SCPI_PKG

The scpi_pkg package provides Python implementations of estimation and inference procedures for Synthetic Control methods.

Authors

Matias D. Cattaneo (cattaneo@princeton.edu)

Yingjie Feng (fengyj@sem.tsinghua.edu.cn)

Filippo Palomba (fpalomba@princeton.edu)

Rocio Titiunik (titiunik@princeton.edu)

Website

https://nppackages.github.io/scpi/

Installation

To install/update use pip

pip install scpi_pkg

Usage

from from scpi_pkg.scdata import scdata
from from scpi_pkg.scdataMulti import scdataMulti
from scpi_pkg.scest import scest
from scpi_pkg.scpi import scpi
from scpi_pkg.scplot import scplot
from scpi_pkg.scplotMulti import scplotMulti

• Replication: Germany reunification example.

Dependencies

• cvxpy (>= 1.1.18)
• dask (>= 2021.04.0)
• ecos (>= 2.0.7)
• luddite (>= 1.0.2)
• numpy (>= 1.20.1)
• pandas (>= 1.5.0)
• plotnine (>= 0.8.0)
• scikit-learn (>= 0.24.1)
• scipy (>= 1.7.1)
• statsmodels (>= 0.12.2)

References

For overviews and introductions, see nppackages website.

Software and Implementation

• Cattaneo, Feng, Palomba, and Titiunik (2022) scpi: Uncertainty Quantification for Synthetic Control Estimators.

Technical and Methodological

• Cattaneo, Feng, and Titiunik (2021): Prediction Intervals for Synthetic Control Methods.<br> Journal of the American Statistical Association.
• Cattaneo, Feng, Palomba, and Titiunik (2022): Uncertainty Quantification in Synthetic Controls with Staggered Treatment Adoption, working paper.

<br><br>