The Economist's Guide to Causal Forests

This repository contains the code and data accompanying the blog article The Economist's Guide to Causal Forests.

Overview

The article demonstrates how to use the grf package in R for estimating heterogeneous treatment effects. The example analyzes the effect of 401(k) eligibility on household savings using data from the 1990 Survey of Income and Program Participation.

Quick Start

Prerequisites

Install the required R packages:

install.packages("grf")
install.packages("foreign")

Running the Code

1. Clone this repository:

   git clone https://github.com/JulianStreyczek/causal-forests-guide.git
   cd causal-forests-guide

2. Run code/grf_application_401k.R using R.

What the Code does

The script performs the following analysis:

• Loads the SIPP 1991 dataset containing information on 401(k) eligibility and household finances
• Trains a causal forest with 12,000 trees to estimate conditional average treatment effects (CATEs)
• Computes the average treatment effect of 401(k) eligibility on net total financial assets
• Generates visualizations:
  • Histogram of estimated treatment effects across all observations
  • Treatment effect by age with 95% confidence intervals

Repository Contents

causal-forests-guide/
├── README.md                         
├── LICENSE                           # MIT license
├── data/
│   └── sipp1991.dta                  # Survey of Income and Program Participation (1990)
├── code/
│   └── grf_application_401k.R        # R script for causal forest analysis
├── outputs/
│   ├── CATEdistribution.png          # Histogram of treatment effects
│   └── CATEbyAge.png                 # Treatment effects by age with confidence intervals
└── paper/
    └── causalforests.pdf             # PDF version of the guide

Main References

This article builds on the following key papers:

• Athey, S., & Imbens, G. (2016). Recursive partitioning for heterogeneous causal effects. Proceedings of the National Academy of Sciences, 113(27), 7353-7360.

• Wager, S., & Athey, S. (2018). Estimation and inference of heterogeneous treatment effects using random forests. Journal of the American Statistical Association, 113(523), 1228-1242.

• Athey, S., Tibshirani, J., & Wager, S. (2019). Generalized random forests. The Annals of Statistics, 47(2), 1148-1178.

Software:

• grf: Generalized Random Forests - R package documentation available at https://grf-labs.github.io/grf/index.html

Dataset source:

• Chernozhukov, V., & Hansen, C. (2004). The effects of 401(k) participation on the wealth distribution: an instrumental quantile regression analysis. Review of Economics and Statistics, 86(3), 735-751.

License

This project is licensed under the MIT License.