This preview shows page 1. Sign up to view the full content.
Unformatted text preview: hips in unaffiliated firms (NXDIRSHIP). Firms with the lowest discretionary
accruals have larger boards, a result that may be explained by the fact that larger firms
generally have larger boards and the possibility for larger boards to have more non-executive
directors. Contrary to our expectation, firms with the lowest discretionary accruals do not
21 have non-executive directors with the highest monitoring incentive as measured by NXOWN.
The highest value is for the high negative discretionary accruals firms.
Out of the motivation control variables only IPO is significantly different across the
earnings management categories, the lowest discretionary accrual firms being less likely to
have an IPO in 1996. The other monitoring mechanisms are all significant, the lowest
discretionary accruals firms are more likely to hire Big 6 auditors (BIG6) and have a larger
percentage of outstanding common stock shares held by blockholders (BLOCK).
Table 3 presents the results of the logistic regression model used to test the relation between
discretionary accruals and governance characteristics. The models are highly significant (p =
0.0001), and the pseudo-R2 varies between 59 and 62 percent.
Audit committee characteristics
We test simultaneously for the effect of all the audit committee characteristics using a
Wald test, and we find that it is statistically significant (χ2=19.7 df=7, p=0.006). As Table 3
shows, contrary to our expectations, the presence of earnings management is not related to
whether the audit committee is composed solely of independent non-executive directors
(ACIND). As alternative measures of committee independence, we use first the percentage of
independent non-executive directors on the committee and then an indicator variable with the
value of 1 if the committee has a majority of independent non-executive directors and the results (not shown in a table) are similar to those in Table 3. Contrary to studies examining situations
of fraud, which found a significant effect for audit committee independence (Abbott et al.
2000), it appears that there is no meaningful direct relation between earnings management
and the proportion of independent non-executive directors or, as proposed by the BRC 22 (1999), only independent directors on the committee. We find, however, that having more
independent non-executive directors who are not managers in other firms (ACNMAN) is
negatively related to both the absolute and positive levels of discretionary accruals. This
result is consistent with the assertion that non-executive directors, if they are at the same time
managers in another company, may be less inclined to criticize the firm’s management and to
exercise their monitoring role (Westphal and Zajac 1997). It also provides archival evidence
supporting the experimental findings of DeZoort and Salterio (2001) that these directors are
more likely to support management in disputes opposing auditors and corporate management.
As expected we find that the proportion of options that can be exercised in the short run
in the independent directors’ holdings (ACOPTION) is positively related to positive
discretionary accruals. While not significant, the negative effect on negative earnings
management also supports the assertion that these options reduce independence. These results
support the Cadbury Committee’s (1992, 4.13) suggested good practice that non-executive
directors should not participate in share option schemes.
The presence of at least one audit committee member with financial expertise
(FNEXPERT), as recommended by the BRC, is negatively related to the level of earnings
management. The effect, however, is only statistically significant for negative discretionary
We find that an active committee is associated with less earnings management. Indeed,
the presence of a clear mandate establishing the committee's responsibility for the oversight
of both the financial statements and external auditing (MANDATE) is negatively related to
the level of both positive and negative discretionary accruals. Hence, as proposed by the
BRC, a formal charter does not only provide guidance to members as to their duties but is a
source of power for the audit committee. This result with archival data supports the findings
from a survey by Kalbers and Fogarty (1993) that a formal written charter plays an important
23 role in audit committee power and that perceived effectiveness of the committee is
significantly related to this power.
Contrary to our expectations, we find that meeting more than twice a year (MEETINGS)
is positively related to positive earnings management (+1.845), but when it is put in
interaction with independence (ACIND*MEETINGS) the combined effect (-2.733) is
significantly negative. This suggests that an audit committee composed solely of independent
non-executive directors who meet more than twice a year reduces positive earnings
management, but that a non-independent committee that meets more...
View Full Document
This document was uploaded on 11/27/2013.
- Fall '13