergency Fund?

The first time I sat down with a client who had no emergency fund — none, zero, not even a couple hundred dollars in a savings account — was in the spring of 2009. He was a software engineer in his early thirties, made decent money, drove a paid-off Honda. He'd just been laid off. His severance was three weeks. His rent was due in eleven days.

I tell that story because the standard advice you read everywhere — "save three to six months of expenses" — sounds reasonable until you're sitting across from someone whose three months disappeared in three weeks. Or, on the other end, until you're talking to a recently retired couple sitting on $80,000 in cash earning nothing, terrified to spend a dollar of it.

The honest answer to "how much should I keep in an emergency fund?" is: it depends on you, and most of the financial press has been giving the same advice for forty years without updating it.

Let me try to do better.

Where the "three to six months" rule actually came from

If you trace this number back, it doesn't come from a study. It comes from personal-finance columnists in the 1970s and 1980s, who picked it because it sounded prudent and was easy to remember. (Sylvia Porter, the columnist who probably did more to shape American household finance than anyone else, used "three months" in her 1975 Money Book. Six months crept in later, around the early-1980s recession, when people stayed unemployed longer.)

It's not a bad rule. But it's a starting point, not a destination.

The Federal Reserve's Report on the Economic Well-Being of U.S. Households — the SHED survey — has been asking Americans for over a decade whether they could cover a $400 emergency. As of the most recent reports, somewhere around 30-37% say they couldn't, or would have to borrow. Whatever number you settle on, you want to be on the right side of that statistic.

The variables that actually matter

Forget "three to six months" for a minute. Here's what I'd actually want to know if you walked into my office:

1. How stable is your income?

A tenured professor and a freelance graphic designer should not be running the same emergency fund. The professor could probably get away with three months. The designer — whose income might swing 40% from one quarter to the next — should be looking at nine to twelve.

2. Do you have one income or two?

If a household has two earners in different industries, the odds of both losing work simultaneously are much lower than for a single-earner household. (This is one of the reasons two-income households recovered faster from the 2008 crisis, on average, than single-income ones — see the work of Elizabeth Warren and Amelia Warren Tyagi from that period.)

3. How specialized is your job?

A nurse can find work in any city in three weeks. A senior product manager at a niche SaaS company might take seven months to find an equivalent role. The more specialized — and the higher-paid — you are, the longer your runway needs to be.

4. What are your fixed costs versus your discretionary costs?

This is the one almost no one calculates correctly. Your "monthly expenses" for emergency-fund purposes are not what you spend now. They're what you'd spend if you were unemployed and being careful. Strip out restaurants, vacations, the second streaming service, the gym you keep promising to use. The number is usually 25-40% smaller than what people think.

5. Do you have other safety nets?

A working spouse. A line of credit you've already opened (you can't open one when you need it). A family member who'd float you a few thousand. None of these replace cash, but they do change the math.

A more honest framework

Here's what I've actually used with clients for the last fifteen years. Pick the row that fits and that's roughly your number.

Situation	Months to keep
Two stable incomes, both employees, low specialization	3
One stable income, employee, low specialization	4-5
One income, employee, high specialization or high salary	6-9
Self-employed or commission-based, household	6-9
Self-employed, single-earner household	9-12
Retired, drawing from a portfolio	12-24 (in cash + short bonds)

The retiree number surprises people. It shouldn't. The single biggest risk to a retiree's portfolio is being forced to sell stocks during a downturn to cover living expenses — what advisors call "sequence-of-returns risk." A larger cash buffer doesn't earn much, but it lets you ride out a bad market without selling at the bottom. Wade Pfau, the retirement researcher at The American College, has written extensively about this; his research on "bucketing" strategies is worth reading if you're within ten years of retirement.

Where to actually keep it

Not in your checking account. (You'll spend it. Everyone spends it. I have spent it.)

Not in stocks. (Emergency funds and the S&P 500 do not go together. The whole point is that the money is there when you need it, and what you need it for is usually correlated with the times stocks are dropping.)

A high-yield savings account at an online bank is the obvious answer for most people. As of late 2024, the better online banks — Marcus, Ally, Discover, Capital One 360, a handful of others — are paying somewhere in the 4-5% range, give or take whatever the Fed has done lately. Money market funds at brokerages like Fidelity or Vanguard often pay slightly more, with slightly more friction to access.

For amounts above, say, $25,000, I'd seriously consider splitting the money: half in a high-yield savings account for instant access, half in a 4-week or 8-week Treasury bill ladder. The yield bump is small but real, and T-bills are about as safe as anything denominated in dollars can be.

What I tell people not to do

Don't invest your emergency fund "to make it work harder." I know, the opportunity cost feels painful. It's supposed to. The cash isn't there to grow your wealth. It's there so that when something goes wrong — and something always goes wrong — you don't have to call your 401(k) provider in tears.

Don't count your Roth IRA as an emergency fund. Yes, technically you can withdraw your contributions penalty-free. Yes, financial bloggers love pointing this out. The problem is that once it's out, it's almost impossible to put back (you're capped by annual contribution limits), and you've permanently lost decades of tax-free compounding. I've watched people do this, and I've never seen them re-fund the account. Not once.

Don't skip the emergency fund to pay off low-rate debt faster. A 4% mortgage and a fully-funded emergency account is a much better position than a 3.5% mortgage and a panic. (High-rate credit-card debt is a different conversation; there I'd at least split the difference.)

How to actually get there if you have nothing

Start with $1,000. That number — popularized by Dave Ramsey, whose investment advice I disagree with on most days but whose behavioral framework is genuinely useful — handles the majority of household emergencies that don't involve job loss. A car repair. An ER co-pay. A surprise vet bill.

Get to $1,000 first. Then, while you're paying down high-rate debt, slow your savings to a trickle but don't stop. Once the credit cards are gone, redirect those payments into the emergency account until you hit the number that fits your situation from the table above.

Most households I've worked with build a real emergency fund in 14 to 22 months, not the 6 months that articles love to imply. That's fine. The point isn't speed. The point is that, eventually, you become someone for whom a $400 surprise is annoying instead of catastrophic. That single shift changes more about a financial life than almost anything else I can name.

A final thought

I used to be more dogmatic about this. In my twenties, I'd quote the three-to-six-months rule with a straight face. Now, after watching enough lives play out, I'm convinced the bigger emergency fund is almost always the right call — even at the cost of a couple of points of long-run return.

You will never, in your life, regret having too much cash on the day you needed cash. I have watched many people regret the opposite.