Essays in Health Economics

This dissertation consists of four chapters. In the first chapter, I examine how pharmacies and pharmacy closures impact patients' prescription drug use. In the second chapter, coauthored with Michelle Zheng, we examine a mechanism states use to extract more federal Medicaid contributions. In the third chapter, I examine how Medicare insurers use price and non-price mechanisms to reduce fraud and waste in home health and post-acute care. In the fourth chapter, I estimate the effects of small prescription drug copays on low-income Medicare beneficiaries' drug consumption.

The first chapter starts from the startling fact that despite potential health benefits of prescription drugs, many patients do not take them as prescribed. While prior work has emphasized the role of prices and insurers in patients' drug consumption decisions, this paper provides novel evidence on a non-price driver of drug consumption: pharmacies. Using a staggered event study design, I find that pharmacy closures cause Medicare patients' drug use to decline in the short-run---especially among low-income patients---but increase in the long run. To explain the long-run increase in drug consumption following a pharmacy closure, I model three potential mechanisms driving the reduced-form effect: temporary disruption/switching costs, permanent changes in patient costs (e.g., copays or travel distance), and permanent shifts to higher- or lower-dispensing pharmacies. To quantify the relative impacts of these mechanisms, I estimate a two-way-fixed-effects model in the style of Abowd, Kramarz, and Margolis (1999) of pharmacies' effects on low-income patients' drug use. Combining the pharmacy effects from the AKM model with my reduced-form closure analyses, I find that the long-run increase in drug consumption following a pharmacy closure is explained by patients switching from lower-dispensing pharmacies (which are disproportionately likely to close) to higher-dispensing pharmacies. More generally, the variation in pharmacy fixed effects is about half that of prescriber fixed effects, indicating that pharmacies matter for drug consumption.

The second chapter, coauthored with Michelle Zheng, studies the impacts of states' responses to federal Medicaid matching incentives. In particular, we examine state efforts to fund Medicaid hospital spending using hospital provider taxes, which allow states to obtain additional federal matching dollars. We use the staggered rollout of these taxes across 11 states to show that they increase Medicaid hospital spending by 20-35%. Furthermore, in a subset of four states where the lack of managed care providers allows us to disaggregate all spending, we find that roughly all of the state tax revenue flows to Medicaid hospital reimbursements. However, hospitals do not use the increased funding to provide additional Medicaid care from their existing beds, either in absolute terms or relative to non-Medicaid care. We conclude by discussing two classes of arguments that could explain this puzzle: The state payments may not change hospitals' incentives to treat the marginal Medicaid patient, or regulatory or reputational considerations may prevent hospitals from responding to real incentives.

The third chapter returns to similar questions in the first chapter, but in the context of Medicare home health and post-acute care. It examines price and non-price costs, especially levied by Medicare Advantage plans, and asks how these impact patient and provider behavior. More broadly, it also examines how price and non-price costs interact with each other and the degree to which they are substitutes for insurers. I show that the geographically concentrated explosions in home health fraud identified in prior work (Einav et al. 2025) were largely confined to Traditional Medicare. Furthermore, I find evidence consistent with the hypothesis that Medicare Advantage plans' provider networks helped prevent fraud in this context. Leveraging a 2009-2010 change in Medicare policy, I also find descriptive evidence suggesting that plans may view non-price costs like prior authorization as substitutes for prices, and switch to the former when policies restrict the latter.

The fourth chapter complements the non-price focus of the first chapter by examining the effects of even small copays for prescription drugs in a low-income Medicare population. I leverage a Medicare policy change in 2012 that eliminated copays for Medicare beneficiaries receiving Medicaid home and community-based services (HCBS). Relative to Medicare-Medicaid dual eligibles not receiving these services (who continued to face small copays), I find that the policy reduced treated patients' drug spending by $8 per month. Patients consume more drugs in response, increasing drug fills by 1.2% and days supply of medication by 0.6%. Relative to prior studies estimating the effects of copays on wealthier patients, I find that low-income patients respond similarly per dollar of copay reduction.