Audit committee financial expertise, audit committee independence, and regulatory oversight on external auditors

ABSTRACT

We empirically investigate how regulatory oversight on external auditors is jointly influenced by audit committee financial expertise and independence. To measure regulatory oversight on external auditors, we use comment letters issued by the Securities and Exchange Organization of Iran. We show that audit committee financial expertise increases (decreases) regulatory oversight on external auditors when audit committee independence is low (high). We further show that this interactive effect is stronger under higher regulatory reviewers’ workload compression. Collectively, our findings suggest that, first, financial expertise and independence of audit committees should be analysed together as independence moderates the benefit of financial expertise. Second, the consideration of regulatory reviewers’ workload compression is important in this analysis.

KEYWORDS:

• Audit committee
• external auditor
• regulatory oversight
• Johnson-Neyman approach

JEL CLASSIFICATION:

• M41
• M42
• G35

Acknowledgement

We deeply appreciate constructive insights from Flora Muiño (associate editor) and two anonymous reviewers.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1. In this paper, the term ‘external auditors’ refers to both audit firms and individual auditors as the separation of regulatory oversight or its determinants and consequences in terms of audit firms and individual auditors is a challenging issue, and not very useful in the context of this paper.

2. In psychology, cognitive processing shortcuts are mental, simple, and efficient cues, which people usually employ to form judgements and make decisions (e.g. Lewis, 2008).

3. Currently, information on CL is only available upon request from the SEO’s Division of Auditing and Corporate Reporting. However, the SEO is now developing a disciplinary opinion dissemination system at http://www.seo.ir.

4. Iranian firms’ financial data are not available in international databases such as Compustat and CRSP. Rahavard Novin provides data coverage similar to CRSP/Compustat in the United States (Aflatooni et al., 2022) and is available under subscription (see https://mabnadp.com/rahavardnovin3). Nevertheless, the data is also fully and freely available on CODAL.

5. Notably, out of all firms whose auditors received (did not received) a regulatory comment letter, 38.7% (31.3%) of them received a regulatory comment letter. This is relatively consistent with prior research (e.g. Ballestero & Schmidt, 2022) indicating that auditor involvement in review process and CL likelihood is higher for firms receiving a regulatory comment letter. Moreover, the correlation between the two types of comment letters, as shown in Table 3, is about 22%. This is because there are many situations – such as vague audit reports (deficiencies in non-audited corporate reports) — in which only auditors (only firms) receive comment letters.

6. In the presence of the interaction term, VIFs for the interaction term and main independent variables, i.e. Expert & Indep, are respectively 28, 17, and 12, and the average (maximum) VIF for the other variable is 1.8 (4.2). The three high VIFs do not refer to a multicollinearity problem, as prior literature (e.g. Disatnik & Sivan, 2014; McClelland et al., 2017) clearly discusses that regression with interaction term does not suffer from a multicollinearity problem when the independent variables are highly correlated with their product since this multicollinearity is simply a matter of interval scaling.

7. We choose these levels because most Iranian audit committees are composed of three members (Oradi & Izadi, 2019).

8. Accordingly, the area between the blue vertical lines is the region where the conditional slope is not significantly different from zero.

9. Esfand is the final month of the Solar Hijri calendar, the official calendar of Iran. It begins in February and ends in March of the Gregorian calendar.

10. The Z-statistic tests for differences across groups (e.g. Cohen, 1983; Holzhacker et al., 2014) and is computed as follows: $Z=\hspace{0.17em}\frac{C_1^{\hspace{0.17em}}-C_2^{\hspace{0.17em}}}{\sqrt[\hspace{0.17em}]{Std.Error{\left(c_1^{\hspace{0.17em}}\right)}^2+Std.Error{\left(c_2^{\hspace{0.17em}}\right)}^2}}$, where c₁ and c₂ are δ₃ in Equation (2)(2) $CL\mathrm{\_}{A}_{it}={\delta }_0+{\delta }_{1}Exper{t}_{it}+{\delta }_{2}Inde{p}_{it}+\hspace{0.17em}{\delta }_3(Expert\times Indep{)}_{\mathrm{i}\mathrm{t}}+Control{s}_{it}+F\mathrm{\_}FE+Y\mathrm{\_}FE+{\epsilon }_{it}^{\prime }$(2) and se refers to standard errors.

11. Audit Organization of Iran is a government entity, which is acknowledged to be the largest audit firm in Iran. Since the international big audit firms (BIG4) have not set up their activities in Iran, following the literature (e.g. Mihret et al., 2020; Oradi et al., 2020), this paper classifies audit firms in two groups of big (Audit Organization) and small (all other audit firms).