Supply Response to Consumer Inertia: Strategic Pricing in Medicare Part D (Job Market Paper)
Abstract: A growing literature has documented evidence that consumers in health insurance markets are inertial, or behave as though they face substantial switching costs in choosing a health insurance plan. I investigate whether the private firms which provide prescription drug insurance through Medicare Part D exploit this inertia when setting prices. I first document descriptive evidence consistent with insurers initially setting low prices in order to “invest” in future demand before later raising prices to “harvest” inertial consumers. I then apply a two-step estimation approach following Bajari, Benkard and Levin (2007) to explore the implications of these invest and harvest incentives for equilibrium pricing, finding that on net, demand inertia reduces equilibrium prices (i.e. the invest incentive dominates the harvest incentive). Finally, I evaluate welfare consequences of policies that could be used to constrain insurers’ ability to conduct such invest-then-harvest pricing patterns. I find, for example, that a policy change to cap premium increases would improve consumer welfare by both lowering average premiums and smoothing prices over time.
Research in Progress
Does the Meaningful-Difference Regulation Make a Meaningful Difference? Welfare Impacts of Prohibiting Strategic Entry and Product Proliferation
Abstract: There has been empirical evidence that firms exploit demand inertia when setting prices a variety of markets including health insurance settings. I investigate strategic entry as another response to consumer inertia in the Medicare Part D market: since price discrimination across new and old enrollees is banned, firms face an incentive to continuously introduce new plans that can be priced low to “invest” in future demand while charging higher premiums to incumbent consumers. I provide descriptive evidence consistent with such strategic product proliferation. A “meaningful-difference” regulation introduced in 2010 attempted to limit such behavior by requiring new plans to be sufficiently differentiated in coverage from existing plans by the same insurer. To quantify its impact on insurance supply and consumer welfare, I will enrich a model of insurers’ dynamic pricing with entry decisions, and use baseline variation in entries and market structure to identify the distribution of entry cost prior to the regulation. In order to assess the general desirability of adopting similar regulations in insurance markets, I will then simulate counterfactual welfare if this regulation had been in place when this market first opened; and if that the same insurer can offer only two plans based on a recent policy proposal.
Market Size and Innovation: The Intermediary Role of Pharmaceutical Licensing (with Manuel Hermosilla)
Abstract: Motivated by prior work on market size spurring innovation, we study the role of increased downstream demand in facilitating inter-firm cooperation in the pharmaceutical industry, where licensing is a common form of collaboration between upstream innovators and downstream commercializers. We propose a simple model of licensing with heterogeneous match quality which predicts that positive demand shocks will increase the likelihood of licensing and improve match quality by reducing the relative importance of transaction costs. We then use the differential impacts of the introduction of Medicare Part D across drug categories targeting different ages of consumers as a source of variation in demand, and document empirical evidence consistent with the model.
Uncertain Plan Quality and Consumer Learning in Medicare Part D (with Manisha Padi)
Abstract: Health insurance plans are complex contracts that function as experience goods, due to consumers' uncertainty over their own health status and over plan generosity. We exploit unique features of Medicare Part D in order to study the role of consumer learning in designing and pricing health insurance products. A novel feature of learning in this setting is differential learning across consumers with different unobserved health types, conditional on selection into plans and observed enrollee risk. Consumers learn upon the arrival of claims, and belief updating increases with frequency of utilization. The uncertainty over match quality of a consumer-plan pair may never be resolved for a healthy consumer. We follow Israel (2005) to identify a demand model with this type of learning, using random arrival of signals due to differential health realizations across similar consumers. The interaction between learning and adverse selection has important implications for insurers’ dynamic adjustments of quality and prices. We will model and estimate supply in the presence of this effect and finally simulate effects of relevant counterfactual policies.