Benchmarking Generalization in Financial Statement Fraud Detection: robust evaluation and novel tasks

2026-07-21

AI Summary

This work addresses critical limitations in existing financial fraud detection methods, which often rely on random data splits that overestimate model generalization and neglect textual information in financial reports. To remedy this, the study introduces a new task—Company-Isolated Financial Statement Fraud Detection (CI-FSFD)—featuring a cross-company isolation evaluation paradigm that more realistically reflects model performance on unseen firms or future periods. The authors propose a unified framework integrating structured financial data with unstructured Management Discussion and Analysis (MD&A) text, leveraging large language models for multimodal fusion. They also release the first public, comprehensive dataset comprising U.S. public company financial statements, MD&A sections, and fraud labels. Experiments demonstrate that incorporating textual information significantly enhances model generalization, establishing a robust new benchmark for the field.

Abstract

Financial statement fraud detection (FSFD) is crucial for market integrity but faces challenges from increasingly sophisticated schemes and under-utilized textual data in financial reports. Existing methods often rely on random data splits, leading to overoptimistic performance estimates that do not reflect real-world generalization to new companies or future periods. To address this recurring problem with the state of the art, we propose a robust FSFD framework leveraging Large Language Models (LLMs) to integrate both structured financial data and unstructured textual information from financial reports. We provide a more realistic evaluation through a novel and challenging benchmark task called Company-Isolated FSFD (CI-FSFD). We construct and make publicly available a comprehensive U.S. company dataset combining financial statements, summarized MD&A text, and fraud labels. Our approach achieves the best performance on the challenging CI-FSFD task, demonstrating the critical value of textual data and robust evaluation for reliable financial fraud detection.

Problem

• Financial Statement Fraud Detection
• Generalization
• Textual Data
• Robust Evaluation
• Benchmarking

Innovation

• Large Language Models
• Financial Statement Fraud Detection
• Company-Isolated Benchmark
• Textual Financial Data
• Robust Evaluation