Sized to the Bracket: A bridge-years case study in sequencing what the window allows

Tom and Diane — a couple in their early 60s — came to me in March with a version of a question I bet you’ve heard a hundred times: when is the right time to start converting to Roth?

The picture: Tom is 61. He took a voluntary exit from a corporate engineering role in February — earned roughly $280,000 in his last twelve months on the W-2, severance bridge runs through summer. Diane is 60, plans to leave her own role in about eighteen months. $1.4M in his 401(k), $380K in her 403(b), $140K in a joint brokerage with embedded long-term gains, $90K in an HSA he’s been overfunding for years. One adult son, financially independent. Health insurance on COBRA through Tom’s former employer at $2,300 a month, running out August 2027. After that, ACA marketplace until Medicare kicks in for Tom in 2030.

Tom’s planning instinct was to make one decisive move. He’d read enough to know the bridge years before Social Security were “the time to convert.” He came in asking whether $200,000 in 2026 was the right number.

Underneath that question was a sharper one, and it’s the one most clients in this window never quite say: how do I get the maximum out of these years without breaking the rest of the plan?

The framing move came first. Before talking conversion size, I mapped what the bracket actually looks like across the bridge — year by year, with every income event in it. Severance through 2026. Diane’s W-2 through August 2027. Then both of them off earned income, with no Social Security until 2032 at the earliest. RMDs not in the picture until Tom turns 75. That’s a six-year stretch where the taxable income on a Form 1040 is whatever they choose to put on it — and almost all of that choice runs through the Roth conversion lever.

A single $200,000 conversion in 2026 would have stacked on top of Tom’s severance income and pushed them well into the 32% bracket. It would also have set the AGI floor for 2028 ACA premium subsidies. And it would have triggered IRMAA on Tom’s first year of Medicare in 2030, with the 2028 conversion driving the 2030 Part B surcharge.

One move. Three downstream costs.

The right shape was a stack. Four steps, sized year by year.

Step 1 — Bracket-fill the severance year.

A 2026 conversion sized to fill the 22% bracket on top of severance, capital gains, and any other income landing that year. Smaller than the $200K Tom proposed, larger than zero. The number set by the bracket headroom, not by an annual target. Action: convert in 2026, sized to the bracket. Why: known federal rates beat unknown future ones, and the marginal cost is the lowest it will be all year. Trade-off: tax bill upfront on the conversion amount.

Step 2 — Hold conversions in 2027 to protect the ACA subsidy.

COBRA runs out in August. From September forward they’re on the ACA marketplace with no other coverage, and AGI in 2027 directly governs the 2027 subsidy. A conversion that lands them above the subsidy threshold costs the subsidy on top of the conversion tax. The math doesn’t work. Action: hold conversions in 2027 — or hold them to within the subsidy envelope. Why: the ACA cost is a real number that often outweighs the conversion savings. Trade-off: one year of the window not used for Roth, but used for something else (see Step 3).

Step 3 — Harvest long-term gains in 2027 at the 0% federal rate.

With no W-2 income, no Social Security, and conversions held inside the subsidy envelope, taxable income for the household lands well below the 0% LTCG threshold. The $140,000 brokerage carries embedded gains — selling those positions, recognizing the gain, and rebuying steps up the basis at zero federal cost. State tax still applies in their state — about $4,000 — but the federal bill is $0 on a basis reset that would otherwise compound forward indefinitely. Action: harvest gains in 2027 to the top of the 0% LTCG band, net of any conversion. Why: the bracket allows it, and the basis reset compounds the same way the Roth dollars would. Trade-off: gives up subsidy headroom that conversion would have used — the planning trade in 2027 is one or the other, not both.

Step 4 — Resume bracket-fill conversions 2028 through 2031, sized to the IRMAA tier as well as the bracket.

Once Diane is off W-2 income and they’re on the ACA marketplace at modest subsidy levels, conversions resume at the bracket headroom — with a second constraint added. The 2028 conversion drives Tom’s 2030 IRMAA tier on Medicare Part B. The 2029 conversion drives Tom’s 2031 tier. Each year, the conversion is sized to the lower of (a) bracket headroom or (b) the AGI level that holds IRMAA at an acceptable tier. Action: annual conversions, sized to both constraints. Why: the bridge is multi-year — the right total comes from cumulative discipline, not a hero year. Trade-off: slower conversion pace than the single-shot approach, materially better outcome net of costs.

The teaching move is why the four steps work together. Step 1 alone is what most plans do, and it leaves dollars on the table. Step 1 plus Step 4 ignores the ACA window. A single big conversion ignores both. The stack is what produces the result — meaningful Roth balance built across six years, ACA subsidies preserved when they matter, IRMAA tiers managed, and a basis reset on the brokerage account paid for in tax dollars they were never going to owe.

Two takeaways for our work.

The first is the bracket vs. the calendar. Most conversions get planned as annual decisions in isolation. The bridge years reward a different discipline — a multi-year conversion budget, allocated year by year against the bracket headroom that’s actually available. The number changes every year. The discipline is what doesn’t change.

The second is the second-order math. A bridge-years conversion plan that doesn’t account for ACA subsidies, IRMAA tiers, the 0% LTCG window, and state-level interplay is a plan that’s solving for one thing and breaking three others. Most of the plans I see touching this window solve for the conversion. Few solve for the system around it.

The window closes when Social Security starts. Until then, every year is a sizing decision. Get the sizing right and the result compounds tax-free for the rest of the client’s life.

This is a hypothetical case study based on a real planning engagement; identifying details have been changed. Outcomes shown are illustrative and depend on assumed rates of return, tax law, and household-specific circumstances. Tax brackets, ACA subsidy thresholds, IRMAA tiers, and LTCG rates referenced are as of the publication date and are subject to change. Past performance is not indicative of future results.