본 연구의 목적은 계속 감사기간과 감사품질의 관계에 대해서 규제기관적 관점(auditor entrenchment hypothesis)과 감사인의 관점(auditor expertise hypothesis)에 따른 대립적인 가설로 분석하고, 의무적인 감사인 교체 정책(mandatory auditor rotation policy)의 유효성에 대해서 평가하는데 있다. 감사품질은 산업별 횡단면 수정 Jones 모형과 성과조정 수정 Jones모형에 의해서 재량적 발생액으로 측정하였다. 계속 감사기간과 감사품질의 관계는 단일변량분석, OLS분석, 재량적 발생액과 계속 감사기간의 내생성을 통제한 2SLS(two-stage least squares) 분석과 piecewise linear regression을 통한 비선형 분석을 통해서 검증하였다. 거래소 상장법인에 대해서 분석한 결과, 계속 감사기간과 재량적 발생액이 유의한 음의 상관관계를 가지는 것을 발견하였다. 계속 감사기간이 길어지면 감사품질이 향상되는 것으로 해석할 수 있으며, 감사인의 관점(auditor expertise hypothesis)을 지지하는 결과이다. 이러한 관계는 재량적 발생액과 계속 감사기간이 가지는 내생성을 통제한 2SLS분석 결과로도 지지되며, 감사인 지정기업을 통제한 분석, 감사품질을 재량적 발생액의 절대값으로 측정한 경우, 연도별 분석과 표본을 달리한 분석에서도 발견되었다. Piecewise linear regression으로 계속 감사기간에 대한 비선형성 분석을 한 결과 계속 감사기간이 6년 이하일 때와 2년에서 8년 사이일 때 감사품질과 유의한 양의 관계를 가지는 것을 확인하였다. 계속 감사기간과 감사품질간에 비선형성이 존재하면 감사인의 관점에서도 의무적인 감사인 교체 정책이 지지될 수 있다는 정책적 시사점을 얻을 수 있었다.
The purpose of this study is to examine the relation between auditor tenure and audit quality and to estimate the effect of mandatory auditor rotation policy. The debate on the mandatory auditor rotation policy is related to the research on the auditor tenure and the audit quality. Despite the importance and relevance of the topic, academic studies have only recently started examining the effect of the duration of auditor-client relationships on the quality of financial reporting. In the US, there is research consistent with both auditor entrenchment and auditor expertise hypotheses. However, recently major research is consistent with only the auditor expertise hypothesis. The auditor entrenchment hypothesis suggests that audit quality is decreases in a long auditor tenure. The auditor`s complacency and economic dependence impairs its independence. In this argument, a mandatory auditor rotation policy is a major vehicle to enhance auditor independence and audit quality. On the other hand, the auditor expertise hypothesis argues that audit quality improves the longer the audit tenure because the auditors are unfamiliar with the client`s business, operations, systems, controls, and accounting policies in their early years of tenure. As there is the information asymmetry between audit firms and auditors in the early stage audit engagement, sufficient time is necessary for the auditors to obtain knowledge and experience of audit firms for good audit quality. Therefore, the auditor expertise hypothesis opposes mandatory auditor rotation policy based on the information asymmetry argument. Since these two hypotheses have conflicting views on auditor tenure and audit quality, this issue needs empirical testing. As audit quality increases, the auditor is willing to detect and correct aggressive and questionable financial statement treatment. Earnings management by managers, therefore, will decrease, so audit quality has a negative relation with discretionary accruals. The audit quality is measured by the magnitude of discretionary accruals based on thew cross-sectional modified Jones model and the performance-matched modified Jones model. Univariate results(Table 4: Univariate analysis) and multivariate ones(Table 6: Regressions of discretionary accruals on auditor tenure and control variables) show that audit quality has a positive relationship with auditor tenure, consistent with the auditor expertise hypothesis. The results are also supported by the 2SLS (two-stage least squares regression) analysis that addresses endogeneity of the auditor tenure and audit quality (Table 8: Two-stage least squares regressions of discretionary accruals on auditor tenure and control variables: Control for the endogeneity of auditor tenure), the analysis controlled the auditor designation policy (Table 9: Regressions of discretionary accruals on auditor tenure and control variables: Control for auditor designation policy in Korea) and other additional analyses(
). I find anonlinear relation between auditor tenure and audit quality, which is significantly positive below 6 years and between 2 and 8 years of auditor tenure(
). The nonlinear results have an implication that the mandatory auditor rotation policy can be supportive in light of the auditor expertise hypothesis . Furthermore, the 6-year rotation requirement enforced in Korea from 2006 will have some policy effectiveness. This paper contributes to the literature in several important ways. First, we try to test the nonlinearity features in auditor tenure and audit quality more precisely. Though Ghosh and Moon (2002) tests the nonlinearity using the regression with a dummy variable, this paper examines the nonlinearity using graph analysis and piecewise linear regression methodology. I find that if there is a nonlinear relationship between auditor tenure and audit quality, a mandatory auditor rotation policy is acceptable in view of the auditor expertise hypothesis also. Second, I examine the relation robustly using 2SLS (two-stage least squares) analysis. Because auditor tenure and audit quality (discretionary accruals) have an endogeneity problem according to the auditor change literature, general OLS analysis can create biased results. Third, I try to measure auditor tenure and audit quality with few measurement errors. Previous literature such as Ghosh and Moon (2002), Myers et al. (2003) and Mansi et al. (2004) measure auditor tenure by setting 1 to the beginning year of sample period. But I calculate auditor tenure more accurately by defining those firms which change auditors after 1994 as sample. Also, I enhance the audit quality variable using both the cross-sectional modified Jones model (1995) and the performance- matched modified Jones model (2005). As this paper addresses external auditor change which affects auditor tenure, it can not control internal partner and staff rotation effects which may influence audit quality. Regulation on external audits in Korea requires the same partner not to audit for four consecutive years, and two-thirds of the staff to rotate after three consecutive years of engagement for a listed company.
