Many of my readers will know what adverse selection is, but to reiterate, adverse selection occurs in health insurance markets when sicker patients decide to choose specific types of plans. If a health insurer believes they will have a representative sample of enrollees but instead they have a disproportionate share of sick individuals, they will need to raise premiums to cover their cost. When insurers raise their cost, however, more healthy individuals will leave the plan causing futher premium increases, enrollee exit and so on until the plan become unsustainable. This is the classic adverse selection death spiral.
However, a paper by Geruso and Layton (2017) considers other, more nuanced forms of adverse selection as well. Consider the following case:
But what if we notice that neither plan offers good coverage for cancer treatments? It might be that both plans believe that offering such coverage would attract especially costly patients and would drive the plan to insolvency. Both plans thus attempt to screen out these patients by offering coverage that is unappealing to cancer patients. Because the two plans act identically, neither succeeds in avoiding cancer patients, and both plans get an equal share of such patients. But the result is that cancer patients cannot find good coverage in the market, and currently healthy consumers cannot find a plan to protect them against the possibility of needing cancer care in the future. Despite the fact that that we observe no systematic sorting of sick consumers between the available plans, this too would be a selection-driven distortion: there is a missing market for cancer coverage due to the anticipation of how the sick would sort themselves if a certain kind of coverage were offered.
The authors consider two types of adverse selection. In the first case (which they name ‘fixed contracts’) they assume that the supply side (insurers/health plan) are static, but adverse selection may occur if sicker patients select more generous plans. The authors also consider a second case where contracts are endogenous. In this case, just the threat of adverse selection will affect the generosity of coverage offered (e.g., see the cancer example above). The term ‘cream skimming’ is used to define the supply-side response in terms of health plan benefit coverage designed to exclude the sickest patients from purchasing coverage.
The authors also propose some ways to get around adverse selection including:
- premium rating regulation, (e.g., “community rating”)
- consumer subsidies or penalties to influence insurance take-up;
- risk adjustment
- contract regulation (e.g., minimum benefits package)