Poster
Learning Treatment Effects in Panels with General Intervention Patterns
Vivek Farias · Andrew Li · Tianyi Peng
Keywords: [ Causality ]
Abstract:
The problem of causal inference with panel data is a central econometric question. The following is a fundamental version of this problem: Let be a low rank matrix and be a zero-mean noise matrix. For a `treatment' matrix with entries in we observe the matrix with entries where are unknown, heterogenous treatment effects. The problem requires we estimate the average treatment effect . The synthetic control paradigm provides an approach to estimating when places support on a single row. This paper extends that framework to allow rate-optimal recovery of for general , thus broadly expanding its applicability. Our guarantees are the first of their type in this general setting. Computational experiments on synthetic and real-world data show a substantial advantage over competing estimators.
Chat is not available.