What Counts Toward Marketplace MAGI for Early Retirees

If you retire before 65 and buy health insurance on the ACA Marketplace, one number decides what that insurance costs: your modified adjusted gross income, as the Marketplace defines it.

Get it wrong and the consequence is not a rounding error. It is the difference between a heavily subsidised silver plan and paying something close to full freight — potentially thousands of dollars, triggered by a single decision you made in December.

The trouble is that "MAGI" means at least three different things in US tax rules, and the Marketplace version is its own animal.

The definition

Marketplace MAGI starts from AGI and adds back three things:

marketplace_magi = AGI
                 + non-taxable Social Security benefits
                 + tax-exempt interest
                 + excluded foreign earned income

That is the whole formula. What makes it consequential is what AGI already contains, and what it does not.

What counts

• Wages and self-employment income. Including part-time or consulting work in an otherwise low-income year.
• Roth conversions. In full. A conversion is ordinary income in the year you make it — this is the single largest controllable input for most early retirees.
• Realised capital gains. Long-term gains are taxed at preferential rates, but they are still part of AGI. A 0% federal tax rate on a gain does not make it invisible to the Marketplace.
• Dividends and interest. Qualified and non-qualified alike.
• Traditional IRA and 401(k) withdrawals.
• Rental and business income, net.
• Tax-exempt municipal bond interest. Excluded from AGI, added back here. Munis do not shelter you from the ACA calculation.
• Non-taxable Social Security. The Marketplace counts the full benefit, not just the taxable portion.

What does not count

• Roth withdrawals, including converted principal after the five-year clock.
• Selling shares to return your own basis. Only the gain is income; a $50,000 sale of stock with $40,000 of basis contributes $10,000, not $50,000.
• Spending down cash. Withdrawing from savings is not income. This is why a cash bridge is so valuable in ACA years — it funds spending without touching MAGI at all.
• Loan proceeds, including a HELOC draw.
• Gifts and inheritances.

That second and third bullet are the heart of early-retirement income planning. Two people can spend exactly $70,000 in a year and report wildly different MAGI, depending entirely on which accounts they spent from.

Why the number matters so much

Marketplace subsidies are calculated against the federal poverty level for your household size, and the FPL used is the prior year's table. For 2026 coverage, that is the 2025 guidelines — where a single-person household is $15,650.

So your position is usually expressed as a percentage:

fpl_percent = marketplace_magi / fpl_amount × 100

A single filer with $39,125 of MAGI is at 250% of FPL. A household of four with $64,300 is at 200%.

Two things are tied to that percentage, and only one of them is widely known.

Premium tax credits phase down gradually as income rises. Gradual is the key word — one extra dollar of income costs you a small fraction of a dollar in credits, not a cliff.

Cost-sharing reductions are the ones people miss. These are not premium discounts; they silently upgrade the actuarial value of a silver plan — lower deductible, lower out-of-pocket maximum, lower copays. They are tied to hard thresholds, commonly at 150%, 200%, and 250% of FPL. Crossing one by a single dollar drops you to the next tier, and the value forfeited can be thousands of dollars of protection in a year you actually get sick.

That is why the sensible move is to pick a guardrail — a percentage of FPL you will not exceed — and plan your income against it, rather than optimising against tax brackets alone and finding out in April what it cost.

The mistake this causes

Here is a concrete case, and it is the most common error in early-retirement tax planning.

A single filer, 2026 coverage, holding a 250% FPL guardrail. Their year looks like this:

Item	Amount
Part-time earned income	$12,125
Ordinary dividends and interest	$1,000
Qualified dividends	$3,000
Long-term capital gains	$15,000
Planned Roth conversion	$8,000
Marketplace MAGI	$39,125

The 250% ceiling on a $15,650 FPL is exactly $39,125. They are sitting precisely on the line with $0 of headroom.

Now look at the same person through a purely tax-focused lens:

• Taxable income is $39,125 − $16,100 standard deduction = $23,025
• The 2026 single 0% long-term capital gains ceiling is $49,450, leaving $26,425 of room to realise more gains at a 0% federal rate
• Room to the 24% ordinary bracket is $105,700 + $16,100 − $21,125 = $100,675 of additional ordinary income

A tax-bracket view says this person has enormous room. Six figures of conversion capacity and $26,425 of free capital gains, apparently sitting on the table.

The Marketplace view says they have nothing. Both readings are correct. They are measuring against different ceilings, and in an ACA year the lower one is the one that binds.

Convert another $1,000 and MAGI hits $40,125 — 256.4% of FPL, over the guardrail. Harvest another $10,000 of gains and you are at 313.9%.

This is the entire reason the Roth Conversion & Capital Gains Headroom Planner reports the binding constraint rather than three separate numbers. Additional conversion room is the smaller of your MAGI headroom and your bracket room. When the guardrail is exhausted, the answer is zero regardless of how much bracket room remains.

Practical implications

Estimate MAGI before December, not in April. Almost every lever — conversions, harvesting, when to sell — is only available before the year closes.

Sequence your accounts deliberately. Cash and Roth withdrawals fund spending without touching MAGI. Traditional withdrawals, conversions, and realised gains all consume it. In a year where you are near a guardrail, spending from cash is not a neutral choice; it is the whole strategy.

Watch the year you sell property. A large capital gain can blow through every threshold at once. If a sale is coming, that is a year to plan no conversions at all.

Munis do not help here. Tax-exempt interest is added straight back. If you hold munis specifically to manage ACA income, the strategy does not do what you think.

Reconcile at filing. Subsidies are advanced based on your estimate and trued up on Form 8962. Underestimate and you repay; overestimate and you get credit back. The estimate is not a formality.

Verify current figures. Poverty guidelines are updated annually, and ACA subsidy rules have changed repeatedly and may change again. Treat any figure here — including 2026 brackets and the 2025 FPL table — as something to confirm for your own year.

Where to start

Work out where you actually sit before deciding what to do about it. The headroom planner estimates Marketplace MAGI from your income sources, converts it to a percentage of FPL, and shows how much room is left under a guardrail you choose — plus a sensitivity table for what $1,000 or $5,000 more of conversion does.

If you are still deciding whether the bridge years are even funded, start with the Can I FIRE Yet? Calculator. And if two people are involved, the Couples FI Planner handles the staggered dates that determine how many ACA years you are actually planning for.

This is an explanation of how the calculation works, not tax advice. Confirm any conversion or harvesting decision with a qualified tax professional — the interaction between subsidies, brackets, and state tax is genuinely intricate, and the rules change.