The Welfare Consequences of Risk Adjustment in Competitive Health Insurance Markets

Adverse selection is a common problem in competitive health insurance markets due to asymmetric information and price regulations. When contracts are fixed (as with the actuarial values required by the Affordable Care Act), selection leads to suboptimal pricing of health insurance plans which reduces welfare due to inefficient sorting of individuals between plans. Various tools are employed to combat selection in these environments. The most prominent of these is risk adjustment, which redistributes health plan revenue from plans with high expected cost enrollees to plans with low expected cost enrollees. In this paper, I study how risk adjustment affects welfare in competitive health insurance markets.

First, I build on the model of Einav et al. (2010) to show graphically and in a theoretical model that when market participation is compulsory, risk adjustment alters the average cost curve for plan alternatives while leaving the demand curve unchanged. This shifting of the average cost curve results in a new competitive equilibrium that can be either better or worse than the no risk adjustment equilibrium, depending on the correlation between demand and the component of consumers’ health care costs not explained by the risk adjustment model.

Second, I show that when individuals can opt out of the market, risk adjustment can cause a “death spiral” of the entire market. Risk adjustment has some of the same effects on equilibrium premiums as does “community rating,” and can cause a death spiral for a similar reason.  When the price of the cheapest plan in the market is low because it is advantageously selected, risk adjustment raises the price of this plan. This causes the healthiest individuals to exit the market, resulting in an increase in the average cost of the group of individuals remaining in the market. This causes the prices of all of the plans in the market to increase, eventually, in the extreme case, leading to a death spiral of the entire health insurance market. In other words, while risk adjustment can limit adverse selection problems within a market, it may simultaneously worsen adverse selection problems with respect to the choice of entering the market versus going uninsured.

Finally, using data on individual plan choice and claims from a large employer, I estimate a structural model of insurance demand. I use the estimates of risk preferences and costs to simulate how welfare changes in competitive equilibria with different risk adjustment models. I find that when participation is compulsory risk adjustment improves welfare significantly, with more aggressive concurrent risk adjustment models causing larger welfare improvements than weaker prospective and age/gender models.