synthetic control

Here are my study notes on the matrix completion method for causal panel data models proposed by Athey et al. (2021). 1. Setup and Notation Let $Y(0),Y(1)$ be potential outcome matrices, $W$ be a matrix of binary treatment, $W_{it}\in \{0,1\}$ .

Motivation Abadie et al. (2010) motivate the synthetic control method with a model that generalizes the difference-in-differences (fixed-effects) model commonly applied in the empirical social science literature by allowing the effect of unobserved confounding characteristics to vary over time.