Overlap in Observational Studies with High-Dimensional Covariates

Abstract

Estimating causal effects under exogeneity hinges on two key assumptions: unconfoundedness and overlap. Researchers often argue that unconfoundedness is more plausible when more covariates are included in the analysis. Less discussed is the fact that covariate overlap is more difficult to satisfy in this setting. In this paper, we explore the implications of overlap in observational studies with high-dimensional covariates and formalize curse-of-dimensionality argument, suggesting that these assumptions are stronger than investigators likely realize. Our key innovation is to explore how strict overlap restricts global discrepancies between the covariate distributions in the treated and control populations. Exploiting results from information theory, we derive explicit bounds on the average imbalance in covariate means under strict overlap and show that these bounds become more restrictive as the dimension grows large. We discuss how these implications interact with assumptions and procedures commonly deployed in observational causal inference, including sparsity and trimming.

To appear in Journal of Econometrics

References in corpus (3)

• Semiparametric efficiency in GMM models with auxiliary data
• Finite-Sample Optimal Estimation and Inference on Average Treatment Effects Under Unconfoundedness
• Asymptotic causal inference with observational studies trimmed by the estimated propensity scores

Cited by in corpus (11)

• Support and Invertibility in Domain-Invariant Representations
• Conformal Inference of Counterfactuals and Individual Treatment Effects
• Generalization Bounds and Representation Learning for Estimation of Potential Outcomes and Causal Effects
• Removing Hidden Confounding by Experimental Grounding
• A Survey on Causal Inference
• Transfer Learning for Estimating Causal Effects using Neural Networks
• Characterization of Overlap in Observational Studies
• From controlled to undisciplined data: estimating causal effects in the era of data science using a potential outcome framework
• Deconfounding Scores: Feature Representations for Causal Effect Estimation with Weak Overlap
• Target Consistency for Domain Adaptation: when Robustness meets Transferability
• Learning Robust Decision Policies from Observational Data