Liberation Day and After
Whether the April 2025 tariff shock durably changed how U.S. equity sectors move together, and whether a forecast that says 70% is right 70% of the time.
Author · GTSF quant mentorship · Spring 2026 · GTSF (Georgia Tech Student Foundation)
Demo walkthrough coming soon
Why I built it
The April 2, 2025 tariff announcement was a clean natural experiment. I wanted to see whether the correlations between equity sectors actually broke in a structural way, not just noisily, in the days after. I ran structural-break tests, GARCH(1,1) volatility models, and a 2,000-iteration bootstrap over 11 sector ETFs and 323 trading days. Then I recalibrated the ensemble forecasts, because a model that's confidently wrong is a different problem from one that's honestly uncertain.
Results
- Energy beta, conditioned on regime
- 0.66 → 0.08
- Expected calibration error
- 0.050
- Brier score
- 0.2115
Stack
- Python
- pandas
- Jupyter
- Yahoo Finance
- FRED