Heterogeneous-Endogenous-Effects-in-Networks

Matlab code to implement the two stage LASSO process as proposed in Peng (2019)

Code to replicate the empirical application in the paper is in the folder: empirical replication
The micro finance dataset is downloaded from https://web.stanford.edu/~jacksonm/Data.html

This package requires CVX solver which can be downloaded from here

HeterogeneousEndogenousEffects: Main script that executes an example with Erdos Renyi network or Small-World network.

estimateHEEM: Main function that implement the two stage LASSO estimator.

- Inputs:  y: outcome of interest. 
           X: exogeneous variable.
           M: adjacency matrix of network.
           ertol: error tolerance for the LASSO estimator.
- Output:  hateta: debiased estimator.
           betaEsts2: direct two stage LASSO estimator (without debiasing).
           V: standard error of the debiased estimator.