회계이익과 과세소득의 차이가 회사채 신용등급에 미치는 효과에 대한 재검토 및 새로운 증거 (A Review and Some New Evidence on the Effect of Book-Tax Differences on Bond Rating)

서지정보

· 발행기관 : 한국회계학회
· 수록지 정보 : 회계학연구 / 39권 / 3호 / 1 ~ 55페이지
· 저자명 : 박종일, 기은선, 권수영

초록

본 연구는 회계이익과 과세소득의 차이(BTD)가 회사채 신용등급(RATING)에 미치는 영향을 분석하였다. 본 연구는 이전 연구들의 주장과 같이 경영자의 기회주의적 이익조정으로 보고이익이 상향조정되면 회계이익이 과세소득보다 커질 수 있어 이익의 질은 낮아질 것으로 보고, 이런 경우 신용평가기관이 신용등급을 더 낮출 것으로 기대했다. 실증분석기간은 2001년부터 2010년까지 총 10년간이고, 총 1,929개 기업/연 자료가 분석에 이용되었다.
실증결과는 다음과 같다. 첫째, 소수의 순위등급변수(fractional ranks variable)로 측정한 BTD 변수가 기존 연속변수의 경우보다 BTD와 RATING간에 음(-)의 관련성을 보다 잘 나타내 주는 것으로 나타났다. 둘째, BTD를 더미변수로 측정한 경우에도 두 변수 간에 음(-)의 관련성이 나타났다. 이는 BTD 값이 음(-)보다 양(+)인 경우를 신용평가기관은 정보위험이 더 높다고 평가함을 의미한다. 셋째, 두 변수간 음(-)의 관계는 시장별로 차이를 보였는데, 특히 KOSDAQ보다 KOSPI시장에서 더 뚜렷한 결과로 나타났다. 또한 KOSPI 표본에서 BTD는 재량적 발생액(DA) 이상의 정보가치를 제공하지만, KOSDAQ 표본에서는 이와 반대의 결과를 보였다. 넷째, 이상의 결과들은 주로 세전이익이 음(-)인 경우보다 양(+)인 표본의 결과였다. 다섯째, 이상의 결과들의 경우 이익의 질(DA)에 따라 표본을 나누어 분석하면 두 변수 간 유의한 음(-)의 관계는 이익의 질이 낮은 양(+)의 DA 표본에서, 또한 낮은 품질의 감사인(non-Big 4) 표본에서 더 뚜렷한 결과를 보였다. 이는 이익의 질이나 감사품질이 낮은 경우 BTD 정보는 신용평가기관의 신용등급에 주로 반영된다는 발견이다.
이상의 연구결과를 종합하면 이전 연구들에서 나타난 BTD와 RATING 간의 평균적인 유의한 음(-)의 결과는 BTD 측정에서 변수정의와 표본구성의 방법에 따라, 그리고 이익의 질, 감사품질 및 시장특성에 따라서도 연구결과에 차이가 있음을 본 연구는 보여주었다는 점에 의미가 있다. 이상과 같은 본 연구의 실증적 증거는 BTD 관련연구에 새로운 증거로서 추가적인 공헌을 할 것으로 기대된다.

영어초록

Prior research suggests that differences between book and taxable incomes reflect firms’ underlying economic fundamentals and their opportunistic earnings management activities. A number of recent studies investigate whether differences between book income and taxable income are a source of information useful to financial report users, and several key findings have emerged, including: book-tax differences (BTDs) are useful in detecting certain types of earnings and tax management (e.g., Phillips at el. 2003); large magnitude positive and negative BTDs reflect differences in earnings persistence (Hanlon 2005); the ratio of taxable income to book income is useful in predicting earnings growth and explaining security returns (Lev and Nissim 2004); and taxable income provides information about a firm’s performance and its value incremental to book income, and vice versa (Hanlon et al. 2008). Overall, previous research suggests that BTDs arises (i) from differences in book and tax reporting requirements that reflect aspects of firms’ underlying fundamentals not reflected in either income measure by itself, and (ii) from firms’ earnings and tax management activities. That is, firm value depends on both financial reporting and tax reporting decisions, and BTDs combine the effects of these decisions. Thus, we note that with regard to the quality of accounting information, BTDs can reflect firms’ earnings management as well as their tax management strategies. Therefore, we extend this literature by investigating the relation between BTDs and credit ratings to gain a deeper understanding of the information reflected in the differences between these two earnings measures and its impact on rating agencies. Meanwhile, credit ratings are determined by rating agencies’ assessments of the probability distribution of future cash flows to bondholders, which in turn, depends on the future cash flows to the firm. A firm’s creditworthiness is determined by assessing the likelihood that its future cash flows will be sufficient to cover debt service costs and principal payments. A credit rating implications include signaling, maintaining relationships with third parties such as suppliers and customers, and maintaining firms’ credit ratings in line with competitors. A credit rating represents a signal of overall quality in the capital market. So, rating agencies may face fewer incentives to detect earnings management by issuers.
Hence, this study investigates the effects of firms’ BTDs on bond’s credit ratings using firms listed on KOSPI and KOSDAQ. Especially, we are trying to differentiate this study from previous studies by focusing on two aspects. One is to examine how different measures of BTDs (measured in various way) affect the relation between BTDs and credit ratings. The other is to examine whether this relation between two variables varies with the methods for sampling partition, which may represent a conditional situation. As these issues have not been well-studied, this paper is expected to give additional information to the emerging literatures on the BTDs.
As in the previous studies, we anticipate that firms would report higher pre-tax book income than taxable income when managers manipulate earnings upward. In this case, credit analysts may interpret positive BTDs (i.e. where book income is more than taxable income) as a signal of decreased earnings quality, which results in less favorable credit ratings. To test this hypothesis, we use credit ratings as a dependent variable and measure of BTDs, a variable of interest, following Jeon and Kim (2008). We use 1,929 firm-year observations for the period of 2001-2010.
The empirical results are as follows. First, in the tests using clustering-adjusted ordered probit regression to consider ranked order attribute of credit ratings, the negative relation between BTDs and credit ratings exhibits more evidently when we measure BTDs based on the fractional ranking of the values of variables rather than continuous values. Second, this negative relation does not change where measuring BTDs as a indicator variable. This result suggests that credit analysts consider firms with positive BTDs have higher credit risk than firms with negative BTDs. Third, these findings show a difference across the listed market. Specifically, the negative relation between BTDs and credit ratings is more pronounced for KOSPI firms than for KOSDAQ firms. Also, for KOSPI firms it is revealed that BTDs provide incremental information content beyond discretionary accruals, while we get the opposite result from KOSDAQ firms. Fourth, the above results are mostly observed in firms with positive taxable income, not in firms with negative taxable income. Fifth, when we partition full sample into sub-samples based on earnings quality or audit quality, the negative relation between two variables are more noticeable for firms with positive discretionary accruals or for firms audited by non-BIG4 auditors. These results mean that credit analysts negatively much more reflect BTDs into the bond’s credit ratings for firms with low earnings quality or low audit quality. Finally, in the additional tests, we divide the samples into deciles and examine in which decile the negative relation between two variables. We find that the negative relation exhibits just one decile of many, but also only firms from the lowest negative BTDs, irrespective of listed market.
The findings of this study have contributions in that they suggest the negative relation between BTDs and credit ratings may be affected by measures of BTDs, sample partition, earnings quality, audit quality and market characteristics, while previous studies focus on investigating the relation between two variables on average. Above empirical results of this study are expected to make additional contribution to studies on the BTDs as a new evidence. Therefore, understanding the credit ratings consequences of BTDs is important and interesting in its own merit. Our findings will thus be of interest to academicians as well as managers, credit rating agencies, bondholders, and regulators alike.

참고자료

· 없음