Propensity Score Matching with Replacement Using Weighted Least Squares: SAS Code and Macros

Abstract

While propensity score matching (PSM) is increasingly common in finance and accounting research (Roberts and Whited 2013; Shipman, Swanquist, and Whited 2017), coding PSM in SAS can be tedious, with limited coding resources available to researchers. This is especially true when undertaking PSM with replacement, which is preferred in many settings (Roberts and Whited 2013) and which also serves as a worthwhile robustness test (Shipman et al. 2017). In this paper, I provide comprehensive macros for PSM in SAS to significantly increase the speed of coding, with a particular focus on PSM with replacement. While this paper provides the tools for undertaking PSM, I encourage researchers to carefully consider the appropriateness of PSM for their study (e.g., King and Nielsen 2016) before using these tools.

Keywords: Propensity Score Matching, Weighted Least Squares, SAS

JEL Classification: C10