pyxtabond2 0.2.0

Dynamic panel GMM estimation replicating Stata's xtabond2

PyXtabond2: Dynamic Panel Data Estimation in Python

pyxtabond2 is a comprehensive Python package for estimating dynamic panel data models using the Generalized Method of Moments (GMM). It aims to faithfully replicate the functionality, matrix algebra, and robust diagnostics of Stata's highly popular xtabond2 command developed by David Roodman.

Ideal for applied econometrics and macroeconomic research, this package bridges the gap between Python's data science ecosystem and advanced dynamic panel methodologies.

---

🌟 Key Features

• Difference GMM (Arellano-Bond 1991): Standard first-differenced GMM estimation.
• System GMM (Blundell-Bond 1998): Combined level and differenced equations for highly persistent series.
• Interactive Fixed Effects (PCA-GMM): Automatically detect and purge unobserved common factors using Bai & Ng (2002) and Ahn & Horenstein (2013) criteria.
• Windmeijer (2005) Correction: Exact finite-sample correction for two-step GMM standard errors.
• Forward Orthogonal Deviations (FOD): Arellano-Bover (1995) transformation, maximizing sample size in unbalanced panels with gaps.
• Instrument Collapsing: Prevents instrument proliferation and the weakening of overidentification tests.
• Comprehensive Diagnostics: Arellano-Bond AR(1)/AR(2) tests, Sargan/Hansen J-tests, and Difference-in-Hansen tests for instrument exogeneity.
• Direct Export: Export publication-ready tables directly to LaTeX or Microsoft Word.

---

📦 Installation

pip install pyxtabond2

For export to Word/Excel and example datasets:

pip install "pyxtabond2[export]"

Development install from source:

git clone https://github.com/Kahindo048/pyxtabond2.git
cd pyxtabond2
pip install -e ".[dev,export]"

🚀 Quick Start pyxtabond2 comes with integrated example datasets so you can start experimenting immediately.

from pyxtabond2.data_utils import PanelData
from pyxtabond2.api import PyXtabond2
from pyxtabond2.load_data import load_dataset

# 1. Loading the data
# Make sure the df_panel.xlsx file is in the same folder
try:
    df = load_dataset('df_panel.xlsx')
    print(f"Data loaded successfully: {df.shape[0]} observations.")
except FileNotFoundError:
    print("Error: The file 'df_panel.xlsx' could not be found.")


# 2. Data preparation
panel = PanelData(df, id_col='Country', time_col='Year')
panel.data['L1_Growth'] = panel.get_lag('Growth', 1)
df_ready = panel.data.reset_index()

id_col = 'Country'          
time_col = 'Year'           
dep_var = 'Growth'          
x_vars = ['Capital', 'Labor', 'Wage', 'Investment', 'Ide'] 
gmm_vars = ['Growth', 'Capital'] 
iv_vars = ['Ide']           

modele = PyXtabond2(df_ready, 
                    id_col = 'Country', # Group identifier (country, firm)
                    time_col = 'Year', # Time identifier
                    dep_var = 'Growth', # Dependent variable
                    x_vars = ['L1_Growth', 'Capital', 'Labor', 'Wage', 'Investment', 'Ide'], # Explanatory variables
                    gmm_vars =['Growth', 'Capital'], # Variables for Arellano-Bond instruments
                    iv_vars = ['Ide'], # Variables for standard instruments
                    model_type='difference')

# 3. Estimation
result = modele.fit()
result.summary()

# 4. Export results for publication
result.to_latex("gmm_results.tex", full_output=False)
result.to_word("gmm_results.docx", full_output=False)

See examples/example_basique.py for a full workflow with six specifications and comparative export.

Performance

Version 0.2.0 includes vectorized NumPy/Numba implementations of the heaviest routines (panel transforms, instrument construction, GMM engine). On a standard macro panel (~2 000 obs.), a full six-model workflow is roughly 40× faster than v0.1.x.

Tests

pytest tests/

---

📖 References & Methodology

This package implements the algorithms and corrections outlined in the following seminal papers:

Arellano, M., & Bond, S. (1991). Some tests of specification for panel data: Monte Carlo evidence and an application to employment equations. The Review of Economic Studies.

Arellano, M., & Bover, O. (1995). Another look at the instrumental variable estimation of error-components models. Journal of Econometrics.

Blundell, R., & Bond, S. (1998). Initial conditions and moment restrictions in dynamic panel data models. Journal of Econometrics.

Windmeijer, F. (2005). A finite sample correction for the variance of linear efficient two-step GMM estimators. Journal of Econometrics.

Roodman, D. (2009). How to do xtabond2: An introduction to difference and system GMM in Stata. The Stata Journal.

Bai, J., & Ng, S. (2002). Determining the number of factors in approximate factor models. Econometrica.

Ahn, S. C., & Horenstein, A. R. (2013). Eigenvalue ratio test for the number of factors. Econometrica.

---

🤝 Contributing

Contributions, issues, and feature requests are welcome! Feel free to check the issues page on the GitHub repository.

License

MIT License. See LICENSE if included in the distribution.