The ACA subsidy cliff after 2025

Last reviewed July 2026 · 5 min read

For early retirees, health insurance before Medicare at 65 is often the scariest line in the budget — and since the start of 2026, it comes with a cliff. The enhanced ACA premium subsidies created in 2021 (and extended through 2025) expired, and the marketplace reverted to the original statutory rules: premium tax credits are available only up to 400% of the federal poverty level. One dollar of income above that line, and the entire subsidy is gone.

How the subsidy is computed

Marketplace subsidies cap your required contribution toward the benchmark (second-cheapest silver) plan at a percentage of income that slides from 2.1% at the bottom of the scale to 9.96% at 300–400% of the poverty level (2026 figures). Your subsidy is the benchmark premium minus that required contribution. Coverage-year 2026 uses the 2025 poverty guidelines: $15,650 for a household of one and $21,150 for a household of two (48 contiguous states), so the 400% cliff sits at $62,600 and $84,600 of MAGI respectively.

MAGI: the income that counts

The ACA uses its own definition of modified adjusted gross income: your AGI plus tax-exempt interest, untaxed foreign income, and the non-taxable portion of Social Security. For a retiree, almost every planning lever moves it: pre-tax IRA/401(k) withdrawals, Roth conversions, realized capital gains, dividends, and interest all count. Roth withdrawals and spending down principal (basis) in a taxable account generally do not. Note that this is a different definition from the MAGI used for IRA deduction limits or Medicare IRMAA — same acronym, different add-backs.

Why $1 can cost thousands

Below the cliff, subsidies phase out smoothly. At the cliff, they vanish. A 60-year-old couple at $84,600 MAGI has a required contribution of about 9.96% of income — roughly $8,400 — and the subsidy covers the (age-rated, often very large) remainder of the benchmark premium. At $84,601, the subsidy is $0 and they pay the full sticker price. Because insurers can charge 64-year-olds three times the premium of 21-year-olds, the cliff is steepest exactly where early retirees live: KFF estimates a 60-year-old just over the line can pay thousands more per year than a neighbor earning $2,000 less.

Subsidized: pay at most 2.1–9.96% of MAGI+$1 → entire subsidy gone$0 subsidy$84.6KHousehold MAGI, couple (2026 coverage; 400% of the 2025 FPL for two)
The post-2025 cliff for a two-person household, 2026 coverage: up to 400% of the poverty level ($84,600 MAGI), the premium tax credit caps your benchmark-plan contribution at 2.1–9.96% of income; at $84,601 the credit is zero and you pay full sticker price.

Managing MAGI in the bridge years

Retirees with mixed account types have real control over MAGI. Spending from cash, taxable-account basis, or Roth contributions generates little or no income; harvesting gains, converting to Roth, or drawing pre-tax dollars generates a lot. The common pattern is a deliberate trade-off: years optimized for ACA subsidies (low MAGI) versus years optimized for cheap Roth conversions (filling brackets raises MAGI). There is no universal answer — it depends on balances, ages, and premiums — which is exactly why it is worth simulating.

Timing matters too. MAGI is an annual, calendar-year number, so a single lumpy event — a large Roth conversion, a home-sale gain beyond the exclusion, a big mutual-fund capital-gain distribution in December — can blow through the cliff for that one year even if your average income is modest. Some retirees deliberately alternate: a “clean” low-MAGI year to capture the subsidy, then a conversion-heavy year where the subsidy is written off anyway. Others simply hold MAGI a comfortable margin below the cliff every year, accepting slower conversions as the price of predictable premiums.

Try it in Deorbit Plan

In the Spending panel, turn on Use ACA marketplace before 65 and set the benchmark premium; the simulator applies the 2026 applicable-percentage table, the 400%-FPL cliff, and the 3:1 age curve each year. Every Strategy Lab result also reports the closest cliff approach — how near any year came to an ACA or IRMAA threshold — and the Strategy panel’s cliff safety margin makes fill-bracket Roth conversions stop short of the cliff instead of blundering over it.

Educational content only — not financial, tax, or investment advice.

References