Corporate Risk Disclosures and Audit Fees: A Text Mining Approach

• In Event: Auditing 3

Abstract

This study examines the association between textual risk disclosures in the risk factor section from 10-K filings and audit fees using a text mining approach. Specifically, we use Natural Language Processing (NLP) techniques to extract firms’ self-identified risks including financial, strategic, operational, and hazard risks based on an enterprise risk management framework. This novel methodology enables us to detect firms’ risk explicitly disclosed in their filings based on a domain-specific wordlist generated by a superior statistical approach. We find that audit fees are significantly and positively related to firm-specific textual disclosures in financial, strategic, and operational risks after controlling for endogeneity. In contrast to prior studies arguing that firm risk disclosures are likely to be boilerplate, our findings suggest the informativeness of corporate textual risk disclosures in that they are associated with the risk perceived by auditors. This study also provides direct support for the recent U.S. reporting regulatory requirement of adding a new section on risk factors in corporate annual reports.

Authors

• Rong Yang, Rochester Institute of Technology

• Manlu Liu, Rochester Institute of Technology

• Yang Yu, Rochester Institute of Technology

• Kean Wu, Rochester Institute of Technology