AKM Model Translation from R to Python

This repository contains a Python implementation of the Abowd, Kramarz, and Margolis (1999) AKM two-way fixed effects model, translated from an original R implementation by Stephen Tino (University of Toronto). The AKM model is widely used in labor economics to estimate worker and firm contributions to earnings inequality using matched employer-employee data.

📌 Project Goals

• Convert the R-based AKM estimation to Python
• Maintain equivalent model estimation, variance decomposition, and fixed-effects computation
• Extend the model to support gender-specific AKM estimation

📂 Repository Structure

• r_version/ – Original R scripts for reference
• python_version/ – Translated Python implementation
• data/ – Simulated employer-employee dataset for testing
• README.md – Project documentation

🔗 Original R-based Implementation

The original R implementation was developed by Stephen Tino (University of Toronto).
Author: Stephen Tino, PhD Candidate in Economics, University of Toronto, s.tino@mail.utoronto.ca
GitHub Repository: https://github.com/stephentino/estimate_akm

📄 License

This project retains the LGPL-2.1 license as required by the original repository.

👥 Contributors

• Chiayi (Aiden) Wu
• Mia Zhang
• Brenda Noh

🚀 Getting Started

1. Clone the Repository

git clone https://github.com/cwu423/akm_python_translation.git
cd akm_python_translation