External Auditor Reassessment of Client Business Risk Following the Issuance of a Comment Letter by the SEC

Abstract

Following Arthur Andersen's conviction for obstructing justice, auditors faced a one-time significant change in their regulatory environment because it was clear that (i) major audit partnerships could be closed and (ii) post Sarbanes-Oxley Act (SOX), regulators would take a far more attentive (aggressive) role. In response auditors considered whether the pricing of audits should be revised to take account of the increased risk of regulatory intervention and litigation. Obviously such re-pricing would need to be targeted at those firms for which the risks were greatest. One early warning signal of such events occurring is the issuance by the Security Exchange Commission (SEC) of a Comment Letter (CL). We investigate whether there is any evidence that if a client receives a CL this is used to re-price audit services. Specifically, we investigate whether issuance resulted in upward pressure on audit fees, and whether this effect was simply transient around the issuance period or alternatively persisted some years into the future. This research finds that after a client receives a CL, auditors adjust audit fees upwards in the period in which the CL is received. In addition it is shown that for subsequent periods in which the auditor does not spend time assisting the client respond to a specific CL, an initial rise in audit fee persists. This is consistent with the hypothesis that auditors reassess the reputation and litigation risk of the client on the basis of the SEC issuance of a Comment Letter.

Acknowledgements

We thank Salvador Carmona (Past Editor), Ann Vanstraelen (Associate Editor) and the two anonymous reviewers for their helpful comments and suggestions. We also thank Erroll Stewart and Lauren Dreher for comments at the January 2011 mid-year AAA Audit Conference and Clive Lennox for additional comments in June 2011.

Notes

^†“Best practices for managing unresolved SEC Comment letters”, in SEC Comment Letters on Domestic Registrants. Deloitte (2009).

1 ‘SEC insight’, filed and recorded on the SEC site http://www.sec.gov/news/press/s72804/secinsight093004.pdf

2 In a very small number of cases firms explicitly disclose what additional fees they pay auditors to help them deal with responding to Comment Letters. See Appendix C for an example.

3 By significant we mean excluding those Comment Letters that are simply asking for minor clarification. We provide more details in Section 4.

4 Theoretical support for existence of such correlations is presented in Subsection 2.3.

5 While the receipt of an SEC comment letter might potentially affect also client risk (especially in cases of a subsequent restatement resulting from the letter, the ability of the company to obtain financial funding might be significantly impaired), we believe it is mostly the engagement aspect of business risk (i.e. engagement risk) that plays a role, through increased reputation and litigation risk. This is explained later in Subsection 2.1.

6 We are aware that any potential increase in audit fees after the receipt of a CL might be caused by either an increase in price per audit hour (p) or by an increase of effort (q) or by both (p×q). As explained in the conclusion section we comment that public data limitation does not permit us to disentangle such effects. Moreover, we are more interested in the final impact of regulatory intervention on total audit fees rather that where does this increase origin from.

7 Additional information about the SEC Comment Letters process is provided in Appendix 4.

8 Appendix C gives more detail on this particular filing.

9 The sample includes also firms which did not receive any Comment Letter in the sample period (923 firms).

10 That is a sub-sample of CLs referring to accounting issues also referred to at least one of ‘other’ or ‘risk’ issues.

11 The same firm can actually receive more than one letter in the sample period. In particular, in our sample, around 1800 firms received more than one letter in the sample period.

12 Charles et al. (2010) report similar negative return on assets for the 2000–2003 period.

13 The percentage of increase is computed on raw audit fees, using the formula e^z−1, where the z exponent value is the coefficient value of .029.

14 LET_ACC, LET_RISK and LET_DISC partially overlap (see previous section on ‘Data Sources and Model Specification’). For this reason, we perform the variance inflator factor and verify that multicollinearity is not biasing our results.

15 We eliminated cases where subsequent letters were received from this analysis.

16 Using a 2SLS regression when one of the endogenous variables is dichotomous still leads to efficient and consistent coefficients (Heckman, 1978; SOX, 2002). In order to further control for this issue, however, we re-estimate the model using a two-stage probit least square (in which the first-stage regression is run using a probit model, Maddala, 1983). Results are confirmed.