Second, our sample period (2000–2003) includes the year of the 2001 Enron debacle, the subsequent Andersen collapse, and the passage of the SOX in 2002. In an attempt to control for the potential effects of these time-specific factors on our regression results, we include year dummies in Equation (4), as reported in Table 4. To further check whether our results are sensitive to these year-specific events, we also estimate both Equation (1) and Equation (4) for each sample year (without including year dummies). The results of annual regressions are qualitatively similar to those reported in Table 4 for all years except 2000, in which the coefficients of the variable of interests are insignificant.