This preview shows page 1. Sign up to view the full content.
Unformatted text preview: firm has an earnings-based bonus plan and is in
the highest 10% of its industry for its leverage (total debt to total asset) and 0 otherwise. The
firms that had an IPO in 1996 are identified with the indicator variable IPO.
Monitoring mechanisms other than corporate governance practices may reduce the level
of earnings management. Past studies have shown that clients of Big 6 auditors report lower
levels of discretionary accruals than firms employing non-Big 6 auditors (Becker et al. 1998;
Francis et al. 1999). We use the indicator variable BIG6 which takes a value of 1 if the
auditor is a BIG 6 firm and 0 otherwise to control for this effect. The presence of outside
shareholders owning large blocks of the firm’s shares provides a monitoring mechanism that
may prevent earnings management. The control variable BLOCK is the cumulative
percentage of outstanding common shares held by all shareholders who own at least 5% of
the firm’s shares and who are not affiliated with management.
Finally, firm size may affect the board and audit committee characteristics as well as the
level of earnings management (Becker et al. 1998). We measure it as the log of the firm’s
total assets (LNSIZE).
We examine the relation between discretionary accruals and governance characteristics by
estimating the coefficient in the following logistic regression model: 19 EARNMAN = β 0 + β1 ACIND + β 2 ACNMAN + β 3 ACOPTION + β 4 FNEXPERT + β 5 MANDATE + β 6 MEETINGS +
β 7 ACIND * MEETINGS + β 8 BOARSIZE + β 9 BOARDIND + β10 CEOCHAIR + β11 NOMCOM + β12 NXOWN +
β13 NXTENURE + β14 NXDIRSHIP + β15 AGENCY + β16 IPO + β17 BIG6 + β18 BLOCK + β19 LNSIZE with three definitions of EARNMAN:
EARNMANP is an indicator variable with the value of 1 if the firm is in the HIGH_POS
category and 0 if it is in the LOW category;
EARNMANN is an indicator variable with the value of 1 if the firm is in the HIGH_NEG
category and 0 if it is in the LOW category;
EARNMANH is an indicator variable with the value of 1 if the firm is in the HIGH_POS
or the HIGH_NEG categories and 0 if it is in the LOW category.
All the other variables are defined in Table 1.
These three definitions of the dependent variables allow us to examine whether incomeincreasing and income-decreasing discretionary accruals have the same relationship with
corporate governance practices or whether they are affected differently. 5. Empirical results
Descriptive Statistics and Univariate Analysis
Table 2 provides descriptive statistics, by earnings management category, for the financial
and corporate governance characteristics of our sample. The last column compares the three
categories using Krustal-Wallis tests for the continuous variables and Chi-Square statistics
for the dichotomous variables.
The results of these tests indicate that the financial characteristics are significantly
different between the groups. Firms with the highest discretionary accruals are smaller in
terms of total assets than those with the lowest accruals. Because of the size difference, all
subsequent financial variables are scaled by total assets and the regression model controls for
the log of size. Firms with the highest positive discretionary accruals have a positive net
income and negative cash flows from operations, while firms with the highest negative 20 discretionary accruals have a negative net income and negative cash flows from operations,
and firms with the lowest discretionary accruals have both a positive net income and positive
cash flows from operations. Not surprisingly, the total and discretionary accruals are positive
and negative for the highest positive and the highest negative discretionary accruals firms,
respectively. By design, discretionary accruals are close to zero for the lowest discretionary
accruals firms and total accruals are slightly negative.
Comparing audit committee characteristics between the three categories, we find
significant differences for all of them. As expected the lowest discretionary accrual firms
have better audit committee characteristics than the other firms. Their committee is more
likely to be composed solely of independent non-executive directors (ACIND), they have
more independent non-executive directors who are not managers in other firms (ACNMAN),
they are more likely to have at least one member with financial expertise (FNEXPERT), they
are more likely have a clear mandate for the oversight of both the financial statements and
external auditing (MANDATE), and are more likely to meet more than twice a year
Comparing the board of director’s characteristics we find significant differences for all of
them except for the combination of the roles of chair and CEO (CEOCHAIR). As expected,
firms with the lowest discretionary accruals have a higher percentage of board members who
are independent non-executive directors (BOARDIND), are more likely to have a nominating
committee composed in majority of non-executive directors (NOMCOM), and have nonexecutive directors with more years of board service (NXTENURE) and with a larger number
View Full Document
This document was uploaded on 11/27/2013.
- Fall '13