Estimating causal effects: considering three alternatives to difference-in-differences estimation

• DOI: 10.1007/s10742-016-0146-8
• Publication Date: 2016-05-07T14:45:01Z
• Authors: Stephen O’Neill; Noémi Kreif; Richard Grieve; Matthew Sutton; Jasjeet S. Sekhon
• Keywords: Difference in differences; Average treatment effect; Instrumental variable; Randomized experiment

Difference-in-differences (DiD) estimators provide unbiased treatment effect estimates when, in the absence of treatment, average outcomes for treated and control groups would have followed parallel trends over time. This assumption is implausible many settings. An alternative that potential are independent status, conditional on past outcomes. paper considers three methods share this assumption: synthetic method, a lagged dependent variable (LDV) regression approach, matching Our motivating empirical study an evaluation hospital pay-for-performance scheme England, best practice tariffs programme. The conclusions original DiD analysis sensitive to choice approach. We conduct Monte Carlo simulation investigates these methods' performance. While produces when holds, approaches less biased effects it violated. In cases, LDV approach most efficient least estimates.