초과감사보수가 신용등급에 미치는 영향 : 상장기업과 비상장기업의 비교분석

This study examines whether and how abnormal audit fees paid out by firms affect its credit ratings. In addition, we test the effect of abnormal audit fees on credit ratings on subsamples of public and private firms because listed firms are larger in size and are likely to pay higher audit fees compared to non-listed firms. We hypothesize that credit-rating agencies evaluate differently the effect of abnormal audit fees on credit ratings because the size and audit fees are very different between public firms and private firms. Prior research implies that audit quality is affected by audit fees. To date, extant research, however, focuses on the relation between audit quality, proxied by the magnitude of discretionary accruals, and abnormal audit fees (Choi et al. 2010; Hoitash et al. 2007; Coulton et al. 2007; Kim and Choi 2009). Several previous studies investigate the relation between cost of capital and abnormal audit fees (Hope et al. 2009; Kim and Park 2010; Oh et al. 2008). However, there are few extant studies that examine the effect of audit quality on debt market participants. In this paper, we seek to fill this gap by investigating the relation between abnormal audit fees and firms' credit ratings. Examining how abnormal audit fees are evaluated by capital market participants offers additional evidence on whether audit information incorporated in audit fees are a signal of auditor’s independence in appearance or economic bonding with the client. In addition, prior studies on the quality of accounting information focus on equity market participants while empirical evidence on debt market participants is relatively scarce. In these respects, our research has implications for academics and practitioners. Especially, regulators could refer to our findings in making policy decisions on audit fees. By using announced credit ratings by professional rating agencies, we avoid measurement error or bias from estimating probability of bankruptcy from an empirical model. In addition, we compare the effect of abnormal audit fees on firm's credit ratings between public firms and private firms. In these respects, this study makes direct and indirect contributions to our understanding of the effects of audit quality. The sample period spans 7 years from 2002 (2001) to 2008 (2007) based on dependent (explanatory) variable. We exclude firms in the financial industry, firms without audit-related information, firms without credit ratings information, and firms without unaudited financial statements. Audit fee data is obtained from the KICPA database. Credit ratings data are extracted from the KIS-Value III database provided by KIS (Korea Investor Service). We obtain a final sample that consists of 10,507 firm-years for public firms and 51,337 firm-years for private firms. The empirical test results are as follows. In the sample of public firms, we find the relation between abnormal audit fees and credit ratings to be significantly negative. In contrast, the relation between abnormal audit fees and credit ratings for private firms is positive and significant. The findings imply that credit rating agencies recognize abnormally high audit fees paid by public firms as impairment to auditor independence and downgrade their credit ratings. The excess audit fees paid by large firms can be interpreted as evidence of economic bonding between auditors and client firms. On the other hand, for private firms that has smaller size and pay lower audit fees, credit rating agencies evaluate excess audit fees as a sign of higher audit quality and upgrade credit ratings. These results reveal that credit rating agencies differently recognize abnormal audit fees and incorporate their evaluations of credit ratings depending on firm's listing status because the amount of audit fees is very different between public firms and private firms. The results for public firms are consistent with prior studies on the relationship between abnormal audit fees and discretionary accruals (Choi et al. 2010; Hoitash et al. 2007; Coulton et al. 2007) although the proxies are different. They are also consistent with studies that examine the relation between abnormal audit fees and cost of capital (Hope et al. 2009; Kim and Park 2010). However, the findings are different for private firms. Interestingly, the results are different from Oh et al. (2008) which investigates the relationship between abnormal audit fees and cost of debt capital. These results are robust to controlling for various factors including auditor characteristics such as auditor size and auditor tenure, firm characteristics such as firm size, leverage, liquidity, and profitability. We also account for industry fixed effects and year fixed effects To reduce measurement error in estimating abnormal audit fees, we employ two different proxies – the residual from the audit fee estimation model and an indicator variable that equals 1 if the residual from the audit fee estimation model is non-negative and 0 if it is negative. The results are consistent using either definition of abnormal audit fees. However, for private firms, the results of this indicator variable show statistically weaker significance relative to those of continuous variable. In an additional test, we divide total sample into four sub-samples based on the level of abnormal audit fees and perform the sample analyses for each sub-sample. In this test, we find a significant and positive relation between credit rating and abnormal audit fees for only the highest abnormal fee sub-sample. These results show that the positive relation between abnormal fees and credit ratings could be due to the extremely high level of abnormal fee cases. Therefore, we are very cautious in generalizing these results. Our study provides an understanding of how credit rating agencies differently reflect abnormal audit fees into credit ratings depending on firm's listing status that reflects environmental difference in capital markets and audit service market. In addition to enhancing our understanding of the information used by credit rating agencies, this finding contributes to the emerging literature on the use of abnormal audit fees by various financial statement users (e.g., investors, creditors, analysts, and auditors). Among other implications, this study’s evidence may be useful to researchers or others interested in modeling the credit rating process. In these points, we expect this paper to have useful implications for scholars, practitioners, accounting standard setters, and regulators. Still, there remain some limitations to this study. Our measure of abnormal audit fees relies on the precision of the statistical model used in estimation. There is a possibility of potential omitted variables in the estimation. Another concern is potential omitted variables in the regression model examining the relationship between credit ratings and abnormal audit fees. However, we minimize problems of model specification by including control variables used in prior literature. In addition, all empirical studies are subject to these limitations.