Abstract
We examine the association between audit fees and accounting misstatements and test the moderating effect of auditor reputation on this association in China, where auditors’ legal liability is essentially weak. We find that current-year audit fees are negatively and significantly associated with the likelihood of the current-year financial statement being misstated. This result is consistent with prior work, which has indicated a negative relation between audit fees and future misstatements in a strong litigation environment such as in the United States. Our finding indicates that higher audit fees can reflect audit effort, even in a setting where litigation plays essentially no role. The prior literature suggests that audit firms have reputational incentives; thus, we use auditor irregularities and audit firm rank as proxies for auditor reputation and find that the negative relationship between misstatements and audit fees weakens as auditor reputation decreases.
Disclosure statement
No potential conflict of interest was reported by the authors.
Funding
We gratefully acknowledge National Natural Science Foundation of China (71502138, 71102095, 71572144) and China Postdoctoral Science Foundation (2014M550505; 2015T81041), and Natural Science Foundation of Shaanxi (2015JQ7281).
Notes
1. Recently, arguably the most publicized restatement cases in China involved the Wanfu Biotechnology Agricultural Development Company in 2013 and the Yunnan Green-Land Biological Technology Company in 2011. These firms’ auditors were Zhonglei and Pengcheng, respectively. The CSRC administered the most severe penalties to these two culprits, that is, suspension of securities services. However, Zhonglei and Pengcheng merged into Daxin and Kunitomi Hiroka, respectively, before enforcement of the penalties.
2. Please find more from the website of CICPA: http://www.cicpa.org.cn/Column/swszhpm/.
3. Lower rank means the auditor is a lower-ranking one, whereas higher rank means the auditor is a higher-ranking one.
4. The number becomes smaller due to audit firm mergers.
5. The CICPA is required by the MOF to seek those individual talented CPAs and to let them become leaders of CPA industry to make the Chinese CPA industry develop well.
6. Disclosed misstatements are obtained through financial restatements in later years.
7. URL: http://www.cninfo.com.cn/.
8. This sample period is chosen due to audit fee data availability, because audit fee disclosure was not mandatory until 2001 in the Chinese capital market.
9. We also utilized a logit regression model. The results are very similar.
10. The CSRC may disclose and punish those auditors helping the listed firms to cook the numbers. Furthermore, the CICPA began to examine CPA’s work in 1999 to ensure that audit firms adhere to the Chinese Auditing Standard. Inspection announcements will be disclosed to public when each examination ends. By the end of 2013, CICPA has issued 28 inspection announcements. These announcements specify auditors’ irregularities and the resulting punishments. Based on these, we classify those auditors’ behaviors punished by CICPA or CSRC as auditor irregularity.
11. Thanks for the suggestions of one anonymous reviewer.
12. Available upon request.
13. The special treatment (ST) companies are those with a high risk of being delisted, which includes companies that reported consecutive annual losses in the two most recent years, those that have serious errors and irregularities in their financial statements or indications of fraud, companies that missed the deadlines for the release of annual or semi-annual reports, and companies with other problems as determined by Chinese regulations.
14. In accordance with Blankley, Hurtt, and MacGregor (Citation2012), we generate abnormal audit fees. Our results remain the same.
15. We are very grateful for this suggestion from the anonymous reviewer.
16. We form a new variable LVIO, equals one if the audit firm violated audit standards or committed fraud during our sample and was punished publicly by the CSRC and zero otherwise. Then we use this variable to replace AGGVIO and run the regression. The regression results are quite similar.
17. We are very grateful for this suggestion from the anonymous reviewer.
18. We also use Top 30 and Top 40, and get the quite similar results.