scpi-pkg 3.0.1

The package computes point estimates and prediction intervals for Synthetic Control methods with multiple treated units and staggered adoption as proposed in Cattaneo, Feng, Palomba, and Titiunik (2024).

SCPI_PKG

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

Authors

Matias D. Cattaneo ([e-mail])

Yingjie Feng ([e-mail])

Filippo Palomba ([e-mail])

Rocio Titiunik ([e-mail])

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 (2024+) scpi: Uncertainty Quantification for Synthetic Control Estimators. forthcoming at Journal of Statistical Software.

Technical and Methodological

• Cattaneo, Feng, and Titiunik (2021): Prediction Intervals for Synthetic Control Methods. Journal of the American Statistical Association.
• Cattaneo, Feng, Palomba, and Titiunik (2023): Uncertainty Quantification in Synthetic Controls with Staggered Treatment Adoption, working paper.