Roth Conversion Ladders: Timing, IRMAA, and the NIIT Tradeoff

The conversion window is the stretch of years when your taxable income is lower than it will be later. For most households it opens when paychecks stop and closes when RMDs start at 73 (born 1951 to 1959) or 75 (born 1960 or later), with Social Security somewhere in between. Someone who retires at 62 and delays Social Security to 70 can have eight to thirteen years of unusually low income, which is exactly the time to move pre-tax dollars to Roth at known rates. A ladder simply means you plan the whole window at once rather than deciding each December in isolation. Year one might convert more because there is no Social Security yet; later years might convert less once benefits start and 85% of them become taxable. The goal is a smoother lifetime tax curve, not the lowest possible tax bill this year.

Bracket filling is the starting point. For a married couple in 2026, the 12% bracket ends at $100,800 of taxable income, the 22% bracket at $211,400, and the 24% bracket at $403,550. Add the $32,200 standard deduction (plus $1,650 per spouse 65 or older) and you get the gross income ceiling for each target. The gap between 12% and 22% is large, but the gap between 22% and 24% is only two points, which is why many ladders aim for the top of 24% when the IRA is large. The jump from 24% to 32% is where the math usually stops working unless the future RMD picture is severe or the heirs are in high brackets themselves.

IRMAA is the first complication. Medicare sets your Part B and Part D premiums using MAGI from two years earlier, so a conversion at 63 sets the premium you pay at 65, and a conversion at 70 sets the premium at 72. In 2026 the first married tier starts at $218,001 of MAGI and costs about $1,148 per person per year in added premiums; the second tier starts at $274,001 and costs about $2,885 per person; the fourth tier, from $410,001 to $749,999, costs about $6,355 per person. These are cliffs, not phase-ins: one dollar over the line triggers the full surcharge. For a couple, the first tier alone adds roughly $2,300 a year, which on a $25,000 slice of conversion is an extra 9% on top of the bracket rate.

The net investment income tax is the second complication, and it is widely misunderstood. Conversion income is not net investment income, so the 3.8% never applies to the conversion itself. But the tax is charged on the lesser of your net investment income or the amount by which MAGI exceeds $250,000 for married couples ($200,000 single), and conversion income counts in MAGI. A couple with $30,000 of dividends and interest who converts enough to push MAGI to $300,000 will pay 3.8% on the full $30,000 of investment income, about $1,140, that they would not have owed without the conversion. The threshold is not indexed for inflation, so more households cross it every year. If you are also realizing capital gains in the same year, the NIIT exposure grows accordingly.

Because IRMAA is a step function, the 'one big jump versus many small' question has a real answer. Once you have decided to cross into a tier, every dollar up to the top of that tier carries no additional IRMAA cost, so stopping $20,000 into a tier wastes the surcharge you already paid. The efficient sizes are just under a threshold or near the top of a tier. It can also be efficient to concentrate conversions: paying the tier-four surcharge in one or two years to convert a large block can cost less in total IRMAA than paying the tier-one surcharge every year for ten years, depending on the size of the IRA. The only way to know is to run both paths through to RMD age and compare total tax plus total premiums, not one year at a time.

Deciding whether a tier is worth paying comes down to a simple comparison. Divide the extra IRMAA by the extra conversion dollars the tier lets you convert, add that to the bracket rate and Arizona's 2.5%, and compare the result to the marginal rate you expect those same dollars to face later. If a couple's future RMDs plus Social Security will keep them in the 24% bracket and in IRMAA tiers anyway, or if a surviving spouse will face single brackets on the same income, paying a 31% or 33% all-in rate today on a slice that would otherwise be taxed at 32% or 35% later is reasonable. If the future rate looks like 22%, it is not. Other hidden rate bumps belong in the same calculation: the 2025 to 2028 senior deduction of up to $6,000 per person phases out above $150,000 of MAGI for joint filers, and for pre-Medicare retirees on ACA coverage, conversion income can wipe out the premium tax credit entirely at 400% of the poverty level.

Two mechanical rules round out the picture. The five-year rules for Roth IRAs are narrower than people think: each conversion has its own five-year clock, but that clock only matters for penalty purposes if you are under 59 1/2 when you withdraw the converted principal, and earnings are tax-free once you are over 59 1/2 and your first Roth IRA has been open five tax years. For most people converting in their 60s, the practical rule is to avoid spending converted money in the first few years if the account is brand new. The second rule is about paying the tax. Paying conversion tax from a taxable brokerage account lets the full converted amount grow tax-free and effectively moves more money into the Roth. Paying it by withholding from the IRA shrinks the conversion, and under 59 1/2 the withheld amount can be hit with the 10% penalty.

Map the whole window before converting a dollar. List every year from retirement to RMD age, the income you expect in each (Social Security start date, pension, part-time work, dividends, planned capital gains), and the brackets and IRMAA thresholds that apply. The conversion amount for each year falls out of that map rather than a rule of thumb.

Start treating IRMAA as live at 63. Conversions at 62 or earlier do not affect Medicare premiums at all, which makes the early 60s the cheapest conversion years for many households, especially if ACA subsidies are not in play because you have retiree coverage or are paying full price. From 63 on, every conversion needs a two-year IRMAA check.

If your investment income is meaningful, add the NIIT to the ceiling list. A couple with $60,000 of dividends and gains who converts past $250,000 of MAGI is paying up to 3.8% extra on that $60,000 each year. Sometimes that is still worth it, but it should be a decision rather than a surprise on the tax return.

When a tier is worth paying, use all of it. If the plan says crossing the first IRMAA tier is justified, the default conversion size should land a few thousand dollars under the next tier threshold, leaving a cushion for year-end dividends and capital gain distributions you cannot fully predict. Doing the conversion in November or December, after most of the year's income is known, makes that precision much easier.

Keep cash or a taxable account earmarked for conversion taxes. For a ladder that converts $150,000 to $250,000 a year, the federal and Arizona tax can run $30,000 to $60,000 annually. Planning that funding source in advance, often from taxable holdings with high basis or from cash reserves, keeps you from raiding the IRA to pay the bill and undercutting the strategy.

Revisit the ladder every year. Tax law changes, markets move the IRA balance, and Social Security or health decisions shift the window. A plan built at 62 should be re-run at 63, 64, and every year after, with each year's conversion re-sized to the current facts rather than locked in.

• Filling the 22% bracket every year without noticing that the top of that bracket, about $243,600 of gross income for a couple under 65, sits above the $218,000 IRMAA threshold.
• Crossing an IRMAA tier by a few thousand dollars, paying the full surcharge for two people, and converting nothing more to make the surcharge worthwhile.
• Assuming the 3.8% NIIT applies to conversion income, and skipping conversions that would have been worthwhile, or assuming it never matters, and missing the tax on dividends and gains once MAGI tops $250,000.
• Paying conversion tax by withholding from the IRA when taxable cash was available, shrinking the amount that actually lands in the Roth.
• Forgetting that ACA premium tax credits and the senior deduction phase-out can add several points to the marginal rate on conversion dollars before 65 or above $150,000 of MAGI.
• Planning conversions one December at a time instead of across the full window to RMD age, which usually produces uneven brackets and more total IRMAA.
• Converting heavily in a year with a large capital gain or RSU vest already on the return, stacking income that could have been spread across lower years.