Working Papers

"Insurance Choice with Non-Monetary Plan Attributes: Limited Consideration in Medicare Part D''

Abstract: I propose an empirical model of demand for prescription drug plans where non-monetary plan attributes stochastically determine the composition of the set of plans that an individual considers, and monetary plan attributes determine the individual's expected utility over contracts in her consideration set. This model reconciles the classic view of insurance contracts as lotteries with purely monetary outcomes with the empirical finding that choice among insurance plans is driven by their non-monetary attributes and financial attributes beyond their impacts on costs. I estimate the model using data from Medicare Part D allowing for unobserved heterogeneity in risk aversion and in consideration sets. I find that the latter plays a crucial role in plan choices: although 46 plans are available in the market, more than 90% of individuals consider no more than 5 plans. While the majority of available plans include a deductible, nearly 75% of all plans considered have no deductible. Just three firms account for over 60% of plans considered, while three other firms account for fewer than 0.5%. In contrast to previous literature that assumes full consideration of all plans, I uncover an important role for risk aversion in determining individual choices. My results inform the debate on how to refine market design for prescription drug plans in Medicare to improve the match between beneficiaries and plans.

"Heterogeneous Choice Sets and Preferences'' with Levon Barseghyan, Francesca Molinari, and Joshua C. Teitelbaum

(Econometrica, 2021)

Abstract: We propose a robust method of discrete choice analysis when agents' choice sets are unobserved. Our core model assumes nothing about agents' choice sets apart from their minimum size. Importantly, it leaves unrestricted the dependence, conditional on observables, between agents' choice sets and their preferences. We first establish that the model is partially identified and characterize its sharp identification region. We also show how the model can be used to assess the welfare cost of limited choice sets. We then apply our theoretical findings to learn about households' risk preferences and choice sets from data on their deductible choices in auto collision insurance. We find that the data can be explained by expected utility theory with relatively low levels of risk aversion and heterogeneous choice sets. We also find that a mixed logit model, as well as some familiar models of choice set formation, are rejected in our data.