사채 발행시점에서 회계보수성 변화와 사채발행 후 기업성과

We examine (1) whether accounting conditional conservatism temporally changes around the year of bond issues and (2) whether the post-issue firm performance is systematically related to the level of conservatism at the issue year. There are conflicting predictions about the temporal variation in the level of conservatism around bond issues. First, to the extent that conservatism reduces information asymmetry between debtholders and firms, managers may have incentives to increase the level of conservatism at the time of a bond issue in order to reduce information asymmetry between debtholders and firms, thereby decrease the cost of debt. On the other hand, previous studies suggest that the firm’s accounting performance during the period when the firm issues a bond is negatively associated with cost of debt. To the extent that conservative accounting depresses the firm’s accounting performance, it is plausible that managers may decrease the level of conservatism around bond issues in order to report more positive accounting performance. Given such opposing predictions, the pattern of the temporal changes around a bond issue perhaps is an empirical issue. Employing a sample consisted of the bond issues in 2000-201 and using the popular Basu(1997) model and the relatively recent Khan and Watts(2009) model that allows firm-year estimation of accounting conservatism, we find that conservatism becomes increasingly higher as the year of a bond issue approaches, culminating at the year of the issuance. It then starts decreasing, reaching the pre-issue level of year –3 in year +3. This pattern is consistent with the interpretation that firms that issue a bond purposefully increase conservatism, expecting that higher conservatism levels would benefit them when issuing bonds. This, in turn, is consistent with the argument that conservatism benefits firms by reducing information asymmetry and dampening managerial opportunism. The results are similar if the issuing firms are compared with the non-issue firms. That is, the level of conservatism of the issuing firms is higher than that of the non-issuing firms. Post-issue performance of the firms with higher conservatism on average is better than that with lower conservatism both in terms of operating and stock return performance. More specifically, the post-year operating performance measured by industry mean adjusted ROA is positively associated with the issue year conservatism level. Similarly, the matching portfolio adjusted stock return in the post-issue period up to three years is positively related to the degree of accounting conservatism at the bond issue year. Overall, our finding is consistent with the literature that suggests that conservatism benefits firms for instance by reducing information asymmetry between debtholders and managers. Our work has several contributions to the related literature. First, while the extant literature largely focuses on the cross-sectional differences with respect to the level of conservatism, our study provides evidence that conservatism changes temporally around the period of certain corporate events during which the benefits from and the costs associated with the degree of conservatism a firm adopts changes from other “normal” period. As such, our study adds insights into the extant literature. Second, while previous studies at least implicitly assume that conservatism is beneficial to firms, direct evidence to validate or refute this assumption is surprisingly scarce. We provide some direct evidence that the level of conservatism at the bond issue year is positively related to the post-issue operating and stock return performance. Given the absence of direct evidence on this issue, our work provides valuable additional support to the implicit assumption in the literature concerning the benefits of conservatism. Our work, however, is also subject to several caveats. First, conditional conservatism in particular at the firm-year level is notoriously difficult to measure. Although we employ the popular Basu(1997) method and the Khan and Watts(2009) method, it is plausible that our results might be subject to measurement errors whose extent is difficult to gauge. Second, the relationship between conservatism and future operating performance is somewhat difficult to interpret because conservatism by definition affects operating performance, suggesting that the relationship we document can be largely mechanistic. Hence, the finding that conservatism is positively associated with future operating performance should be interpreted with care even in the limited bond issue context.