Sorting on Plan Design: Theory and Evidence from the ACA

Americans are increasingly facing choices about their health insurance coverage. This raises questions about what economic forces shape the health plans that are available in the market. Prior literature focuses on how selection patterns by different risk types can distort the price or level of coverage available. This paper identifies a new dimension of sorting in health insurance: sorting by plan design. Health insurance plans often have multi-dimensional cost-sharing attributes that can vary across plans with the same expected level of coverage. Classic theory suggests that plans with optimal risk protection have a straight-deductible design. I show that under risk-adjusted prices, only high-risk types will sort into such plans. The low-risk types will prefer plans that trade off higher maximum out of pocket (MOOP) for lower deductibles because such plans provide more expected coverage to them (but also more variance in out-of-pocket expenditure) holding fixed the premium level.

Consistent with theory, I find that ACA markets a) are populated with plans with very different levels of risk protection within the same coverage tier and b) high MOOP plans appear to attract healthier enrollees. This opens up the question of whether and how to regulate plan designs. I discuss conditions under which plan standardization policies and design-specific subsidy will improve welfare, and estimate the potential benefits of such policies.