경영자-감사인간의 의견불일치와 감사인교체

Auditor changes have been an important issue to practitioner, regulators, and academics. But the empirical researches finds little evidence of auditor changes. This study investigates whether the disagreement between auditor and client causes the auditor change. Currently, client(auditee) selects the auditor based on the market mechanism. Under the current system, since the auditor is chosen by client, there is a chance for opinion shopping. Disagreement between these two parties has not been fully examined even though it is very essential to understand the nature of this relationship because it is not observed and at the same time, it's not easy to find a proxy. While the preliminary earnings number reflects the intention of the management, the reported earnings number is the outcome of negotiation between two parties even though it can't be observable. We believe that this difference represents the auditor-client disagreement. The sample firm is composed of non-financial companies which are listed in Korean Stock Exchange for 5 years from 2000 to 2004. We collect from DART system(Data Analysis, Retrieval and Transfer system) on preliminary earnings. Financial data is collected from TS2000 and Kis-Line database. We eliminate samples which do not disclose preliminary earnings. Finally we use 574 firm/year samples. To determine whether the auditor change is associated with auditor-client disagreement, we estimate the logistic regressions because the dependent variable is dummy variable(auditor change or not). We measure auditor-client disagreement using the difference between preliminary earnings and actual earnings. And we control for Altman's z-score, accrual change, audit opinion and audit quality, proxied with big4 or non-big4. Our empirical results show that disagreement causes the auditor change. We believe that this empirical evidence signifies a meaningful policy implication. For example, when these two numbers are significantly different, regulatory body may exercise caution if there is an auditor change. Measurement of the disagreement has been an issue. Sometimes, this was measured with the discretionary accrual. However, it has a serious measurement problem. Our measure for auditor-client disagreement could be used extensively in the auditing literature.