What Does Ranking Marketplace Plans by Premium Actually Cost You?

1. Everyone sorts by the wrong number, and nobody has measured what it costs

Healthcare.gov sorts by premium. So does every broker, every comparison site, and every "cheapest health insurance in [state]" article on the internet. That is a defensible default — the premium is the only number you are certain to pay, twelve times a year, whether or not you ever see a doctor.

The problem is that the premium is roughly half the bill. The other half is the plan's cost-sharing: the deductible, the copays, the coinsurance rates, the out-of-pocket maximum. Those are filed with CMS, plan by plan and service by service, and published. They are simply never joined to the premium, because doing it requires deciding how much care the buyer will use — and a comparison site that guesses wrong for you has made your decision worse, not better.

So instead of guessing once, we ran it three times, at three honest levels of use.

For a fracture year, ranking by premium is sound. In no county of the 409 does the cheapest-premium plan cost $1,000 a year more than the best plan, the median penalty is $0 in Texas, and in the median Texas county the cheapest-premium plan is the cheapest plan outright ^[1,2]. If you are 40, healthy, and expect to use the system once, the standard advice is sound and this page has nothing to sell you.

For a diabetes year, the picture splits by geography. A gap exists in 96% of counties, but it only clears $1,000 in 44% of them, and where you live decides which side you fall on: 77% of Ohio counties, 58% of Florida's, 28% of Texas's ^[1,2]. Texas's median penalty is $150 because the cheapest-premium plan there is typically the second-cheapest overall. That is the honest qualifier on this whole page and we would rather print it than bury it.

For a birth year, there is no ambiguity. The cheapest-premium plan loses in all 409 counties, by a median of $6,310 in Texas, $5,017 in Florida and $4,546 in Ohio, and by at least $1,000 in 96% of them ^[1,2]. The median penalty exceeds what many of these buyers pay in premium for the entire year.

4. Why the expensive tier keeps winning

The mechanism is not subtle once you see it. A low-premium plan buys its low premium with a high deductible — $9,800 in the Harris County example above ^[1,2]. A year of real care blows straight through that deductible, so the buyer pays it in full, and the premium saving is spent several times over.

Which is why, for a year containing a birth, a Gold plan is the cheapest plan overall in 40% of all 409 counties ^[1,2] despite Gold carrying the highest premiums on the marketplace, and why the winning plan has a $0 deductible in 77% of counties ^[1,2]. Those two facts are the same fact. A $0-deductible plan charges more every month precisely so that it charges nothing when you use it, and a list sorted by premium is guaranteed to put it near the bottom.

Franklin County, Ohio is the clearest single case. The cheapest-premium plan there is Oscar's Bronze Classic Standard at $4,767 a year with a $7,500 deductible. For a year containing a birth it costs $14,087 all in. Oscar's own Gold Elite Saver Plus, at $7,040 a year with a $0 deductible, costs $8,740 — $5,348 less, from the same insurer, in the same county ^[1,2].

5. The subsidy does not save you, and the reason is worth understanding

The obvious objection is that most marketplace buyers do not pay full price — the large majority receive a premium tax credit — so perhaps this gap is an artifact of pricing plans at a sticker almost nobody is charged.

It is not, and the arithmetic is unusually clean. The credit is set by the second-lowest-cost Silver plan in your county — the benchmark — and not by the plan you choose. It is a fixed dollar amount for your household. So it subtracts the same number from every plan on the list, which moves the whole field down together and changes neither the ordering nor the size of any gap between two plans. The median penalty is identical at full price and at a $55,000 income in all three states ^[1,3,4].

We measured this rather than assuming it. The median diabetes-year penalty is $1,438 in Florida, $2,107 in Ohio and $150 in Texas — at full price, at a $55,000 income and at a $40,000 income alike. The same figure to the dollar, in all nine cases ^[1,3,4]. A subsidy changes what you pay. It does not change which plan you should buy, and it does not shrink the cost of choosing wrongly by a single dollar.

One 2026-specific warning belongs here. The enhanced credits of the American Rescue Plan and the Inflation Reduction Act expired on 31 December 2025, so the 400%-of-poverty cliff is back: a single person earning one dollar over $62,600 receives nothing at all ^[3,4]. For anyone near that line, the difference between a good plan choice and a bad one is no longer partly absorbed by the government.

6. What to do with this

Decide what kind of year you are likely to have before you look at a single price, because that decision — not the plan list — is what determines which number you should be sorting on.

If you expect to use very little care, sort by premium and stop. The data says you will be right.

If you are managing a chronic condition, check your own county rather than trusting a national rule: the gap is large in most of Ohio, over half of Florida and under a third of Texas ^[1,2]. If you are planning a pregnancy, ignore the premium column almost entirely, look hard at plans with a $0 or low deductible, and price the Gold tier properly before dismissing it ^[1,2].

And in every case, the number that decides it is your plan's own filed cost-sharing, which is public and which nobody shows you next to the price. That is what the county table below is for. The state-by-state versions, with every county and the named plans, are at cheapest health insurance in Texas, in Florida and in Ohio.

7. The full table, and its limits

Every county we priced, both plans named, all component figures, as a CSV:

Download the full 409-county table (CSV, 1,227 rows)

Four limits travel with it, and they are real.

It covers 28 states, not 50. The CMS Public Use Files describe the federally-facilitated marketplace only. California, New York, Washington, Illinois and Georgia run their own exchanges and publish nothing through this route ^[1] — 22 states and the District of Columbia in total, so each would have to be sourced separately rather than recompiled. We publish three states because three is what we can stand behind, not because the others are uninteresting.

It prices one buyer. A single 40-year-old. Family pricing, tobacco rating and age-banding all move these numbers, and the direction is not uniform — the diabetes penalty shrinks with age as premium differences come to dominate out-of-pocket ones.

The three scenarios are stylised. They are CMS's own, applied uniformly, which makes counties comparable to each other. They are not a forecast of your year.

It ignores networks and drug formularies. Two plans at the same total cost are not equivalent if one of them does not cover your doctor or your prescription. Check both before buying anything on this page.

And the filed figures are the issuers' own, including where they disagree. A plan can file a $0 medical deductible and a separate drug deductible behind it; 157 of the 193 plans in that shape file $0 for a diabetes year and 36 charge the drug deductible ^[1]. We publish what each issuer filed rather than imposing one reading, which matters in five Texas counties where such a plan wins — Harris among them, noted at that example above. Every figure is capped at the plan's own out-of-pocket maximum, which 3,230 filed scenario totals exceed ^[1].