Collard-Wexler, AllanJiang, Yilin2019-06-072021-05-212019https://hdl.handle.net/10161/18782<p>I study the industrial organization of public health insurance markets with competition of insurance companies, focusing on the context of the Medicare Part D program. In the first chapter, I study how the the relations between insurance companies and pharmacy benefits managers affect the formulary design of Medicare Part D prescription drug plans. I construct an original data set of the contractual relations between pharmacy benefits managers and insurers, and find that plans that work with larger pharmacy benefits managers do not seem to have better coverage. Formularies are the most similar for plans offered by the same insurers, which suggests that formulary design might be conducted at the insurer level.</p><p>In the second chapter, I study whether beneficiaries in the Medicare Part D program are confused about the multiplicity and similarity of plans offered when they make decisions about program enrollment and plan selection, and whether the meaningful difference requirement helps beneficiaries improve their results. I use the 1% random sample of Medicare Part D enrollment and claims data from 2006 to 2013 to construct measures of beneficiaries’ quality of choice, and find that their choices of prescription drug plans only improve minimally when there are fewer plans and when the plans are more differentiated. However, new beneficiaries are less likely to participate in the program when are fewer available plans, and recurring beneficiaries are less likely to switch plans.</p><p>In the third chapter, I study how Medicare beneficiaries’ incomplete search of prescription drug plans affects their plan choice and spendings, and evaluate policy interventions that might help beneficiaries choose better plans at lower costs. I estimate the incomplete search model of beneficiaries using the 20% random sample of Medicare Part D data as well as insurance companies’ advertising spending data. My estimates suggest that beneficiaries are more elastic to plan premium than previous estimates, and they are more likely to consider plans offered by insurers with higher amount of advertising spending. Applying my model and estimates to the analysis of counterfactual policies, I find that suggesting the lowest-cost insurer to beneficiaries would increase their overall spendings, while nudging beneficiaries to consider all insurers could help them and Medicare achieve moderate savings.</p>EconomicsDissertation