Policy Learning under Endogeneity Using Instrumental Variables

TL;DR

This paper develops a new method for learning optimal policies in observational studies with endogenous treatments by leveraging instrumental variables and encouragement rules, demonstrated through a tuition subsidy case in Indonesia.

Contribution

It introduces a framework that uses encouragement rules and the marginal treatment effect to identify and estimate optimal policies under endogeneity.

Findings

Established identification of social welfare with instrumental variables.

Proposed EWM method for estimating optimal encouragement rules.

Applied approach to real-world tuition subsidy policy in Indonesia.

Abstract

I propose a framework for learning individualized policy rules in observational data settings characterized by endogenous treatment selection and the availability of an instrumental variable. I introduce encouragement rules that manipulate the instrument. By incorporating the marginal treatment effect (MTE) as a policy invariant parameter, I establish the identification of the social welfare criterion for the optimal encouragement rule. Focusing on binary encouragement rules, I propose to estimate the optimal encouragement rule via the Empirical Welfare Maximization (EWM) method and derive the welfare loss convergence rate. I apply my method to advise on the optimal tuition subsidy assignment in Indonesia.