Essays in Industrial Organization

Abstract

This dissertation studies environmental and health policy-relevant issues using cutting-edge applied microeconomic, especially industrial organization, tools and related econometric developments. The first chapter considers policies incentivizing renewable power investors to achieve the policymaker’s environmental objectives. The second chapter relaxes econometric assumptions in dynamic games—a workhorse in industrial organization—to enhance the robustness of policy suggestions implied by the structural estimates. The third chapter explores how physicians and patients interact and how these organic interactions affect the productivity and disparities of the healthcare system. In the first chapter, I propose a structural framework of policymakers providing financial incentives to support renewable investors in building new capacity. Widespread investment in renewable energy is seen as a critical tool in mitigating the impacts of climate change. However, renewable energy projects face substantial risk because they sell their electricity into volatile wholesale electricity markets, and this risk may hinder investments that otherwise appear profitable. Policymakers looking to encourage renewable investment can choose between direct subsidies for renewable investment and power purchase agreements that assume the risk of future electricity sales. The relative value of these two approaches depends critically upon investors’ risk premium. This chapter investigates this trade-off in the context of Brazilian wind energy actions that award winners purchase agreements for a share of their production. I develop and estimate a structural auction model that separately recovers the investors’ risk aversion and private costs. I find that investors are substantially risk averse: investors require an additional risk premium of $20.16/MWh to accept the risk of selling all electricity into the wholesale market, where revenues have a standard deviation of $5.44/MWh. For 3% of Brazil’s generation capacity auctioned, full share purchase agreements will be expected to cost $20 billion less than subsidies. In the second chapter, my coauthors and I study the identification of dynamic games when the underlying information structure is unknown to the researcher. To tractably characterize the set of Markov perfect equilibrium predictions while maintaining weak assumptions on players’ information, we introduce Markov correlated equilibrium, a dynamic analog of Bayes correlated equilibrium. The set of Markov correlated equilibrium predictions coincides with the set of Markov perfect equilibrium predictions that can arise when the players might observe more signals than assumed by the analyst. We propose computational strategies for estimation and counterfactual analysis to deal with non-convexities that arise in dynamic environments. In the third chapter, my coauthor and I study the role of racial concordance among doctors, specialists, and patients in doctors’ referral decisions. Using a 20% random sample of Medicare beneficiaries from 1999–2010 in the United States, we find that doctors refer to the same race specialist more than otherwise, especially when the patient is also the same race. We propose a structural framework to identify the racial concordance effect net of outcome benefits on doctors’ referral decisions, accounting for endogenous specialist choice set within the local market.