FinCARE: Financial Causal Analysis with Reasoning and Evidence

ArXiv ID: 2510.20221 “View on arXiv”

Authors: Alejandro Michel, Abhinav Arun, Bhaskarjit Sarmah, Stefano Pasquali

Abstract

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 optimization (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 algorithms: 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 counterfactual predictions and perfect directional accuracy for intervention effects. It also addresses critical limitations of existing methods by grounding statistical discoveries in financial domain expertise while maintaining empirical validation, providing portfolio managers with the causal foundation necessary for proactive risk management and strategic decision-making in dynamic market environments.

Keywords: causal discovery, knowledge graph, large language models, counterfactual analysis, portfolio management, Equity

Complexity vs Empirical Score

• Math Complexity: 7.5/10
• Empirical Rigor: 4.0/10
• Quadrant: Lab Rats
• Why: The paper presents substantial mathematical complexity through formal causal discovery algorithms (PC, GES, NOTEARS), constraint formulations, and optimization objectives. However, its empirical validation is limited to a synthetic dataset, lacking real-world backtesting or implementation-heavy data pipelines required for Street Traders.

  flowchart TD
    A["Research Goal: Discover True Causal Drivers<br>in Financial Performance"] --> B["Data & Inputs"]
    B --> C["Computational Hybrid Framework"]
    C --> D["Key Outcomes & Findings"]
    
    subgraph B ["Data & Inputs"]
        B1["Synthetic Dataset<br>500 Firms, 18 Variables"]
        B2["Financial Knowledge Graph<br>from SEC 10-K Filings"]
        B3["Large Language Model Reasoning"]
    end
    
    subgraph C ["Computational Hybrid Framework"]
        C1["Statistical Causal Discovery Algorithms"]
        C2{"Enhance with Domain Knowledge"}
        C1 --> C2
        C2 --> C3["KG+LLM Enhanced Methods"]
    end
    
    subgraph C1 ["Statistical Causal Discovery Algorithms"]
        C1a["Constraint-based: PC"]
        C1b["Score-based: GES"]
        C1c["Continuous Optimization: NOTEARS"]
    end
    
    subgraph C2 ["Enhance with Domain Knowledge"]
        C2a["Algorithmic KG Constraints"]
        C2b["LLM Conceptual Reasoning"]
    end
    
    subgraph D ["Key Outcomes & Findings"]
        D1["Significant F1 Score Improvements"]
        D2["Reliable Scenario Analysis"]
        D3["Proactive Risk Management Capability"]
    end
    
    subgraph D1 ["Significant F1 Score Improvements"]
        D1a["PC: 0.622 vs 0.459<br>+36% improvement"]
        D1b["GES: 0.735 vs 0.367<br>+100% improvement"]
        D1c["NOTEARS: 0.759 vs 0.163<br>+366% improvement"]
    end
    
    subgraph D2 ["Reliable Scenario Analysis"]
        D2a["Mean Absolute Error: 0.003610<br>for Counterfactual Predictions"]
        D2b["Perfect Directional Accuracy<br>for Intervention Effects"]
    end
    
    D3["Framework grounded in<br>domain expertise with<br>empirical validation"]