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

The Effects of Relaxing the Reconciliation Requirement in Foreign Private Issuers’ SEC Filings on Earnings Management Strategies: IFRS Adopters versus U.S. GAAPAdopters 94 6. Empirical Results 6.1 Descriptive Statistics As stated earlier, we use a probit model (i.e., equation (5)) to analyze our sample firms’ decision to voluntarily switch to U.S. GAAP/IFRS from domestic standards. Panel A (B) of Table 2 provides descriptive statistics for U.S. GAAP (IFRS) and matched domestic samples regarding the explanatory variables in the analysis. The two rightmost columns report mean and median differences between the two sample groups using a two- sample t -test and a nonparametric Wilcoxon Mann-Whitney test, respectively. Panel A shows that the mean and median differences for ROA it , LEV it , and CS it are both significant at better than p = 10% based on the two tests, respectively. The results suggest that U.S. GAAP firms are more likely to experience poor financial performance, have higher leverage ratios, and have a lower propensity to raise capital than firms applying domestic GAAP. On the other hand, Panel B reveals that IFRS firms are more likely to exhibit better financial performance and higher leverage than firms complying with local GAAP. Panel C of Table 2 presents descriptive statistics related to the independent variables used in our probit model for our U.S. GAAP and IFRS test samples. The results show that the mean and median values of NMKT it ( ROA it and MV it ) for the U.S. GAAP sample are all significantly higher (lower) than for its counterpart, suggesting that U.S. GAAP adopters receive more information demands from foreign investors and are smaller and more likely to experience poor financial performance than IFRS adopters. LEV it , CS it , and LT_DEBT it , on the other hand, do not statistically differ between the two test samples. Panel D (E) of Table 2 provides descriptive statistics for U.S. GAAP (IFRS) and matched domestic samples regarding the explanatory variables used in the simultaneous equations (6) and (7) to explain the relation between the magnitudes of discretionary accruals and real earnings management. Panel D shows that the mean and median differences for the proxies for real activities manipulation (i.e., RM_SUM it , RM_PROD it , and RM_DISX it ), BBATH it , DEBT it , and ST_DEBT it are all significant at better than p = 10% based on t -tests and Wilcoxon tests, respectively. Specifically, the results suggest that U.S. GAAP adopters exhibit a higher magnitude of real activities manipulation, have a greater likelihood to take a “big bath,” have a higher debt-to-asset ratio, and have a smaller proportion of short-term debt than firms applying domestic GAAP. On the other hand, Panel E reveals that IFRS firms are more likely to exhibit a higher debt-to-asset ratio, have a smaller proportion of short-term debt, and experience better financial performance than