COBRA: Adverse Selection in Action

1102_p032-cobra_170x170The November 2 Forbes suggests that health insurance under COBRA provides a clear example of adverse selection in action. COBRA is the law that allows workers who leave a job (either voluntarily or not) to continue participating in the health insurance they were getting from their employer. To do so, however, they have to pay the full monthly premium—both the employee and the employer portions—plus a 2% administrative fee.

That sticker shock means that many eligible individuals decide not to continue their coverage under COBRA. Not surprisingly, those people tend to be healthier than average. The folks who use COBRA, on the other hand, tend to be less healthy—and, therefore, more expensive—than average. As a result, insurance companies report that COBRA coverage is a money loser:

Citi analyst Charles Boorady says health plans lose a considerable amount of money on Cobra policies. He estimates that the loss ratio–the amount spent on care compared to the premium collected–is around 200%.

Earlier this year, the stimulus bill created a federal subsidy that pays up to 65% of COBRA premiums for laid-off workers who meet certain income limits. That boosted COBRA enrollment and, according to the article, worsened the hit on insurers. It will be interesting to see whether the insurance industry raises any objections if Congress considers extending the COBRA subsidy (eligibility currently expires on December 31, 2009).

Bonus: Here’s a question I might ask my students in the spring, when we study adverse selection: Would insurers feel differently about a 100% federal subsidy for COBRA coverage for laid-off workers?

8 thoughts on “COBRA: Adverse Selection in Action”

  1. I would think that the subsidies would have made things better for insurers. If healthy people refuse COBRA because of costs, then “subsidizing” the costs should bring in healthier people.

    1. Hi Steve – My intuition is that there are two effects here. First is the one you mention: a higher subsidy rate should, all else equal, increase the average healthiness of COBRA enrollees and, thereby, reduce their average costs. Second, a higher subsidy rate should increase the overall number of COBRA enrollees. The hard question (which one would need data to address) is the net of those two effects. In principle it could go either way. If you consider the case of comparing no subsidy to a small subsidy, I would imagine that the insurers are made worse off, since even the “healthier” people attracted by the subsidy would still have costs in excess of premiums. At a 100% premium, on the other hand, you are presumably getting close to the entire population of laid-off workers. As long as they aren’t too different from the pool of all workers, the insurance company should be able to cover its costs. So maybe the right answer for the insurance companies is to encourage a higher subsidy rate …

  2. The most interesting aspect of anti-selection is not mentioned by Forbes (at least I didn’t see it). Specifically, the COBRA election period gives the individual 60 days after receiving the COBRA notice to sign up (and an additional 45 days to pay premiums). If they do sign up, their benefits are *retroactive*. Therefore, many people will wait until they develop a health condition and then retroactively signup. If they end up staying healthy, then they’ll never sign up.

    As to your question, I would answer that the greater the subsidy, the less the problem. The part of the selection problem due to the 45 day premium grace period would be entirely eliminated.

    But insurance companies and self-insured employers would still view the problem as selective. That’s primarily because exiting your job is generally associated with higher than average claim costs, meaning that COBRA individuals are still higher than average cost. Further, premature exit from COBRA itself is more rapid for healthy individuals who are able to obtain coverage elsewhere (COBRA coverage terms when you obtain alternative group coverage).

    Notice that sometimes you can predict which employers are going to be subject to greater COBRA selection, meaning that you can pre-emptively rate the entire group for the anti-selection impact. Obviously you frequently can’t do that in the small group market because of rate laws, but as the employer gets larger, the costs become increasingly transferred from the insurer to the employer.

    1. Thanks Victor, those are very good points. I had forgotten about the 45-day lookback, which is indeed a strong invitation to adverse selection. And I have been wondering whether the laid-off have systematically higher costs. I can imagine scenarios in which the laid-off skew sicker and are thus more expensive. Or, for that matter, scenarios in which they have comparable health, but now have more time to go to the doctor. On the other hand, if there are plans with significant co-pays or deductibles, a sicker laid-off worker might cost less if they demand less care because of the out-of-pocket costs. Consistent with what you said, my intuition is that the higher cost scenario predominates, although I personally don’t have any data on that point.

  3. Rather than a 100% subsidy, what if it essentially held the individual harmless. In other words s/he would still be responsible for the same share of premiums as when they were employed. Top add another twist, what if the default rule was automatic enrollment with the individual share of the premium deducted from unemployment compensation?

    1. Auto enrollment is an interesting idea. You could eliminate the adverse selection of the 45 day period during which individuals can choose. But you’d still get adverse selection in terms of who decides to opt-out.

Comments are closed.