감사시간의 의무 공시 이후 감사시간에 대한 관심이 높아지고 있다. 감사시간이 감사품질의 대용치 역할을 할 수 있기 때문이다. 그러나 감사시간에 대한 관심에 비해 감사시간의 의무 공시 자체에 대한 관심은 부족한 것으로 보인다. 즉 감사시간의 의무 공시가 감사시간에 미치는 영향은 아직 주목받지 못하고 있다. 또한 감리 대상 선정 요소의 하나로 감사투입시간이 활용되는 새로운 감리제도도 보고되는 감사시간에 영향을 미쳤을 수 있다. 본 연구에서는 감사시간의 의무 공시, 새로운 감리제도 등 제도적 요인이 보고되는 감사시간에 미치는 영향을 살펴보았다. 이때 감사시간의 변화는 비기대감사시간을 통해 분석하였는데, 비기대감사시간은 감사시간에 영향을 미치는 요소들을 회귀식을 통해 통제한 잔차로 계산하였다. 그리고 의무 공시가 감사보수에 미치는 영향 역시 추가 분석하여 감사시간의 공시행태 변화에 대한 추가적인 해석을 시도하였다. 논문접수: 2008. 9 게재확정: 2009. 3 *연세대학교 경영대학 박사과정, Ph. D. Student, Yonsei University, School of Business, Seoul, Korea (E-mail: sunh@yonsei.ac.kr) ** 연세대학교 경영학 석사, 한영회계법인 공인회계사, CPA, Ernst & Young Hanyung, Seoul, Korea (E-mail: nayahami@hanmail.net) *** 연세대학교 경영대학 교수, Professor, Yonsei University, School of Business, Seoul, Korea (E-mail: hylee@yonsei.ac.kr) 실증결과, 감사시간의 의무 공시 이후 비기대감사시간의 분산과 절대값은 전기대비 및 시계열적으로도 줄어들었고, 이는 감사시간이 유사한 특징을 가진 기업들 간에 서로 수렴하고 있음을 의미한다. 그러나 2004년부터 실질적으로 적용된 새로운 감리제도가 감사시간에 미치는 추가적인 영향은 없는 것으로 나타났다. 이는 새로운 감리제도 이전부터 이미 비정상적인 감사시간을 감독기관에서 관찰하고 있었으며, 감사인 및 피감사기업은 불필요한 감독기관의 주의를 끌거나 투자자들의 감사품질에 대한 의심을 받지 않기 위해 감사시간을 조정하려는 유인이 있었던 것으로 해석된다. 반면 감사보수의 수렴현상은 감사시간의 경우만큼 확연히 나타나지는 않았다. 이는 감사시간의 수렴현상에는 감사품질에 대한 의구심을 피하기 위한 기회주의적 공시 행태가 영향을 미치고 있다는 주장을 뒷받침 해준다.
The Financial Supervisory Service (FSS) of Korea promulgated a new regulation enhancing the transparency of corporate disclosures in 2001. One of major changes from this regulation includes the disclosure of audit hours spent by incumbent auditors. The mandatory disclosure of audit hours provides information users with an opportunity to estimate the quality of audit. However, the disclosure of audit hours was not standardized and were difficult to interpret because the regulation did not provide detailed guidelines for the disclosure. In 2003, FSS, therefore, renewed the regulation providing more detailed guidelines for the disclosure in firms’ annual report. Audit hours may indicate the amount of audit effort and, therefore, the quality of audit. With the recognition of importance of audit hours ensuring audit quality, the FSS stated that audit hours reported in annual reports can be used as one of the selection criteria for audit review in finding audit failures performed by external auditors. Spending abnormally long audit hours may imply auditors' additional effort maintaining a good quality audit. On the other hand, longer audit hours may imply problems in clients’ internal control systems and books. Also, spending abnormally short audit hours may imply less efforts and poor quality audit. The above arguments imply that audit hours can be an important piece of information for earnings quality and investment decisions. Thus, investors and regulators are interested in information contained in audit hours. With this in mind, FSS, in particular, decided to use audit hours as one of selection criteria for audit review in monitoring audit quality in 2005. Auditors may be suspected as conducting low quality audit when they report abnormally long or short audit hours relative to firms with similar characteristics. In order to avoid any concerns raised by investors and/or regulators over the audit quality, both auditors and clients have incentives to manipulate audit hours disclosed in their annual reports. However, reporting behavior with respect to audit hours has not been examined and information users do not understand underlying meanings of audit hour and firms’ reporting behavior. This study, therefore, examines if reporting behavior pertaining to audit hours shows any systematic trend(s) over time and around the introduction of audit hour as the selection criterion for FSS audit review. Auditors and/or client managers who have strong incentives to minimize the possibility being investigated by FSS are likely to reduce the variance of abnormal audit hours and the absolute value of abnormal audit hours after the requirement of audit hour disclosure. This study also hypothesizes the negative association between prior period abnormal auditor hour and change in audit hour from prior period. This study examines whether a change in audit review policy affects clients’ firms audit hour reporting behavior. In 2005, FSS announced that they use audit hours and public financial data as selection criteria for audit review. Both clients and auditors do not want to be investigated by FSS because of direct and/or indirect costs of investigation associated with audit review and they have incentives to change their reporting behavior systematically. In practice, FSS has used audit hour as one of the selection criteria for their audit review since 2004. Thus, this study expects that the systematic patterns with respect to the variance of abnormal audit hour and the absolute value of abnormal audit hour would be more pronounced after the use of audit hour as an audit review selection criterion by FSS than before. The sample of this study covers firms having December fiscal year from year 2001 to 2005 listed in either the Korea Stock Exchange or KOSDAQ markets. After considering the availability of financial, audit fee, and audit hour data, the final sample includes 4,689 firm-years. Financial date are collected from Kis-Value and DataGuide Pro of FnGuide databases while ownership and subsidiaries data are collected from TS2000. Audit hour and audit fee or nonaudit fee data are manually collected from annual reports available at the FSS's Dart system. Considering prior studies (e.g., Simunic 1980; Palmrose 1989; O'Keefe et al. 1994; Choi and Paek 1998; Kwon and Kim 2001; Park 2005; Choi et al. 2006; Caramanis and Lennox 2008) in determining audit hours, this study develop an abnormal audit hour model. Control variables include the natural log of number of subsidiaries, inventory and receivables, proportion of foreign sales over total sales, issuance of stock and bonds, current ratio, leverage, ROA, sales growth, largest shareholder ownership, foreign shareholder ownership, nonaudit fee, Big4(5) auditors, audit opinion, first year audit, and designated auditor. Consistent with our expectation, we find that the variance and the magnitude of unexpected audit hours have been deceasing over time. This result suggests that firms observed other firms with similar characteristics in audit hour disclosures have tried to minimize the deviation from audit hours reported by other firms. The results of variance and Kruskal-Wallis analyses also provide evidence that the absolute value of abnormal audit hours has been reduced over time and the mean values are statistically significantly different across years. Regression analyses on the association between prior period abnormal audit hour and changes in audit hour also show that prior period abnormal audit hours affect current period audit hour. The analyses on the effect of change in audit review selection policy on reported audit hour reveal that the change in audit hour has decreased after the policy change. As an additional analysis, this study examines reporting behavior of audit fees to see whether audit fee shows a similar pattern to audit hour. The result shows that standard deviations of audit fees have increased after 2002. The absolute value of abnormal audit fees have also increased since 2002. This result is inconsistent with Francis and Wang (2005) who finds a decreasing pattern after audit fee disclosure requirement. We interpret the results from two different perspectives: static and time-series perspectives. Unlike the results with audit hours, audit fees show increasing deviation over the sample years suggesting that auditors are getting more sensitive to audit risk as litigation risk increases. Overall, the results of this study suggest that auditors and/or client firms tend to avoid reporting abnormal audit hours by adjusting actual audit hours or possibly opportunistically manipulating reported audit hours. A disclosure of information may not be useful if the disclosure is not reliable. Thus, further study is needed to investigate various aspects of audit quality and their associations with audit hour.
