How the credit is actually constructed
The premium tax credit under IRC §36B works backwards from an affordability target rather than forwards from a discount. Congress decided what share of income a household at each income level should pay for health insurance, then made the credit whatever it takes to get there against a reference plan. Three quantities are needed and only three.
Your income as a percent of the federal poverty level. Household modified AGI divided by the poverty line for your tax household size. MAGI here is adjusted gross income plus tax-exempt interest, excluded foreign income and the untaxed portion of Social Security benefits — a broader measure than AGI alone.
The applicable percentage. A sliding scale that converts that FPL percentage into the share of income you are expected to contribute. It rises with income, so higher earners pay more of their own premium.
The benchmark premium. The monthly premium of the second-lowest-cost silver plan available to your household in your rating area, adjusted for the ages of everyone seeking coverage. It is not the plan you buy — it is the yardstick.
The credit is the benchmark minus your expected contribution, floored at zero. Because it is anchored to the benchmark, your plan choice changes your net cost but never your credit. Take a bronze plan $200 cheaper than the benchmark and you keep the whole $200. Take a gold plan $150 dearer and you pay the whole $150.
The credit is also advanceable. Most enrollees take it monthly as advance payments sent straight to the insurer, which is why it is reconciled on Form 8962 at filing against what your income turned out to be.
The sliding scale, and the cliff that sits at its top
The applicable percentage schedule used here runs from 0% of income at or below 150% of the poverty level to 8.5% at 400% and above, interpolating linearly within each band. That schedule was enacted by the American Rescue Plan Act for 2021 and 2022 and extended through 2025 by the Inflation Reduction Act. Whether it applies to a later year is a question of legislation, so confirm the current schedule with your marketplace before relying on the number.
Underneath it sits the original structure of §36B(b)(3)(A), which had two features the enhanced schedule removed. Its percentages were higher at every income level, and it cut off entirely at 400% of the poverty level — the subsidy cliff. Under the cliff, a household at 399% of FPL receives a substantial credit and the same household at 401% receives nothing. For a family of three that boundary sits at about $106,600, and crossing it by one dollar could cost several thousand dollars a year. The checkbox on this page models that structure.
The 8.5% cap replaces the cliff with a taper. Above 400% of FPL, your contribution is fixed at 8.5% of income, so the credit is benchmark minus 8.5% of income and falls smoothly to zero as income rises. Note where zero actually arrives: it depends on your benchmark premium, not on a fixed income level. A household with a $900 benchmark stops receiving a credit once 8.5% of income reaches $10,800 a year, which is an income of $127,059. A household with a $500 benchmark reaches zero at $70,588. Older enrollees and high-premium rating areas therefore keep a credit far up the income scale, which is exactly what the taper was designed to do.
At the bottom the schedule has its own boundary. Below 100% of FPL there is generally no premium tax credit at all. In states that expanded Medicaid the household qualifies for Medicaid instead, and the transition is seamless. In states that did not expand, there is a coverage gap between the Medicaid ceiling and the credit floor in which a household qualifies for neither.
Worked example: a household of three at $55,000
A married couple with one child projects $55,000 of MAGI for the year. Their benchmark silver premium is $900 a month, and they are considering a bronze plan at $720.
- Poverty line for three. 15,650 + 2 × 5,500 = $26,650.
- Income as a percent of FPL. 55,000 ÷ 26,650 = 2.06379, so 206.38%.
- Locate the band. 206.38% falls in the 200%–250% band, where the applicable percentage runs from 2.0% to 4.0%.
- Interpolate. (206.379 − 200) ÷ 50 = 0.12758 of the way through the band. 0.12758 × 2.0 = 0.2552, so the applicable percentage is 2.0 + 0.2552 = 2.2552%.
- Expected annual contribution. 55,000 × 0.022552 = $1,240.34.
- Expected monthly contribution. 1,240.34 ÷ 12 = $103.36.
- Premium tax credit. 900.00 − 103.36 = $796.64 a month, or $9,559.66 a year.
- Net cost of the silver benchmark plan. 900.00 − 796.64 = $103.36, which by construction equals the expected contribution.
- Net cost of the bronze plan. 720.00 − 796.64 is negative, so the credit is capped at the premium and the plan costs $0.00 a month. The unused $76.64 is not refundable — the credit can never exceed the premium actually charged.
Before taking the free bronze plan, note what it gives up. At 206% of FPL this household qualifies for cost-sharing reductions, which are only available on silver plans and which cut the deductible and out-of-pocket maximum substantially. Paying $103.36 a month for a cost-share-reduced silver plan is frequently better value than paying nothing for bronze, and the comparison turns on expected claims — which is what the deductible and coinsurance calculator is for.
Applicable percentage schedule and what it means for a household of three
| Income as % of FPL | Applicable percentage | Household income | Expected monthly contribution |
|---|---|---|---|
| Up to 150% | 0.0% | up to $39,975 | $0.00 |
| 200% | 2.0% | $53,300 | $88.83 |
| 250% | 4.0% | $66,625 | $222.08 |
| 300% | 6.0% | $79,950 | $399.75 |
| 400% | 8.5% | $106,600 | $755.08 |
| Above 400% | 8.5% (capped) | above $106,600 | 8.5% of income ÷ 12 |
Percentages interpolate linearly between the listed points — the 150–200 band runs 0% to 2%, 200–250 runs 2% to 4%, 250–300 runs 4% to 6% and 300–400 runs 6% to 8.5%. Each contribution is income × percentage ÷ 12.
Reading the result, and the reconciliation risk that follows it
The credit you see is an estimate built on a projection of income you have not earned yet, and that is the source of nearly every unpleasant surprise. Advance payments are made monthly on the strength of the income you reported at enrolment; Form 8962 then compares the advance against the credit your actual income supported. Earn less than projected and the difference comes back as a refund. Earn more and you repay.
Repayment is capped for households below 400% of FPL, on a scale that rises with income, but it is uncapped at or above 400%. That asymmetry is the practical danger. A household that projected 380% of FPL, took a large advance credit, and finished the year at 405% may have to repay the entire year's advance in one lump. Unexpected income arriving late in the year — a bonus, a Roth conversion, capital gains, a spouse returning to work — is the usual trigger.
Two responses are worth knowing. First, report income changes to the marketplace during the year; it adjusts the advance payments forward and the reconciliation stays small. Second, remember that MAGI is reduceable by above-the-line items. A deductible traditional IRA or HSA contribution reduces MAGI dollar for dollar, and near a threshold that can be worth far more than the deduction itself. The HSA contribution limit calculator and the taxable income calculator help size that move.
On plan choice, treat the credit as fixed and shop on price and structure. Because the credit is anchored to the benchmark, a plan $100 cheaper is $100 cheaper to you — the full saving passes through. The exception is the 100%–250% band, where cost-sharing reductions attach to silver plans only and can be worth more than the premium difference. Above 250% that consideration disappears and the metal-level choice becomes a straightforward trade between premium and expected claims.
What this calculator does not do
- It does not price plans. The benchmark and chosen-plan premiums are inputs; get them from your marketplace or from Form 1095-A columns A and B, since they vary by age, county, tobacco use and household composition.
- It does not apply the employer-coverage bar. If you have an offer of employer coverage that is affordable and provides minimum value, you are barred from the credit however low your income is.
- It does not compute cost-sharing reductions. Those change the deductible and out-of-pocket maximum on silver plans below 250% of FPL, not the premium.
- It does not model the repayment caps. Reconciliation on Form 8962 limits repayment for households below 400% of FPL and does not limit it at or above.
- It does not handle mid-year changes. The credit is computed month by month on Form 8962, so a change in household size, income or benchmark partway through the year needs a per-month calculation.
- It uses one poverty guideline table. Alaska and Hawaii have higher guidelines, and marketplaces use the guidelines published in the year before the coverage year.
- It assumes you file a joint return if married. Married taxpayers filing separately are generally ineligible, with an exception for survivors of domestic abuse and spousal abandonment.
An affordable employer offer removes eligibility entirely
If you are eligible for employer-sponsored coverage that is both affordable — measured against a statutory percentage of household income for the employee's own self-only premium — and provides minimum value, you cannot claim the premium tax credit for a marketplace plan, no matter how much cheaper the marketplace would be. Since 2023, family members have been tested against the cost of family coverage rather than the employee's self-only cost, which restored eligibility for many families previously caught by the so-called family glitch. The employee's own eligibility is still tested on the self-only figure. If you are weighing marketplace coverage against a COBRA continuation offer instead, note that COBRA eligibility does not bar the credit unless you actually enrol; the COBRA premium cost calculator prices that alternative.
Why the benchmark is a silver plan, and what that implies
Silver was chosen as the reference tier because it sits in the middle of the actuarial value scale — roughly 70% of expected costs covered by the plan, against 60% for bronze, 80% for gold and 90% for platinum. Anchoring the credit to the second-lowest silver plan rather than the lowest gives the market a little competitive slack while still tying the subsidy to a genuinely available price.
One consequence is worth understanding because it looks like a pricing error. In many states, insurers load the cost of cost-sharing reductions onto silver premiums alone, a practice known as silver loading. That inflates the benchmark, which inflates everyone's credit, which can make bronze and even gold plans unusually cheap after the credit. A gold plan priced below a silver plan net of credit is not a mistake; it is silver loading working through the formula.
A second consequence concerns age. Premiums may vary with age by up to a 3:1 ratio for adults, and the benchmark varies with them. An older household therefore has a higher benchmark, a higher credit at the same income, and a credit that persists to a higher income under the 8.5% taper. Two households with identical income can receive very different credits purely because of age and county.
Finally, note that the credit is fully refundable and is claimed on the tax return whether or not it was advanced. A household that pays full price all year and then finds it qualified can claim the whole amount on Form 8962 at filing. Doing it that way costs cash flow but eliminates reconciliation risk entirely, which is a defensible choice for anyone with volatile income near the 400% line. Sizing that decision means knowing your marginal rate, which the marginal vs effective tax rate calculator sets out.