An early retiree who buys health coverage on the ACA marketplace before Medicare lives inside a sharp income constraint. Cross 400% of the federal poverty level by a single dollar in 2026 and the premium tax credit does not taper — it stops. That makes the "bridge years" between retirement and age 65 a planning window with a hard ceiling: how much income to realize, and how much of a traditional balance to convert to Roth, without tipping over the edge. This study hands six representative households to QuantCalc's ACA-bridge optimizer and reports the recommended year-by-year plan for each, next to two reference points: never converting, and a cliff-unaware plan that fills the 12% tax bracket every year.
The 2026 rules this rests on
The cliff returned for 2026. The enhanced premium tax credits — the 2021 American Rescue Plan schedule that capped premiums at 8.5% of income and removed the 400% FPL limit, later carried through 2025 — lapsed at the end of 2025. A bill to continue them, S. 3385, did not pass in December 2025. For 2026 the credit reverts to the indexed §36B schedule: a required contribution rising from 2.10% of income at the low end to 9.96% near the top, and terminating entirely above 400% FPL. There is no contribution cap above the line — a household one dollar over pays the full unsubsidized premium, which is the structural reason the ceiling matters so much.
The applicable-percentage figures are the 2026 values in Rev. Proc. 2025-25 (retrieved 2026-07-22, https://www.irs.gov/pub/irs-drop/rp-25-25.pdf); the income-tax brackets the engine applies come from Rev. Proc. 2025-32. If Congress were to restore the enhanced schedule for 2026 retroactively, the cliff modeled here would not apply and the plans would change; the engine carries a switch for that schedule, and every figure below is dated to the current-law 2026 regime.
What the data shows. Across all six households the recommended plan holds every bridge year's income under the household's FPL ceiling. A cliff-unaware plan that fills the 12% bracket instead crosses 400% FPL, forfeits the entire premium subsidy, and costs $65,243 to $234,441 more over a three-to-five-year bridge. For most households, converting is not required at all — the never-convert plan already holds the subsidy (or already sits in Medicaid), so the discipline is the income ceiling, not a conversion target. And at the Medicaid threshold the trade-off runs the other way: for the New York · single · 5-year bridge, whose 250% FPL ceiling leaves ample room to reach the marketplace, the plan stays in Medicaid and converts nothing — because the ACA premium credit nets against the premium (it is not a cash rebate on top), a Medicaid year's out-of-pocket ($4,500 over the bridge) comes in below a subsidized-marketplace year, so forcing income up would raise net bridge cost. We surface the coverage regime year by year.
Why this study is different
QuantCalc already publishes a companion on the cost of the cliff and one on the conversion decision. This one adds the year-by-year income and conversion schedule that keeps a real household under the line.
Six households, plan by plan
Each household retires before 65 and buys marketplace coverage until Medicare. "Net bridge cost" is lifetime federal and state tax, plus net healthcare (premiums after subsidy), plus any IRMAA — the total the plan controls. "% FPL" is each year's income as a share of the federal poverty level for the household size; the coverage column reports the regime that year's plan actually assumes.
California · single · 3-year bridge
The recommended plan keeps every bridge year's income under the 395% FPL ceiling and captures $43,849 in premium credits — matching the never-convert plan, which also stays subsidy-eligible. Converting is not required to hold the subsidy here; the discipline is the ceiling, not a conversion target. A cliff-unaware plan that fills the 12% bracket every year ($125,405 converted) pushes income across 400% FPL, forfeits $43,849 of subsidy the recommended plan keeps, and costs $65,559 more over the bridge.
| Plan | Converted | Lifetime tax | Net healthcare | Net bridge cost | vs. recommended |
|---|---|---|---|---|---|
| Recommended plan | $0 | $807 | $12,384 | $13,191 | — |
| Never convert | $0 | $807 | $12,384 | $13,191 | +$0 |
| Fill the 12% bracket | $125,405 | $22,517 | $56,232 | $78,750 | −$65,559 |
Recommended plan, year by year:
| Age | Conversion | MAGI | % FPL | Coverage | Premium subsidy | Net healthcare |
|---|---|---|---|---|---|---|
| 62 | $0 | $24,699 | 158% | Marketplace | $14,249 | $4,128 |
| 63 | $0 | $24,699 | 158% | Marketplace | $14,671 | $4,128 |
| 64 | $0 | $24,699 | 158% | Marketplace | $14,928 | $4,128 |
Texas · couple · 5-year bridge
The recommended plan keeps every bridge year's income under the 250% FPL ceiling and captures $175,410 in premium credits — matching the never-convert plan, which also stays subsidy-eligible. Converting is not required to hold the subsidy here; the discipline is the ceiling, not a conversion target. A cliff-unaware plan that fills the 12% bracket every year ($561,918 converted) pushes income across 400% FPL, forfeits $175,410 of subsidy the recommended plan keeps, and costs $234,441 more over the bridge.
| Plan | Converted | Lifetime tax | Net healthcare | Net bridge cost | vs. recommended |
|---|---|---|---|---|---|
| Recommended plan | $0 | $0 | $17,706 | $17,706 | — |
| Never convert | $0 | $0 | $17,706 | $17,706 | +$0 |
| Fill the 12% bracket | $561,918 | $59,030 | $193,116 | $252,147 | −$234,441 |
Recommended plan, year by year:
| Age | Conversion | MAGI | % FPL | Coverage | Premium subsidy | Net healthcare |
|---|---|---|---|---|---|---|
| 60 | $0 | $25,771 | 122% | Marketplace | $33,148 | $3,541 |
| 61 | $0 | $25,771 | 122% | Marketplace | $34,340 | $3,541 |
| 62 | $0 | $25,771 | 122% | Marketplace | $35,122 | $3,541 |
| 63 | $0 | $25,771 | 122% | Marketplace | $36,102 | $3,541 |
| 64 | $0 | $25,771 | 122% | Marketplace | $36,698 | $3,541 |
California · couple · 5-year bridge
This household sits at the Medicaid threshold, and the recommended plan keeps it there every bridge year — total out-of-pocket healthcare $4,500 over the bridge, with no Roth conversion. Lifting income into the subsidized marketplace instead does not pay off here: the ACA premium credit reduces the premium you owe, it is not a cash rebate on top, so once it is counted correctly a marketplace year costs more out of pocket than a Medicaid year. On a net-cost basis the recommendation at this income is to stay in Medicaid, not to convert into the marketplace. (Medicaid's non-premium trade-offs — asset tests, provider access — sit outside this cost model.) A cliff-unaware plan that fills the 12% bracket every year ($533,822 converted) even captures $30,912 in premium credits in the years it stays under 400% FPL, but crosses the cliff in the rest and still costs $205,644 more over the bridge than the recommended Medicaid plan — once taxes and the cliff are counted.
| Plan | Converted | Lifetime tax | Net healthcare | Net bridge cost | vs. recommended |
|---|---|---|---|---|---|
| Recommended plan | $0 | $153 | $4,500 | $4,653 | — |
| Never convert | $0 | $153 | $4,500 | $4,653 | +$0 |
| Fill the 12% bracket | $533,822 | $72,613 | $137,683 | $210,297 | −$205,644 |
Recommended plan, year by year:
| Age | Conversion | MAGI | % FPL | Coverage | Premium subsidy | Net healthcare |
|---|---|---|---|---|---|---|
| 60 | $0 | $14,465 | 68% | Medicaid | $0 | $900 |
| 61 | $0 | $14,465 | 68% | Medicaid | $0 | $900 |
| 62 | $0 | $14,465 | 68% | Medicaid | $0 | $900 |
| 63 | $0 | $14,465 | 68% | Medicaid | $0 | $900 |
| 64 | $0 | $14,465 | 68% | Medicaid | $0 | $900 |
Texas · single · 3-year bridge
The recommended plan keeps every bridge year's income under the 395% FPL ceiling and captures $52,028 in premium credits — matching the never-convert plan, which also stays subsidy-eligible. Converting is not required to hold the subsidy here; the discipline is the ceiling, not a conversion target. A cliff-unaware plan that fills the 12% bracket every year ($130,628 converted) pushes income across 400% FPL, forfeits $52,028 of subsidy the recommended plan keeps, and costs $65,243 more over the bridge.
| Plan | Converted | Lifetime tax | Net healthcare | Net bridge cost | vs. recommended |
|---|---|---|---|---|---|
| Recommended plan | $0 | $0 | $11,745 | $11,745 | — |
| Never convert | $0 | $0 | $11,745 | $11,745 | +$0 |
| Fill the 12% bracket | $130,628 | $13,215 | $63,773 | $76,988 | −$65,243 |
Recommended plan, year by year:
| Age | Conversion | MAGI | % FPL | Coverage | Premium subsidy | Net healthcare |
|---|---|---|---|---|---|---|
| 62 | $0 | $22,958 | 147% | Marketplace | $16,916 | $3,915 |
| 63 | $0 | $22,958 | 147% | Marketplace | $17,407 | $3,915 |
| 64 | $0 | $22,958 | 147% | Marketplace | $17,705 | $3,915 |
New York · single · 5-year bridge
This household sits at the Medicaid threshold, and the recommended plan keeps it there every bridge year — total out-of-pocket healthcare $4,500 over the bridge, with no Roth conversion. Lifting income into the subsidized marketplace instead does not pay off here: the ACA premium credit reduces the premium you owe, it is not a cash rebate on top, so once it is counted correctly a marketplace year costs more out of pocket than a Medicaid year. On a net-cost basis the recommendation at this income is to stay in Medicaid, not to convert into the marketplace. (Medicaid's non-premium trade-offs — asset tests, provider access — sit outside this cost model.) A cliff-unaware plan that fills the 12% bracket every year ($276,601 converted) pushes income across 400% FPL and forfeits the premium subsidy entirely, costing $160,817 more over the bridge than the recommended Medicaid plan.
| Plan | Converted | Lifetime tax | Net healthcare | Net bridge cost | vs. recommended |
|---|---|---|---|---|---|
| Recommended plan | $0 | $2,021 | $4,500 | $6,521 | — |
| Never convert | $0 | $2,021 | $4,500 | $6,521 | +$0 |
| Fill the 12% bracket | $276,601 | $42,262 | $125,076 | $167,338 | −$160,817 |
Recommended plan, year by year:
| Age | Conversion | MAGI | % FPL | Coverage | Premium subsidy | Net healthcare |
|---|---|---|---|---|---|---|
| 60 | $0 | $18,152 | 116% | Medicaid | $0 | $900 |
| 61 | $0 | $18,152 | 116% | Medicaid | $0 | $900 |
| 62 | $0 | $18,152 | 116% | Medicaid | $0 | $900 |
| 63 | $0 | $18,152 | 116% | Medicaid | $0 | $900 |
| 64 | $0 | $18,152 | 116% | Medicaid | $0 | $900 |
New York · couple · 3-year bridge
Every bridge year resolves to medicaid. The recommended plan runs $41,051 of Roth conversions at essentially zero added bridge cost (net cost within $0 of never converting) — room the never-convert plan leaves unused, because at this income the conversions fit under the standard deduction and do not erode any subsidy. A cliff-unaware plan that fills the 12% bracket every year ($371,100 converted) pushes income across 400% FPL and forfeits the premium subsidy entirely, costing $193,027 more over the bridge than the recommended Medicaid plan.
| Plan | Converted | Lifetime tax | Net healthcare | Net bridge cost | vs. recommended |
|---|---|---|---|---|---|
| Recommended plan | $41,051 | $0 | $2,700 | $2,700 | — |
| Never convert | $0 | $0 | $2,700 | $2,700 | +$0 |
| Fill the 12% bracket | $371,100 | $51,327 | $144,400 | $195,727 | −$193,027 |
Recommended plan, year by year:
| Age | Conversion | MAGI | % FPL | Coverage | Premium subsidy | Net healthcare |
|---|---|---|---|---|---|---|
| 62 | $13,684 | $27,634 | 131% | Medicaid | $0 | $900 |
| 63 | $13,684 | $27,634 | 131% | Medicaid | $0 | $900 |
| 64 | $13,684 | $27,634 | 131% | Medicaid | $0 | $900 |
How to read the tables. The recommended plan is highlighted. "vs. recommended" is that plan's net bridge cost measured against the recommended plan — a − value means the plan costs more. The year-by-year table shows the recommended plan's realized income (MAGI), its position against the FPL ceiling, the coverage regime, and the premium subsidy it captures. Because the engine is deterministic, these are exact figures, not sampled estimates.
What it means
- The ceiling is the lever, not the conversion. Holding the subsidy is about keeping realized income under 400% FPL (or under the household's chosen ceiling). For most of these households that is achievable while converting little or nothing; a plan that converts aggressively to "fill a bracket" is what crosses the line and forfeits the credit.
- The Medicaid threshold is a genuine fork. Below roughly 138% FPL in an expansion state, a household is in Medicaid rather than a subsidized marketplace plan. Nudging income just above that line trades Medicaid for a credited marketplace plan; the model values the captured subsidy, but Medicaid's non-premium features — provider networks, cost-sharing, asset rules — sit outside it. We surface the per-year regime so that fork is visible rather than buried in a single number.
- This is a bridge-window plan, not a lifetime-conversion verdict. The plan minimizes cost during the pre-Medicare bridge — tax, net premiums, IRMAA — and treats a conversion as an in-window cost. It does not price the post-65 payoff of converting (lower future required distributions or Medicare surcharges), so the recommended conversion amounts are deliberately conservative: convert where it is free or captures subsidy, not to chase a lifetime Roth target. Sizing the full lifetime conversion is a separate calculation.
- The figures are dated to current 2026 law. Every number assumes the restored 400% FPL cliff and the 2.10%–9.96% schedule. A retroactive extension of the enhanced credits would change the picture, and the engine can switch schedules if that happens.
CC-BY-4.0 — free for any use including republication and journalism, with attribution to QuantCalc Research.
Plan your own bridge years
The engine behind this study is the year-by-year ACA-bridge planner inside the QuantCalc Roth Conversion Planner. Enter your own balances, spending, state, and coverage, and it lays out the income and conversion schedule that holds your subsidy through to Medicare — and sizes the conversions that pay off after 65.
See the Roth Conversion Planner →Methodology
Engine. The recommended plan for each household comes from QuantCalc's ACA-bridge optimizer, which chooses a per-year Roth-conversion schedule to minimize lifetime bridge cost (federal and state tax, net healthcare after ACA premium credits, and IRMAA) subject to a hard per-year cap that keeps ACA-basis income under the household's chosen FPL ceiling. It is a seeded joint refinement over a per-year warm-start, with a deterministic policy for the Medicaid/marketplace boundary (the boundary belongs to the below-threshold side, so the plan never claims a subsidy it is not eligible for). Because it is deterministic, re-running it reproduces the same schedule and every figure is exact.
Cross-checks. The engine's output is cross-checked against exhaustive income (conversion) sweeps and an independent re-evaluation built only from public tax and subsidy functions, in the verification rig backend/tools/aca_verify.c — it confirms the recommended schedule respects its constraints, that its stated objective reproduces under the plan's own per-year coverage regime, and that a brute-force sweep does not beat it. Every dollar on this page is emitted directly by the engine via backend/tools/aca_study_dump.c.
Constants. The 2026 federal poverty levels, the applicable-percentage schedule (2.10%–9.96%, cliff above 400% FPL), the Silver-benchmark premiums, and the age-rating curve are the current-law 2026 values — applicable percentages from Rev. Proc. 2025-25 (retrieved 2026-07-22), income-tax brackets from Rev. Proc. 2025-32 — and are pinned identically between the engine and the calculator's own subsidy code by a parity test.
Reproducibility. The first persona's recommended plan was solved twice and returned a bit-identical schedule and objective (objective difference 0, largest per-year conversion difference 0). The generator also recomputes each plan's lifetime components from its per-year figures and confirms every recommended year stays under its ceiling, aborting on any mismatch, so the page can never drift from the data snapshot.
Model-dependence. These are recommended plans under a specific model of 2026 tax and ACA rules, a constant real return, and modeled healthcare costs — not personalized advice. Actual premiums, Silver benchmarks, and Medicaid features vary by county and household, and a plan that minimizes modeled bridge cost is not automatically right for your circumstances. Consult a qualified tax professional before acting.
Changelog
- v2026.1 — initial release (2026-07-22). 6 households × 3 plans, deterministic ACA-bridge engine, generated from
backend/tools/aca_study_dump.cviascripts/gen_aca_bridge_study.mjs. Current-law 2026 regime (400% FPL cliff, 2.10%–9.96% schedule).
Dataset license: CC-BY-4.0. QuantCalc is an independent retirement-planning research project. Model-dependent results under current 2026 tax and ACA rules; not financial, tax, or legal advice.
Cite this research study
QuantCalc Research (2026). The ACA Bridge Income Plan (2026). https://quantcalc.app/research/aca-bridge-income-plan-2026/ (accessed <date>).
BibTeX
@misc{quantcalc2026theacabridgeincomeplan2026,
title = {The ACA Bridge Income Plan (2026)},
author = {{QuantCalc Research}},
year = {2026},
url = {https://quantcalc.app/research/aca-bridge-income-plan-2026/},
note = {Accessed <date>}
}
Machine-readable citation metadata (schema.org identifier and citation fields) is embedded in this page's JSON-LD, at the stable identifier https://quantcalc.app/research/aca-bridge-income-plan-2026/.