Amihud Stress and ETF Return Predictability: Evidence from Pre- and Post-COVID Volatility Regimes

This dataset and code package supports the study titled "Liquidity Stress, ETF Returns, and Structural Breaks: Evidence from Pre- and Post-COVID Markets."It investigates liquidity stress measures, ETF trading performance, and structural breaks around the COVID-19 shock, using daily data from 2010–2025. Included in this deposit: /src/full_analysis.py: Complete Python script that replicates all results, figures, and statistical tests in one file. /data/: F-F_Research_Data_5_Factors_2x3_daily.CSV: Fama-French 5-Factor daily dataset. F-F_Momentum_Factor_daily.CSV: Fama-French UMD (Momentum) factor daily dataset. Figures: Figure 1.png: Alternative Liquidity Stress Measures. Figure 2.png: Sharpe Ratios by ETF: Pre- and Post-COVID Comparison. Figure 3.png: SPY Cumulative Returns (Pre/Post COVID). Figure 4.png: Approximate Bai-Perron Structural Break Detection (SSR vs Break Date). README.md: Basic project instructions and file structure overview. LICENSE (CC BY 4.0).txt: License file permitting reuse with attribution. requirements.txt: Python package requirements for full reproducibility. Key methods implemented: Amihud illiquidity calculation Rolling stress detection (top 5% extreme illiquidity) Simple ETF trading strategy backtests across holding periods (5, 15, 30 days) Pre- vs. Post-COVID Sharpe ratio comparisons Fama-French 5-Factor + Momentum factor regressions (Newey-West HAC standard errors) Chow Test for structural breaks at COVID onset Approximate Bai-Perron search for best breakpoints using Sum of Squared Residuals (SSR) This work is licensed under Creative Commons Attribution 4.0 International (CC BY 4.0). Researchers, practitioners, and students are welcome to reuse, adapt, or extend this dataset and code for academic, professional, or instructional purposes.