Retiring at 55 with $1.2 million? Watch the tightrope at $62,600.

"I want to retire at 55 with $1.2 million. Will it last?" The forums will argue withdrawal rates. The number that decides the first decade of the plan is $62,600.

Between 55 and 65 there is no Medicare. You buy insurance on the marketplace and the government prices it off your income; the difference between managing that number well and badly is five figures a year, so for an early retiree, income planning is health insurance planning. I built the household and ran it through the app's Monte Carlo engine, 5,000 simulated lifetimes per strategy.

What happens to health insurance at 55?

The ACA premium tax credit is keyed to your MAGI as a percentage of the federal poverty level; for a single adult in 2026 that creates a corridor with two hard edges.

The cliff sits at $62,600, 400 percent of the federal poverty level for a household of one. It is not a phase-out: one dollar of MAGI past the line and the entire subsidy is gone. The pandemic-era enhanced credits softened the cliff into a slope from 2021 through 2025; the 2026 rules in Rev. Proc. 2025-25 restore the original geometry.

The floor sits at $15,650, and it only exists in some states. This household lives in Texas, one of the ten states that never expanded Medicaid: below 100 percent of the poverty level there is no Medicaid and no marketplace subsidy either, the coverage gap. In an expansion state, too little income lands you in Medicaid; in Texas it means paying the full premium.

So the constraint for the next ten years: keep MAGI above $15,650 and below $62,600 every year, while still finding roughly $60,000 of actual cash to live on.

The household

A single 55-year-old in Texas with $1.2 million: $1,050,000 in a traditional 401(k), $100,000 in a taxable brokerage account carrying $40,000 of unrealized gains, and $50,000 in cash. He spends $48,000 a year, his marketplace plan runs about $10,800 a year at full price, and his Social Security will be about $26,000 a year at 67. He plans to age 95, a 40-year problem.

Almost everything is locked in the 401(k), which the IRS penalizes 10 percent for touching before 59½, and his liquid assets cover barely three years of a ten-year bridge. The standard tool for opening the account early is the SEPP.

How big should the SEPP be?

Exactly as big as it has to be, and no larger. The payment comes from a formula based on the account balance: run the SEPP on the entire $1,050,000 and the fixed-amortization method produces about $66,800 a year of forced, fully taxable income. His spending needs only about $60,000 of cash, but $66,800 of MAGI is past the $62,600 cliff before a single capital gain is realized, so the subsidy for the first years of the program is zero.

The IRS allows splitting the account first and running the SEPP on only the piece you need. Move $700,000 into a separate IRA and SEPP that: the formula produces about $44,500 a year, spending is topped up with small brokerage sales, and MAGI settles near $50,000, inside the corridor. Across 5,000 lifetimes:

One more detail sits where the gold line meets the pink one. The poverty level rises with inflation but a SEPP payment is frozen at the start, so the cliff climbs past the oversized payment around the third year and partial subsidies return: nothing at 55 and 56, most of the premium by 59. An oversized SEPP is a three-to-four-year, five-figure mistake rather than a permanent one.

Skipping the SEPP and eating the 10 percent penalty does worse. Living on cash and the brokerage account for two years drops documented income under the floor, which in Texas means two years of full premiums rather than Medicaid; when the liquid money runs out, penalized 401(k) withdrawals begin, with federal tax of roughly $9,200 at 58 and $11,600 at 59 against about $2,900 in the split-SEPP plan. Total cost: about $58,000 more lifetime tax and a success rate 2.5 points lower.

What does one splurge cost?

About $5,000, and the bill arrives a year late. Subsidies are paid in advance, priced on the income documented for the prior year, so this year's credit is keyed to last year's MAGI. Give the split-SEPP retiree a $35,000 truck at 58, paid for by selling brokerage shares, and the extra realized gains push that year's MAGI to about $67,800, over the cliff. The subsidy at 58 is untouched, because it was priced on 57's income; at 59 the marketplace reads the 58 paperwork and the subsidy drops from $12,866 to $7,903.

Nothing in the plan fails and the success rate barely moves, but this is the machinery behind the forum habit of spreading large purchases across two calendar years.

What about Roth conversions?

Not before 65. After 65 is a close call.

Every dollar converted from a 401(k) to a Roth is MAGI in the year of the conversion. Before 65 that MAGI feeds the subsidy formula; after 65 it feeds IRMAA, the Medicare surcharge that raises Part B and Part D premiums when income from two years earlier crossed about $109,000 (the SSA explains the brackets). The conventional advice says convert in your low-income years, and for an early retiree the low-income years are exactly the subsidy years.

Converting during the bridge, filling the 22 percent bracket each year from 55 to 64, removes the subsidy corridor: about $4,000 of subsidies collected across the decade instead of $120,000, success falling from 77.6 to 65.6 percent, and median net worth at 95 down almost $470,000. That strategy also has the lowest lifetime tax bill of anything in the run, about $323,000 less than the no-conversion plan. Paying six figures of tax in the first decade, out of a portfolio that has to survive four more, is how sequence risk arrives.

There is no gentle version: filling only the 12 percent bracket, the $16,100 standard deduction plus the $50,400 bracket top puts income at $66,500, past the cliff before the SEPP and the gains stack on top. During the bridge years there is no subsidy-safe conversion for this household.

Moving the same conversions to ages 65 through 74 preserves every subsidy dollar; the cost shows up as IRMAA instead, roughly $5,600 to $7,700 a year during the conversion decade, against a subsidy worth $10,000 to $16,000 a year, so IRMAA is the cheaper sacrifice. The conversions also shrink his RMDs at 75 from about $72,000 to $19,000, and from age 77 onward his total healthcare costs run lower than with no conversions at all, because the never-convert plan's larger RMDs trip IRMAA permanently.

Even the well-timed conversion lands at 76.3 percent success versus 77.6 for doing nothing, because tax paid early is compounding foregone. The case for converting is not the success rate; it is $211,000 less lifetime tax, tax-free money for heirs, and insurance against future tax rates.

The scoreboard

Means across 5,000 trials per strategy, engine v1.45.1. Median net worth in today's dollars; subsidy and tax totals summed in the dollars of the years they were paid.

The plan with the smallest lifetime tax bill is the worst on the board by eleven points of success, and the best plan pays more than double its tax. Minimizing taxes and maximizing the chance the money lasts are different objectives, and the decade from 55 to 65 is where they separate most.

What this means if you are eyeing the exit

Before 65 the government prices your health insurance off your income; after 65 it prices your Medicare off your income; an early-retirement plan decides which decade absorbs which income. The stakes in the corridor are roughly $10,000 to $16,000 a year, every year, for a decade.

Three rules fall out of the numbers: size the SEPP to your spending rather than your balance, splitting the account first if needed; treat every large one-time sale as a two-year decision, because the marketplace reads this year's income into next year's premium; and hold Roth conversions until after 65 unless the reason is worth more than the subsidy.

This is one household, one state, one set of assumptions, and a model. Texas has no state income tax, which flattered every strategy; a California version would add a tax layer and a different floor, because California expanded Medicaid. Congress may restore the enhanced credits and soften the cliff again, and these numbers would need a rerun if it does.