FSA Use-It-or-Lose-It: What Unspent Money Actually Costs You

You put $2,400 into a health FSA last January. On December 31 there’s $610 sitting in it, and your plan doesn’t carry anything over. That $610 disappears. But you didn’t lose $610 of spending power, because the money never got taxed on the way in. At a 22% federal bracket plus 7.65% FICA plus 5% state, you’d have kept about 65 cents of every pre-tax dollar anyway, so the after-tax damage is roughly $397. That’s the number that matters, and it’s the one nobody puts on the forfeiture notice.

Key points

• A $610 forfeiture at a combined 34.65% marginal rate costs about $397 in after-tax money, not $610, because the contribution was never taxed.
• The break-even is simple: as long as you spend more than (1 − your marginal rate) of what you contributed, the FSA still beat taking the cash. At 34.65%, that break-even is 65.35% of contributions.
• On $2,400 contributed and $1,790 spent, you’re still ahead by about $434 versus paying for that same care with after-tax dollars.
• Forfeiting 100% of contributions is the only way an FSA can lose outright, and even a 25% forfeiture rate leaves you ahead at a 34.65% marginal rate.
• Your plan may offer a carryover or a grace period, but not both, and neither is required by law, so the plan document is the only place to check.

Why does forfeited FSA money cost less than the sticker amount?

Because it went in before tax. That’s the whole mechanism, and it’s worth being precise about it.

Meet Dana. Single filer, $78,000 salary, elects $2,400 into a health FSA for the 2026 plan year, $100 per semi-monthly paycheck across 24 paychecks. Her plan has no carryover and no grace period. Hard December 31 deadline.

When payroll pulls that $100, it comes out of gross pay. It never appears in box 1 of her W-2, and it never appears in Social Security or Medicare wages either. So each $100 that goes into the FSA would have shown up in her checking account as:

• Federal income tax at 22%: $22.00 gone
• FICA (6.2% Social Security + 1.45% Medicare = 7.65%): $7.65 gone
• State income tax at 5%: $5.00 gone
• Landing in her account: $65.35

Combined marginal rate: 34.65%. Every pre-tax dollar in the FSA is worth 65.35 cents of take-home pay.

Now the forfeiture. Dana spent $1,790 on copays, dental work, and contact lenses. She leaves $610 on the table.

$610 × 0.6535 = $398.64.

Round it: about $399 of real, spendable, after-tax money vanished. Not $610. The other $211 was never hers. It was the government’s, and it stayed the government’s either way.

A sceptic’s objection here is fair: doesn’t reducing Social Security wages cut her future benefit? Technically yes. Practically, $2,400 shaved off one year of a 35-year earnings record moves the benefit calculation by a trivial amount, and only if she’s below the wage base. Real, but not worth a line in the household budget.

What is the actual break-even on an FSA election?

You break even when the tax you saved equals the money you forfeited. Written out:

Spend more than (1 − marginal rate) of what you contributed, and the FSA won.

At Dana’s 34.65% combined rate, that’s 65.35%. She contributed $2,400. Her break-even spend is:

$2,400 × 0.6535 = $1,568.40.

She spent $1,790. She cleared the bar by $221.60 of pre-tax dollars.

Here’s the direct comparison, all in after-tax dollars, which is the only unit that makes sense for a household.

Path A: the FSA. $2,400 left her gross pay. In after-tax terms that reduced her take-home by $2,400 × 0.6535 = $1,568.40. In exchange she received $1,790 of medical care and forfeited $610 of face value.

Path B: no FSA at all. She keeps the full $2,400 of gross pay, which arrives as $1,568.40 after tax. Then she pays the same $1,790 medical bill out of that after-tax money. She’s $1,790 − $1,568.40 = $221.60 in the hole relative to having taken nothing.

Compare: Path A leaves her out of pocket $1,568.40 with the care fully covered. Path B has her out of pocket $1,568.40 in reduced pay and then another $1,790 in medical bills, offset by the $1,568.40 she kept. Net position under Path A versus Path B, on the same $1,790 of care:

• Path A total after-tax cost of the care: $1,568.40
• Path B total after-tax cost of the care: $1,790.00
• Advantage to Path A: $221.60

She forfeited $610 and still came out $221.60 ahead. That’s the part people get wrong. The forfeiture feels like a loss because it’s a visible, dated, itemised loss. The tax saving is invisible because it happened 24 times in $34.65 slivers.

How bad does a forfeiture have to get before the FSA loses?

At a 34.65% marginal rate, she has to forfeit more than 65.35% of the election before she’s worse off than taking the cash. Below is her $2,400 election under different spending outcomes, every column in dollars.

Read the “FSA advantage” column. It’s positive until spending drops to $1,568, which is exactly the 65.35% break-even. Below that, the election starts costing her.

The shape of that column is the real lesson. The penalty for over-electing is gradual and starts from a cushion. The penalty for under-electing is that you pay medical bills with taxed money, which is a guaranteed 34.65% premium on every dollar of care you were always going to buy.

One caveat on the table: it holds Dana’s spending as the independent variable. In practice, a large election nudges people to spend more, sometimes on things they’d have skipped. That’s a behavioural cost the arithmetic can’t see.

Does a higher tax bracket make forfeiting cheaper or more expensive?

Cheaper per forfeited dollar, and it widens the cushion. This is counterintuitive, so work it.

Take Dana’s colleague at the same firm, higher salary, 32% federal, same 5% state, and above the Social Security wage base so only the 1.45% Medicare portion applies. Combined marginal rate: 32 + 1.45 + 5 = 38.45%.

For that person, each pre-tax FSA dollar is worth 61.55 cents of take-home. Forfeit $610 and the after-tax hit is $610 × 0.6155 = $375.46, about $23 less painful than Dana’s $398.64.

And the break-even drops. On a $2,400 election, the colleague only needs to spend $2,400 × 0.6155 = $1,477.20 to stay ahead. Dana needed $1,568.40.

Higher marginal rate, bigger subsidy, more room to be wrong. It also means the person most likely to shrug off a forfeiture is precisely the person who lost the least in real terms.

Run it the other way and it tightens fast. Someone in the 12% federal bracket with no state income tax has a combined rate of 12 + 7.65 = 19.65%. Each pre-tax dollar is worth 80.35 cents. A $610 forfeiture costs them $490.14, and on a $2,400 election they need to spend $1,928.40 to break even. That’s 80% of the election. Far less margin for a bad estimate.

What does the forfeited money cost if you think of it as savings instead?

$399 is not a rounding error, and the honest way to size it is to ask what it would have become.

Dana’s $398.64 of after-tax loss, if it had instead gone into a broad diversified account and compounded at 6% annually for 20 years, would be:

$398.64 × 1.06^20 = $398.64 × 3.2071 = $1,278.40.

That 6% is an assumption, not a forecast, and I’m using it only to show the shape of compounding rather than to predict anything. You can run any rate and horizon yourself with the SEC’s compound interest calculator on Investor.gov, which is the same arithmetic without my hand on the inputs.

Two things that number does not account for. It’s in nominal dollars, so $1,278 in 2046 buys less than $1,278 today, and how much less depends on cumulative inflation over those two decades, which you can track through the BLS Consumer Price Index. And it assumes she’d actually have invested the $399 rather than spent it, which for most people is the weaker half of the assumption.

Still, the exercise reframes the loss usefully. A single year’s over-election isn’t a catastrophe. Repeating the same over-election for ten straight years is where it becomes a number worth naming.

Why do people over-elect in the first place?

Because the election is a forecast made 11 months before the last dollar is spent, and forecasts made under uncertainty skew high when the downside feels asymmetric.

Dana elected $2,400 in November 2025 for the 2026 plan year. At that moment she knew about a planned dental crown and her usual prescriptions. She didn’t know she’d switch jobs in August, change insurance, and stop seeing the physical therapist she’d budgeted 20 visits for.

There’s a second, stranger force at work. The health FSA has a uniform coverage rule: the full annual election is available on day one, before you’ve funded it. If Dana had needed $2,400 of care in February, the plan pays $2,400 even though she’d only contributed $400 by then. If she’d then quit in March, she’d generally keep the benefit and the employer eats the difference.

That asymmetry is genuinely valuable, and it rationally pushes people toward larger elections. The insurance-like upside is real. The forfeiture risk is the premium you pay for it.

Third force: the deadline is a cliff, not a slope. December 31 at midnight, the balance is worth $610. One second later, zero. Cliffs produce panic spending in the last two weeks of December, which is its own form of waste, just one that doesn’t show up as a forfeiture.

What does an employer actually do with forfeited FSA money?

The forfeited balances don’t go to the government. Under the cafeteria plan rules, unused amounts stay with the plan sponsor, which in practice means the employer. Employers may generally use them to offset plan administration costs or reallocate them among participants, subject to what the plan document allows.

Practical consequence: nobody at your company is incentivised to call you in November and tell you to spend your balance. Some do it anyway as good practice. Many don’t. The alert, if it comes, arrives as a form letter in late December when your options have narrowed to whatever your plan’s eligible-expense list still allows.

Also worth knowing: your plan may offer a carryover of a limited amount into the next year, or a grace period of up to two and a half extra months to incur expenses. Plans can offer one or the other, not both, and they aren’t required to offer either. The annual contribution limit and any carryover amount are set by the IRS and adjust for inflation, so check the IRS website for the current year’s figures rather than trusting a number you remember from a previous open enrollment.

What this does not tell you

The $398.64 figure rests on Dana’s specific tax situation, and yours almost certainly differs.

The marginal rate is the whole ballgame, and it’s easy to get wrong. I used 22% federal, 7.65% FICA, 5% state. If your state has no income tax, drop 5 points. If your income is above the Social Security wage base, only the 1.45% Medicare portion applies, not the full 7.65%. If your FSA contribution straddles a bracket boundary, part of the deduction saves at one rate and part at another, and the blended figure sits between them.

Every dollar in this article is an illustration. Dana isn’t real. The $2,400 election, the $1,790 spend, the $610 forfeiture are constructed so the arithmetic is checkable, not because they represent typical behaviour. I have no data on typical FSA forfeiture rates and I’m not going to assert one.

The 6% compounding rate is arbitrary. It’s there to show the mechanism of compounding over 20 years, not to suggest any particular return is available. Markets don’t deliver a smooth 6%, or any smooth number.

Dependent care FSAs work differently. The contribution limits differ, the eligible expenses differ, the interaction with the dependent care tax credit is a separate calculation entirely, and the uniform coverage rule does not apply the same way. None of the arithmetic above transfers.

HSAs are a different instrument. They don’t forfeit, they roll over indefinitely, and they carry an eligibility requirement tied to your health plan. If you’re comparing the two, the forfeiture arithmetic here answers none of the relevant questions.

Plan-specific rules override general rules. Carryover amount, grace period, run-out deadline for submitting claims on expenses already incurred, what happens if you leave mid-year. All of it is in your summary plan description, and variation between employers is wide.

The behavioural cost is invisible to the arithmetic. If a large election causes you to buy care you’d otherwise have skipped, the tax saving is partly illusory. The table can’t measure that.

FAQ

How do I calculate my own break-even?

Add your federal marginal rate, your FICA rate (7.65% if you’re below the Social Security wage base, 1.45% if above), and your state income tax rate. Subtract that total from 100%. Multiply the result by your election. That’s the dollar amount you need to spend during the plan year to come out ahead of not electing at all. Example: 24% + 7.65% + 0% state = 31.65%. Keep-rate 68.35%. On a $3,000 election, break-even spend is $2,050.50.

Is it better to under-elect or over-elect?

At a combined marginal rate above 50%, over-electing is mathematically cheaper per dollar of error, because each forfeited dollar costs less than half its face value while each dollar of unfunded care costs full price. Below 50%, the two errors are closer to symmetric and under-electing is the safer mistake. At Dana’s 34.65%, forfeiting $1 costs 65 cents while paying a $1 bill with after-tax money costs $1.53 in gross pay. Over-electing is still the cheaper error for her, but the gap is narrower than most people assume.

Does forfeited money go back to the IRS?

No. Unused balances remain with the plan sponsor, generally the employer, subject to the cafeteria plan rules. The employer may apply them against plan administration costs or reallocate them among participants, depending on plan terms. It’s not a tax payment and it’s not returned to you.

If I quit mid-year, do I have to pay back money I already spent?

Generally not, for a health FSA. The uniform coverage rule means the full annual election is available from the start of the plan year, and if you’ve spent more than you’ve contributed by your departure date, the plan absorbs the shortfall. The mirror image is that contributions already made for expenses not yet incurred are typically forfeited once your coverage ends, unless you elect COBRA continuation for the FSA. Your plan document governs this and the details vary.

Should I front-load spending early in the year?

The arithmetic doesn’t say. What it does say is that spending early carries less deadline risk than spending late, because a December scramble narrows your options to whatever’s still purchasable and eligible. It also says that if you spend the full election in January and leave the job in February, you’ve captured the uniform coverage benefit. Neither of those is a recommendation about your health care. Both are observations about how the plan’s timing rules work.

Does the annual limit change every year?

Yes. The health FSA contribution limit is indexed for inflation and the IRS announces the following year’s figure ahead of open enrollment, usually in the autumn. Any carryover amount your plan offers is indexed separately. Don’t carry last year’s number into this year’s election form.

What to look at next

Three documents, in this order.

Your summary plan description, specifically the sections on carryover, grace period, and the run-out deadline for submitting claims. Those three dates determine whether a December 31 balance is actually lost or merely inconvenient.

Your most recent pay stub, to find your actual marginal rates rather than the ones you assume. The federal withholding line won’t give you the marginal rate directly, but your gross-to-net breakdown plus a bracket table will get you close enough to run the break-even.

Twelve months of your own medical spending, pulled from insurance statements rather than memory. Memory systematically undercounts small recurring costs and overcounts the one big event you remember. The election is a forecast, and forecasts built from records beat forecasts built from impressions.

This article is general information, not financial advice. See our disclaimer.

Read next

• 401(k) Vesting: What You Forfeit by Leaving at Year 2
• HSA Triple Tax Advantage: What It’s Actually Worth in Dollars
• 401(k) Match: The Return You Forfeit Under 6%
• Car Loan vs Cash: The Total Interest You Pay to Drive

Also worth reading: Credit Utilization Ratio: How Much a High Balance Actually Costs Your Score

More in Money Tips · all topics · calculators · how this was checked