Machine Learning Instrument Variables for Causal Inference

EC '20: Proceedings of the 21st ACM Conference on Economics and Computation

46 Pages Posted: 6 Apr 2019 Last revised: 18 Jul 2023

See all articles by Amit Gandhi

Amit Gandhi

University of Pennsylvania

Kartik Hosanagar

University of Pennsylvania - Operations & Information Management Department

Amandeep Singh

University of Washington - Michael G. Foster School of Business; University of Pennsylvania - The Wharton School

Date Written: March 15, 2019

Abstract

Instrumental variables (IVs) are a commonly used technique for causal inference from observational data. In practice, the variation induced by IVs can be limited, which yields imprecise or biased estimates of causal effects and renders the approach ineffective for policy decisions. We confront this challenge by formulating the problem of constructing instrumental variables from candidate exogenous data as a learning problem. We provide formal asymptotic theory and show root-n consistency and asymptotic efficiency of our estimators hold under very general conditions. We show that for linear models with homoskedasticity, this translates to a standard learning problem with cross-fitting. Simulations and application to real-world data demonstrate that the algorithm is highly effective and significantly improves the performance of instrumental variable estimators from observational data. Finally, we look at recent research that critiqued the use of political cycles as an instrument for advertising. Specifically, the authors test the strength of the first stage category-by-category for 274 product categories. The authors find that for most categories, the first-stage F-statistics are less than 10 (221 of 274 product categories) in their benchmark. We demonstrate most of the issues found by the authors can be resolved using MLIVs.

Keywords: Econometrics, Machine Learning, Causal Inference, Empirical Industrial Organization

Suggested Citation

Gandhi, Amit and Hosanagar, Kartik and Singh, Amandeep, Machine Learning Instrument Variables for Causal Inference (March 15, 2019). EC '20: Proceedings of the 21st ACM Conference on Economics and Computation, Available at SSRN: https://ssrn.com/abstract=3352957 or http://dx.doi.org/10.2139/ssrn.3352957

Amit Gandhi

University of Pennsylvania ( email )

Philadelphia, PA 19104
United States

HOME PAGE: http://web.sas.upenn.edu/agandhi/cv/

Kartik Hosanagar

University of Pennsylvania - Operations & Information Management Department ( email )

Philadelphia, PA 19104
United States

Amandeep Singh (Contact Author)

University of Washington - Michael G. Foster School of Business ( email )

Box 353200
Seattle, WA 98195-3200
United States

University of Pennsylvania - The Wharton School ( email )

3641 Locust Walk
Philadelphia, PA 19104-6365
United States

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Downloads
880
Abstract Views
3,763
Rank
53,631
PlumX Metrics