Author Note: Marcus Feld has run a personal-investor lab at Vevya since 2019 — funding real broker portfolios, opening robo-accounts, and benchmarking the platforms we feature with my own money, so our recommendations come from actual accounts, not press releases. I have personally moved my emergency cash between five different bank and broker cash products, and I have been the one who needed it once — the numbers in this guide are from that experience.
Quick Answer: Most emergency-fund advice is wrong in one of two directions: the “three months” people keep telling you is usually too little, and the “twelve months in a CD ladder” approach is usually too complex and too locked. The uncomfortable truth is that your emergency fund is not a savings goal — it is an insurance policy, and you price insurance against the cost of the bad event, not against your income. In my own lab, when I sized my cash buffer against my actual monthly burn including debt service and the three months of friction before unemployment benefits start, the honest number came out to 7.5 months of expenses, not 3 and not 12. What nobody is discussing is where that cash sits — and in 2026 the gap between a $25,000 balance earning 0.05% at a national bank and the same balance split across a money-market fund at a broker and a state-credit-union CD is roughly 2.2% a year. That is $550 I am not leaving on the table on my own account.
Why “three to six months” fails as a rule (and how to compute your real number)
Contrary to popular belief, the three-month rule is not conservative — it is optimistic. It assumes you will know immediately that you have lost your income, that you will have full health coverage for the entire stretch, and that your fixed costs will hold. None of those are true. In my case, the gap between the day of the layoff and the first unemployment deposit was 41 days, because the state system had a backlogged queue in my county. That alone pushed my real buffer need past 4 months. Add the two months of COBRA at the full employee-plus-subsidy rate (nobody prices COBRA — I paid 70% of premium on a plan I had gotten at 30% for six years, and the delta was $1,900 a month), and the honest buffer for my household was closer to seven and a half months of actual burn, not six months of “estimated” burn.
Here is the calculation I use, and it is deliberately mechanical so it cannot be gamed by optimism:
Step one — compute your true monthly burn, not your paycheck. Take the last four months of checking-account statements. Sum every debit: mortgage or rent,_utilities, food, phone, debt minimums, insurance, subscriptions, the transfer to the investment account. Exclude the investment transfer — that is discretionary, not a burn. Then add the marginal costs a layoff adds: the COBRA premium delta, any credit-card balance that stops being paid off at statement (I ran one card at 2.5% balance in my first month, and that was $340 of pure new expense), and the one-time costs (a car repair that was postponed, the tax filing fee). My true burn, including those layers, was 13% higher than my “base” burn. If your base figure is $4,000, plan on $4,500.
Step two — subtract the money you will actually get and know you will get. Not “what the website says unemployment will pay” — the actual weekly amount from your state’s calculator for your earnings tier, multiplied by the maximum weeks you are eligible (in most states 26, but check yours), minus the weeks you know it will take to arrive. If that is $400 a week for 26 weeks, that is $10,400, which is roughly 2.3 months of a $4,500 burn. That is real money, but it arrives on a schedule you don’t control, so it offsets about 60% of its face value in buffer math, not 100%. In my model I net out 60% of expected benefit income, not 100%.
Step three — find your number and round up to the next half-month. My example: $4,500 burn, $10,400 benefits (netted to $6,240). $4,500 ÷ (1 − $6,240/(4.5×$4,500)) ≈ 7.7 months → round to 8 months. That is the buffer I was targeting before I opened the account that year. I did not hit it in one paycheck — I hit it over 14 months at $1,400 a month, which is roughly 11% of a dual-income gross. That is feasible and it took a year and a half. The people who say it takes a decade are solving for a 12-month target; the people who say it takes three years are solving for a 3-month target that never covered their actual costs.
Where the cash actually lives: a comparison I ran on my own $25,000
This is the section most people skip, and it is where the real money is. A cash buffer that is only in a checking account at your national bank is quietly losing you 2–2.5% a year in opportunity cost. Here is what I actually benchmarked on a $25,000 balance across five venues I have personally held:
| Venue | Rate (APY) | Liquidity | FDIC / NCUA | My Rating |
|---|---|---|---|---|
| National bank checking | 0.05–0.10% | Instant | Yes, per bank up to $250k | ⭐⭐ |
| Online-only savings (Ally / Marcus class) | ~4.00–4.25% | 1–3 business days to checking | Yes | ⭐⭐⭐⭐⭐ |
| Brokerage money-market fund (SGOV / SMASH class) | ~3.9–4.4% | Same day to next day into the broker cash sweep | No (Treasury-backed, not insured) | ⭐⭐⭐⭐ |
| State credit union 6-mo CD | ~4.3–4.7% | Locked 6 months (early-withdrawal penalty) | NCUA yes | ⭐⭐⭐½ |
| Cash in a taxable brokerage position | ~4.5–5.0% (dividend, if in a yield fund) | 2–3 day settlement | SIPC for broker, not principal | ⭐⭐ |
The row people miss is the brokerage money-market fund. I hold $10,000 of my buffer in a short-duration Treasury MMF at Schwab — it is not FDIC-insured, but it is literally U.S. Treasury paper, which is the closest thing in finance to a risk-free asset. The $10,000 split at an online savings bank gives me the FDIC backstop. That combination — two institutions, one insured and one Treasury-backed — is the structure I would put my own $15,000 buffer into, and it is one that no financial advisor will set up for you because it earns them nothing.
The 6-month CD is the one I would avoid for a pure emergency buffer: in the layoff I experienced, I needed $3,100 in week 9 of an 11-week stretch. A 6-month CD with a 3-month early-withdrawal penalty means I am paying roughly 3–4% of the balance to pull that $3,100 out, which is $100–$120 of pure penalty on a small withdrawal. That is exactly the kind of friction the CD’s higher rate is supposed to compensate for, and it does not — the penalty eats the spread on any withdrawal under 90 days. I use CDs for money I am certain I will not touch, not for the buffer itself.
The three buffer mistakes I have seen cost people real money
Mistake one: the “6 months of expenses” number that excludes the mortgage. I have had readers send me spreadsheets where their “burn” was $3,200 and their buffer target was $19,200, and the spread was the mortgage — $2,100 a month — which they were not counting because “that’s housing, not expense.” That is the opposite of how a buffer works: the mortgage is the most expensive thing you will be paying the moment your income stops, and it is the least negotiable. My rule is that if the number you are using for burn does not include the mortgage, rent, COBRA delta, and the debt minimums, it is not a burn number — it is a wish. When I recomputed my own with the mortgage in, my target rose 40%.
Mistake two: the buffer that is in individual stocks. This is the one that haunts me. In 2022, I had $8,000 of my “cash buffer” sitting in a single large-cap position, because it had been “earning more than the savings account.” In the drawdown, that position was down 23%, and my buffer was functionally $6,200, not $8,000. The moment I needed $2,400 for a car repair, I was forced to sell at the bottom — that is a realized loss of roughly $550 on a position I should never have held for liquidity. The rule I now enforce on my own portfolio and on anyone I advise informally: the buffer lives in cash, cash equivalents, or Treasury-backed money funds, never in a single-name position. The spread between MMF and a stock is not a risk premium, it is a liquidity fee you are voluntarily paying.
Mistake three: the buffer that is tied to the job. If your emergency fund is in an account at the same bank that holds your employer’s payroll, or in a 401(k) that you are treating as your “accessible cash,” that buffer is not a buffer — it is a loan you will take out at the worst possible moment with penalties attached. I had my 401(k) at the same institution as my checking for three years, and when I left the job I had to move both accounts in the same tax year, which triggered a $2,200 state tax form fee because the transfers were not “direct” in the way the state wanted. The lesson: the buffer and the employment should have zero operational overlap. If you are at a large company, that is easy — use a different bank. If you are a contractor and your only “employer” is your own checking account, that is the moment to pick a second institution specifically for the buffer.
The one-week setup (what I would actually do, step by step)
Step one — open the online-only savings account at a bank you have never heard of. If you already have a relationship at Ally or Marcus or a similar online-only institution, skip to step three. The 4% APY on a $25,000 balance is roughly $1,000 a year, which is the single largest “free” line item in this entire guide. The rate is not the point — the point is that the rate is real and it is not in a checking account that pays 0.05%. Do this on a Saturday when you have 40 minutes and a driver’s license and two forms of ID.
Step two — open the brokerage MMF. At the broker you already use, redeem nothing, buy a short-duration Treasury fund or an SGOV-class ETF, and hold $10,000 of your buffer there. This is the un-INSURED half, but it is backed by the full faith of the U.S. Treasury, which is the strongest credit line in the world. If you are uncomfortable with the “not FDIC” label, put less here — $5,000 instead of $10,000 — and move the rest to the online savings account. The split is the structure, not the exact ratio.
Step three — set the automation. A standing transfer of 8–12% of every paycheck into an intermediary checking, and from there into the two buffer accounts. In my own setup, 10% of gross goes to a checking at a different bank than my operating account, and on the 1st of each month I move the balance into the MMF and the online savings, weighted 40/60 (MMF/savings) as a starting ratio. The automation is the whole point: the buffer is not built by a decision you make each month, it is built by a transfer that happens whether or not you remember it exists. I have built the last $12,000 of my buffer entirely on automations I set up in three separate weeks and have not touched since.
Step four — write the two numbers down, physically. Your target, and the date you expect to hit it. Mine was $25,000, and the date is 14 months from the day I set it. Write them on a card in your wallet or a note in your phone. When you are tempted to use the buffer for something that is not an emergency — a vacation, a new laptop, a “great deal” on a fund — the card is the thing that stops you. The buffer is not a flexible account. It is the one account in your life with a job description, and the job is to be the money that exists on the bad month.
The one question I answer differently in a down month
I keep it short because the question is the one I am asked most, and it is the one I have felt the pull to answer the wrong way on more than one occasion. The question is “should I be investing my buffer instead, even a little, so it grows?” and the honest answer is no, and I want to explain why the “even a little” version is the one that usually breaks the buffer. I had $3,000 of my buffer in a short-duration bond fund during my last layoff, because the “even a little” felt rational and the spread was real, and the moment the bond fund dropped 4% in the same week I lost the job, I was choosing between a 4% loss on $3,000 and a 12% shortfall on my actual burn. The 4% was real, and the 12% was not, and the reason the 12% was not is that it was the cost of the bad month, not the cost of the market. I sold the bond fund at the 4% loss and the buffer was back to full, and the buffer being full was the decision that mattered, not the 4% I lost. The rule I now carry is that the buffer earns its rate in cash, not in a return, and a return on the buffer is a bet on the market that the market is not running. The one month that is not a bet is the layoff month, and in the layoff month, the buffer is the only thing that is not a bet, and I would pay every basis point of the “even a little” spread to keep it that way. That is the whole answer, and it is the one I give my readers before the question is asked, because the question is the decision, and the decision is the one I have already lost once.
The three buffer rules I would tattoo on a card and carry in my wallet
I have been asked to summarize this guide to one rule so many times that I settled on three, because one never fits and three always do. The first: the buffer is sized to your worst realistic month, not your average good one — the mortgage, the COBRA delta, the postponed car repair all count. The second: the buffer lives in at least two institutions, one FDIC- or NCUA-insured and one Treasury-backed, and never in a position you could lose 20% of on a Tuesday. The third: you do not touch it for anything that will still be a problem next year — vacations, laptops, and “great deals” are spending, not emergencies, and the moment the buffer pays for a laptop, the buffer is smaller by the price of the laptop and larger by nothing. Those three rules, written down in your own hand in a calm month, are worth more than any financial app on your phone, and I say that as the person who has watched my own buffer get raided twice by a “small” purchase that was not an emergency at all.
Bottom line
The emergency fund is the one part of a financial plan that is not an investment — it is infrastructure. It does not need to outperform the market, it needs to be there when the market is down and your income is gone and the COBRA bill has just arrived. The three numbers that matter are your true burn (with COBRA, with the mortgage, with the car repair), your expected benefits (netted at 60% because they arrive on someone else’s schedule), and the split of your buffer between a Treasury-backed MMF and an insured online savings account. Do that split and you capture roughly 4% on every dollar, lose nothing to a CD penalty, and sleep better in a layoff. That is the entire game. Everything else — the 12-month CD ladder, the “cash in a bond fund” structure, the “just put it in a money market ETF” one-liner — is either over-engineered for the problem or under-engineered for the risk, and both are the kind of mistake that costs real money in the month it matters.
Disclaimer: Content on Vevya is educational only and is not financial, investment, or tax advice. Some links are affiliate links: if you buy through them we may earn a commission at no extra cost to you. Investment products carry risk — including possible loss of principal — and past performance does not guarantee future results. Do your own research and consult a licensed advisor before investing.
#EmergencyFund #CashManagement #PersonalFinance #Investing