Fixed effects counterfactual estimator (fect) based on rpy2. This is a package for implementing counterfactual estimators in panel fixed-effect settings. It is suitable for panel/TSCS analysis with binary treatments under (hypothetically) baseline randomization. It allows a treatment to switch on and off and limited carryover effects. It supports linear factor models—hence, a generalization of gsynth—and the matrix completion method.
Repo: GitHub (1.0.0)
Examples: The original R tutorial can be replicated in Python with this tutorial. You can also find a markdown version tutorial here
Reference: Licheng Liu, Ye Wang, Yiqing Xu (2021). A Practical Guide to Counterfactual Estimators for Causal Inference with Time-Series Cross-Sectional Data. American Journal of Political Science, conditionally accepted.
The original R package of fect can be found here: GitHub (1.0.0). This package is based on rpy2, which makes R objects available in Python environments.
You can install fect from github by the following command:
pip install git+https://github.com/xuyiqing/fect_python.git
- Python 3.7+
- rpy2 3.5+
- numpy 1.1+
- pandas 1.1.2+
Although fect_python works on the latest version of rpy2, we strongly recommend installing rpy2 3.5.4 in case of potential conflits.