Financial Statement Analysis with Large Language Models

ArXiv ID: 2407.17866 “View on arXiv”

Authors: Unknown

Abstract

We investigate whether large language models (LLMs) can successfully perform financial statement analysis in a way similar to a professional human analyst. We provide standardized and anonymous financial statements to GPT4 and instruct the model to analyze them to determine the direction of firms’ future earnings. Even without narrative or industry-specific information, the LLM outperforms financial analysts in its ability to predict earnings changes directionally. The LLM exhibits a relative advantage over human analysts in situations when the analysts tend to struggle. Furthermore, we find that the prediction accuracy of the LLM is on par with a narrowly trained state-of-the-art ML model. LLM prediction does not stem from its training memory. Instead, we find that the LLM generates useful narrative insights about a company’s future performance. Lastly, our trading strategies based on GPT’s predictions yield a higher Sharpe ratio and alphas than strategies based on other models. Our results suggest that LLMs may take a central role in analysis and decision-making.

Keywords: Financial Statement Analysis, Large Language Models (LLMs), Fundamental Analysis, Earnings Forecasting, Algorithmic Trading, Equities

Complexity vs Empirical Score

• Math Complexity: 1.5/10
• Empirical Rigor: 7.0/10
• Quadrant: Street Traders
• Why: The paper’s core methodology relies on prompting and analyzing outputs from a pre-trained LLM (GPT-4), lacking dense mathematical derivations or complex statistical theory, resulting in low math complexity. Conversely, it employs extensive real-world financial datasets, rigorous benchmarking against human analysts and traditional ML models, and evaluates live trading strategies with metrics like Sharpe ratio and alpha, indicating high empirical rigor.

  flowchart TD
    A["Research Goal<br>Can LLMs replicate expert<br>financial statement analysis?"] --> B
    B["Methodology<br>Provide standardized anonymous<br>financial statements to GPT-4"] --> C
    C["Data Input<br>Standardized Financial<br>Statements w/o Narrative"] --> D
    D["LLM Process<br>GPT-4 analyzes data to<br>predict earnings direction"] --> E
    E["Key Findings & Outcomes"]
    
    subgraph E
        F["LLM outperforms<br>human analysts"]
        G["Performance on par<br>with specialized ML models"]
        H["Trading strategies yield<br>higher Sharpe ratio & alphas"]
    end