Early Warning Signals for Financial Collapse: Replication Dataset and Code (2001–2024)

This dataset and code repository supports the paper "Early Warning Signals for Financial Collapse: A Critical Slowing Down Framework Applied to Equity Dynamics" (Malone, 2026). The package contains split- and dividend-adjusted daily equity price data for 46 financial collapse events and 29 stable control firms spanning 2001–2024, covering corporate bankruptcies, bank failures, accounting frauds, and industry shocks across U.S. and international markets. Macroeconomic series include FRED ICE BofA High Yield and BBB Option-Adjusted Spreads (1997–2026) and the St. Louis Fed Financial Stress Index (1993–2026). All Python analysis code is included, implementing rolling variance, lag-1 autocorrelation, Gaussian kernel detrending, Kendall's tau trend tests, Mann-Whitney U, and AUC-ROC with bootstrap confidence intervals — all from scratch in NumPy for full reproducibility without proprietary dependencies. Scripts reproduce all 9 figures and 3 tables in the paper.