A Machine Learning Approach to Assessing Audit Quality in Company with Non-Switching Auditors: Extra Trees Classifier Model

Document Type : Research Paper

Authors

1 Department of Accounting, Khomein Branch, Islamic Azad University, Khomein, Iran.

2 Department of Economic and Administration Science Faculty, Lorestan University, Lorestan, Iran.

10.22059/ijms.2025.384690.677133

Abstract

In this study, the authors utilize machine learning techniques to investigate the likelihood of a company switching auditors and examine whether the increased likelihood of switching is associated with audit quality in Tehran stock exchange (TSE). This study aims to understand the impact of auditor switching on audit quality and uses adjusted restatements of financial statements (AudFailA and AudFailB) and a new modified report (NMR) as proxies to measure adjusted audit quality, based on the environmental conditions of the research. These findings indicate that companies with a higher likelihood of switching auditors but ultimately deciding to stay with incumbent auditors exhibit poor audit quality. It is worth noting that this study focuses on larger companies, and the results show a stronger association between audit quality and the probability of switching auditors in this group. This finding suggests that the relationship between audit quality and the decision to switch auditors is more significant among larger companies than among smaller ones. The methodology employed in this study is designed to be implemented by investors, audit firms, and regulators. It aims to identify companies with a higher probability of switching auditors, allowing for proactive measures to address the potential deterioration in audit quality.

Keywords

Main Subjects