Engagement partners participating in auditing standard setting and audit quality

* Accepted by Jeong-Bon Kim upon recommendation by Jong Hag Choi.

Abstract

In this study, we investigate whether engagement partners who participate in the national auditing standard setting can provide better audit quality, and explore the underlying cause in the context of Taiwan, where the name of engagement partners can be identified in audit reports. The empirical results indicate that firms audited by incumbent Auditing Standards Committee (ASC) member auditors are associated with lower discretionary accruals and a lower probability of having small profits. The results support the view that concerns about the impairments of increased reputational capital may be the most likely cause that explains the ASC member auditors’ better audit quality. This study also provides the benefits that can arise from the identification of engagement partners.

Keywords:

• Auditing standard setting
• partner-level characteristics
• reputation risks
• earnings management
• audit quality

Disclosure statement

No potential conflict of interest was reported by the authors.

Funding

This work was financial support of the Ministry of Science and Technology, Taiwan, ROC [Project No. MOST 105-2410-H-006 -024 -MY2].

Acknowledgements

We are grateful for the helpful comments from Jeng-Fang Chen, Wu-Chun Chi, Ping-Hsun Huang, Jan-Zan Lee, Stephen Taylor, Laurence van Lent, and participants at the Joint Conference of 26th Asian-Pacific Conference on International Accounting Issues & 2014 Accounting Theory and Practice Conference in Taipei and the 2017 British Accounting and Finance Association in Edinburgh. Hua-Wei Huang gratefully acknowledges the financial support of the National Science Council, Taiwan, ROC (Project No. MOST 105-2410-H-006 -024 -MY2).

Notes

1. In this study, we use the term ‘Big 4’ uniformly to represent the largest international Big ‘N’ audit firms in each period. For example, during 1998–2002, Big 4 comprises Arthur Andersen, Deloitte & Touche, Ernst & Young, KPMG, and PricewaterhouseCoopers.

2. Of course, although we present these three arguments based on the best of our knowledge, we cannot rule out other possible causes that may lead to our results.

3. In note 9, we discuss the composition of other national auditing standard setters.

4. For example, negative abnormal returns of client firms, client firms switch their audit firms, and the audit firm loses market share.

5. On May 9, 2016, the SEC (Securities and Exchange Commission) approved the Proposed Rules filed by PCAOB (SEC 2016). It requires that ‘for each audit report it issues for an issuer, a registered public accounting firm must file with the Board a report on Form AP that includes the following:

• The name of the engagement partner and Partner ID;

• For other accounting firms participating in the audit for which the responsibility for the audit is not divided’.

The new rules are effective for audit reports issued on or after January 31 (June 30), 2017 for the engagement partner (other accounting firms).

6. Regulations Governing Approval of Certified Public Accountants to Audit and Attest to the Financial Reports of Public Companies, Article 6. In fact, the Taiwan SASs also include audit-related services, such as examinations, reviews, agreed-upon procedures, and compilations.

7. The main differences between the members and consultants are that the consultants are set primarily for consultation and do not have the right to vote on the passage of auditing standards. Considering the core function in promulgating auditing standards, we focus on the association between the ASC member auditors and the audit quality provided by them.

8. According to article 9 of the Constitution of the ARDF, committee members and consultants are nominated by the chairperson of the committee and approved by the board of standing directors. Terms of committee members and consultants are same as that of the committee chairperson. Before the end of each term (3/31 every three years), the ASC committee chair will invite the heads (the actual title depends on each firm’s organizational structure) of the Big 4 audit firms to be ASC members. Generally, the heads of the Big 4 audit firms either accept the offer, or recommend their senior auditing partners as members of the ASC. According to our investigation, 60% (i.e. six engagement partners) of the ASC member auditors also held important positions in their respective audit firms, such as CEO, chief risk officer, chief information officer, and assurance service leader, in our sample period. This raises the possibility that the higher audit quality found in this study is actually due to concerns about reputational capital accompanied by the important positions held by these ASC member auditors. In Section 6, we use additional research designs to deal with this concern.

9. Some countries, such as the USA and China, also allow practicing CPAs to become members of their auditing standard setting bodies. In the USA, the independent Auditing Standards Board (ASB) of the AICPA promulgates standards for auditing, attestation, and quality control service for non issuers. Among the 19 members, five are from local, regional and other non-Big 4 national firms; five are nominated by the National Association of State Boards of Accountancy (NASBA); four are from the Big 4 firms; and five are users and public members. (http://www.aicpa.org/Research/Standards/AuditAttest/ASB/Pages/ASBbiography.aspx). In China, the Auditing Standards Committee is one of the 12 special committees under the Council of the Chinese Institute of Certified Public Accountants (CICPA). The main composition of its 31 members is similar to that in Taiwan (http://www.cicpa.org.cn/leaders/zmzywyh/201012/W020101214376848488603.doc).

10. In general, these studies have shown that the size of audit firms are positively related to audit quality, longer audit firm tenure does not compromise audit quality, and client firms’ comparability of earnings is higher when they are audited by the same Big 4 audit firm. Based on the argument that larger offices have greater in-house expertise so that they can provide better audit quality, Francis and Yu (2009) find that instead of industry expertise, the size of the Big 4 practice office is the fundamental driver of audit quality.

11. Using partner-level data from Taiwan, Aobdia, Lin, and Petacchi (2015) find that high quality audit partners can provide informational value to the capital market participants beyond audit firms. Firms audited by high quality audit partners therefore enjoy higher earnings response coefficients, less IPO underpricing, and better debt contract terms.

12. Of course, we admit that firm (or office) level inputs, such as audit firms' internal control, review, decision, and general governance mechanisms are of substantial importance.

13. We acknowledge that the validity of our results hinges on the accuracy of our proxies to measure discretionary accruals. In Section 6, we use various accruals models to confirm the robustness of our results.

14. Managers can manipulate unmanaged earnings by using accrual items (Dechow, Ge, and Schrand 2010); particularly, they have the most discretion over current accruals (e.g. Ashbaugh, LaFond, and Mayhew 2003; Becker et al. 1998; Chi et al. 2009). In addition, prior literature indicates that accruals are correlated with firm performance (Dechow, Sloan, and Sweeney 1995; Kothari, Leone, and Wasley 2005), therefore we control for firm performance (ROA) in Equation (1.1(1.1) $$C{A}_{i,t}={\mathit{\beta }}_1(1/A_{i,t-1})+{\mathit{\beta }}_2((\mathrm{\Delta }RE{V}_{i,t}-\mathrm{\Delta }A{R}_{i,t})/A_{i,t-1})+{\mathit{\beta }}_{3}RO{A}_{i,t-1}+u_{i,t}$$(1.1) ).

15. We also use net income as the numerator to define small profits. The results are unchanged from those of the main analysis.

16. This is the first year that the TEJ database begins to calculate consecutive tenure (in years) audited by the latest audit firm.

17. We define our variables of interest, for example, DURING, as 1 if any one of a firm’s engagement partners is the incumbent ASC member auditor. We use this criterion for two reasons. First, some of our arguments, such as reputation risks, should not make obvious differences between lead and concurring CPAs. Second, we do not need to assume the first CPA of an audit report as the lead CPA and the second CPA as the concurring CPA (e.g. Chi and Chin 2011).

18. Potential measurement errors arise because we calculate the value of ASBEXP since 2000. In fact, some engagement partners may also be ASC members before 2000. However, as detailed in Section 4 and note 20, year 2000 is a turning point for the financial and audit reporting environment in Taiwan, and thus may provide a good opportunity for well-prepared CPAs to distinguish their service from others. Limiting the sample period of the Accumulated Knowledge Model to 2009–2011 may partially mitigate concerns about measurement errors. We also conduct a sensitivity test in Section 6.

19. We do not control for ROA and LOSS in Equation (2.2)(2.2) $$\begin{array}{c}S{P}_{i,t}={\mathit{\beta }}_0+{\mathit{\beta }}_{1}BEFOR{E}_{i,t}+{\mathit{\beta }}_{2}DURIN{G}_{i,t}+{\mathit{\beta }}_{3}AFTE{R}_{i,t}+{\mathit{\beta }}_{4}SIZ{E}_{i,t}+{\mathit{\beta }}_{5}LE{V}_{i,t}+{\mathit{\beta }}_{6}AGE{M}_{i,t}\\ +{\mathit{\beta }}_{7}OC{F}_{i,t}+{\mathit{\beta }}_{\text{8}}INDEXP{T}_{i,t}+{\mathit{\beta }}_{\text{9}}PRE1C{A}_{i,t}+{\mathit{\beta }}_{\text{10}}TENUR{E}_{i,t}+{\mathit{\beta }}_{\text{11}}SDSALE{S}_{i,t}+{\mathit{\beta }}_{1\text{2}}CIM{P}_{i,t}+\\ INDUSTRY+YEAR+u_{i,t}\\ \end{array}$$(2.2) and (3.2)(3.2) $$\begin{array}{c}S{P}_{i,t}={\mathit{\beta }}_0+{\mathit{\beta }}_{1}ASBEX{P}_{i,t}+{\mathit{\beta }}_{2}SIZ{E}_{i,t}+{\mathit{\beta }}_{3}LE{V}_{i,t}+{\mathit{\beta }}_{4}AGE{M}_{i,t}+{\mathit{\beta }}_{5}OC{F}_{i,t}+{\mathit{\beta }}_{\text{6}}INDEXP{T}_{i,t}+{\mathit{\beta }}_{7}PRE1C{A}_{i,t}+\\ {\mathit{\beta }}_{8}TENUR{E}_{i,t}+{\mathit{\beta }}_{9}SDSALE{S}_{i,t}+{\mathit{\beta }}_{10}CIM{P}_{i,t}+INDUSTRY+YEAR+u_{i,t}\\ \end{array}$$(3.2) because a firm with negative ROA or that reports a loss can never report small profits at the same time, so ROA is mechanically positively correlated with small profits, while LOSS is mechanically negatively correlated with small profits.

20. With regard to auditing standards, Taiwan SASs No. 33 was issued in 1999. Unqualified audit opinions with explanatory paragraph(s) are separated from the qualified opinion; hence, the type of audit opinion is expanded to five. Since then, Taiwan SASs are issued or revised per Taiwan SASs No. 33. Moreover, in response to globalization and the need to increase the comparability of financial reporting between domestic and foreign companies, the Taiwan Financial Accounting Standards Committee (TFASC) decided to issue its twSFAS based on IFRS (i.e. to converge with IFRS). Accounting standards with new topics are based on contemporaneous IFRS (with some degree of carve-outs) since 2000 (Wu et al. 2017).

21. During 2000–2011, the ASC issued or revised 20 Taiwan SASs, in almost every year except 2002. The main advantage of our collection method is that we can even identify irregular turnovers (e.g. unexpected resignations).

22. To avoid the concern that our results are driven by the removal of outliers, we also try to rerun all equations (i.e. Equations (2.1.a) to (3.2)) without trimming outliers. The results are unchanged from those of the main analysis.

23. For brevity, we also only provide the results of Equations (2.1.a) and (2.2) in Table 4. The results of other equations are available upon request.

24. Similarly, the TWSE increased its industry classifications from 20 to 29 since 2007, with the difference being that it further split the electronics industry into eight new classifications.

25. Due to the collapse of Arthur Anderson in 2002, its member firm in Taiwan merged with the member firm of Deloitte Touche Tohmatsu in Taiwan in 2003.

26. According to the ‘Accounting Firm Service Survey’ issued annually by the Financial Supervisory Commission (FSC) of Taiwan, the percentage of female practicing CPAs during 2006–2011 is between 31and 34 percent. See website: http://www.fsc.gov.tw/ch/home.jsp?id=136&parentpath=0,4.

27. The results of all equations in this study are available upon request.

28. Untabulated results do not find any significant relationship between firms audited by the incumbent ASC member auditors and negative discretionary accruals in all equations of this study. This result is also consistent with the argument of Lennox, Wu, and Zhang (2016) that the implication of a change in negative discretionary accruals is not clear from the perspective of audit quality.

29. Similar to Carcello and Li (2013), we calculate economic implications by multiplying the DURING coefficient by (probability of a small profit) × (1 – probability of a small profit). Given the mean probability of a small profit (Table 4) is 0.0694, so the economic implication is (−0.7997) × (0.0694)*(1 – 0.0694) = −0.052.

30. We thank the referee’s advice that reminds us to further analyze the differences in coefficients of the variables of interest (i.e. BEFORE, DURING, and AFTER)

31. As another matching criterion, we allow for matching firm-years audited by different and same audit firms. This lead to a larger sample size for ABSCDA and SP (N = 206) and POSCDA (N = 54). The empirical results are unchanged. However, we decide not to match firm-years audited by the same audit firms, because the number of matched pairs is too small to make reliable inferences.

32. However, a caution regarding the small size problem should be made when using a logistic regression model (Long 1997). Therefore, the results of univariate analysis may be more representative when comparing SP.

33. We do not use PSM (propensity score matching) to deal with functional form misspecification in this study given that our variables of interest, BEFORE, DURING, and AFTER, are relatively rarely coded 1 (see Table 2). When proceeding with the conventional PSM research designs, we find that the sample size is further reduced substantially. According to Shipman, Swanquist, and Whited (2017, 220): ‘The smaller the sample size, the less likely it is that the subsample generalizes to the population.’ However, we still acknowledge that there may be some unknown functional forms not captured by our expanded model.