5.2. Correlation matrix
The results for the Pearson correlation are reported in Table 6. The table shows a significant negative correlation between accounting expertise (ACC) and the level of discretionary accruals (|DA|), suggesting earnings management is lower for firms with audit committees when accounting expertise is present. Consistent with Dhaliwal et al. (2006), we also find that financial (FIN) and supervisory (SUP) expertise are not related to the level of earnings management. No significant correlation is reported between the adoption of IFRS (dummyPP) and the level of discretionary accruals (|DA|), which indicates the adoption of IFRS had an insignificant effect on accounting quality in Australia. In contrast, the dichotomous variable dummyPP is positively correlated with the number of audit committee members. Furthermore the frequency of meetings signifies the prominence of audit committees in post-IFRS adoption.
Table 7 shows the Pearson correlation results for the standard deviation of accrual residuals (sdRES) and the independent variables. Table 7 reports a statistically significant negative correlation with audit committees in the presence of accounting experts at the 1% level. The table also shows that the standard deviation of accrual residuals are significantly and negatively correlated (5% level) with audit committees that employ at least one finance expert. This indicates audit committees that employ finance experts report higher accrual quality.
Past research such as Fields et al. (2001) and Street and Bryant (2000) suggest that multicollinearity can affect the credibility of a regression analysis when it exceeds the critical level of 0.8. Table 6 and Table 7indicate that no correlation coefficient between independent variables is higher than 0.8. This suggests that multicollinearity does not constitute a problem in the models.3 To further examine the possibility of multicollinearity the variance inflation factor (VIF) is analysed and no problems were found for the independent variables.