Software
We build open-source software for structural-change econometrics and forecasting. Our tools are designed for researchers and practitioners who need to detect instability, estimate models across regimes, and build forecasting workflows that adapt when relationships shift.
regimes — Structural-Change Econometrics for Python
regimes is a Python package for regime detection and estimation in time-series econometrics. It provides:
- Models: OLS, AR, and ADL models with structural breaks, rolling and recursive estimation
- Break detection: Bai-Perron sequential testing, Chow tests, CUSUM, Andrews-Ploberger
- Regime switching: Markov switching models
- Model selection: GETS indicator saturation (impulse, step, trend, and mixed indicators)
- Visualization: Regime-aware plots, coefficient paths, break diagnostics
Current version: v0.4.0 (in development). The package follows the statsmodels API pattern and is designed to integrate with the standard Python data science stack.
Coming in Future Versions
- Documentation and PyPI release
- VAR models, cointegration, panel data
- End-to-end forecasting workflows
Get Involved
regimes is open source. View the code and contribute on GitHub.