Large Sample Properties of Matching for Balance
arXiv:1905.11386 · doi:10.5705/ss.202020.0343
Abstract
Matching methods are widely used for causal inference in observational studies. Among them, nearest neighbor matching is arguably the most popular. However, nearest neighbor matching does not generally yield an average treatment effect estimator that is -consistent (Abadie and Imbens, 2006). Are matching methods not -consistent in general? In this paper, we study a recent class of matching methods that use integer programming to directly target aggregate covariate balance as opposed to finding close neighbor matches. We show that under suitable conditions these methods can yield simple estimators that are -consistent and asymptotically optimal.
arXiv comment32 pages
References in corpus (4)
- Entropy balancing is doubly robust
- Nonparametric Estimation of an Additive Model With a Link Function
- Minimal Dispersion Approximately Balancing Weights: Asymptotic Properties and Practical Considerations
- Stronger instruments via integer programming in an observational study of late preterm birth outcomes