Bandwidth Selection for Treatment Choice with Binary Outcomes

TL;DR

This paper introduces a new bandwidth selection method for treatment choice with binary outcomes, optimizing minimax regret to improve decision-making accuracy in nonparametric kernel regression.

Contribution

It proposes a novel bandwidth selection approach based on minimax regret specifically for binary outcome treatment rules, extending existing methods.

Findings

The method effectively minimizes maximum regret in treatment decisions.

Optimal bandwidth choices differ between binary and normally distributed outcomes.

Numerical analysis demonstrates improved treatment rule performance.

Abstract

This study considers the treatment choice problem when outcome variables are binary. We focus on statistical treatment rules that plug in fitted values based on nonparametric kernel regression and show that optimizing two parameters enables the calculation of the maximum regret. Using this result, we propose a novel bandwidth selection method based on the minimax regret criterion. Finally, we perform a numerical analysis to compare the optimal bandwidth choices for the binary and normally distributed outcomes.