| In this thesis we study consumer behavior and market outcomes in health insurance from several perspectives. The linking feature is that we use a novel dataset with detailed health, financial, and demographic information for all employees at a large firm to study consumer choice patterns and relate them to market outcomes.;In the first chapter, we investigate consumer switching costs in the context of health insurance markets, where adverse selection is a potential concern. Though previous empirical work has studied the problems of adverse selection and consumer choice inadequacy in isolation, these phenomena interact in a way that directly impacts market outcomes. We use the data described above to show that switching costs (i) are large and (ii) significantly impact the degree of adverse selection in equilibrium. We estimate a structural choice model to quantify switching costs, risk preferences, and health risk in the population and use the output to study the welfare impact of an information provision policy that nudges consumers toward better decisions by reducing switching costs. We find that, in a partial equilibrium setting where observed plan prices are held fixed, this policy improves consumer welfare by 10% while in a general equilibrium setting where insurance prices adjust welfare is reduced by 6%.;In the second chapter, we use different features of the dataset to study utilization of Flexible Spending Account (FSAs), an incremental insurance product tied to employer-based health insurance plans. There is significant money at stake for individuals who enroll in FSAs, firms that offer them, and the governments that forgo tax revenue when FSAs are used. We study choice under uncertainty from the ex ante perspective using individual-specific health distributions and investigate the causes consequences of deviation from the optimal contribution policy.;In the third chapter, we analyze the determinants of health plan choice and utilization and investigate the extent to which different insurer information sets imply different levels of selection based on unobservables. We present results showing the relative strengths of different health and choice relevant variables for predicting consumer behavior and link our framework to standard asymmetric information tests in the literature. |
