臺大管理論叢 NTU Management Review VOL.30 NO.2

103 NTU Management Review Vol. 30 No. 2 Aug. 2020 6.3 The Relation between Accrual-Based Earnings Management and Real Activities Manipulation Table 5 reports 2SLS regression results for the simultaneous equations (6) and (7) estimated on the samples of voluntary U.S. GAAP and IFRS adopters, respectively. The empirical results in Panels A, B, and C use RM_SUM it , RM_PROD it , and RM_DISX it , respectively, as the measures for real activities manipulation. We find that the coefficients on DA it , RM_SUM it , RM_PROD it , and RM_DISX it are all negative and significant at better than p = 10% for both the U.S. GAAP and IFRS samples in all the panels of Table 5, consistent with H 1 . As mentioned before, to test for the existence of endogeneity formally, we use the Hausman (1978) test for contemporaneous correlation between the error term and discretionary accruals/real activities manipulation, and we reject the null of no endogeneity at the p = 5% significance level or better across all of the models. Taken together, the results suggest that discretionary accruals and real activities manipulation are partial substitutes for earnings management, and that their magnitudes are determined simultaneously for foreign firms voluntarily switching to U.S. GAAP/IFRS from domestic standards after the SEC waived the reconciliation requirement. 19 The results presented in Table 5 offer further insights into whether executives’ incentives play a substantial role in determining the magnitudes of accrual-based earnings management and real activities manipulation. In accordance with H 2 , the coefficients on BBATH it and the proxies for “income smoothing” reporting (i.e., SM_SUM it , SM_PROD it , and SM_DISX it ) are all negative and significant at better than p = 5% for both U.S. GAAP and IFRS adopters in all the panels of Table 5, suggesting that our sample U.S. GAAP/ IFRS firms with a “big bath” or income smoothing incentive have a higher propensity to use discretionary accruals as well as real activities manipulation to reach their earnings targets. Moreover, except for DEBT it in equation (6) with RM_PROD it as the measure for real earnings management, the results document that the coefficients on DEBT it and ST_DEBT it are all significantly positive at better than p = 10% for the U.S. GAAP sample in all the panels. In general, this finding is consistent with the assertion that U.S. GAAP firms with 19 Note that if the extents of discretionary accruals and real activities manipulation are determined sequentially, then the negative relation will occur in equation (6) or equation (7), but not in both.