Book Description

This dissertation consists of three essays that evaluate how consumers make decisions in settings where the researcher may not know the set of alternatives from which observed choices were selected. Many empirical analyses in economics presume the researcher knows the full set of alternatives an individual compared when selecting their most preferred. In practice, this assumption may fail to hold for a variety of reasons. In the first chapter, I introduce the economic setting of unobserved choice sets and consideration sets defining to this work. In the second chapter, my coauthors and I 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 then apply our theoretical findings to learn about households' risk preferences and choice sets from data on their deductible choices in auto collision insurance. The third chapter evaluates the prescription drug insurance choices of Medicare beneficiaries. 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, and in contrast to previous literature that assumes full consideration of all plans, I uncover an important role for risk aversion in determining individual choices.