Table 2 provides the regression results, with LN(IABUDGET) as the dependent variable. The overall
model is significant (F = 10.24, p < 0.001; adjusted R-square of 0.47). The coefficients on all control
variables, except SRVIND and RESTATE, are significant at conventional levels with the expected signs.7
For the audit committee control variables, the coefficient on ACSIZE and ACIND are not significant.8 Consistent
with CHR (2005), the coefficient on ACBUDGET is positive and significant.
For our test variables, the coefficient on ACCEXP is not significant but the coefficient on AUDEXP is
negative and significant, suggesting that the presence of an auditing expert on the audit committee is
associated with a lower internal audit budget. This observation is consistent with the view that the
presence of an auditing expert increases the ‘‘comfort zone” and expertise, and thus the effectiveness
of the overall audit committee. The auditing expert has the monitoring effect of substituting the need
for increased assurance required from the internal audit. While the coefficient on DIRBUSY is not