A1412
Title: Policy learning with distributional welfare
Authors: Sukjin Han - University of Bristol (United Kingdom) [presenting]
Abstract: Optimal treatment allocation policies that target distributional welfare are explored. Most literature on treatment choice has considered utilitarian welfare based on the conditional average treatment effect (ATE). While average welfare is intuitive, it may yield undesirable allocations, especially when individuals are heterogeneous(e.g., with outliers), which is why individualized treatments were introduced in the first place. This observation motivates proposing an optimal policy that allocates the treatment based on the conditional quantile of individual treatment effects (QoTE). Depending on the choice of the quantile probability, this criterion can accommodate a policymaker who is either prudent or negligent. The challenge of identifying the QoTE lies in its requirement for knowledge of the joint distribution of the counterfactual outcomes, which is generally hard to recover even with experimental data. Therefore, minimax policies are introduced that are robust to model uncertainty. A range of identifying assumptions can be used to yield more informative policies. For stochastic and deterministic policies, the asymptotic bound is established on the regret of implementing the proposed policies. In simulations and two empirical applications, optimal decisions are compared based on the QoTE with decisions based on other criteria. The framework can be generalized to any setting where welfare is defined as a functional part of the joint distribution of potential outcomes.
