Ensuring Stability, Accuracy and Meaningfulness in Synthetic Control Methods: The Regularized SHAP-Distance Method

Abstract

The synthetic control method (SCM) has been increasingly adopted to evaluate causal effects under quasi-experimental designs. However, SCM suffers from sound weaknesses that compromise its accuracy, stability and meaningfulness. The Regularized SHAP-distance synthetic control method (RSD-SCM) is proposed as solution. We evaluate the economic effect of the government formation deadlock in Spain, 2016. The deadlock did not negatively affect economic growth, as the economy grew 1.58% more without full government; standard SCM method overestimates the effect by 0.23 pp. We show that RSD-SCM offers higher accuracy and stability, while ensuring the economic meaningfulness of covariates used in building the counterfactual.