Should There be Vertical Choice in Health Insurance?

A key choice made by health insurance market regulators is which types of plans to offer, or allow to be offered, to consumers. A key dimension of this choice is whether to offer consumers a choice over coverage levels, which we term “vertical choice.” The welfare effects of vertical choice are theoretically ambiguous. The socially optimal level of coverage, at which the marginal benefit of risk protection equals the marginal social cost of moral hazard, may vary across consumers. Providing more choice over coverage levels may allow consumers to match more closely with their socially optimal contract via revealed preference, but it may also lead to the opposite if private and social optima are not aligned. That is, a consumer’s socially optimal coverage level may, at certain prices, not be the same as their privately optimal coverage level. Efficient prices will best align social and private optima in the population, which may or may not involve vertical choice. In this paper, we show that in regulated health insurance markets, vertical choice should be offered if and only if consumers with higher willingness to pay for insurance also have a higher socially optimal level of coverage.

We test for this condition empirically using rich administrative data from the employer-sponsored health insurance market of public-school teachers in Oregon. Our data include plan choice sets, premiums, plan choices, and health insurance claims from over 45,000 households. The setting features the availability of many vertically differentiated plan choices as well as exogenous variation in plan premiums, allowing us to recover the joint distribution of risk, risk preferences, and moral hazard in the population. With estimates of these key primitives, we construct each household’s willingness to pay for marginally more generous insurance, as well as the social surplus generated by allocating a given household to a marginally more generous insurance plan. We find that due to heterogeneity in risk, risk aversion, and moral hazard, some households indeed should, from a social welfare perspective, have more generous insurance than others. Moreover, we find that households with higher willingness to pay for insurance have a higher socially optimal level of insurance, and thus, vertical choice should be offered. We find that optimal prices do not align with perfectly competitive prices, and that they involve cross-subsidizing from less generous to more generous plan options. While the employer may prefer these prices from an efficiency standpoint, we also investigate their impact on distributional outcomes.