Roth Conversion Timing to Stay Below the ACA Cliff

A Roth conversion — moving pre-tax dollars from a Traditional IRA, 401(k), or SEP-IRA into a Roth IRA — is treated as a distribution under IRC § 408A(d)(3)(A) and is therefore included in gross income in the year of conversion. Because the Premium Tax Credit under IRC § 36B is computed on Modified Adjusted Gross Income, every dollar of converted basis-less Roth conversion adds one dollar to MAGI. For households between roughly 300% and 400% of FPL, this can push them across the cliff that was reinstated for tax year 2026 under Rev. Proc. 2025-25 — eliminating the entire PTC.

The cliff problem is not the income tax on the conversion itself; that is a known, predictable cost. The problem is the discontinuity: at 399% FPL the household still receives a PTC capped at 8.5% of income (under the post-ARPA pre-cliff structure that ended in 2025; the 2026 ceiling reverts to 9.12% per Rev. Proc. 2025-25 Table 2 applicable percentages). At 400.01% FPL the PTC is zero. For a Household of 2 at the 400% FPL threshold (~$84,960 in 2026), a $1 conversion can trigger an $8,000–$15,000 PTC loss depending on age, geography, and the SLCSP benchmark. The effective marginal rate on that final dollar exceeds 100,000% in extreme cases.

The right framework is a multi-year Roth ladder. Suppose a 60-year-old couple has $60,000 in baseline MAGI (Social Security + interest + small RMDs) and wants to convert $30,000 from a Traditional IRA to fill up the 12% tax bracket before age 73 RMDs hit. A one-year $30,000 conversion brings MAGI to $90,000 — about $5,000 over the 400% FPL cliff for HH=2 — destroying perhaps $12,000 of PTC. A two-year ladder of $15,000 per year keeps MAGI at $75,000 each year, comfortably under the cliff, and preserves the PTC both years. The total ordinary income tax on $30,000 is roughly the same either way (both years are in the 12% bracket); the $12,000 PTC is preserved.

The coordination opportunities are where this gets interesting. Above-the-line deductions reduce MAGI on a dollar-for-dollar basis and create slack room above any baseline income. HSA contributions ($4,300 single / $8,550 family for 2025; verify 2026 amounts in IRS Rev. Proc. 2024-25) are deductible if made to a non-cafeteria-plan HSA. A traditional 401(k) deferral reduces MAGI by the deferred amount — useful for early retirees with consulting income. Deductible Traditional IRA contributions (subject to phase-out if covered by a workplace plan, IRC § 219(g)) work similarly. A Donor-Advised Fund gift does NOT reduce MAGI (charitable deductions are itemized, below-the-line for AGI purposes). The hierarchy: above-the-line deductions create cliff slack; itemized deductions do not.

The arbitrage worth understanding: the marginal "rate" of crossing the cliff combines ordinary income tax (10-22% for typical early-retiree households), state income tax (0-13.3%), the lost PTC ($8-$15K on a base of perhaps $5,000 over the cliff, yielding 160-300% on that overage), and — for taxpayers over 63 in a state with IRMAA exposure two years later — Medicare IRMAA brackets. Stacked, the marginal cost of the cliff-crossing dollar can exceed 200%. By contrast, the marginal cost of converting one dollar within the bracket (without crossing the cliff) is just the ordinary tax rate. The decision rule: convert only the amount that keeps MAGI below the cliff (or below the next IRMAA bracket if applicable), and continue the ladder over multiple years.

Roth conversions are the single largest discretionary MAGI lever available to retirees aged 59½ to 73 (between the early-withdrawal age and the RMD start age under SECURE 2.0). For PTC-eligible households, the conversion decision is no longer just a tax-bracket question — it is a tax-bracket question stacked on top of a PTC-cliff question, and the latter usually dominates for households between 300% and 400% FPL.

The effective marginal rate on the cliff-crossing dollar is a function of three things: (1) the ordinary federal+state tax rate on the conversion (typically 12–24% for early retirees); (2) the size of the lost PTC divided by the amount by which MAGI exceeds the cliff; (3) any IRMAA bracket triggered two years later for taxpayers over 63. The PTC loss alone can produce effective marginal rates above 200% in worked examples — meaning the household pays more in lost subsidy than the value of the dollar earned.

The multi-year ladder is not just about staying under the cliff in any single year — it is also about smoothing across the Roth conversion window (typically the 5–10 years between early retirement and RMD age). Converting $20k/year for 5 years often produces the same total tax outcome as $100k in one year (assuming bracket stability) while preserving PTC every year and avoiding any single-year IRMAA bracket trigger.

Pub 974 Worksheet W (the SEHI iterative solve) is irrelevant to Roth conversions specifically — Roth conversions are not part of the SEHI circularity — but the same MAGI definition applies. The conversion dollar enters MAGI directly under IRC § 36B(d)(2), so Worksheet W users who are self-employed AND doing Roth conversions in the same year must add the conversion to gross income before beginning the Worksheet W iteration.