IDEAS home Printed from https://ideas.repec.org/p/arx/papers/2506.08762.html

EDINET-Bench: Evaluating LLMs on Complex Financial Tasks using Japanese Financial Statements

Author

Listed:
  • Issa Sugiura
  • Takashi Ishida
  • Taro Makino
  • Chieko Tazuke
  • Takanori Nakagawa
  • Kosuke Nakago
  • David Ha

Abstract

Large Language Models (LLMs) have made remarkable progress, surpassing human performance on several benchmarks in domains such as mathematics and coding. A key driver of this progress has been the development of benchmark datasets. In contrast, the financial domain poses higher entry barriers due to its demand for specialized expertise, and benchmarks remain relatively scarce compared to those in mathematics or coding. We introduce EDINET-Bench, an open-source Japanese financial benchmark designed to evaluate LLMs on challenging tasks such as accounting fraud detection, earnings forecasting, and industry classification. EDINET-Bench is constructed from ten years of annual reports filed by Japanese companies. These tasks require models to process entire annual reports and integrate information across multiple tables and textual sections, demanding expert-level reasoning that is challenging even for human professionals. Our experiments show that even state-of-the-art LLMs struggle in this domain, performing only marginally better than logistic regression in binary classification tasks such as fraud detection and earnings forecasting. Our results show that simply providing reports to LLMs in a straightforward setting is not enough. This highlights the need for benchmark frameworks that better reflect the environments in which financial professionals operate, with richer scaffolding such as realistic simulations and task-specific reasoning support to enable more effective problem solving. We make our dataset and code publicly available to support future research.

Suggested Citation

  • Issa Sugiura & Takashi Ishida & Taro Makino & Chieko Tazuke & Takanori Nakagawa & Kosuke Nakago & David Ha, 2025. "EDINET-Bench: Evaluating LLMs on Complex Financial Tasks using Japanese Financial Statements," Papers 2506.08762, arXiv.org, revised Mar 2026.
  • Handle: RePEc:arx:papers:2506.08762
    as

    Download full text from publisher

    File URL: http://arxiv.org/pdf/2506.08762
    File Function: Latest version
    Download Restriction: no
    ---><---

    References listed on IDEAS

    as
    1. van der Heijden, Hans, 2022. "Predicting industry sectors from financial statements: An illustration of machine learning in accounting research," The British Accounting Review, Elsevier, vol. 54(5).
    2. Messod D. Beneish, 1999. "The Detection of Earnings Manipulation," Financial Analysts Journal, Taylor & Francis Journals, vol. 55(5), pages 24-36, September.
    3. Monahan, Steven J., 2018. "Financial Statement Analysis and Earnings Forecasting," Foundations and Trends(R) in Accounting, now publishers, vol. 12(2), pages 105-215, July.
    4. Shijie Wu & Ozan Irsoy & Steven Lu & Vadim Dabravolski & Mark Dredze & Sebastian Gehrmann & Prabhanjan Kambadur & David Rosenberg & Gideon Mann, 2023. "BloombergGPT: A Large Language Model for Finance," Papers 2303.17564, arXiv.org, revised Dec 2023.
    5. Patricia M. Dechow & Weili Ge & Chad R. Larson & Richard G. Sloan, 2011. "Predicting Material Accounting Misstatements," Contemporary Accounting Research, John Wiley & Sons, vol. 28(1), pages 17-82, March.
    Full references (including those not matched with items on IDEAS)

    Most related items

    These are the items that most often cite the same works as this one and are cited by the same works as this one.
    1. Alexander Dyck & Adair Morse & Luigi Zingales, 2024. "How pervasive is corporate fraud?," Review of Accounting Studies, Springer, vol. 29(1), pages 736-769, March.
    2. Anas Alqudah & Safa Jaradat & Lara Al-Haddad, 2025. "Financial Consequences of Fraud in Amman Stock Exchange Firms," Advances in Decision Sciences, Asia University, Taiwan, vol. 29(1), pages 83-111, March.
    3. repec:osf:socarx:48g5c_v1 is not listed on IDEAS
    4. Dan Amiram & Zahn Bozanic & Ethan Rouen, 2015. "Financial statement errors: evidence from the distributional properties of financial statement numbers," Review of Accounting Studies, Springer, vol. 20(4), pages 1540-1593, December.
    5. Zhang, Yi & Liu, Tianxiang & Li, Weiping, 2024. "Corporate fraud detection based on linguistic readability vector: Application to financial companies in China," International Review of Financial Analysis, Elsevier, vol. 95(PB).
    6. Bidisha Chakrabarty & Pamela C. Moulton & Leonid Pugachev & Xu (Frank) Wang, 2025. "Catch me if you can: In search of accuracy, scope, and ease of fraud prediction," Review of Accounting Studies, Springer, vol. 30(2), pages 1268-1308, June.
    7. Zhang, Zejun & Wang, Zhao & Cai, Lixin, 2025. "Predicting financial fraud in Chinese listed companies: An enterprise portrait and machine learning approach," Pacific-Basin Finance Journal, Elsevier, vol. 90(C).
    8. Su, Huishui & Jiang, I-Ming & Liu, Duan, 2025. "Detecting financial fraud risk using machine learning: Evidence based on different categories and matching samples," Finance Research Letters, Elsevier, vol. 85(PA).
    9. Li, Guowen & Wang, Shuai & Feng, Yuyao, 2024. "Making differences work: Financial fraud detection based on multi-subject perceptions," Emerging Markets Review, Elsevier, vol. 60(C).
    10. Ilia Dichev & Edward Owens, 2025. "Accrual duration," Review of Accounting Studies, Springer, vol. 30(1), pages 611-649, March.
    11. Lokanan, Mark & Sharma, Satish, 2025. "Reprint of: The use of machine learning algorithms to predict financial statement fraud," The British Accounting Review, Elsevier, vol. 57(1).
    12. Shiqi Wang & Zhibo Zhang & Libing Fang & Cam-Tu Nguyen & Wenzhong Li, 2025. "Corporate Fraud Detection in Rich-yet-Noisy Financial Graph," Papers 2502.19305, arXiv.org, revised May 2025.
    13. Sonika Gupta & Sushil Kumar Mehta, 2024. "Data Mining-based Financial Statement Fraud Detection: Systematic Literature Review and Meta-analysis to Estimate Data Sample Mapping of Fraudulent Companies Against Non-fraudulent Companies," Global Business Review, International Management Institute, vol. 25(5), pages 1290-1313, October.
    14. Rahman, Md Jahidur & Zhu, Hongtao, 2024. "Detecting accounting fraud in family firms: Evidence from machine learning approaches," Advances in accounting, Elsevier, vol. 64(C).
    15. Slavko ?odan, 0000. "Can Accrual-based Metrics Indicate Material Accounting Misstatements? Evidence on Audit Adjustments," Proceedings of Economics and Finance Conferences 14416287, International Institute of Social and Economic Sciences.
    16. Cheng-Wen Lee & Mao-Wen Fu & Chin-Chuan Wang & Muh. Irfandy Azis, 2025. "Evaluating Machine Learning Algorithms for Financial Fraud Detection: Insights from Indonesia," Mathematics, MDPI, vol. 13(4), pages 1-35, February.
    17. Zhao, Dongshuai & Wang, Zhongli & Schweizer-Gamborino, Florian & Sornette, Didier, 2025. "Polytope Fraud Theory," International Review of Financial Analysis, Elsevier, vol. 97(C).
    18. repec:osf:osfxxx:kecmv_v1 is not listed on IDEAS
    19. Nguyen Huu Anh & Nguyen Ha Linh & Sung Wook Yoon, 2018. "Earnings Manipulation Benchmark For Nonfinancial Listed Companies In Vietnamese Stock Market," Accounting & Taxation, The Institute for Business and Finance Research, vol. 10(1), pages 39-49.
    20. Wei, Lang & Zhang, Yiling, 2023. "Nonfinancial indicators in identifying stock price crash risk," Finance Research Letters, Elsevier, vol. 52(C).
    21. Mohammed-Khalil Ghali & Cecil Pang & Oscar Molina & Carlos Gershenson-Garcia & Daehan Won, 2025. "Forecasting Commodity Price Shocks Using Temporal and Semantic Fusion of Prices Signals and Agentic Generative AI Extracted Economic News," Papers 2508.06497, arXiv.org.
    22. Xiaowei Chen & Cong Zhai, 2023. "Bagging or boosting? Empirical evidence from financial statement fraud detection," Accounting and Finance, Accounting and Finance Association of Australia and New Zealand, vol. 63(5), pages 5093-5142, December.

    More about this item

    NEP fields

    This paper has been announced in the following NEP Reports:

    Statistics

    Access and download statistics

    Corrections

    All material on this site has been provided by the respective publishers and authors. You can help correct errors and omissions. When requesting a correction, please mention this item's handle: RePEc:arx:papers:2506.08762. See general information about how to correct material in RePEc.

    If you have authored this item and are not yet registered with RePEc, we encourage you to do it here. This allows to link your profile to this item. It also allows you to accept potential citations to this item that we are uncertain about.

    If CitEc recognized a bibliographic reference but did not link an item in RePEc to it, you can help with this form .

    If you know of missing items citing this one, you can help us creating those links by adding the relevant references in the same way as above, for each refering item. If you are a registered author of this item, you may also want to check the "citations" tab in your RePEc Author Service profile, as there may be some citations waiting for confirmation.

    For technical questions regarding this item, or to correct its authors, title, abstract, bibliographic or download information, contact: arXiv administrators (email available below). General contact details of provider: http://arxiv.org/ .

    Please note that corrections may take a couple of weeks to filter through the various RePEc services.

    IDEAS is a RePEc service. RePEc uses bibliographic data supplied by the respective publishers.