Everyone agrees you should have an emergency fund. Almost nobody tells you where to put it, so most people leave it in a checking account earning nothing, or they get talked into investing it and discover the problem the first time the market is down twenty percent in the same month their car dies.
The place you keep it matters more than the advice suggests, and the gap between the worst reasonable option and the best one is several hundred dollars a year on a normal-sized fund.
What the fund is actually for #
The purpose is narrow and worth stating, because it determines everything else.
An emergency fund exists so that an unexpected expense does not become debt. A car repair, a medical bill, a broken laptop you need for work, or the big one, a job loss. Without a buffer, those go onto a credit card at 20 to 27 percent, and a $2,000 repair becomes a $2,600 repair paid over eighteen months.
That single sentence sets the requirements. The money has to be available in days, not weeks. It has to be worth roughly what you put in, regardless of what happened in the market last Tuesday. And after those two constraints are satisfied, it should earn whatever it can.
Yield is the third priority, not the first. Any option that compromises the first two to get a better rate has misunderstood the job.
How much, briefly #
Three to six months of expenses is the standard answer and it is a reasonable one. Note that it is expenses, not income, which for most people is a meaningfully smaller number.
The right end of that range depends on how replaceable your income is. Two earners in stable fields can sit at three months. A single earner in a volatile industry, a contractor, or anyone with variable income should be at six or more. In San Francisco, where the cost of a bad month is high and tech layoffs arrive in waves, I lean toward six.
Start with one month, or even $1,000, before you do anything else with your money. That first thousand dollars is what stops the small emergencies from becoming credit card balances, and it comes before investing and before extra debt payments in the standard order of operations.
The options, compared #
A checking account #
Yield: essentially zero, often literally zero.
This is where most emergency funds live, and it is the single most expensive choice on this list. $15,000 sitting in checking at 0.01% earns $1.50 a year. The same money at 4% earns $600.
That is a six hundred dollar annual difference for a decision that takes fifteen minutes to change and costs you nothing in convenience. It is probably the highest hourly rate available anywhere in personal finance.
A high-yield savings account #
Yield: tracks short-term rates, and has recently been in the range of 3.5% to 5%.
This is the default answer and it is a good one. Online banks pay meaningfully more than brick-and- mortar banks because they have no branches to fund. Deposits are FDIC insured up to the standard limit, currently $250,000 per depositor per bank, which for an emergency fund is not a binding constraint.
Liquidity is same day to two business days, which is fast enough for essentially any emergency. You can generally link it to your checking account and pull money with a transfer.
The catch worth knowing: the rate is not a promise. High-yield savings rates float with the short-term rate environment, and the same account that paid 5% can pay 1.5% eighteen months later without asking you. Some banks also run teaser rates that drop after an introductory period, and others quietly stop being competitive and count on you not noticing. Check your rate against current market rates once a year.
Interest is taxed as ordinary income at both federal and state level.
Money market funds #
Yield: similar to or slightly better than high-yield savings, because they hold short-term government and corporate paper directly.
Available inside a brokerage account. These are not FDIC insured, which sounds alarming and is mostly a technicality for government money market funds, which hold Treasury securities and repurchase agreements backed by them. The risk is not zero, and it is small enough that the yield difference usually wins.
Liquidity is one to two business days to sell and transfer.
A government money market fund holding mostly Treasuries has a useful tax property: the portion of its income derived from Treasury obligations is generally exempt from state income tax. In California, with a 9.3% or higher marginal state rate, that is worth real money. The fund publishes the Treasury percentage each year for exactly this purpose.
Treasury bills #
Yield: the market rate on short-term government debt, and typically the highest of these options.
T-bills are short-term US government debt with maturities from 4 weeks to 52 weeks. You buy at a discount and receive face value at maturity, and the difference is your return. They are backed by the US government, which is the reference point for credit risk rather than a thing you evaluate against it.
The tax treatment is the part people miss. Treasury interest is exempt from state and local income tax. In California that is worth 9.3% to 13.3% of the interest, which means a T-bill at 4.3% can beat a savings account at 4.6% on an after-tax basis for a high earner. If you live in a state with no income tax, this advantage disappears entirely.
You can buy them directly through TreasuryDirect, or more conveniently through a brokerage. Liquidity is good, since there is a deep secondary market and you can sell before maturity, though you take whatever price the market gives you that day.
The friction is real: they mature, and you have to do something with the money. Which leads to the next option.
A T-bill or CD ladder #
The ladder solves the maturity problem. Split the fund into four or six pieces and buy bills or CDs maturing at staggered intervals, say every three months. Something matures regularly, so you always have cash coming available, and when nothing is needed you roll it into a new one at the back of the ladder.
You get most of the yield of a longer maturity with most of the liquidity of a short one.
The honest assessment: this is more machinery than most emergency funds need. The yield difference between a ladder and a plain high-yield savings account is usually well under one percentage point, which on a $20,000 fund is under $200 a year. If you enjoy this sort of thing, it is a reasonable hobby. If you do not, the savings account is fine and you should spend the attention elsewhere.
I bonds #
Yield: an inflation-linked rate that resets twice a year.
Worth mentioning and mostly not the answer, because of the liquidity rules. You cannot redeem an I bond at all in the first twelve months, and redeeming before five years forfeits three months of interest. A fund you cannot touch for a year is not an emergency fund.
They are a reasonable place for the tier of savings just behind the emergency fund, and there is an annual purchase limit per person that keeps the amounts modest anyway.
What not to do #
Do not put your emergency fund in stocks. The correlation is the problem. Recessions cause both layoffs and market declines, so the scenario where you need the money is disproportionately the scenario where it is down. Selling equities at a 30% loss to cover rent is the exact failure this fund exists to prevent.
Do not put it in crypto. Same argument, with more volatility and no deposit insurance.
Do not count a credit card as an emergency fund. A card is a way to pay for the emergency at 24% interest, and cards get closed or limits get cut precisely when your finances deteriorate, which is when you would need it.
Do not count your Roth IRA contributions as your only emergency fund. You can withdraw Roth contributions penalty-free at any time, which is a genuinely useful property, and money pulled out of a retirement account cannot be put back beyond the annual limit. It is a legitimate backstop behind a real emergency fund. It is a bad primary one.
Where I actually keep mine #
A high-yield savings account at an online bank, linked to my checking account, holding about six months of expenses. Boring, liquid in a day, and I do not think about it.
I have run the T-bill version and the after-tax difference in California was real but small, maybe a hundred and fifty dollars a year on my balance, in exchange for maturity dates I had to manage. I went back to the savings account and spent the attention on things with bigger numbers attached.
That is the honest ranking of effort against payoff. Moving from checking to high-yield savings is worth several hundred dollars a year for fifteen minutes of work. Moving from high-yield savings to an optimized T-bill ladder is worth perhaps a hundred more for ongoing effort. The first move is obviously correct and the second is a preference.
Building it in the first place #
The hard part is not the account, it is the balance. Six months of expenses is a large number when you are starting from zero, and the gap between $0 and $18,000 is where most people give up.
Two things help. Automate the transfer for the day after payday so the money leaves before you see it, because automatic beats intentional in every study anyone has run on this. And set it as an explicit goal with a date rather than a vague intention, because a target you are 40% of the way toward behaves differently in your head than an open-ended pile.
That is one of the things I built Retire Goals to handle. It holds multiple goals rather than one enormous retirement number, so an emergency fund target sits alongside the long-term one and actually completes, which matters because completing things is what keeps people engaged with a system. It is free, there are no accounts, and the data stays on your device. The build write-up covers why the multiple-goal thing turned out to matter more than the projection math.
A spreadsheet does the same job. The mechanism matters less than having one.
Once the fund is full, the money that was going into it should go somewhere it compounds, which is the retirement arithmetic, and the target you are eventually aiming at is how much you actually need.