FREE NEWSLETTER
IF YOU’RE IN YOUR early 60s and retired, you probably have a lot of financial questions on your mind. The next few years may be among your lowest-income and lowest-tax-paying years. Your salary and bonus years are behind you. Social Security and required minimum distributions from your IRAs and 401(k)s have not started yet. You are hearing advice about doing Roth conversions during this low-tax window, and the arguments are compelling. You may also be thinking about consulting or part-time work to stay active and bring in some income.
This article is about the hidden cost of those decisions: how income choices you make now can affect both your health insurance costs today and your Medicare premiums later. If you don’t understand the interaction, the surprise can cost thousands of dollars.
The ACA cliff is back… and it’s steep
The enhanced ACA subsidies that softened premium costs from 2021 through 2025 expired at the end of last year. Congress didn’t extend them. That means the hard cliff is back in full effect for 2026.
The cliff sits at 400% of the federal poverty level. Cross it by even $1 and you lose your entire premium tax credit. It’s not a partial reduction; it’s all of it. If you aren’t prepared, that can create real cashflow problems.
For 2026 coverage, based on the 2025 federal poverty guidelines, those thresholds are:
Per KFF’s analysis, a 60-year-old earning $62,000 pays roughly $515 a month in health premiums, about 10% of income. The same person earning $64,000, or just $2,000 more, pays around $1,244 a month, roughly 23% of income. That’s not a typo. Two thousand dollars of extra income triggers roughly $8,750 in extra annual premiums.
The income figure that determines your eligibility is your MAGI. It includes everything you might be doing in retirement to manage your finances: Roth conversions, capital gain realizations, dividends, interest, part-time income and Social Security if you’re already drawing it.
The IRMAA clock starts when you’re 63, not 65
The ACA cliff is only part of the issue. Medicare uses a two-year lookback to set your premiums. Your 2028 Medicare Part B and Part D costs will be determined by your 2026 income, the same year you’re managing your ACA cliff right now.
The 2026 IRMAA thresholds reflect 2024 income for those already on Medicare. They give us a reasonable proxy for what 2028 will likely look like, as the Centers for Medicare and Medicaid Services won’t publish the actual 2028 brackets until late 2027. The first IRMAA tier kicks in at $109,000 for single filers and $218,000 for couples. Cross that threshold in 2026, and when you turn 65 in 2028, you’ll be looking at roughly an extra $81.20 per month per person in Part B premiums or $974 per person per year, on top of the standard $202.90/month premium.
That’s the first tier. The surcharges climb from there. And both Part B and Part D carry their own IRMAA surcharges, so couples can easily see $2,000 to $4,000 in added annual Medicare costs from a single income year that was too high.
It is ironic but the income year most likely to push you over an IRMAA threshold is often one of your last years before Medicare when you might be selling an asset, doing a large Roth conversion, or drawing down a pre-tax account to fund living expenses. Why do these two cliffs need to be planned together?
Put these two together and you can see the problem clearly. Take a 63-year-old couple with $80,000 of MAGI: they’re under the $84,600 cliff, subsidies intact. Now add a $20,000 Roth conversion. That one decision pushes them to $100,000 and it wipes out the entire ACA subsidy this year.
The same conversion, sized larger or stacked with a capital gain that crosses $218,000, would also raise their Medicare premiums starting in 2028. That is why the two cliffs need to be modeled together, not checked separately after the fact.
Where the $30,000 comes from:
| Scenario | Estimated Cost |
|---|---|
| Couple crosses the ACA cliff in 2026, full subsidy lost | ≈ +$21,500/yr |
| Same 2026 MAGI over the first IRMAA tier triggers the 2028 Medicare surcharge (Part B + D, couple) | +$2,297 |
| If 2027 income also stays over the ACA cliff | ≈ +$21,500 more |
| Combined two-year exposure from the same income pattern | Potentially $45,000+ |
The chart below plots 2026 MAGI against both costs at once: the bars are your annual ACA premium (indigo while subsidized, red past the cliff), and the line is the annual Medicare surcharge that same income locks in for 2028.
If you’re 63 in 2026:
Too much income this year and you lose ACA subsidies, costing potentially $10,000 to $25,000 more in health premiums in 2026 and 2027. Too much income this year and you trigger IRMAA, paying $2,000 to $8,000+ more in Medicare premiums annually starting in 2028.
Both cliffs draw from the same income year at once, not in sequence. Your 2026 MAGI sets your ACA subsidy right now, and that same 2026 return sets your 2028 Medicare premium through the two-year lookback. Because the two systems are run separately (one by the IRS and the Department of Health and Human Services, the other by Social Security and the Centers for Medicare and Medicaid Services) most people never see the combined exposure until it’s already locked in.
What you can do about it
The goal is to keep your 2026 MAGI below both cliffs where possible, or at least to be deliberate about which cliff you’re willing to cross and why.
For people with earned income, deductible Traditional IRA contributions can be one of the most direct MAGI reducers. If you or your spouse has earned income, you can contribute to a Traditional IRA and deduct it, reducing MAGI dollar-for-dollar. The 2026 limit is $7,500 per person, or $8,600 if you’re 50 or older.
For a couple where one spouse is still working, that’s potentially $17,200 off your MAGI. One catch: if you’re covered by a workplace retirement plan, the deduction phases out at higher incomes. For 2026, between $81,000 and $91,000 of MAGI for single filers, or $129,000 and $149,000 for joint filers when the contributing spouse is covered.
The counterintuitive part: you’re putting money into a pre-tax account when your tax rate is relatively low, with the understanding that you’ll pay taxes on it later and possibly at higher rates. For some people, that trade doesn’t pencil out. For others, protecting a $10,000 ACA subsidy this year is worth the future tax cost. The math depends on your specific situation, and it’s worth modeling rather than assuming.
Health savings account contributions work similarly. Pre-tax contributions reduce MAGI directly. The catch is that you must be on an HSA-eligible high-deductible health plan to contribute. If your ACA marketplace plan qualifies, and you’re not yet on Medicare, this can be a meaningful lever. The 2026 limits are $4,400 for self-only coverage and $8,750 for family coverage, plus an extra $1,000 catch-up if you’re 55 or older. Plan to stop contributions before Medicare begins. Medicare’s Part A coverage can backdate up to six months, which can turn recent contributions into excess contributions, so watch that timeline carefully.
Capital gain timing is often the biggest swing. If you’re planning to sell appreciated assets, a taxable brokerage position, a rental property, anything with embedded gain, the year you do it matters enormously. Deferring a large realization from 2026 to 2029, after Medicare begins, sidesteps both the ACA cliff and the IRMAA lookback simultaneously. That’s not always possible, but it’s worth asking whether the transaction needs to happen this year.
Roth conversions don’t reduce MAGI, they add to it. If you’re in the pincer zone, aggressive Roth conversion in 2026 can push you over the ACA cliff and set your 2028 IRMAA tier at the same time. That’s not an argument against Roth conversions generally. It’s an argument for sizing them carefully relative to where you are on both cliff structures. If you’re already below both thresholds with room to spare, a modest conversion can make sense. If you’re hovering near either line, the math changes quickly.
One longer-horizon point, separate from the two-year window this article is about: if you’re in the pre-pincer years, your late 50s or early 60s, modest Roth conversions now can reduce the size of your future RMDs. Smaller RMDs mean less forced taxable income in your late 60s and beyond, which means less pressure on the IRMAA tiers you’ll face once you’re on Medicare. That is a multi-decade trade, not a fix for the immediate cliff, and it works best when you have a decade or more of runway before Medicare enrollment.
Plan this out
The two-year lookback means you lose the ability to affect your 2028 Medicare premiums after December 31, 2026. You can’t file an amended return and get a different IRMAA. There is an appeal process through Social Security, but it’s designed for genuine life-changing events like retirement or divorce, not for voluntary income decisions that turned out to be more expensive than expected.
For ACA purposes, 2026 is the year in question. January 1, 2027 starts a new calculation. That means the window for planning is now. Not 2027, when you’re closer to Medicare.
________________________________________________________________________________
John Urban is the founder of RetireSmartIRA, a retirement tax-planning app. Earlier, he founded GT Nexus, a supply-chain software company acquired by Infor in 2015. He lives in Northern California with his wife, Kathy, and enjoys time with family, travel, reading, Bay Area sports, and the occasional deep dive into the fine print of the tax code.
Why must finances be so darn complicated?
Thank GOD I never had to deal with this crap…and I never will.
I am retired; I have ZERO IRMAA concerns, ZERO ACA concerns, and ZERO Income Tax concerns.
It’s all in the planning.
I am just the opposite – I am retired, and I pay a lot of income tax, I pay IRMAA, I pay NII.
Which of us has the better approach? There’s a lot to be said for both ways.
This is one of those things that many tax advisors love to site for reasons why you shouldn’t handle things yourself. I’m good at math and handle things myself, so I was quite conscious of and concerned about this as I went through these years.
However, as I ran scenarios I realized this was a bit of a false flag for me. Why? In the short term I was concerned about keeping my marginal tax rate at 12%, but I realized in the long term I would sooner or later be happy to keep it at 24% (using today’s married tax brackets). That meant that worrying about these cliffs was insignificant long term compared to getting those tax-deferred dollars out sooner.
At the same time, I do always wait to decide how much to place in traditional IRAs vs Roth IRAs until I’m ready to file. That gives me a $17,200 cushion (in 2026) to dance around the edges of any cliffs when I actually file. Of course those traditional contributions are just going to be converted to Roth “next year.”
When considering ROTH conversions I am always reminded of the age old conundrum of how many angels can dance on the head of a pin. Meaning, it really doesn’t matter. For most people well prepared for retirement $30,000 is pocket change and the number of unverified assumptions about the future required to base a conversion strategy on renders it more like playing poker than playing chess. In my case it never mattered because I was unable to keep my income below six figures even after retiring. A situation that I believe probably exists for the majority of readers in this space. However I do appreciate the analysis, which is no doubt helpful for the minority of people for which this decision might make a material difference.
Steve, I think that’s a fair challenge. For many well-prepared retirees, especially those who are going to remain comfortably above the key thresholds no matter what, the Roth conversion decision may not move the needle as much as people hope.
Where I think the issue still matters is income sequencing. Even if someone can’t avoid higher income altogether, they may still be able to control which year income lands in: whether a capital gain is realized before or after Medicare, whether a QCD replaces a taxable withdrawal, whether a conversion crowds an IRMAA or NIIT tier. Those choices may not matter for everyone, but when they do, the dollar impact can be surprisingly large.
So I agree with your larger point: this is not universal advice. For some households it is noise. For others, especially those near ACA, IRMAA, NIIT, or RMD-driven thresholds, it is worth modeling before acting. Appreciate the thoughtful comment.
Yes, there is a huge IRMAA bracket for singles between $200K and $500K, and for couples between $400K and $750K. Once you are well in it, you can push to increase your income.
We will never have to concern ourselves with IRMAA, but at other times in our lives we have dealt with the ACA and IRA contributions. I’m just saying that there is plenty of great information in this post for those of us with income on the lower end of the HD readership.
I appreciate your comments Dan and I’m really pleased that article had value for you. Thanks!
Several years ago, after leaving a company-sponsored health plan, my wife and I needed ACA coverage for the two years before we became eligible for Medicare. We received little or no subsidy as I recall because of our income and were paying roughly $2,200 per month for the least expensive Bronze plan, which also carried about a $16,000 family deductible. I understand comparable coverage can cost considerably more today.
It was a huge shock. At the time, most of my frustration was directed at the insurance companies because we were paying enormous premiums for coverage we rarely used. But that is, of course, how insurance works. You are protecting yourself against a potentially catastrophic medical expense, even if you hope never to collect on the policy.
Then, once we entered Medicare, the two-year IRMAA lookback kicked in—and not in a good way.
As much as I dislike paying the additional premiums, the basic concept is difficult to argue with: people with higher incomes are asked to pay more. If you have been a diligent saver and investor and now receive a steady stream of dividends, interest, capital gains and other income, there is a good chance you will be affected by the IRMAA tables. Even tax-exempt municipal-bond interest counts when IRMAA income is calculated.
I would still rather be in that position than the alternative.
David, the muni-bond point is the one that catches most people off guard. Tax-exempt at the federal level, but it still counts toward IRMAA MAGI, so a portfolio heavy in munis can push someone into a higher tier without a single dollar of taxable income. Your $2,200/month Bronze experience is exactly the kind of real-world data point that makes the abstract thresholds feel concrete. And your closing line is the right frame. Better to be paying IRMAA than nowhere near it. Appreciate you sharing it.
I’ve been working through different scenarios in the Boldin software, and the best option (for us) seems to be to go ahead with Roth conversions in the next few years, take the ACA and IRMAA hit for a while, but then drastically reduce RMDs and income taxes in the future.
David, that’s a sound approach, and running it through a few scenarios is the get to a plan that’s right for you. One subtlety worth checking as you model it: the ACA hit and the IRMAA hit don’t land on the same clock. ACA subsidy loss is same-year, based on this year’s MAGI. IRMAA runs on a two-year lookback, so a conversion in 2026 doesn’t surface as a surcharge until 2028. In the pre-Medicare conversion years, that means you can be absorbing an ACA clawback and pre-loading a future IRMAA surcharge from the same dollar, in the same year. It doesn’t change your overall plan, but it can change the sequencing: how much to convert in which year, and where the bracket-fill line sits, so you’re not stacking both hits harder than you need to in any single year. It’s worth watching the two timelines separately as you run it.
For anyone who is self employed and dealing with the ACA cliff, there are a number of actions that can appreciably lower your AGI: business deductions including 1/2 of self employment tax, medical premiums, and solo 401k’s which allow more generous contributions than traditional IRAs. Hate having to put more $ into tax deferred accounts, but that’s the new game.
Excellent information! And very tricky! One point, I’ve been thinking a lot about is the ACA cliff when younger than 65. If you can take itemized deductions is it still a hard cliff? Isn’t it a soft descent that depends on how much of your medical expenses, which includes premiums, you can deduct?
Isn’t your personal itemization situation a big toggle switch in pre-65 years?
We Americans are so healthcare expense adverse, I sometimes think we work harder to avoid paying for healthcare than staying healthy?
Deductions are “below the (AGI) line”
John points out in the article that the ACA cliff is determined by your Modified Adjusted Gross Income (MAGI). Itemized deductions are subtracted after your MAGI is figured, so they don’t help you avoid the ACA cliff (or the IRMAA cliff).
Yes, but if your itemizations (of which healthcare premiums are a big portion) eliminate your AGI such that you aren’t paying taxes or are in the lowest tax brackets; haven’t you realized a discount on the face value of you health care premiums? Isn’t this 1 reason Republicans did away with the credits? Some were double dipping?
Maybe I’m not following you, but MAGI (AGI + tax exempt interest, etc., added back in) is calculated before deductions are subtracted. ACA uses MAGI, not taxable income (which is what you later get after subtracting deductions).
So your itemized deductions, no matter how large, don’t reduce or eliminate your MAGI.
Check your 1040, line 11a (AGI), and then farther down, line 15 (taxable income).
John, thanks for this article. One question about this sentence:
Deferring a large realization from 2026 to 2029, after Medicare begins, sidesteps both the ACA cliff and the IRMAA lookback simultaneously.
I understand how the deferral avoids the ACA cliff since in 2029 you’ll be on Medicare. But as to sidestepping the IRMAA lookback, it will simply be deferred rather than avoided, won’t it? That is, in 2031, the IRMAA lookback will include the 2029 gain. What am I missing?
In any event, starting several years ago, every November I sit down and figure what our approximate income will be for that year. In the earlier years I was figuring how big a Roth conversion I could make and stay under the IRMAA threshold. These days, since I’m taking RMDs, I’m figuring how large of a Qualified Charitable Distribution (QME) we should make in order to avoid IRMAA.
Andrew, you’re not missing anything. You’re right, and my wording was sloppy there. The 2029 gain does land in the 2031 IRMAA lookback, so it isn’t avoided. It’s paid, just later.
What the deferral actually buys is narrower than “sidesteps both.” In a pre-Medicare year, a large realization can hit twice at once: the ACA subsidy clawback and the setup for a future IRMAA surcharge, stacked in the same income. Moving it to 2029 clears the ACA cliff entirely (you’re on Medicare) and turns the IRMAA piece into a single, known, budgeted surcharge year in 2031, rather than one hit compounding on top of another. So the benefit is separating and planning the two events, not escaping IRMAA. I should have said that.
Your November routine is exactly right. Sizing the QCD each year to hold MAGI under the next threshold is the right move. It counts toward your RMD but stays out of MAGI, which is what makes it such a clean IRMAA lever once RMDs begin. Sounds like you’re already on it.
Thanks, John, for your comment and clarification. Much appreciated.
John thanks for your article today. I have never found it possible to convert to Roth without large tax consequences and now at 80 years old, I pay as I go. I try to maximize my RMD, to not trigger IRMMA and higher tax brackets. Last year I had to go in the next bracket due to RMD.
William, thank you for reading, and for sharing this. You’re describing exactly why the conversion math changes with age. Once RMDs are large and there’s no long runway left for the lower rates to pay off, converting often just pulls the tax forward without a real payback. Paying as you go can be the best answer at that point.
The one lever still worth a look at 80, if you give to charity is the QCD. It comes straight out of the IRA, counts toward your RMD, but never lands in your income. So it’s the rare move that lowers the RMD dollars hitting your MAGI without a conversion. Even a modest QCD can be enough to keep you off the next bracket or IRMAA tier in a close year like the one you just had. You may already be doing this. If not, it might be worth a look before next year’s RMD.
Good to be aware of these rules and IRMMA does apply to me. At the same time, I consider IRMMA not as a tax, but as greater cost sharing for hMedicare insurance premiums, which is less than what I paid for health insurance during my working years.
While IRMMA is not technically a tax ….
What people need to understand is that IRMAA is a de facto federal income tax,” Reichenstein says [retired Baylor University finance professor William Reichenstein]. “As your income goes up, you owe more to the federal government for Medicare.” Your IRMAA taxes will likely be as high—or higher—for the rest of your life, since the RMD rate gradually increases as you age.
From Barron’s, 6.24.26 article about taking RMDs in conjunction with social security and/or pension payments …
Does this apply to retirees with working income over age 70? I’m on medicare.
The IRMMA bug can bite anyone on Medicare. The cliffs begin with incomes above $109,000 for single, and $218,000 for married taxpayers.
John, thanks for this very actionable information.
Allowing contributions to 401(k) HSAs and SEPs to make one eligible to make deductible IRA contributions never seemed logical to me. They can also make one eligible for the Retirement Savings Contributions Credit. I wonder if that was an intentional or accidental act by the lawmakers.
Thanks for the article, from a fellow NorCal resident (raised in Marin County, spent most of adult life in Davis)!
A related ACA income cliff issue we ran into was where one spouse crosses into Medicare and the other remains under 65. Then, the joint income is invoked on a one person household; greatly lowering the income threshold of the year prior. It’s a crazy part of the law.
This is not correct. The threshold is based on the number of people in the household, not the number enrolled in the ACA.
https://www.healthcare.gov/income-and-household-information/household-size/
“Include your spouse and tax dependents even if they don’t need health coverage.”
Doug, this is one of the least-discussed cliffs out there, and you’ve described it exactly. The moment one spouse ages into Medicare, the ACA household drops from two people to one, and the income threshold that governs the under-65 spouse’s subsidy compresses hard. Same dollars of income, much lower bar to clear.
It’s also a preview of something none of us really want to think about, but all of us eventually face. When one spouse dies, the survivor moves to single brackets and the single IRMAA tiers, and income that was comfortable for a couple can suddenly sit two or three tiers higher. The split-Medicare years are the first quiet rehearsal of that shift, while both spouses are still here to plan around it together. If you’re already modeling the split-household ACA years, you’re most of the way to modeling the survivor transition too. Same lever, same thresholds, just triggered by a harder event.
We’ve been managing our financial lives around this for a while and have two more years to go (DW turns 63 in September, I’m 69 and thus already on Medicare). It’s meant that Roth conversions are out of the question and (since we’re already retired) being very careful to make sure that investments held in our taxable accounts are tax-efficient.
Under “can’t win for losing” though our “reward” for my wife turning 65 will be Medicare + Supplement premiums that are a multiple of her ACA premium…and my RMD’s kick in right afterwards. Just hoping that ACA insurance plans are still available until we reach that finish line, as insurers are dropping out of the marketplace left and right due to the “death spiral” for ACA caused by the actions (or rather inactions) of Congress.