FinCARE: Financial Causal Analysis with Reasoning & Evidence

Portfolio managers rely on correlation-based analysis and heuristic methods that fail to capture true causal relationships driving performance. We present a hybrid framework that integrates statistical causal discovery algorithms with domain knowledge from two complementary sources: a financial knowledge graph extracted from SEC 10-K filings and large language model reasoning. Our approach systematically enhances three representative causal discovery paradigms, constraint-based (PC), score-based (GES), and continuous optimiza- tion (NOTEARS), by encoding knowledge graph constraints algorithmically and leveraging LLM conceptual reasoning for hypothesis generation. Evaluated on a synthetic financial dataset of 500 firms across 18 variables, our KG+LLM- enhanced methods demonstrate consistent improvements across all three algo- rithms: PC (F1: 0.622 vs. 0.459 baseline, +36%), GES (F1: 0.735 vs. 0.367, +100%), and NOTEARS (F1: 0.759 vs. 0.163, +366%). The framework enables reliable scenario analysis with mean absolute error of 0.003610 for counterfac- tual predictions and perfect directional accuracy for intervention effects. It also addresses critical limitations of existing methods by grounding statistical discov- eries in financial domain expertise while maintaining empirical validation, pro- viding portfolio managers with the causal foundation necessary for proactive risk management and strategic decision-making in dynamic market environments.