multi-levelSC 0.1

Package to implement the multi-level SC estimator

Multi-level Synthetic Control Estimator

This package implements the multi-level SC estimator (mlSC) for the treatment effect for a single, treated, aggregated unit in panel data with multiple levels of aggregation, as proposed in Bottmer (2025).

This package is currently in beta and the functionality and interface is subject to change.

Installation

Install this library using pip:

pip install multi-level-sc-estimator

Example

from multi_level_sc_estimator.mlSC import mlSC_estimator

# Define data sets, treated unit, treated period, population weights (w_c) and how to estimate lambda.
mlSC_results = mlSC_estimator(data_s,data_c, idx, n_c, t, w_c, lambda_est = "heuristic")
tau_hat = mlSC_results[0]
lambda_hat = mlSC_results[1]
w_hat = mlSC_results[2]

References

Lea Bottmer. Synthetic Control with Disaggregated Data, 2025. [link]