Summary

This training module provides an overview of the endogeneity problem in linear models, and ways of identifying and correcting for endogeneity.  The focus is on the methods of instrumental variables and two-stage least squares; the control function approach is also briefly introduced.

Objectives

The general objective of this training module is to provide an overview of approaches to testing and correcting for endogeneity in linear models using cross-sectional data.  We will cover the methods of instrumental variables and two-stage least squares, then briefly discuss the control function approach to testing and correcting for endogeneity.

The more specific objectives are to:

1. Provide a review of the problem of endogeneity, what it is and why Ordinary Least Squares may not be an appropriate estimator in the presence of endogeneity.
2. Discuss approaches for testing for endogeneity.
3. Discuss the instrumental variables, two-stage least squares, and control function approaches to correcting for endogeneity.
4. Provide guidance on how to evaluate the performance of the instrumental variables/two-stage least squares estimator.
5. Demonstrate how to implement the procedures in STATA.

Contributors

Paul Samboko, Nicole Mason-Wardell, and Saweda Liverpool-Tasie

Begin Module