본 연구는 피감대상기업의 성과조정 재량적 발생액의 수준으로 측정한 감사인의 보수적 감사와 감사인 교체 사이의 관련성에 대해 조사하였다. 실증분석 결과 감사인 교체 직전 연도의 재량적 발생액이 다른 연도보다 유의적으로 더 낮았다. 이는 감사인 교체 여부를 감사인이 미리 인지한 후 마지막 감사시점에서 보수적으로 회계처리를 수행한다는 의미다. 시기를 구분하여 감사인 강제교체제도가 적용되던 2006-2009년의 기간 동안에 감사인이 강제교체되었던 기업의 경우, 감사인 교체 1년 전에서 3년 전까지 피감대상기업의 재량적 발생액에 유의적인 변화가 없었다. 그러나 동 기간 동안 감사인이 강제교체되기 이전에 자발적으로 교체된 기업의 경우를 보면, 감사인 교체 직전 연도의 재량적 발생액이 다른 연도보다 유의적으로 낮은 현상이 계속 발견되었다. 2006년 이전이나 2009년 이후의 기간을 구분해서 살펴봐도 계속해서 교체 직전 연도의 재량적 발생액이 유의적으로 낮았다. 이 발견은 감사인 교체의사결정이 감사인의 보수성에 미치는 효과와 감사인 강제교체제도 도입의 영향을 알려준다는 의미에서 공헌점이 있다.

This study investigates whether auditor’s conservative audit approach leads to the client firm’s decision for auditor dismissal (i.e., client-initiated auditor change) and whether the auditor dismissal decision, in turn, influences auditor conservatism at the last year of his tenure. In the U.S. and most of other countries, auditor dismissal decision is disclosed publicly and, in most cases, auditor is not aware of the auditor dismissal in advance. Using the U.S. data, DeFond and Subarmanyam (1998) find that discretionary accruals are income-decreasing during the last year with the incumbent auditor. Several studies in Korea also document similar results (Kwon and Kim 2010; Sohn and Kim 2005). However, in Korea, the decision-making process of auditor dismissal is different. Auditor dismissal decision by client firms needs to be approved by audit committee in the board or the auditor selection committee. Subsequently, the auditor dismissal and the appointment of new auditor should be approved in the annual shareholders’ meeting. Before the shareholders’ meeting, the firm should notify the shareholders of the agenda of shareholders’ meeting which includes the information on auditor change. These procedures enable auditors to acknowledge the firm's decision to dismiss or to retain the auditor even before the completion of the annual audit work. When auditors acknowledge the dismissal, they have strong incentives to perform thorough and conservative audit procedures to minimize the potential litigation and regulatory risks in the future. For example, income-increasing discretionary accounting choices can lead to future litigations from investors (St. Pierre and Andersen 1984) and, as a result, the discretionary accruals at year t-1 (i.e., the last year with the incumbent auditor) is expected to be negative. In addition, Kwon and Kim (2010) and Sohn and Kim (2005) document that the discretionary accruals are also negative in year t-2. However, we expect the discretionary accruals in year t-2 to be less negative than those in year t-1 because the auditor will be dismissed in year t-1 if the auditor enforces conservative accounting choices (i.e., income-decreasing accruals) in year t-2 (i.e., two year before the auditor dismissal) and auditors try to lower discretionary accruals in year t-1 intentionally to minimize future risks. This last point is a unique prediction in this study. In addition, this study investigates how the enforcement of mandatory auditor rotation policy in Korea from 2006 to 2009 influences this relation between the conservative audit approach and auditor change. This study empirically examines these predictions using data collected over the period from 2002 to 2012. Among a total of 8,423 firm-year observations, 1,486 observations are the samples that experience auditor changes at the current year, while the remaining 6,937 observations do not experience the auditor change. To measure the level of conservative accounting treatments, we use performance-adjusted discretionary accruals measured by Kothari et al.’s (2005) method. The negative discretionary accruals proxies for the magnitude of conservative accounting treatments. The empirical findings support our predictions. The level of discretionary accruals is significantly negative in year t-1. The slightly different results from those in Sohn and Kim (2005) appears to be their failure to address measurement errors in the estimation of discretionary accruals. In addition, we find that this phenomenon does not exist during the 2006-2009 period for the firms that were required to switch auditors after 6 years of continuous auditor tenure. These findings provide various interesting information to regulators, practitioners, and academics. Although Korea introduced three-year mandatory auditor retention and six-year mandatory auditor rotation policies to promote auditor independence and to improve audit quality, the effect of these regulations is still in debate. High auditor independence and high audit quality are difficult to achieve. One possible way to resolve this issue could be the enhancement of the role of independent audit committee.