Methodology & literature

The Floor

The annual premium — twelve months of payments, before you use any healthcare at all. It is the cost of owning the plan, and it is the one number you cannot avoid paying.

The Maximum Annual Exposure

The Floor plus the plan’s out-of-pocket maximum: the most the year can cost you, whatever happens. If you spend s on care, your total is the premium plus whichever is smaller — your spending or the out-of-pocket max.

Dominance

Plan B dominates plan A when B’s Floor is no higher and B’s worst case is no higher, with at least one strictly lower. That is the whole test — and it is why the conclusion is so strong. A dominated plan is not usually worse or worse on average. It costs you more no matter how the year goes.

The comparison is (premium, worst case), not (premium, out-of-pocket max). Comparing on the out-of-pocket max alone is a real and easy mistake: it treats a cheap plan with a high max as worse than an expensive plan with a low one, even when the expensive plan can never come out ahead. We made that mistake ourselves and corrected it; the counts on these pages are from the corrected method.

The Short List, and the ladder

What survives is the Short List — every plan that can win for somebody. A subset of those are ladder rungs: the plans where stepping up to more protection is priced proportionately. Between rungs we quote the step in cents per dollar of protection bought, so “is the better plan worth it” becomes a number rather than a feeling.

Past a certain point the worst case stops falling and only the premium rises. That is the crossover, and anything past it is just pre-paying for care.

What this method does not do

• It does not judge networks. Two plans can be identical on both numbers and one can exclude your doctor. The Index tells you which plans are worth comparing; it does not tell you the cheap one covers your oncologist.
• It does not model your specific drugs or care. A plan that is dominated on these two numbers is genuinely worse whatever you spend — but a plan on the Short List may still be wrong for you because of a formulary.
• It scores listings, not people. 94.6% of 52,944 listings are dominated. That is not the same as 94.6% of buyers being in a dominated plan — buyer-weighted figures are lower, because the worst plans are not the ones most people pick.
• Subsidies change the answer. Everything here is unsubsidised. If you qualify for a premium subsidy, it applies on-exchange only, and the comparison has to be redone against your subsidised premium.

Where the method comes from

Insure the tail, don’t pre-pay the middle — Arrow (1963), Rabin (2000), Sydnor (2010). And the finding that people routinely choose dominated health plans when offered one: Bhargava, Loewenstein & Sydnor (2017) found about 60% of employees did exactly that, from a menu far smaller than the one a Texan faces on the individual market.

What had not been done was running that analysis across an entire ACA market. We did, for every plan in Texas — the counts are in The Texas Report.