Abstract: I take advantage of the evolution of the regulatory and pricing environment in the first years of a large federal prescription drug insurance program for seniors - Medicare Part D - to empirically explore interactions among adverse selection, switching costs, and regulation. Using detailed administrative data, I document evidence of both adverse selection of beneficiaries across contracts and switching costs for beneficiaries in changing contracts within Medicare Part D. Motivated by this descriptive evidence, I formulate an empirical model of contract choice that allows for both switching costs and private information about expected risk. Using this model, I show that switching costs are large and have quantitatively important implications for the sorting of individuals among contracts. How switching costs affect selection over time depends on the direction of changes in the contract space relative to the initial conditions. I find that in the present environment, on net, switching costs help sustain an adversely-selected equilibrium with large differences in risks between more and less generous contracts. I then simulate how switching costs could affect regulatory interventions that change the generosity of contracts. For example, I consider the tightening of the minimum standard requirement by “filling” the Part D donut hole as implemented under the Affordable Care Act. I find that absent any switching costs, this regulation would have eliminated the differences in risks across contracts; however, in the presence of the switching costs that I estimate, the effect of the policy is largely muted.
Research in Progress
Abstract: As private markets are increasingly introduced into social insurance systems, the government faces the challenge of designing a regulatory framework that would maximize the welfare of these programs subject to the incentive constraints of the private insurers. In this paper we study the regulatory mechanisms through which the government currently administers subsidies in a large prescription drug program for US seniors - Medicare Part D. We show that some of the supply-side regulatory mechanisms, such as the tying of premiums and subsidies to the realization of average “bids” by insurers in a region, as well as the utilization of the same “bids” to determine which plans are eligible to enroll low-income beneficiaries, prove to be welfare-decreasing empirically. Removing these distortions could reduce the cost of the program without worsening consumers' welfare. Using the data from the first six years of the program, we estimate an econometric model of supply and demand that incorporates the regulatory pricing distortions in the insurers' objective functions. We then conduct a counterfactual analysis of what the premiums would be in this market under different (potentially simpler) ways of providing the subsidies to consumers. Preliminary estimates suggest that the reduction in premiums of the affected plans would be substantial.
“Community Rating versus Long-Term Contracts in Health Insurance: Evidence from the German Two-Tier System” (Updated paper coming soon)
Abstract: In this paper, I study two competing systems that comprise the German health insurance landscape. The two systems differ in the ability of insurers to underwrite individual-specific risk. While the insurers in the so-called “statutory” insurance system have to offer the federally set income-adjusted community rating premiums, and face a regulatory lower-bound on the generosity of coverage, the competing private insurance sector is less restricted in coverage levels, and its prices are regulated only to the extent that they have to follow annuity-like long-term contracts. Thus, while individuals with the same income would pay the same premiums in the statutory system independently of their risk, enrollees of the private plans face full underwriting of their individual risk and may be rejected by the insurers. I utilize a survey panel dataset from Germany for the years 2005-2009 to empirically assess to what extent the selection of “good risks” dominates the interaction between these two systems. To isolate the selection effect, I use a regression discontinuity design based on a regulation that restricts access to private insurance for individuals below a certain income threshold. I do not find compelling evidence of substantial selection between the two systems, and I argue that this is due to the long-term annuity-style pricing rules used by the private insurers in combination with the eligibility restriction by the government. Motivated by these findings, I formulate a model of insurance pricing and demand for both systems that allows me to quantify the welfare effects of relaxing the eligibility restriction to lower income levels, which has been central to the policy debate.