Eighty-four thousand, six hundred dollars.

For a couple retiring this year, that number decides whether your health insurance costs $7,600 or $33,700. Same coverage. Same two people. One number — and your withdrawal strategy controls it.

That's 400% of the federal poverty level for a household of two. And in 2026, for the first time in several years, it is a cliff again — not a slope. One dollar over, and your premium tax credit doesn't shrink. It disappears.

The conventional advice is to work until 65 and let your employer handle it, and for plenty of people that's the right answer. But if you're 62 and healthy and the only thing keeping you at your desk is a health plan, you should know what that decision costs — and whether it's a decision you actually have to make. Which accounts you draw from, and in what order, is the biggest lever you have.

I'm Daniel Masuda Lehrman, Certified Financial Planner™ based in Honolulu. Stay with me through all three parts, because the third one is where the Roth conversion advice you've heard everywhere else quietly stops working.

Three Decisions, Not One

Almost everyone lumps three separate decisions together under "figuring out health insurance":

• The coverage decision — what bridges you to Medicare
• The income decision — what you show on your tax return, and what it costs
• The conversion decision — whether to do Roth conversions at all in these years

They pull against each other, and the order matters.

Layer One: The Coverage Decision — Fewer Options Than You Think

Leave a job before 65 and there are four main ways to stay covered. COBRA continues your employer plan, generally up to 18 months — which gets a 62-year-old to about 63 and a half, not to 65. A spouse's employer plan. A marketplace plan through the Affordable Care Act. Or part-time work with benefits.

That's the list. A few of you have something extra — retiree medical, TRICARE or VA benefits, Medicaid if your income lands low enough — but if none apply, that is the list, and I say it bluntly because people spend months looking for options that don't exist.

Notice what that does to your timeline. COBRA is a bridge to a bridge — it runs out before Medicare starts. A working spouse solves the problem only as long as that spouse keeps working. So most couples who both stop in their early 60s will spend at least part of those three years on a marketplace plan.

And there's a COBRA trap almost nobody mentions. Once you elect it, you generally can't drop it mid-year for a subsidized marketplace plan. Losing coverage triggers a special enrollment period; walking away voluntarily doesn't. Only exhausting the full eighteen months opens that door. So decide between COBRA and the marketplace once, deliberately, at the moment you leave.

Which means the ACA is not the backup plan. For most early retirees, it is the plan. So layer one, the bottom line: your options are limited and predictable. The real variable isn't which plan you pick. It's what you tell the IRS you earned in income.

A Quick Word About "I'll Just Work Until 65"

In the Employee Benefit Research Institute's 2026 survey, workers expected to retire at a median age of 65. The retirees they surveyed had actually retired at a median age of 62 — exactly the three-year gap we're talking about. And it wasn't mostly by choice: nearly half left earlier than planned, roughly a third for a health problem and a third because of changes at their company.

So work until 65 for the insurance depends on two things holding for three more years — your health, and your employer's decisions. The two variables you control least. It isn't the safe option. It's a bet, and it should be priced like one.

Layer Two: The Income Decision — Your Withdrawal Strategy Sets the Bill

Marketplace coverage comes with a premium tax credit. From 2021 through 2025 there was no hard income limit on it. Those enhanced credits expired at the end of 2025, so for 2026 the old rule is back: above 400% of the federal poverty level, the credit goes to zero. Not reduced. Zero. For a household of two, that's about $84,600.

There's no asset test; the marketplace looks only at MAGI: adjusted gross income plus tax-exempt interest, excluded foreign income, and your entire Social Security benefit, including the part that isn't taxable. That last one catches people.

Meet Mark and Kelly, a hypothetical couple, both 62, both just retired. About $2.3 million between the IRAs and a brokerage account, house paid off, Social Security delayed so the higher earner's benefit can grow. They need $90,000 a year.

The obvious way: pull all $90,000 from the traditional IRA. That puts their MAGI at $90,000 — $5,400 over the line.

What does $5,400 cost? Assume the benchmark plan for a couple their age runs about $2,807 a month, roughly $33,700 a year. Under the line they'd owe just under 10% of income — about $7,600 — and the credit covers the rest. Over the line the credit is zero, so they pay the full $33,700. That's a difference of about $26,100: an effective marginal cost of nearly 500%. Not a tax bracket. Not IRMAA. Purely a function of which account they pulled from.

Now the version nobody shows you. Same $90,000 landing in checking, but $72,000 from the IRA and $18,000 from the brokerage — and because they've held those shares for years, only about $4,000 of that is capital gain. The rest is their own money coming back, and it isn't income. So their MAGI is $76,000. Under the line, subsidy intact. Identical lifestyle, identical spending, $26,100 of difference. Not a loophole — the same money, correctly reported. The only thing that changed is the order of operations.

Two more things. First, you don't set your income once in January and walk away: the subsidy is paid in advance and reconciled when you file, with no cap on repayment above that line. A December capital gains distribution, a forgotten consulting check, a last-minute conversion — any one can claw the whole twenty-six thousand back, retroactively.

Second, under the line it gets better than you'd expect. At $76,000 the benchmark silver plan costs them $7,600, but that same subsidy applied to a bronze plan covers the entire premium — roughly zero dollars a month. Higher deductible, and silver is still the better buy for many. But the honest contrast isn't $7,600 versus $33,700. It's potentially nothing.

So layer two, the verdict: in the bridge years, withdrawal sequencing isn't a tax optimization. It's the single largest input into your health insurance bill.

Layer Three: The Roth Conversion Decision — Where the Good Advice Collides

You've heard this, including from me: the years after you stop working and before required distributions begin are your Roth conversion window. Paycheck gone, Social Security not started, RMDs years away — possibly the lowest bracket you'll ever see again. That advice is correct, and I stand behind it. (I've written before about which retirees should convert to Roth — and which shouldn't.)

But look at what it asks Mark and Kelly to do. A conversion adds to income dollar for dollar. Convert $50,000 and their MAGI goes from $76,000 to $126,000. They just blew through the cliff to fund a strategy meant to save them money. And the two sides aren't the same kind of money: the subsidy is a real cost, this year, in cash, while the conversion benefit is a projected saving years out, under tax law that may well change first.

There is a legitimate version of converting anyway. If you're looking at seven figures of pre-tax money and a bracket jump you can already see coming, paying full freight for insurance to start early can be the right call. And if a pension puts you over the line no matter what, the cliff isn't a variable for you — convert.

My point isn't that conversions are wrong before 65. It's that these three years are the one stretch where the standard advice has a very expensive exception, and almost nobody checks whether they're in it.

For most people it sorts out like this. From 62 to 65, keep MAGI low and protect the subsidy. At 65, Medicare takes over, the ACA cliff stops applying, and the conversion window opens properly — with Medicare premiums (IRMAA) as the new constraint, and IRMAA's steps are far smaller in dollar terms than the cliff you just cleared.

One wrinkle strengthens the case: IRMAA runs on a two-year lookback, so your 2026 premiums are set by your 2024 return. A conversion at 63 doesn't only cost you the subsidy that year — it can also raise your Part B and Part D premiums at 65. It gets penalized twice.

You're not skipping the conversions. You're doing them in the ten years after 65 instead of the three before it — and if you were born in 1960 or later, your required distributions don't start until 75, so that window is longer than most people think.

So layer three, the verdict: before 65, the subsidy usually wins. After 65, the conversion usually wins. Getting the order backwards is the expensive mistake.

The Hawaii Wrinkle

One note for those of you reading from here in Hawaii: the federal poverty guidelines are higher for Hawaii. For a household of two, that same 400% threshold sits around $97,280 rather than $84,600 — roughly $12,700 of additional headroom before the cliff. It's one of the quieter advantages of planning a retirement in Hawaii.

Permission to Run the Numbers

For thirty or forty years you climbed. Earn more, save more, defer more. Then you reach the summit and the objective quietly inverts — now the job is turning what you built into an income without setting off tripwires nobody warned you about. The descent is the harder half.

This is one of those tripwires. But it's a known one. It has a number, the number is published, and you can plan around it. So if health insurance is the only thing standing between you and your retirement date, that's worth knowing precisely — it's an arithmetic problem, not a life sentence. Get a real quote in your own market, at the income you'd actually show.

The Most Expensive Number in Your Plan

The coverage decision has a short list of answers, and for most early retirees the marketplace isn't the fallback — it's the plan. The income decision is where the money is. And the conversion decision has to wait its turn, because good advice applied in the wrong three years stops being good advice.

Most people treat health insurance in their early 60s as a price — a fixed, unpleasant number they either absorb or avoid by working longer. It isn't. In these three years it's a consequence of decisions you're already making about which account to draw from, and when.

And that matters more than the premium does, because of what people spend to avoid it. I've watched careful, capable people work three extra years for a health plan — three years out of the front of retirement, while they still had the knees for the trail. Retirement may last 30 years, but it is not 30 equal years. Those three don't come off the end. They come out of the good ones.

Between 62 and 65, your income isn't something that happens to you. It's something you set. And for three years, it's the most expensive number in your plan.

If you'd like help building a withdrawal strategy that protects your subsidy — and a Roth conversion plan that waits its turn — I'd be happy to talk. You can schedule a free consultation; I work with clients in Honolulu and virtually across the country. You can also start with my free Retirement Readiness Assessment.

— Daniel Masuda Lehrman, CFP®, Founder of Masuda Lehrman Wealth. Mahalo for reading.

This article is for educational purposes only and is not personalized financial, tax, or legal advice.