End-to-End Python framework reproducing Dolphin et al.'s (2026) grounded event extraction (from SEC 8-K filings) method. Extracts taxonomy-validated, quote-grounded event tags via LLMs, then proves economic significance through variance-standardized abnormal returns (SAR), Kruskal-Wallis tests, and bootstrap variance decomposition over CRSP data.
-
Updated
Jul 18, 2026 - Jupyter Notebook