파트너연고 감사인변경과 이익조정의 관련성

Prior studies (Blouin et al. 2007; Vermeer et al. 2008 etc.) identified the auditor changes following partner familiarity in the incidents of auditor changes caused by the close of Arthur Andersen (AA). They named the AA clients as the “follower”, who have changed auditors to the same audit firms that the regional audit partners have transferred to; the AA clients other than the follower are the “non-follower”. I identified similar incidents in Korean audit firms' organizational change history. Following to global accounting firms’ organizational changes – including close of AA– after the accounting scandals such as Enron, Worldcom, etc., many global network accounting firms in Korea also made organizational changes subsequently. For example, Anjin LLC.(Anjin; the former AA member firm) changed its network relationship to Deloitte, a division of Ahnkwon LLC. (Ahnkwon; the former Deloitte member firm) was spun off to new firm of Hana LLC.(Hana), which subsequently merged to Anjin after Anjin became Deloitte member. Meanwhile, In March 2005, Ahnkwon (after spin off of Hana firm) has substantially closed its audit business and the majority of their partners and audit clients have moved to Ernst & Young Hanyoung (Hanyoung) - the E&Y firm in Korea. Among these Korean audit firm’s organizational changes, I identify the incident that the majority of the former Ahnkwon partners moved to Hanyoung and, simultaneously, large number of the former Ahnkwon clients changed their auditors to Hanyoung. Ahnkwon’s 32 partners (64%) out of 50 partners in 2004 became Hanyoung’s partners in 2005. Moreover, according to New KIS VALUE database, 86 (60%) out of 143 public company audits and 288 (68%) out of 424 private company audits of Ahnkwon in 2004 become Hanyoung’s new audit clients in 2005. In addition, the former Ahnkwon clients consist of 67% of Hanyoung’s 2005 new public and private company audits, respectively. Auditor changes consist of two procedures: (i) dismissal or resignation, and (ii) new appointment. Audit literature to investigate auditor changes does not usually separate such procedures since they can not be discriminated from the auditor change incidents.(Nichols and Smith 1983; Shu 2000; Landsman et al. 2006). However, similar with AA’s close of audit business, audit clients of the audit firms closing its audit business or liquidated should appoint new auditors; this enable us to investigate auditor change incidents focusing on the second stage only. Based on these partner transfer and audit client’s audit changes, I see the above incident as the “auditor changes following partner familiarity” as reported by Blouin et al. (2007) and Vermeer et al. (2008). Samples are 86 firms which are the Ahnkwon's 2004 audit clients who have changed auditors in 2005. Test samples are 460 firm- years from 2002 to 2007, consisting of the experimental samples (follower) of 333 firm-years and the control samples (non-follower) of 127 firm-years. Auditor changes following partner familiarity are likely to associated with earnings management or audit quality. For example, the higher auditor switching costs, the more likely the clients follow partner familiarity. Accordingly, earnings management (audit quality) of the follower clients is likely to higher (lower) than the non-follower clients before the mandatory auditor changes. Also, the audit quality of the followers would be lower than the control samples of the non-follower clients after the auditor change since the partner familiarity would threaten the independence of auditors. However, there is another possibility that the new audit firm would suspiciously see the new client because of Ahnkwon’s bad reputations and takes a harsh quality control monitoring process, which would mitigate the threats against auditor independence. In the case it is estimated that audit quality of the follower clients would not be significantly worse than (or different from) that of the non-follower clients after the auditor change. I found that the follower client’s discretionary accruals is systematically higher than the non-follower client samples (control samples) before auditor changes. It is consistent with the previous study(Blouin et al. 2007) reporting that auditor switching costs of the partner-familarity followers are higher than non-follower clients. Also, I found that there is no significant difference between discretionary accruals of the follower and the non-follower clients. It is consistent with the findings of Cahan and Zhang(2006) and Nagy(2005) who report that new auditors accepting clients of the former AA subsequent to its audit businss closure take relatively harsh quality control process. This result indicate that auditor changes following partner familiarity are associated with high switching costs, and, new auditor’s independence in substance are properly maintained. It has practical implications since higher switching clients (i.e. higher accrual client firms) may be likely to have incentives to avoid harsh audits by auditor changes following partner familiarity; however, the above results indicate that such relatively higher discretionary accruals may not be maintained under the new audit firm's quality control system in spite of the continued partner-client familiarity relationship. It has contributions that, (i) in spite of general perceptions of independence threats after the auditor changes following partner familiarity, I empirically find the audit quality are properly monitored by the new audit firms, and (ii) I identify partner familiarity incidents in Korean audit practice experiences and empirically investigate its relations with audit quality.