IRMAA cliff age 63 + Roth conversions

Updated: July 28, 2026

OPRS may receive compensation when readers open an account through partner links on this page. Our analysis is based on independent research, BBB data, and IRS publications.

For a Family Medicine MD in late career, the Income-Related Monthly Adjustment Amount (IRMAA) lookback rule under 42 USC §1395r(i) turns the age-63 calendar year into a planning fulcrum.

The 2026 Medicare Part B and Part D premium tier each household member pays at age 65 is set by the modified adjusted gross income reported on the federal return filed two years earlier. For a physician turning 65 in calendar year 2028, that lookback year is 2026.

For one turning 65 in 2030, it is 2028. The IRMAA tier structure is not a graduated phase-in. Each bracket is a hard cliff under 42 CFR §418.1115. One dollar of MAGI above a threshold triggers the full incremental surcharge for the year, for both Medicare-enrolled spouses, applied monthly through Social Security check withholdings.

Element I of the sequence is the basis-and-bracket inventory: how much pre-tax 403(b), traditional IRA, and rollover IRA balance the household plans to convert, and which IRMAA bracket each potential conversion size lands in.

Element II is the timing decision: complete the conversion calendar in the years before age 63, throttle it during the 63-and-64 window, or accept the IRMAA premium as a known cost. Element III is the destination-allocation question, including whether a sized gold IRA sleeve under IRC §408(m) belongs inside the Roth wrapper at all.

Before any conversion is executed, it is worth screening any destination self-directed metals custodian and dealer against the 2026 list of gold IRA operators OPRS does not recommend. The framework below covers each element in sequence.

Why the age-63 calendar year is the lookback inflection

Medicare premiums for Part B and Part D are set annually by CMS through the Initial IRMAA determination process documented in 42 CFR §418.1010 and the SSA operations manual at POMS HI 01101.010.

The premium tier each beneficiary pays in a given calendar year is based on the modified adjusted gross income reported on the federal return for the tax year two years prior. There is an option to use three-years-prior data if the two-year return is not yet available at the determination time.

Here is what the two-year offset means in practice. CMS receives the household’s calendar-year-63 federal return from the IRS in 2026 or later. The agency processes that data during the annual fall premium-setting cycle. It then applies the result to the Medicare premium starting the month the beneficiary turns 65.

The mechanics are simpler than they look. A physician who turns 65 in February 2029 will see a Medicare premium set by their 2027 federal return MAGI. A physician who turns 65 in August 2030 will see a premium set by their 2028 MAGI.

The age-63 calendar year is the inflection because it is the first full tax year whose MAGI feeds directly into the first month of Medicare enrollment for a household with a typical retirement-age timeline. A Roth conversion executed during age 63 enters the IRMAA computation.

A Roth conversion executed during age 62 or earlier does not enter the at-65 IRMAA computation, although it can enter later IRMAA computations if the household stays in the workforce.

The 2025 IRMAA cliff schedule from CMS

The most recently CMS-published bracket schedule illustrates the cliff structure clearly. It was used for calendar-year 2025 premiums based on 2023 MAGI. That structure carries forward into 2026 and 2027 with statutory inflation adjustments under 42 USC §1395r(i)(5). Five tier-additions plus the standard premium combine to produce six effective tiers per filing status.

For married filing jointly households, the breakpoints sit at $212,000, $266,000, $334,000, $400,000, and $750,000 of MAGI on the 2023 return.

A household sitting at $211,999 of MAGI in the lookback year owes $0 of Part B IRMAA and $0 of Part D IRMAA. The same household at $212,001 owes $2,092 in combined per-couple annual surcharge under the 2025 schedule. The same household at $266,001 owes $5,261.

A Roth conversion that pushes MAGI from $211,000 to $213,000 carries a hidden marginal cost of $2,092 for that single year. That cost is paid through Medicare premium withholdings two years later, on top of the conversion’s federal income tax. That is the cliff.

It is one of the few places in the federal tax-and-premium code where a $2,000 increment of MAGI carries a four-figure consequence.

Where this matters: CMS publishes the next-year bracket schedule each October. The October 2025 release set the 2026 calendar-year brackets, which then determine premium adjustments for households whose age-65 birthday falls in 2026. The 2026 brackets carry forward the same six-tier cliff structure with statutory inflation adjustments. Any household sizing conversions during 2026 against an age-65 enrollment in 2028 should pull the CMS October 2025 release for the exact MAGI thresholds applicable to the 2026 lookback year.

How a Roth conversion shows up in MAGI

MAGI for IRMAA purposes is defined at 42 USC §1395r(i)(4) as adjusted gross income from line 11 of Form 1040, plus tax-exempt interest from line 2a. It is a narrower MAGI than the one used for Roth contribution eligibility under IRC §408A(c)(3), but it captures the entire dollar amount of any Roth conversion.

A $50,000 conversion of a pre-tax traditional IRA balance adds $50,000 to AGI on Form 1040. It is reported through Form 1099-R distribution code 2 (early conversion exception applies) or code 7 (conversion at or after 59½). The conversion side is reported on IRS Form 8606, and the receiving-side basis is tracked on IRS Form 5498 by the custodian.

For a Family Medicine MD with a $1.2 million 403(b) balance considering a structured Roth conversion ladder, the IRMAA-aware sizing question is straightforward. What is the household’s anticipated MAGI floor in age 63 absent any conversion? And what is the room to the next IRMAA tier boundary?

A household with $180,000 of clinical W-2 income and $25,000 of taxable investment income has roughly $7,000 of room before the $212,000 MFJ Tier 1 cliff. A $7,000 conversion costs federal income tax at the marginal rate (likely 32% or 35% for this household, plus state) and zero IRMAA.

A $7,500 conversion costs the same federal tax plus $2,092 of per-couple IRMAA exposure for the at-65 premium year.

Worth knowing: the IRMAA cliff is asymmetric. A household that has already crossed Tier 1 has no incremental IRMAA cost on further conversions until the Tier 2 boundary at $266,000 MFJ.

The marginal IRMAA cost of conversion dollars therefore swings between zero (inside a tier) and the entire tier increment (at the boundary), unlike the federal income tax marginal rate which is smooth. For households already above Tier 1 in baseline MAGI, the planning question becomes which is the next boundary and how much room remains inside the current tier.

Procedural sequence for sizing conversions around the age-63 window

The sequence the OPRS desk sees executed cleanly when an MD household wants to convert pre-tax balances and avoid an unintended IRMAA cliff at age 65 has four steps. The first three are reversible (no taxable event has occurred). The fourth is the irreversible side.

Step 1. Inventory baseline MAGI for the lookback calendar year. Pull projected W-2 income (clinical plus locums), 1099 self-employment income, taxable interest and dividends, capital gains realized year-to-date, and required tax-exempt interest under §1395r(i)(4).

Add expected balance items: any partial 403(b) in-service distribution the household has elected, any deferred-compensation distribution if a §409A balance becomes payable, and any Social Security benefits that have begun. Social Security is rare at age 63 for high earners, but possible if a spouse has filed earlier. The resulting figure is the conversion-free MAGI projection for the lookback year.

Step 2. Identify the next IRMAA tier boundary and the room to the cliff. Pull the current-year CMS bracket schedule. For MFJ households, the breakpoints under the 2025 schedule are $212,000, $266,000, $334,000, $400,000, and $750,000. Subtract the Step 1 conversion-free MAGI from the next tier boundary to compute the room available. A household projected at $245,000 conversion-free MAGI has $21,000 of room to the Tier 2 cliff at $266,000.

Step 3. Size the conversion to keep MAGI inside the intended tier. The conservative target is to leave a buffer of 5% to 10% of the room, in case year-end taxable items (mutual-fund capital-gain distributions, late 1099 corrections, unexpected bonus) push baseline MAGI higher than projected. A $20,000 conversion against $21,000 of headroom is acceptable; a $19,000 conversion is safer.

For households deliberately stacking conversions into a higher tier, the calculus reverses: the household accepts the next-tier IRMAA cost as a known cost and sizes the conversion against the next-next-tier boundary.

Step 4. Execute the conversion and document the destination. The conversion is a trustee-to-trustee transaction from the source traditional IRA or 403(b) into the destination Roth IRA, reported on Form 1099-R by the source custodian and on Form 5498 by the destination custodian. The taxable amount is reported on Form 8606 with the household’s 1040.

If the destination Roth IRA is a self-directed structure holding IRS-approved metals under §408(m), the destination custodian, depository, and dealer are named at this step. Screen the destination dealer against the 2026 OPRS list before any metals invoice is signed. Dealer choice materially affects markup, depository fee, and buy-back posture, all of which are separate from the IRMAA question but compound the cost of an unfavorable selection.

Common errors when stacking Roth conversions against the age-63 cliff

The first error is treating IRMAA as a recurring annual cost rather than a year-specific consequence of one calendar year of MAGI. The premium adjustment applies only to the years in which the lookback MAGI determines the bracket.

A single $30,000 over-conversion in age 63 produces a single year of elevated Medicare premiums at age 65 (or, more precisely, the 12 months that straddle the IRMAA determination cycle). A multi-year over-conversion produces multi-year elevated premiums.

The fix in practice is to concentrate conversions in years that do not feed the lookback window: age 60 to 62, or age 65 and beyond once the household is already in a stable tier.

The second error is forgetting that tax-exempt municipal-bond interest counts in the §1395r(i)(4) MAGI definition. Households that hold tax-exempt bonds inside taxable accounts for federal-tax-efficiency reasons often see baseline MAGI run several thousand dollars higher than their AGI alone suggests. The Form 1040 line 2a figure is added back. A household with $180,000 of W-2 plus $40,000 of municipal-bond interest has $220,000 of IRMAA MAGI, which is already above the MFJ Tier 1 cliff.

The third error is sizing the conversion against the wrong year. Households sometimes plan the conversion timing against the year of Medicare enrollment rather than the lookback year. A physician turning 65 in 2028 should be planning conversions against the 2026 MAGI window, not the 2028 window. The two-year offset is foundational to §1395r(i) and is the single most common source of IRMAA planning errors at the household level.

The fourth error is failing to coordinate with the spouse’s IRMAA exposure. The MFJ bracket is computed once on the joint return; both Medicare-enrolled spouses pay the resulting IRMAA tier each month.

A $200 increase in MAGI that pushes a couple over the Tier 1 boundary creates a $2,092 combined per-couple annual premium addition. Each spouse pays the tier-1 surcharge on their own Part B and Part D premium. The household-level impact is double the per-person figure. Any planning that ignores the second spouse undercounts the marginal cost.

The fifth error is converting into a destination IRA structure without verifying the custodian’s reporting reconciles to the basis side. A clean Roth conversion is reported on Form 8606 by the household and matched by the receiving custodian’s Form 5498.

Mismatches between the source 1099-R, the household’s 8606, and the receiving 5498 surface in the §1395r(i) MAGI computation only indirectly (through IRS-to-SSA data exchange), but they surface in the IRS notice cycle directly. The defensible practice is a same-tax-year reconciliation between source, household, and destination paperwork before the conversion calendar advances to the next year.

SSA-44 life-changing event and when it actually applies

The Social Security Administration permits beneficiaries to request a reduction in the IRMAA-based premium adjustment by filing Form SSA-44, the Medicare Income-Related Monthly Adjustment Amount Life-Changing Event form.

The form covers eight enumerated life-changing events under 42 CFR §418.1205. Those events are: marriage, divorce or annulment, death of a spouse, work stoppage, work reduction, loss of income-producing property due to disaster, loss of pension income, and employer settlement payment. None of these covers a Roth conversion, by design.

The Social Security Office of the Inspector General clarified in audit reports that voluntary tax-planning conversions are explicitly outside the SSA-44 framework.

The practical implication is that a household cannot retroactively undo an IRMAA cliff caused by a Roth conversion through the SSA-44 process. The form is meaningful only when the lookback-year MAGI is elevated by a life event whose income effect has since reversed.

A physician who retired between the lookback year and the IRMAA determination year, with the lookback MAGI including W-2 income that no longer exists, has a viable SSA-44 claim under the work-stoppage event. A physician who converted $50,000 to a Roth in the lookback year and is otherwise unchanged has no SSA-44 path.

How the age-63 IRMAA question interacts with the rest of household planning

For a physician household weighing a sized gold IRA allocation inside the Roth wrapper, the IRMAA sizing question is upstream of the asset-location question covered in the residency-era Roth IRA gold allocation framework. The conversion year, conversion size, and destination wrapper are decided through the IRMAA-aware sequence above. The asset-location decision (which sleeve sits inside the Roth) is decided separately, against the household’s overall strategic allocation. The two decisions interact only at the conversion calendar.

For households facing a Stark Law practice-acquisition event in or near the lookback year, the compliance sequencing covered in the Stark Law and 403(b) framework can change the timing of any 403(b) plan distribution. A practice-acquisition distribution event that lands in the age-63 calendar year adds to MAGI in the same window as a planned Roth conversion.

The household’s options are to delay the conversion, defer the practice-acquisition distribution into the next plan year, or accept the combined MAGI and the resulting IRMAA tier. Each option carries an arithmetic comparison the CPA can model.

Our view: the age-63 calendar year is one of the most consequential planning windows in a physician household’s late-career sequence. The IRMAA cliff is not the only consideration. State income tax on the conversion, the household’s marginal federal bracket, the projected post-retirement bracket, and the asset-location logic all combine.

But IRMAA is the consideration most often overlooked, because the cost shows up two years later through Medicare premium withholdings, on a different form, processed by a different agency. A household that runs the four-step sequence in advance avoids the surprise.

The IRMAA cliff decision is reversible up to December 31 of the conversion year. The destination decision (which custodian, which dealer if the Roth is funding a self-directed metals structure) is reversible up to the moment the metals invoice is signed.

Before any conversion is executed and any metals are purchased, run the four-step sequence above. Pull the most recent CMS IRMAA bracket schedule. Then screen any destination dealer against the 2026 OPRS list of gold IRA operators we do not recommend.

For households where the destination is a standard Roth IRA holding equities, the dealer-screening step does not apply; the IRMAA sizing step still does.

More on OPRS

For physicians considering the asset-location side of a Roth conversion that funds a metals sleeve, the residency-era Roth IRA gold allocation framework covers the §408A(d)(4) ordering rules and the pro-rata trap. For households coordinating a Stark Law practice-acquisition event with a 403(b) distribution in the same calendar year, the Stark Law and 403(b) framework covers the §411.357 acquisition exceptions.

For physicians weighing a partial 403(b) in-service distribution before the conversion calendar advances, the late-career 403(b) phase-down guide covers the §403(b)(11) mechanics. Before any metals invoice is signed, the 2026 OPRS list of gold IRA dealers we warn readers against is the dealer-screening reference.

Sources cited

1. 42 USC §1395r: Amount of Medicare premiums, including §1395r(i) Income-Related Monthly Adjustment Amount and the two-year MAGI lookback rule
2. 42 CFR §418.1010: Initial IRMAA determination by SSA
3. 42 CFR §418.1115: IRMAA tier thresholds and tier-determination methodology
4. 42 CFR §418.1205: Life-changing event categories qualifying for an SSA-44 redetermination
5. SSA POMS HI 01101.010: IRMAA Program Overview (SSA operations manual)
6. SSA Form SSA-44: Medicare Income-Related Monthly Adjustment Amount Life-Changing Event
7. IRC §408A: Roth IRAs (conversions, ordering rules, qualified distributions, five-year clocks)
8. IRC §408: Individual Retirement Accounts (including §408(m) IRS-approved precious metals)
9. IRS Form 8606: Nondeductible IRAs (Roth conversion basis tracking)
10. IRS Form 5498: IRA Contribution Information (custodian-side reporting of conversions)
11. IRS Publication 590-A: Contributions to Individual Retirement Arrangements
12. IRS Publication 590-B: Distributions from Individual Retirement Arrangements
13. CMS Fact Sheet: 2025 Medicare Parts A and B Premiums and Deductibles (IRMAA bracket schedule)