2027 IRMAA Brackets: Projected Medicare Tiers

The 2027 IRMAA brackets begin at a modified adjusted gross income of $112,000 for a single filer and $224,000 for a married couple filing jointly. Below those amounts a beneficiary pays the standard Medicare Part B premium. Above them, the income-related monthly adjustment amount adds IRMAA surcharges to both Medicare Part B and Part D coverage, and the amount climbs through five tiers.

Medicare measures that income from a tax return filed two years earlier, so the 2025 return sets these brackets. The table below pairs each income tier with the monthly surcharge and the total Part B premium it produces, using the Trustees' standard premium estimate of $209.50.

Every figure above is a projection. The income thresholds are computed from the formula in federal law and are firm within about a thousand dollars; the standard premium underneath them is an actuarial estimate that CMS will not settle until it publishes the official amounts in the fall. For the year already finalized, see the 2026 IRMAA brackets, and for the year before that the 2025 IRMAA brackets.

Every fall, CMS announces two sets of Medicare numbers within a single notice: the standard Part B premium, and the income thresholds that decide who pays a surcharge on top of it. The announcement for next year will land around November.

Most projections you will read between now and then treat the two as one problem — take this year's numbers, apply an inflation guess, publish. They are not one problem. The thresholds are set by a formula written into federal law and computable from published data today. The premium is an actuarial decision that no outsider can compute. One can be calculated; the other can only be bounded.

This piece does both, shows the work, and — in November, when CMS publishes the real numbers — we will grade ourselves against them in public.

Section 1839(i)(5) of the Social Security Act indexes each IRMAA threshold to inflation in a very specific way: take the 12-month average of CPI-U ending in August of the preceding year, divide by the 12-month average ending August 2018, apply that growth to fixed statutory base amounts, and round to the nearest $1,000. For those thresholds, the window is September 2025 through August 2026 — and nine of those twelve months are already published.

Before trusting a formula, test it. We ran the same computation for every year since indexing resumed: 2020, 2021, 2022, 2023, 2024, 2025, and 2026. It reproduces the first threshold CMS actually published exactly, all seven years — $87,000, $88,000, $91,000, $97,000, $103,000, $106,000, $109,000 — and all four indexed 2026 tiers exactly.

Run forward a year: the first threshold lands at $112,000 for a single filer and $224,000 for a joint return, up about 2.75% from $109,000 and $218,000. The later tiers project to roughly $141,000–142,000, $176,000–177,000, and $211,000–212,000 for single filers, with joint amounts double. The remaining uncertainty is confined to the three unpublished window months, and in five of the six scenarios we ran — from zero inflation to an aggressive 0.4% a month — the first threshold does not move.

Two quirks of the computation worth knowing. October 2025's CPI value was never published because of the government shutdown — the first hole in the statutory window since indexing resumed; we computed the average both with the month omitted and interpolated, and the answer is the same either way. And the top tier — $500,000 single, $750,000 joint — is frozen by statute through 2027 regardless of inflation. It first indexes in 2028.

The Part B premium has no inflation formula. Each September, CMS actuaries project the coming year's per-enrollee program cost, set the premium at roughly a quarter of it, and add a margin to keep the program's reserve near its target. No outside model has the claims data to replicate that. What an outsider can do is use the one published projection with a verifiable record: the Medicare Trustees Report.

The current report estimates next year's standard premium at $209.50 a month — up $6.60, or 3.25%, from 2026's $202.90. The Trustees' year-ahead estimates have been close in every normal year: off by −3.1% for 2023, −0.06% for 2024, exactly right for 2025, and −1.7% for 2026. The one big miss, 2022, was +7.3% — a mid-year shock when CMS added a reserve for a new Alzheimer's drug the projection had explicitly excluded. Note the direction of the recent errors: when they miss, they have missed high.

You may also have seen estimates of $215 or more. Those trace to the previous report's two-year-out figure of $218.60 — and two-year-out numbers are the ones with the bad record, missing by −8% to +4%. The same stale-number problem ran the other way a year ago: the 2024 report's two-year figure had 2026 at $186.90, a near-flat year, right before the actual 9.7% increase. The rule we apply: quote the one-year figure, never the two-year figure.

A fair objection: the 2026 premium rose 9.7%, so why believe 3.25% next year? CMS's own rate notice answers it, because the notice publishes the premium's components.

Of the $17.90 increase, about $10.27 was real per-enrollee cost growth — expenditures rose about 5.4%. The other $8.76 was the contingency margin swinging from negative to positive. CMS had over-collected in 2022 for a drug benefit that never materialized, and passed the excess back by running negative margins three years straight — subsidizing the 2023, 2024, and 2025 premiums. In 2026 the excess ran out, the margin flipped to rebuild the reserve, and the subsidized baseline snapped back.

In other words, roughly half of the 9.7% was a one-time event ending, not costs exploding. The rebuild does not repeat, which is exactly why the Trustees' estimate — enough to maintain the reserve rather than grow it — implies a much smaller increase on the same underlying cost trend.

Put the two machines together and the surcharge schedule follows, because the law sets each tier's payment as a fixed multiple of the standard premium. At $209.50, the first-tier surcharge is $83.80 a month and the top tier is $502.80 — on top of the standard premium, per person.

The structural fact advisors should keep in front of clients is that the surcharge is a cliff, not a slope. One dollar of MAGI over the first joint threshold costs a couple an estimated $2,011 for the year; one dollar over the fourth costs $11,062. The size of the crossing is irrelevant — onlywhich side of the line the return lands on. Our IRMAA gap analysis shows how far a household sits from the nearest threshold.

And the two growth rates are not symmetric. On the Trustees' own trajectory the premium reaches $338.50 by 2034 — up 67% from 2026, roughly 6.6% a year — while the thresholds crawl at an inflation pace near 2.75%. Treat the out-years as direction, not destination; but the direction is one-way: every tier's dollar penalty grows with the premium, and brackets that grow slower than retirement income pull more households over the line every year.

Medicare does not use taxable income, and it does not use the number on the bottom of a client's tax return. It uses modified adjusted gross income: the adjusted gross income reported on the return, plus tax-exempt interest added back. That single figure decides which of the IRMAA brackets a household lands in, and it is measured from the return filed two years earlier.

The two-year lookback is what makes this a planning problem rather than a billing problem. Income realized this year does not raise a Medicare premium this year — it raises the premium two years from now, long after the decision that caused it. A client who reads an IRMAA determination letter in the fall is reading the consequence of a tax return they filed almost two years earlier — after that year's taxes were long settled.

What lands in that calculation is broader than most retirees expect, and the parts that surprise people are the ones a retirement income plan controls directly.

Our guide to MAGI for Medicare walks the calculation line by line, and what income determines Medicare premiums covers the edge cases that trip up year-end planning.

Because IRMAA surcharges are a cliff rather than a slope, the entire planning question is which side of a line a tax return lands on. Unlike income taxes, where one extra dollar is taxed at a marginal rate, one extra dollar here reprices every month of Medicare coverage for the year. The levers that move modified adjusted gross income are the ordinary ones, applied with an eye on the threshold rather than on the tax bracket.

Size Roth conversions against the threshold, not the tax bracket. Conversions are the most common cause of unexpected IRMAA surcharges, because a conversion that looks efficient against a client's current taxes can cross an income threshold and add four figures of Medicare premiums two years later. Filling a bracket to the top of a tier — and stopping there — captures the tax benefit without the surcharge.

Use qualified charitable distributions for charitably inclined clients over 70½. A QCD satisfies a required minimum withdrawal without ever entering adjusted gross income, while the same gift claimed as a deduction does not reduce MAGI at all. Same gift, same charity, different Medicare premium.

Watch the years before enrollment. Income at 63 sets the premium at 65, so the planning window opens before a client is on Medicare at all — which is exactly when most households are not yet thinking about Medicare premiums. Our guide to reducing MAGI covers the full set of levers.

Appeal when a life-changing event applies. Retirement itself is the most common one: a working year's income sets a Medicare bill that no longer reflects reality once the paycheck stops. Filing Form SSA-44 asks Social Security to use a more recent year, and appealing an IRMAA determination walks the process end to end. A one-time spike such as a home sale does not qualify.

One structural note: enrollees in Medicare Advantage owe the same income-related amounts as those on original Medicare. A $0 plan premium does not remove IRMAA surcharges — they are billed separately, regardless of how the coverage itself is delivered, usually as a deduction from a Social Security check.

Everything on this page rests on published CMS and Social Security sources. The guides below cover the rules, the history, and the planning moves behind each number.

Why publish a November number in August? Because of the two-year lookback. The tier a client lands in next year was set by their 2025 return — that is already history. The bracket still open is 2028's, and it will be set by 2026 income, with about four and a half months left in the year.

• Project each client's 2026 MAGI now, against the 2028 thresholds the same formula will produce — not against this year's brackets, which is what most software defaults to.
• Size year-end moves — Roth conversion amounts, gain harvesting — against the nearest threshold, remembering the cliff: the cost of crossing is the same whether the line is missed by one dollar or twenty thousand.
• For charitably inclined clients over 70½, model a qualified charitable distribution against a deducted gift — a QCD reduces MAGI, a below-the-line deduction does not.
• For clients near the frozen $500,000/$750,000 line, note that its first indexing in 2028 creates a small, dated planning window.

None of this is a prescription — it is a projection exercise, and the whole point of an August number is that there is still timeto run it against the full IRMAA planning process.

Methodology note: thresholds computed under SSA §1839(i)(5) from BLS CPI-U (not seasonally adjusted, U.S. city average), September 2025–August 2026 window, nine of twelve months published as of August 11, 2026; validated against CMS-published thresholds 2020–2026. Premium and out-year trajectory from the 2026 Medicare Trustees Report, intermediate assumptions; surcharges are the statutory multiples of the standard premium. 2026 actuals from the CMS premium notice of November 2025. Part D surcharges are set by separate machinery and are deliberately excluded here. All 2027 figures are projections; CMS finalizes official amounts in fall 2026, and this page will be graded against them.