Variable importance for causal forests: breaking down the heterogeneity of treatment effects

TL;DR

This paper introduces a new variable importance algorithm for causal forests that quantifies how input variables influence treatment effect heterogeneity, addressing the black-box limitation of existing methods.

Contribution

We develop a novel importance measure for causal forests that handles confounders and extends to groups of variables, improving interpretability of treatment effect heterogeneity.

Findings

Outperforms existing importance measures on simulated and real data

Handles confounders effectively to ensure measure consistency

Efficiently extends to groups of variables for practical insights

Abstract

Causal random forests provide efficient estimates of heterogeneous treatment effects. However, forest algorithms are also well-known for their black-box nature, and therefore, do not characterize how input variables are involved in treatment effect heterogeneity, which is a strong practical limitation. In this article, we develop a new importance variable algorithm for causal forests, to quantify the impact of each input on the heterogeneity of treatment effects. The proposed approach is inspired from the drop and relearn principle, widely used for regression problems. Importantly, we show how to handle the forest retrain without a confounding variable. If the confounder is not involved in the treatment effect heterogeneity, the local centering step enforces consistency of the importance measure. Otherwise, when a confounder also impacts heterogeneity, we introduce a corrective term in the retrained causal forest to recover consistency. Additionally, experiments on simulated, semi-synthetic, and real data show the good performance of our importance measure, which outperforms competitors on several test cases. Experiments also show that our approach can be efficiently extended to groups of variables, providing key insights in practice.