COBRA vs Marketplace 2026: The Self-Employed AGI Math That Changes the Answer

If you Googled "COBRA vs Marketplace" in 2024, half the answer was "just take the Marketplace plan, the subsidies make it cheap." That answer expired on January 1, 2026.

The enhanced premium tax credits — the temporary expansion that ran from 2021 through 2025 — sunset at the end of last year. The pre-2021 rules came back. The biggest change: the 400% FPL cliff is back. One dollar over $62,600 (single) or $128,600 (family of four) and you lose every penny of subsidy. Not phased out. Gone.

The Congressional Budget Office estimated that average premium payments for subsidized Marketplace enrollees would jump roughly 114% — about $1,016 a year — once the enhanced credits expired. That's not premium prices going up. That's the same premium just costing you more because the federal government stopped covering as much of it.

The House passed a three-year extension on January 8, 2026, by a 230-196 vote. As of this writing, the Senate has not passed it, and there's no clear path. The Bipartisan "CARE Act" is in negotiation but uncertain. Plan for the rules as they exist today, not for what Congress might do.

For self-employed people, this rewrites the COBRA-vs-Marketplace decision. The auto-answer of "Marketplace, obviously" doesn't survive 2026. But there's also a flip side most W-2 advice columns miss: if you're self-employed, you have AGI levers W-2 employees don't, and those levers can pull you back under the cliff and rescue the subsidy. That's the math problem of this article.

What Actually Changed January 1, 2026

The 2021 American Rescue Plan and 2022 Inflation Reduction Act made two changes to ACA premium tax credits, both temporary:

The 400% FPL cap was removed. Subsidies were available above 400% of the federal poverty level — anyone who would otherwise pay more than 8.5% of income for the benchmark silver plan got help, regardless of income.

The applicable percentages were reduced across the board. The maximum premium contribution at any income level dropped, with people at lower incomes paying close to zero for the benchmark plan.

Both changes ended December 31, 2025. The 2026 coverage year operates under the original ACA rules: 100-400% FPL eligibility window, with a hard cliff at 401%, and the higher pre-ARPA applicable percentages (capping at roughly 9.83% of income at the top of the slope).

The 2026 numbers — using the 2025 HHS Federal Poverty Guidelines, which is what the IRS uses for 2026 coverage subsidies:

If your projected 2026 modified AGI is below the 400% column, you're subsidy-eligible, with the size of the subsidy depending on where in the slope you land. If you're over by even a dollar at year-end, you owe back every dollar of advance subsidy you received during the year. No cap, no proration.

That last sentence is the one I want freelancers to read twice. The cliff isn't just a planning issue — it's a reconciliation issue. You can claim subsidies all year and end up writing a check at filing time if your business has a strong Q4.

COBRA: The Simple, Expensive Answer

I took COBRA in 2020 when I left my logistics job. The COBRA notice showed up about nine days after my last day — formal letter from the third-party administrator with a 60-day clock and a payment coupon. The number on the coupon was $612 per month for individual coverage. Same Aetna plan I'd had for years. Same network, same deductible, same prescription tier. The only thing that changed was who was paying for it.

I'd been paying around $108 per month for that plan as a payroll deduction. The company had been covering the rest. The rest, as it turned out, was most of it.

That's COBRA in one sentence: the same plan you had, but the full bill, plus a 2% administrative fee, lands on you instead of being split with your employer. The 102% rule (100% of the total premium plus 2% for administration) is the federal cap, not a starting point — most plans charge exactly that.

KFF's 2025 Employer Health Benefits Survey put the average individual full premium at $746 per month and family premiums north of $2,000. Employers covered 83% of individual premiums and 74% of family premiums. So the typical W-2-to-COBRA shock is going from $114/month to $761/month for an individual, or from $520/month to $2,040/month for family coverage. That's the math of why COBRA is a tough first answer for most people.

How COBRA Actually Works

You're eligible if you were enrolled in your employer's health plan and your separation was for any reason other than gross misconduct (which essentially means felony-level). Voluntary quits, layoffs, reduced hours that drop you below the eligibility threshold — all count. Self-employment is the natural follow-on for most freelancers, and COBRA doesn't care what you do next; it cares about what you did before.

The election window is 60 days, measured from the later of your qualifying event date or the date the COBRA notice was provided. Coverage is retroactive if you elect — meaning you can wait, see if you have any medical expenses, and elect at day 58 to retroactively cover them. People call this "the COBRA loophole." It's real, but if you elect retroactively, you owe all back premiums. So the loophole only saves you money if you incur a major expense and were planning to skip coverage; it doesn't get you free care.

Duration is 18 months for the standard separation event. Disability extends it to 29 months. Other qualifying events (death of the covered employee, divorce, dependent aging out) can extend coverage to 36 months for the spouse or dependent.

The hidden value of COBRA — the part that doesn't show up in price comparisons — is continuity. Same plan ID, same network, same prior authorizations, same deductible accumulator. If you've already met half your out-of-pocket max for the year, you keep that progress. If you have a specialist relationship, prescription mid-cycle, or treatment in progress, nothing changes from the provider's side. A switch to a Marketplace plan resets all of that even if you pick a plan with the same insurer.

Marketplace Post-2026: The New Math

The Marketplace under the restored rules is still meaningful, just less generous. Let me walk through what subsidies actually look like for a single self-employed person in 2026 across income brackets.

Under 100% FPL ($15,650). In Medicaid expansion states, you're routed to Medicaid — no Marketplace subsidy needed. In non-expansion states, you fall into the coverage gap with no Medicaid and no Marketplace subsidy. That's a structural problem the ACA never fully solved, and it didn't get worse in 2026.

100-150% FPL ($15,650-$23,475). Subsidies are still substantial. Benchmark silver plan premiums after credits typically run $0-$50/month. The enhanced subsidies had pushed many of these enrollees to truly $0; the post-2026 rules charge them a few percent of income.

150-250% FPL ($23,475-$39,125). This is where the post-2026 hit is most visible. Premiums roughly doubled compared to 2025 — going from $50-$100/month to $100-$200/month for benchmark coverage. Still subsidized, still meaningful, but the math is tighter.

250-400% FPL ($39,125-$62,600). Subsidies remain available but the contribution percentage is higher. For a freelancer at $50,000 AGI, expect to pay 8-9% of income for the benchmark plan — roughly $350-$400/month after subsidies. That's competitive with COBRA only if your COBRA quote is on the high end.

Over 400% FPL ($62,601+). No subsidy. Full sticker price for the benchmark plan, which varies by state but typically runs $500-$900/month for an individual. At this point, you're paying roughly the same as COBRA, with the disadvantages of switching networks and resetting your deductible.

The pattern: Marketplace beats COBRA decisively below 250% FPL, beats COBRA modestly between 250-400%, and ties or loses to COBRA above 400%. The 400% cliff turns "obviously Marketplace" into "depends on where I land."

The AGI Lever: What Self-Employed People Have That W-2 Don't

Most articles comparing COBRA and Marketplace miss the part that matters most for freelancers: a self-employed person's AGI isn't fixed. It's the result of choices made before December 31. Some of those choices can move you several thousand dollars in either direction without changing your actual business income.

The big lever is retirement contributions. A Solo 401(k) lets you contribute up to $72,000 in 2026 ($80,000 if you're 50 or older with the standard catch-up), split between employee deferral ($24,500 / $32,500 with 50+ catch-up) and employer profit-sharing (up to 25% of net SE earnings). Every dollar of traditional contribution reduces AGI by a dollar. If you're projecting $68,000 AGI as a single filer — $5,400 over the cliff — a $7,500 traditional IRA contribution plus a partial Solo 401(k) employee deferral can drop you back under $62,600 with room to spare.

The math: $7,500 of contributions to drop under the cliff, recovering maybe $4,000-$6,000 in annual subsidy. That's a near 1-to-1 return on money you're keeping (it's still in your retirement account) plus the future tax-deferred growth. Compared to losing the full subsidy, this is an obvious move — if you have the cash flow to fund the contribution.

The smaller levers, in rough order of usefulness:

HSA contributions (if you're on a high-deductible plan). $4,400 individual or $8,750 family for 2026. Above-the-line deduction, can be funded up until the tax filing deadline.

SEP-IRA. Up to 25% of net SE earnings, max $72,000 in 2026. Mechanically simpler than Solo 401(k) but no employee deferral, so total contribution room is lower at moderate income levels.

Section 179 / bonus depreciation. If you're buying equipment, software, or vehicles for the business, you can often deduct the full cost in the current year rather than spread over years. Useful for pulling AGI down in a high-income year, but only if the purchase is genuinely needed.

Schedule 1 Line 17 — Self-Employed Health Insurance Deduction. The premiums you pay for yourself, spouse, and dependents (out-of-pocket portion only, after any subsidy) reduce AGI. This is the one that creates the circular calculation problem I'll address in a second.

The Circular Calculation Problem

Here's a thing nobody tells you until you sit down to do the math: if you're claiming both the Self-Employed Health Insurance Deduction (SEHID) and the Premium Tax Credit (PTC), the two interact in a circular way.

Your PTC is based on your AGI. Your AGI is reduced by your SEHID. Your SEHID is the amount of premium you paid out of pocket, which is the total premium minus your PTC. So the credit affects the deduction, which affects the AGI, which affects the credit. It's a math snake eating its tail.

The IRS provides two methods to resolve this. Method 1 is a simplified iteration that gets you close. Method 2 is a true iterative solve that gets you exact. Both are documented in Form 7206 instructions and IRS Publication 974. Both are tedious. Both can change your AGI by enough to flip the cliff.

This is one of the genuine cases where I tell people: if you're within $5,000 of the cliff, run the numbers with a CPA. Not because the IRS guidance is impenetrable, but because the iteration is the kind of thing that's easy to get wrong if you're doing it once a year on TurboTax. A CPA who does this regularly will run it iteratively, get the right answer, and probably charge you less than the subsidy you'd lose by getting it wrong.

I'm not a tax professional. I'm a guy who ran his own iteration in 2022 and double-checked it against my CPA's number for $400. He matched my answer to the dollar. But I would not have known if he'd been off, and the difference between subsidy and no-subsidy in my year was about $4,800. The cost of being wrong was bigger than the cost of being checked.

The Decision Matrix

The choice usually breaks out roughly like this, given the 2026 rules:

The "first year self-employed" recommendation is unpopular but I stand by it. When I left my W-2 job, my income for that year ended up being about 30% lower than I projected — which would have been great for Marketplace subsidies if I'd been on a Marketplace plan. But I'd been bracing for the opposite (income higher than projected, owing the subsidy back), which is the more common freelancer mistake. The first year of self-employment, your AGI projection is unreliable. COBRA is expensive insurance against being wrong.

The Off-Ramps Nobody Talks About

Two paths the COBRA-vs-Marketplace framing tends to miss, and one I want to actively talk you out of.

Spouse's employer plan. If your spouse has access to an employer-sponsored plan, the loss of your job-based coverage is a qualifying event for special enrollment on their plan. This is almost always cheaper than COBRA and often cheaper than Marketplace (since employer subsidies are typically 70%+). The catch: the enrollment window is short (usually 30-60 days from the qualifying event), and your spouse's HR department may push back if the timing isn't clearly documented. Have the COBRA election notice and last day of coverage paperwork ready.

Catastrophic plans. Available if you're under 30 or qualify for a hardship exemption. They cover three primary care visits per year and otherwise function as a backstop for major medical events. Premiums are roughly half of bronze plan premiums in most markets. Useful for healthy young freelancers who want the lowest possible coverage cost and are willing to absorb significant out-of-pocket for routine care.

And the one I won't recommend: healthcare sharing ministries (Medi-Share, Christian Healthcare Ministries, Liberty HealthShare, etc). These are not insurance. They are not regulated as insurance. They have no obligation to pay any specific claim. Pre-existing conditions are typically excluded. Lifestyle requirements (no smoking, no extramarital sex, sometimes religious affiliation) are conditions of membership.

I have two friends who used healthshares. One had a clean experience for three years and felt great about the savings. The other had a kidney stone, the ministry decided the imaging was a "share-ineligible expense" because of an ambiguous documentation issue, and he was holding a $14,000 bill. Both stories are true. Healthshares work until they don't, and when they don't, you have no insurance commissioner to complain to. If you're going healthshare, do it because you're young, healthy, and willing to be wrong about a major medical event — not because someone sold you on the savings.

What I'd Do in Your Shoes

If you're looking at this decision in 2026, here's the order of operations I'd run:

First, get your COBRA quote. The election notice will arrive within 14 days of your qualifying event. The number on it is your floor — it's the simplest, lowest-friction option, and it gives you 60 days to make a real decision.

Second, project your 2026 AGI honestly. Not optimistically, not pessimistically. Use last year's net SE income as a baseline and adjust for what you actually expect to change. If you're within $10,000 of the 400% cliff (above or below), you have AGI levers worth running.

Third, run the Marketplace calculator on healthcare.gov for your projected income. Compare the post-subsidy benchmark silver premium to your COBRA quote. If the gap is more than $200/month, the Marketplace usually wins on pure cost, even accounting for the deductible reset.

Fourth, if you're near the cliff, decide whether to use Solo 401(k) / SEP-IRA / HSA contributions to drop under it. Run the numbers: contribution amount, recovered subsidy, and tax savings on the contribution itself. The ROI is usually obvious.

Fifth, if you decide on Marketplace, consider claiming subsidies conservatively (or skipping advance subsidies entirely and claiming at filing) if your income could land near the cliff. The cost of overclaiming is repaying the full year's advance credits at filing, with no cap.

The 2026 cliff didn't take Marketplace off the table. It just removed the auto-pilot answer. For self-employed people who do the math, Marketplace still wins for most income levels. For people who don't do the math, the math will get done for them at filing time, and not in their favor.

For the broader picture on the self-employment financial transition, see the First 90 Days After Quitting playbook. For the AGI levers in detail, the self-employed retirement accounts guide walks through Solo 401(k), SEP-IRA, and traditional IRA mechanics. And if you're still weighing whether to leave your job at all, the runway formula article includes health insurance as a major hidden cost most calculators get wrong.

Part of: The $100K Side Hustle Financial Upgrade Checklist — the income-threshold decision framework this article fits into.