🤖 AI Summary
This study addresses the challenge of risk-averse treatment assignment when individuals self-select based on unobserved characteristics. It proposes a planner’s optimization framework integrating marginal treatment effects with coherent risk measures, extending endogenous selection models to risk-averse settings. This formulation unifies perspectives on uncertainty aversion, distributional robustness, and worst-case welfare while encompassing the risk-neutral case as a special instance. Leveraging Kusuoka’s representation theorem, the authors characterize optimal assignment rules and establish finite-sample regret bounds for empirical policy learning. Theoretically, this work expands the frontiers of policy learning under causal inference. Empirically, an application to Card (1995) demonstrates that incorporating risk aversion substantially alters optimal college enrollment policies, yielding economically significant implications.
📝 Abstract
This paper studies risk-averse treatment allocation when individuals self-select into treatment based on unobserved characteristics. We develop a framework that combines the marginal treatment effect approach to endogenous selection with a general class of coherent risk measures that capture distributional preferences over welfare outcomes. We show that the planner's problem admits equivalent interpretations in terms of uncertainty aversion, distributional robustness, and worst-case welfare. For law-invariant coherent risk measures, we derive a Kusuoka representation that expresses the planner's objective as a weighted evaluation of different regions of the welfare distribution and characterize the resulting optimal allocation rule. We further establish finite-sample regret guarantees for empirical risk-averse policy learning, showing how the statistical difficulty of learning a policy depends on the planner's sensitivity to adverse welfare outcomes. The framework nests the risk-neutral policy learning model of \cite{Kiatagawa_Tetenov_2018} as a special case. An application to the \cite{Card1995} college proximity data demonstrates that incorporating risk aversion can lead to economically meaningful changes in optimal college admission policies.