How subsidy design amplifies ACA tobacco surcharges

Under the Affordable Care Act, insurers selling coverage in the individual market are permitted to charge tobacco users up to 50 percent more in premiums than they charge non-users. The provision sits somewhat oddly within the rest of the law, which established what is known as modified community rating, under which premiums may vary based on only four factors: geographic region, family size, age, and tobacco use. Tobacco use is therefore the only modifiable behavioral factor that insurers may use to differentiate premiums, and the level at which surcharges are set could meaningfully affect whether tobacco users purchase coverage at all.

When our group first collected complete premium data for exchange plans across 36 states, we found that surcharges were not uniform across the country as was commonly expected, but instead varied substantially from one place to another. The plan with the median surcharge charged tobacco users only about 10 percent more than non-users, and nine in ten plans charged less than the federal cap allowed. Taken on its own this looks like a reassuring result, since insurers were generally declining to price at the maximum the law permitted.

The affordability picture nevertheless remained worse for tobacco users. Even with these lower-than-allowed surcharges, tobacco users lacked access to affordable coverage, which we defined as access to at least one plan with premiums below 8 percent of income after subsidies, in more states than non-users did. Understanding why a relatively modest surcharge produces a substantially worse affordability outcome requires looking at how the subsidy is calculated rather than at the behavior of insurers.

The subsidy is calculated before the surcharge is applied

Marketplace premium subsidies are pegged to a benchmark plan, and the benchmark premium used in that calculation is the premium before any tobacco surcharge has been added. The subsidy is sized to bring the base premium down to a set share of household income, and the surcharge is then applied to the amount the enrollee owes after the subsidy has been taken into account. The practical consequence is that the surcharge is entirely unsubsidized, so that for a household receiving substantial premium assistance, a surcharge of 10 percent on the full premium can represent a very large percentage increase in what that household actually pays. In work with Karina Manz and Teresa Waters, we showed that this feature of subsidy design is why marketplace premiums rise faster for tobacco users than the headline surcharge percentages would suggest.

What happens to enrollment

Because states are permitted to set limits more restrictive than the federal cap, and several have eliminated tobacco surcharges altogether, it is possible to estimate what surcharges actually do to coverage. Using data from the Current Population Survey linked to marketplace plan premiums, and comparing smokers with nonsmokers in states that did and did not allow surcharges, we found that allowing a tobacco surcharge reduced insurance enrollment among smokers by 4.0 percentage points. Among smokers who did not have access to insurance through an employer or a public program, the group with the fewest available alternatives, the reduction was 9.0 percentage points. The size of the surcharge mattered as well, with enrollment among smokers falling by 3.4 percentage points for every 10 percentage point increase in the surcharge.

This line of research has had some direct influence on state policy. New Mexico and Virginia have both passed legislation disallowing tobacco surcharges in their marketplaces, Maryland has considered similar legislation, and our findings have been cited in the course of those debates. The work has also been covered in Kaiser Health News, Reuters, The Conversation, and a number of syndicated outlets.

Why this matters beyond tobacco

Tobacco use remains the largest single cause of preventable illness in the United States, and tobacco users as a group have a correspondingly high need for primary care, cancer screening, and cessation support. A rating rule intended to make tobacco users bear more of the cost associated with their risk therefore appears to leave a measurable share of them without insurance, and consequently without access to the care that would help address that risk. Whether that tradeoff is worth making is ultimately a policy judgment rather than an empirical question, but the evidence suggests that the enrollment consequences are real and that they fall on a population with substantial unmet need.

More generally, this case illustrates something that recurs throughout my work on benefit design, which is that the affordability of coverage depends less on any single headline price than on how a given provision interacts with everything else in the benefit structure. A surcharge that appears modest when considered in isolation becomes substantial once the mechanics of the subsidy calculation are taken into account, and evaluating any one parameter on its own will tend to understate its effect.