The Influence of Independent and Effective Audit Committees On Earnings Quality

THE INFLUENCE OF INDEPENDENT AND EFFECTIVE AUDIT
COMMITTEES ON EARNINGS QUALITY
Daniel Bryan
M.H. Carol Liu
Samuel L. Tiras*
Department of Accounting and Law, School of Management

State University of New York at Buffalo
January 6, 2004
*Corresponding author: Sam Tiras, stiras@buffalo.edu

(716) 645-3215
We wish to acknowledge the helpful comments of the seminar participants at University of

Alabama, Nanyang Technical University-Singapore, State University of New York at Buffalo,
and the 2003 American Accounting Association annual meeting. Additionally, we wish to thank
Larry Abbott, Kareen Brown, Todd DeZoort, Aretha Hall, Susan Hamlen, Troy Janes, Emad
Mohd, Susan Parker, Joe Piotroski and Hun-Tong Tan for their helpful comments and assistance.
THE INFLUENCE OF INDEPENDENT AND EFFECTIVE AUDIT
COMMITTEES ON EARNINGS QUALITY
Abstract
The Blue Ribbon Committee (BRC) on Improving the Effectiveness of Corporate Audit
Committees (1999) argued that audit committees would enhance the financial reporting process,
when comprised of members that are independent, financially literate, commit sufficient time to
the committee and meet regularly. We investigate whether the BRC recommendations would
likely increase the quality of reported earnings by testing whether earnings informativeness and
transparency are greater for those firms that currently employ audit committees meeting the BRC
recommendations over those that do not. Our proxy for informativeness is the magnitude of the
earnings response coefficient (ERC) and for transparency is the degree of mispricing
(overpricing) found with accruals. We find stronger ERCs when the audit committee members
are independent and financially literate, and less overpricing of accruals when the audit
committee members are independent and meet regularly. These results are not sensitive to other
factors that could influence the audit committee’s oversight (CEO influence, CEO shareholdings
and large outside blockholders) nor to other variables the literature has indicated could influence
ERCs or market mispricing. Overall, these results suggest that independent and effective audit
committees enhance the quality of reported earnings, lending evidence supporting the BRC
recommendations and the Sarbanes-Oxley Act of 2002.
Key Words: Audit Committee, Corporate Governance, Earnings Quality, Earnings Response
Coefficients, Market Mispricing.
Data Availability: The data used in this study are available from the sources indicated in the
text. A list of sample firms is available from the authors.
1. Introduction
This paper examines the relationship between audit committee characteristics and the
quality of corporate earnings. In Release No. 34-42266, the Securities and Exchange
Commission (SEC) indicates that audit committees play a critical role in the financial reporting
system by overseeing and monitoring the participation of management and the independent
auditor in the financial reporting process (SEC 1999). Although the entire board of directors
bears the ultimate responsibility for monitoring the financial reporting process, the board
typically delegates this responsibility to the audit committee. In the SEC’s view, the audit
committee must be ‘first among equals’ in the financial reporting process, since the audit
committee is an extension of the full board and hence the ultimate monitor of the process (SEC
1999).
The responsibilities of audit committees include appointing the external auditors, meeting
with the auditors to evaluate the corporation’s financial statements, interacting with the internal
financial managers and internal auditors, and reviewing the firm’s internal controls. Recent
accounting scandals and corporate bankruptcies, however, have raised questions about the
independence and effectiveness of audit committees.
In 1998, then SEC chairman Arthur Levitt expressed his “[f]ear that we are witnessing an
erosion in the quality of earnings, and therefore, the quality of financial reporting (Levitt 1998).”
In his plan for enhancing financial reporting quality, he advocated strengthening a firm’s
corporate governance by improving the independence and effectiveness of audit committees.
Levitt expressed his position as follows:
“[Q]ualified, committed, independent and tough-minded audit committees represent the

most reliable guardians of the public interest. Sadly, stories abound of audit committees
whose members lack expertise in the basic principles of financial reporting as well as the
mandate to ask probing questions. In fact, I’ve heard of the audit committee that
convenes only twice a year before the regular board meeting for 15 minutes and whose
duties are limited to a perfunctory presentation. Compare that situation with the audit
committee which meets twelve times a year before each board meeting; where every
member has a financial background; where there are no personal ties to the chairman or
the company; where they ask tough questions of management and outside auditors; and
where, ultimately, the investor interest is being served (Levitt 1998).”
Levitt called for a special committee to address his concerns about the deteriorating
quality of financial reporting. The resulting Blue Ribbon Committee (BRC) on Improving the
Effectiveness of Corporate Audit Committees forwarded ten recommendations on improving
oversight of the financial reporting process. Their recommendations called for audit committees
to be more effective in overseeing the financial reporting process. The BRC report raised the
awareness of certain attributes expected of audit committee members as particularly important,
specifically, recognition of the significance of the audit committee’s responsibility, time
commitment, financial literacy, and above all, independence (BRC 1999). Among other
recommendations, the BRC recommended that the audit committee be comprised of at least three
directors, each of whom is independent and financially literate, or becomes financially literate
within a reasonable period of time after his appointment to the audit committee.
Since earnings quality is not well defined, we investigate whether mandating the BRC
recommendations would likely increase the quality of reported earnings by performing two types
of earnings quality tests that address the concerns expressed by the SEC and Congress. The first
set of tests follows the common theme of the SEC’s call for higher quality earnings. This theme
is exemplified by Lynn Turner’s (former SEC Chief Accountant) 1999 speech on improving the
quality of financial reporting, in which he suggests the impact of higher quality earnings would
be more informed (better) decisions by users.
“Through the continued team efforts of financial management, auditors, and audit
committees, high quality financial statements will continue to provide the necessary
information required for investors to make informed decisions (Turner 1999).”
2
If independent and effective audit committees improve the quality of reported earnings by
providing better information to users, as Turner claims, then we would expect the
informativeness of reported earnings for those firms that currently employ independent and
effective audit committees (those that meet the BRC recommendations) to be greater than for
those that do not.1 Similar studies that examine the influence of auditors, boards and other
factors on earnings informativeness conclude these factors improve informativeness when they
are associated with a higher earnings response coefficient (ERC). For instance, Teoh and Wong
(1993) evaluate whether the ERC for firms employing BIG 8 auditors is higher than for firms
employing NONBIG 8 auditors, concluding auditor size does influence earnings quality.
By measuring earnings informativeness as the ERC, we find that audit committees
comprised of independent and financially literate directors improve earnings informativeness.
We find no evidence, however, that audit committees that meet more frequently or comprised of
directors free to commit more time to the committee improve earnings informativeness. These
results are not sensitive to other factors that could influence the audit committee’s oversight,
such as CEO influence, CEO shareholdings and large outside blockholders serving on the audit
committee, nor to other variables the literature has indicated could influence ERCs.
More recently, Congress, through the Sarbanes-Oxley Act of 2002, expressed concerns
about the transparency of reported earnings. Lynn Turner defines transparency in financial
reporting as, “...the extent to which financial information about a company is visible and
understandable to investors, other market participants, and regulators (1999).” Turner further
1
By defining independence and effectiveness in terms of the BRC’s definition, our tests are essentially tests of
the characteristics that the BRC identifies as being independent and effective. Since our results will show that
not all of the characteristics are associated with higher quality earnings, one possible conclusion of our study is
that the BRC may have mis-specified or over-specified those characteristics that may be important to
independent and effective audit committees.
3
claims that, “[t]ransparent financial information allows market participants to better evaluate
counterparty risks and adjust the availability and pricing of funds in ways that can promote
efficient financial markets (2001).”
Together, Turner’s comments suggest that a lack of transparency would likely lead to the
types of market mispricing documented by Sloan (1996), Xie (2001) and others, where the
market systematically misprices securities by overvaluing the accruals component of reported
earnings. If independent and effective audit committees enhance earnings quality by improving
the transparency of reported earnings, as Congress and others claim, then we would expect to
observe a reduction in the mispricing documented in these studies. Our second set of tests of
earnings quality, therefore, investigates whether independent and effective audit committees
reduce this mispricing.
In our tests of mispricing, our results are consistent with the prior literature in that we
find accruals are systematically overvalued. The degree we find that accruals are overpriced is
significantly lower when audit committees are independent and meet regularly. We find no
evidence of such a reduction, however, when audit committee members are financially literate
and able to dedicate more time to the committee by serving on no more than three other boards.
As with our other set of tests, our results are not sensitive to other factors that could influence the
audit committee’s oversight nor to other variables the literature has indicated could influence the
mispricing. Together, our findings that independent and effective audit committees enhance
earnings informativeness and transparency suggest that the BRC recommendations and
Congress’s actions in passing the Sarbanes-Oxley Act of 2002 are likely to meet their goal of
improving the quality of reported earnings.
4
We organize the remainder of this paper as follows. Section two reviews the prior
literature on the association of certain board and audit committee characteristics with financial
reporting quality. We formulate our models and formal hypotheses in section three. Section
four discusses the sample selection and describes our empirical measures. We present our results
in section five, and section six concludes our study.
2. Related Literature
2.1 THE RELATION BETWEEN THE EARNINGS-RETURNS RELATION AND BOARDS
The prior literature provides some evidence concerning the influence of the board and the
audit committee on earnings quality. In testing the influence of audit committees on earnings
informativeness, Wild (1996) finds that the market reaction to reported earnings issued
subsequent to the formation of the audit committee is higher than in prior periods. Menon and
Williams (1994) point out, however, that the mere formation of an audit committee does not
mean that the board of directors explicitly relies on the audit committee to enhance its
monitoring ability. Menon and Williams (1994) consider two characteristics of the audit
committee, independence and meeting frequency, to determine if the board explicitly relies on
the audit committee as a mechanism to control management. Our study, therefore, follows
Menon and Williams’ (1994) argument in the sense that we test whether specific characteristics
of the audit committee (those recommended by the BRC) influence earnings quality.
Vafeas (2000) investigates whether the earnings-returns relation varies with board
independence and board size. He finds that the earnings-returns relation is stronger (albeit
weakly) for firms with smaller boards, but he finds no significant association between the
5
earnings-returns relation and board independence. Since board independence is most important
when the board serves in a monitoring capacity, the lack of results in Vafeas (2000) may be due
to the relegation of the monitoring function to the audit committee, as suggested by the focus on
audit committees by the BRC and the Sarbanes-Oxley Act of 2002.
In summary, both Wild (1996) and Vafeas (2000) provide evidence suggesting a relation
between the board and the audit committee and earnings quality, measured through the earnings-
returns relation. The nature of the relation, however, is not clear from their studies. Our study
attempts to better characterize the nature of this relation.
2.2 THE INDEPENDENCE OF THE CORPORATE BOARD AND AUDIT COMMITTEES
The BRC regards audit committee independence as paramount, above all other
characteristics. Several studies provide evidence of the relation between board or audit
committee independence and financial reporting quality. Beasley (1996) presents evidence that
the likelihood of fraud is higher when the percentage of outside directors is low. He does not
find, however, that the presence of an audit committee reduces this likelihood.2 In contrast,
Dechow et al. (1996) find that in addition to board independence, the presence of an audit
committee reduces the likelihood of a firm being subject to enforcement actions by the SEC.
Abbott et al. (2000) confirm Dechow et al.’s (1996) result by finding that firms with independent
audit committees are less likely to be sanctioned by the SEC. Further, Abbott et al. (2002) find
that audit committee independence is inversely related to earnings restatements. Finally, Klein
(2002a) documents a negative relation between board and audit committee independence and
earnings management. Together, these studies document the influence of audit committee
2
Again, Menon and Williams (1994) point out that the formation of an audit committee does not mean that the
board relies on the committee to oversee the financial reporting process.
6
independence on financial reporting quality, but only for the extreme cases of financial reporting
fraud and earnings restatements.
2.3 THE EFFECTIVENESS OF THE CORPORATE BOARD AND AUDIT COMMITTEES
The BRC specifies other important characteristics, namely, meeting activity, financial
literacy and the time commitment of committee members as key determinants of an effective
audit committee. These characteristics, according to Levitt (1998) and the BRC, likely influence
the financial reporting process. Abbott et al. (2000) provide evidence consistent with this claim
by finding that firms whose audit committees meet less than the minimum threshold are more
likely to restate earnings. DeZoort et al. (2002) interpret meeting frequency as a measure of the
audit committee’s due diligence. Further, Abbott et al. (2002) document greater incidences of
earnings restatements and fraud when the audit committee members lack financial literacy.
Together, these studies document the influence of audit committee effectiveness for extreme
cases of financial reporting. Finally, Core et al. (1999) argue that audit committee effectiveness
deteriorates when its members serve on ‘too many’ boards. They point out that experience from
serving on another board is expected to initially enhance the effectiveness of the audit committee
members, but that benefit quickly reverses when directors serve on too many (more than three)
boards.
2.4 OTHER FACTORS POTENTIALLY INFLUENCING AUDIT COMMITTEES
Other factors identified in the literature that may influence the audit committee’s
oversight of the financial reporting process include: the CEO’s influence over the board, the
CEO shareholdings in the firm and the existence of large outside blockholders of corporate stock
serving on the audit committee.
7
CEO Influence
The ability of audit committees to enhance the oversight of the financial reporting process
is likely impacted by the degree of influence exerted by the CEO. Shivdasani and Yermack
(1999) document that when the CEO serves on the nominating committee or when no
nominating committee exists, firms appoint fewer independent outside directors to the board.
Further, Dechow et al. (1996) find a greater incidence of earnings management that is subject to
SEC enforcement action when the CEO also serves as the Chairman of the Board, in particular
for those firms lacking an audit committee. These studies provide evidence that the greater the
influence of the CEO, the less likely that the audit committee would enhance the oversight of the
financial reporting process.
CEO Stock Ownership
Warfield et al. (1995) find that higher CEO stock ownership helps alleviate some of the
agency problems that arise in corporations. To the extent that CEO stock ownership aligns the
CEO’s incentives and serves as a substitute for monitoring, the influence or demand for
independent and effective audit committees would likely be diminished. In contrast, Klein
(2002a) finds a positive correlation between CEO shareholdings and earnings management,
consistent with recent events and accounting scandals. Together, these studies suggest that CEO
stock ownership is likely to influence (positively or negatively) the financial reporting process
and the role of the audit committee in overseeing this process.
Outside Blockholders
Large outside blockholders of corporate stock (owning at least five-percent of the firm’s
common stock) typically exert a degree of influence and control on the firm. This influence
would likely supplement the monitoring activities relegated to the audit committee, thus
8
influencing the relation between the audit committee characteristics and the earnings quality.
Klein (2002b) provides evidence of this trade-off between outside blockholders and audit
committee independence, when large outside blockholders serve on the committee. Further,
Klein (2002a) finds a reduction in earnings management in the presence of large outside
blockholders on the committee. Together, these studies suggest audit committee oversight may
be influenced by the presence of large outside blockholders on the committee.
In sum, prior research provides evidence that audit committee independence and
influence exerted by the CEO are correlated with the amount of restatements and fraud in the
financial reports. Prior research also provides evidence that audit committee activity, financial
literacy and time commitment are correlated with accounting restatements and fraud. These
results suggest that earnings quality is higher when firms have independent and effective audit
committees. This evidence, however, is necessarily indirect and addresses only the extreme
elements of financial reporting. Our study contributes to the literature by providing additional
evidence that focuses directly on the influence of independent and effective audit committees on
earnings informativeness and transparency, documenting that the ERC is stronger and mispricing
less severe for those firms with independent and effective audit committees.
3. Model and Hypothesis Development
To test whether strong and effective boards improve earnings quality by enhancing
earnings informativeness and transparency we develop two separate empirical models that test
whether ERCs are higher and mispricing lower, respectively.
9
3.1 EARNINGS INFORMATIVENESS TESTS
We estimate the ERC (our informativeness metric) by regressing cumulative abnormal
returns on unexpected earnings, as follows:3
CAR t = α 0 + α1 UE t + ε t (1)
where:
CAR t = holding period returns cumulated from the day after the prior year’s earnings
announcement date for year t-1 through the day current earnings
announcement date for year t, less market holding period returns for the
same period;
UE t = unexpected earnings reported for year t, proxied as the actual reported
earnings less the consensus analyst forecasts of annual earnings measured as
the first forecast after the prior year’s earnings announcement date;
αi = regression coefficients, i ∈ [0, 1];
εt = error term.
Our test window is one year, since the focus of both the BRC and Sarbanes-Oxley Act of
2002 includes the audit committee’s responsibility in overseeing annual earnings through the
appointment of the auditor and the continued interaction with the auditor during the year-end
audit. A shorter window surrounding the annual earnings announcement alone would capture
only the fourth quarter ERC, rather than the ERC for annual earnings (Manry et al. 2003).
Further, the long windows approach reduces the noise in ERCs and mitigates the concern that the
returns-earnings window could be mis-aligned (Ali and Zarowin 1992; Collins and Kothari
1989; Kothari 2001).4
3
For clarity, we suppress firm specific subscripts for all models presented. We deflate our regressions by price,
set at the beginning of the returns window, to control for heteroskedasticity and size (see Brown et al. 1999).
4
As an alternative specification, we also perform our tests by regressing returns on annual earnings and changes
on annual earnings, where the sum of the coefficients on levels and changes serve as the measure of unexpected
10
To test our expectation that independent and effective audit committees increase the
ERC, we create a dummy (0, 1) indicator variable for audit committee independence (INDt) and
three indicator variables for audit committee effectiveness (EFFi,t, where i ∈ [0, 2]) representing
the BRC’s recommended levels for financial literacy (FINLITt), meeting frequency (MEETt) and
time commitment (NOBUSYt).5 We set each of these variables to one when the audit committee
meets the BRC recommendation for independence, financial literacy, frequent meetings
(minimum four meetings) and time commitment (members serve on no more than three boards),
respectively, otherwise zero.
We interact the indicator variables with our measure of unexpected earnings (UEt) to test
our predictions. We also include the indicator variables individually (intercept effect) in the
regressions to control for other possible cross-sectional differences across our sample firms.
Additionally, we include in our model three indicator variables representing other factors
(OTHERj,t, where j ∈ [0, 2]) that the prior literature suggests could influence the oversight of the
financial reporting process. These variables are set to one when the CEO does not serve both as
the Chairman of the Board and as a member on the nominating committee (NOINFt) or hold
above average quantities of corporate stock (CEOSHRt), and when a member of the audit
committee is a large outside blockholder (BLOCKt), otherwise zero.6
In addition to the factors we include to control for other influences into the oversight of
the financial reporting process, we include additional control variables (CNTLk,t, where k ∈ [0,
earnings (Ali and Zarowin 1992; Ohlson and Shroff 1992). The results and conclusions from this specification
did not vary in any material way from using analyst forecasts to approximate unexpected earnings.
5
Consistent with the intent of the BRC and Congress, we classify ‘gray’ directors as not independent.
6
The results from Dechow et al. (1996) suggest that CEOs that founded the firm would also exert significant
influence. For our sample of Fortune 500 firms, only a small proportion of the CEOs were the firm’s founder,
so any inference regarding founder from our study may be misleading. However, when we redefine NOINFt to
include founders that also serves as the Chairman of the Board or on the nominating committee as also being
influential, the proportion of influential CEOs in our sample does not significantly change nor do our results.
11
4]) that the prior literature has identified as potentially influencing the relation between
unexpected earnings and cumulative abnormal returns. These factors identified by Collins and
Kothari (1989) and Teoh and Wong (1993), among others, include market beta (BETAt), analyst
expectation of future growth (GROWTHt), analyst precision (PRECISIONt), earnings persistence
(PERSt), and negative earnings (LOSSt).7 Finally, we also include an additional variable
( UE t UE t ) that allows for a nonlinear relation between unexpected earnings and cumulative
abnormal returns, as noted in Freeman and Tse (1992), by interacting unexpected earnings with
the absolute value of unexpected earnings. Our informativeness test model is thus as follows:
CAR t = α 0 + α1 UE t + α 2 IND t + α 3 UE t * IND t
(α OTHER j, t + α 11+ 2 j UE t * OTHER j, t )

2 2
+ (α 4 + 2 i EFFi , t + α 5 + 2 i UE t * EFFi , t ) + 10 + 2 j (2)
i =0 j= 0
4
+ (α 16 + 2 k CNTL k , t + α 17 + 2 k UE t * CNTL k , t ) + α 26 UE t UE t + ν t
k =0
where:
αi = regression coefficients, i ∈ [0, 26];

νt = error term.
All other variables are defined as above. A feature of our informativeness test model is
that we construct indicator variables rather than continuous variables to measure the
independence and effectiveness of audit committees. The merit of this approach is that it allows
direct comparisons across each governance characteristic.
7
While Teoh and Wong (1993) also identify auditor size as a variable associated with ERCs, the choice of
auditors is attributable to the audit committee, thus controlling for the auditor selection would in fact be
controlling for a primary function of the audit committee. Further, from our sample of Fortune 500 firms, little
variation exists across our sample firms for auditor size.
12
If independent and effective audit committees enhance earnings informativeness, such
that the ERC is greater for those firms with these characteristics, we will find positive
coefficients in Equation (2) for UEt, interacted with INDt or EFFi,t. We formally state this in our
first two hypotheses (in alternative form) below:
H1a: The association of unexpected earnings with cumulative abnormal returns

is greater with independent audit committees (α3 > 0).

is greater with effective audit committees (α5 > 0; α7 > 0; and, α9 > 0).
We separately test the effects of independence and effectiveness in H1a and H2a,
respectively, to identify the influence of each characteristic on earnings informativeness. We
also test H1a and H2a jointly and test whether the effects of independence and effectiveness are
incremental to each other. The BRC, however, positions independence above all other audit
committee characteristics. This priority placed on independence would be confirmed if α3 is
greater than the coefficients on the effectiveness characteristics (α5, α7 or α9) in our joint test of
Equation (2). We formally stated this in our third hypothesis (in alternative form) below:

is greater with independent committees than with effective committees (α3
> α5, α7 or α9).
3.2 EARNINGS TRANSPARENCY TESTS
Before we can develop our tests of mispricing (our proxy for transparency), we must first
estimate a firm’s intrinsic value. We employ Aboody et al.’s (2002) measure of intrinsic value
13
that estimates intrinsic value as the future cum-dividend price, discounted at one plus the risk-
induced, conditional expected return (CRSP size-based decile return), as follows:8
Pt +1 + DIVt +1
Vt = (3)
1 + rt*+1
where:
Vt = intrinsic value per share;
Pt +1 = market price per share as of the day after the year t+1 earnings
announcement date;
DIVt +1 = dividends per share for year t+1;
rt*+1 = CRSP sized-based decile return for year t+1.
Aboody et al. (2002) then regress Vt on net book value and current earnings to test the
relation between these accounting measures and intrinsic values. To test our hypotheses that
independent and effective audit committees reduce the overpricing of accounting accruals, we
modify Aboody et al.’s base model by including operating cash flows as an additional regressor.
We also omit net book value from our tests (as does Sloan 1996, among others) to focus on
earnings quality through the mispricing of accruals.9 The resulting model is as follows:
Vt = β 0 + β1 OCF t + β 2 NI t + µ t (4)
where:
8
Aboody et al.’s (2002) measure of intrinsic value is not restricted to one-year-ahead cum-dividend market
prices, rather the measure can be extended to multiple years. We chose one-year-ahead market prices as our
basis for intrinsic value to be consistent with other mispricing studies (e.g., Sloan 1996, among others).
9
When we include book value in our regressions, our results are consistent with the results in Aboody et al.
(2002) in that the market significantly undervalues book value. We find that independent and effective audit
committees do not consistently reduce the underpricing of book value across our mispricing tests. The
inclusion of book value in the regression, however, does not alter our findings with respect to the mispricing
accounting accruals.
14
OCFt = operating cash flows per share for year t, calculated from the Statement of
Cash Flows;
NI t = earnings per share before extraordinary items reported for year t;
βi = regression coefficients, i ∈ [0, 2];
µt = error term.
All other variables are defined as above. By regressing reported earnings and operating cash
flows jointly in equation (4), β1 represents the incremental value relevance (intrinsic pricing) of
cash flows over accruals and β2 represents the value relevance (intrinsic pricing) of accruals
alone (Jennings 1990).10
In contrast, if we regress market price (Pt) in Equation (4) as the dependent variable,
rather than intrinsic value, the regression coefficients would represent the actual pricing by the
market. The differences in these coefficients, therefore, represent the market mispricing. To test
how the coefficients differ between the two dependent variables, Vt and Pt, we set the dependent
variable as Vt − Pt, as follows:11
Vt − Pt = γ 0 + γ 1 OCF t + γ 2 NI t + η t (5)
where:
γi = regression coefficients, i ∈ [0, 2];
ηt = error term.
10
Since we focus on the transparency of earnings to the securities markets, rather than earnings management, we
focus on total accruals (as in Sloan 1996) rather than abnormal accruals (as in Subramanyam 1996 or Xie 2001).
11
As above, we deflate our regressions by price, set as of the day after the year t-1 earnings announcement date, to
control for heteroskedasticity and size.
15
All other variables are defined as above. In equation (5), γ1 therefore represents the mispricing
of operating cash flows relative to accruals, and γ2 represents the mispricing of accruals alone
(our test metric).
To test our expectation that independent and effective audit committees reduce the
overpricing of accruals, we interact NIt with our indicator variables for independence (INDt) and
effectiveness (EFFi,t, where i = FINLITt, MEETt and NOBUSYt).12 We again include the
indicator variables individually (intercept effect) in the regressions to control for other possible
cross-sectional differences across our sample firms. We also include in our model the three
indicator variables representing other factors (OTHERj,t, where j = NOINFt, CEOSHRt and
BLOCKt) that could influence the oversight of the financial reporting process.
We also incorporate into our regressions control variables that Ali et al. (2003) and
Fairfield et al. (2003) find account, in part, for the mispricing of accruals. These additional
control variables are the same as three of the additional control variables (CNTLk,t) included in
the prior set of regressions: market beta (BETAt), analyst expectation of future growth
(GROWTHt) and analyst precision (PRECISIONt). Our test model is thus as follows:
Vt − Pt = γ 0 + γ1 OCFt + γ 2 NI t + γ3 IND t + γ 4 NI t * IND t
(γ OTHER j, t + γ 12+ 2 j NI t * OTHER j, t )

2 2
+ (γ 5 + 2 i EFFi , t + γ 6 + 2 i NI t * EFFi , t ) + 11+ 2 j (6)
i =0 j= 0
2
+ (γ 17 + 2 k CNTL k , t + γ 18+ 2 k NI t * CNTL k , t ) +ξ t
k =0
where:
12
Sloan (1996) and Xie (2001) find the market systematically undervalues operating cash flows relative to
accruals. When we interact our independent and effective variables with OCFt in addition to NIt we find that
the audit committee independence and effectiveness do not reduce the relative underpricing of operating cash
flows across our mispricing tests. The inclusion of the additional interaction terms in the regression, however,
does not alter our findings with respect to the mispricing accounting accruals.
16
γi = regression coefficients, i ∈ [0, 22];
ξt = error term.
All other variables are defined as above. If accruals are overpriced, as the prior literature
indicates, we would observe in Equation (6) a negative coefficient on NIt (γ2 < 0) when tested
without the indicator variables or interaction terms.
If independent and effective audit committees enhance earnings transparency, such that
the overpricing of accruals is lower for those firms with these characteristics, we will find
positive coefficients on NIt, interacted with INDt or EFFi,t, from Equation (6). We formally state
this in our next two hypotheses (in alternative form) below:
H1b: The mispricing of accounting accruals is lower with independent audit

committees (γ4 > 0).
H2b: The mispricing of accounting accruals is lower with effective audit

committees (γ6 > 0; γ8 > 0; and, γ10 > 0).
We again separately test the effects of independence and effectiveness in H1b and H2b,
respectively, to identify the influence of each characteristic on earnings transparency. We also
test H1b and H2b jointly and test whether the effects of independence and effectiveness are
incremental to each other. If independence increases earnings transparency more than any other
audit committee characteristic, as implied by the BRC’s positioning of independence above all
else, we would expect to observe the overpricing of accruals will be lower for independent
committees than for effective committees. This would be confirmed if γ4 is greater than the
coefficients on the effectiveness characteristics (γ6, γ8 or γ10) in our joint test of Equation (6).
We formally stated this in our final hypothesis (in alternative form) below:
17
H3a: The mispricing of accounting accruals is lower with independent
committees than with effective committees (γ4 > γ6, γ8 or γ10).
4. Sample Selection and Descriptive Statistics
4.1 SAMPLE SELECTION CRITERIA
We collected board and audit committee data for firms listed on the 1996 Fortune 500,
over the period 1996 to 2000. We limited our sample to only the 500 largest firms since board
and audit committee data is most likely available for these firms and these firms are currently
under the most scrutiny by the SEC and other groups.13 Of the possible 2,500 firm-years, we
deleted those firm-years from the financial institution and utility industries, reducing our sample
by 460 firm-years and 205 firm-years, respectively. We deleted these observations from our
sample since these are from regulated industries such that governmental auditors, compliance
officers, and others also oversee the financial reporting process, beyond the oversight performed
by the board or the audit committee.14
Over our sample period, these selection criteria yielded 1,835 firm-years. To be included
in our final sample, each firm-year had to have sufficient data for our hypothesis tests, as well as
for our sensitivity tests. We thus required proxy data on a firm’s audit committee, including its
members’ affiliations, background and other board activities, and the frequency of meetings
throughout the year. Data on large outside blockholders was also required. As detailed in Table
1, these selection criteria resulted in the loss of 343 firm-years. Further, we lost 12 observations
for missing earnings announcement dates.
13
In comparison, Vafeas (2000) limits his sample to only the top 350 firms in the Fortune 500 and Klein (2002a,
2002b) limits her samples to the S&P 500.
18
______________________________
Insert Table 1 Here

______________________________
From the I/B/E/S database, we required analysts’ forecast of growth and analyst
following for both the informativeness and transparency samples, resulting in a loss of seven
observations. We also required from I/B/E/S for the informativeness sample data on reported
earnings and analysts’ consensus forecasts, reducing our informativeness sample by two
additional observations.
From Compustat, we require for both samples the earnings announcement dates, current
year’s income before extraordinary items, operating cash flows and dividends, resulting in a loss
of 34 observations. For the informativeness sample, we also required the prior eight years of
data on income before extraordinary items to calculate the control variable of earnings
persistence. This additional time-series requirement resulted in a loss to the informativeness
sample of 121 observations.15
From CRSP, we required data for both samples on the prior year’s daily returns to
calculate market beta (and current year’s daily returns to calculate cumulative abnormal returns
for the informativeness sample), resulting in a loss of 22 observations. For the transparency
sample, we also required data on size-based decile returns, market price on the day after the
current earnings announcement date and after the next year’s earnings announcement date,
resulting in an additional loss of 117 observations.
14
For similar reasons, Manry et al. (2003) also deleted financial institutions and utilities from their study of the
BRC recommendation to mandate timely reviews of quarterly earnings.
15
We also calculated earnings persistence as the earnings/price ratio, as in Lundholm and Myers (2002), thus
eliminating the loss of the additional 121 observations. Our results were substantially unchanged by the
alternative specification for persistence and the increased sample size.
19
Finally, we deleted three observations from the informativeness sample and five
observations from the transparency sample that the DFFIT test identified as unduly influencing
our results (see Neter et al. 1989), yielding a final sample of 1,291 and 1,295 for the
informativeness and transparency sample, respectively.16
4.2 DESCRIPTIVE STATISTICS
Table 2 presents descriptive statistics of the data. Panel A presents the full sample along
with the sample partitioned on whether the audit committee is comprised completely of
independent directors (henceforth, 100% independent group) or comprised of at least one
director directly or indirectly related to the company (henceforth, <100% independent group).
Panel B presents the sample partitioned on whether the audit committee meets at least two of the
three BRC recommendations (financial literacy, meeting frequency and time commitment) for
effective audit committees (henceforth, >66% effective group) or fails to meet at least two of
these recommendations (henceforth, <66% effective group). Each panel separately reports the
informativeness and transparency samples.
______________________________
Insert Table 2 Here

______________________________
In comparing the financial statistics for the informativeness sample, we find that neither
cumulative abnormal returns nor unexpected earnings significantly differ across the
16
The distributions of observations across the industries for both samples are similar to the overall distribution of
the Fortune 500. Within both samples, the industry representation of firms with independent and non-
independent audit committees as well as effective and non-effective audit committees is substantially balanced,
suggesting that industry differences would not account for the empirical regularities we observe. As an
additional test, we include industry dummy indicator variables and retest our regressions by including these
variables both as intercepts and as interactions with either unexpected earnings (informativeness tests) or
reported earnings (transparency tests). The inclusion of these variables does not alter our results or conclusions.
20
independence groups. We also find that across the transparency sample, neither mispricing (Vt –
Pt) nor reported earnings significantly differ across the independence groups, however, operating
cash flows for the 100% group are lower than for the <100% group (means of 0.117 and 0.131,
respectively).
With respect to the audit committee characteristics for the informativeness sample, the
100% group exhibits a significantly greater proportion of independent directors (by definition)
than the <100% group (means of 1.000 and 0.712, respectively). Of the effectiveness
characteristics for the independence groups, we find significant differences in both financial
literacy and number of directors serving on no more than three boards (means of 0.847 and 0.670
for the 100% group and 0.784 and 0.623 for the <100% group, respectively). The same
differences in financial literacy and number of boards exist for transparency sample. Together,
this suggests interdependency in the independence and effectiveness characteristics.
For the other factors that may influence earnings informativeness, we find that the 100%
group exhibits significantly higher growth forecasts and greater analyst precision than the <100%
group (means of 0.130 and 0.075 for the 100% group and 0.126 and 0.065 for the <100% group,
respectively). Again, we observe the same differences across the growth forecasts and analyst
precision for the transparency sample.
Across the effectiveness groups, we find no significant differences in cumulative
abnormal returns, unexpected earnings, mispricing or reported earnings. For the transparency
sample, however, we find that operating cash flows are significantly lower for the >66% group
than <66% group (means of 0.112 and 0.128 for the more >66% and <66% groups, respectively).
With respect to the audit committee characteristics for the informativeness sample, the
>66% group exhibits a significantly greater proportion of independent directors than the <66%
21
group (means of 0.917 and 0.879, respectively), again suggesting interdependency in the
independence and effectiveness characteristics. By construction, we observe more financially
literate directors, more meetings and more members serving on no more than three boards for the
>66% group than <66% group, for both the informativeness and transparency sample.
For the other factors that may influence earnings informativeness, we find that the >66%
group exhibits significantly higher growth forecasts, greater analyst precision and higher
percentages of losses than the <66% group (means of 0.136, 0.079 and 0.130 for the >66% group
and 0.125, 0.067 and 0.070 for the <66% group, respectively). We also observe the same
differences across the growth forecasts and analyst precision for the transparency sample.
Overall, the descriptive statistics suggest interdependence in independent and effective
audit committees, and that firms with independent and effective audit committees tend to exhibit
higher growth expectations and greater analyst precision. Further, the significant differences in
operating cash flows, but not earnings, suggest that operating cash flows comprise a smaller
proportion of earnings for those firms with independent and effective audit committees than
those without.
Table 3 presents a correlation matrix of our audit committee characteristics and other
factors that may influence the audit committee’s role in the financial oversight process. For the
informative sample, we find significant positive correlations between our measure for audit
committee independence (INDt) and our measures for financial literacy (FINLITt) and meeting
frequency (MEETt), again suggesting interdependence in independence and effectiveness (p-
values of 0.0001). Further, we find a significant positive correlation between MEETt and
FINLITt, but MEETt is significantly negatively correlated with our measure of time commitment
22
(NOBUSYt), suggesting the effective characteristics may be, in part, substitutes for each other
(p-values of 0.0886 and 0.0311, respectively).
______________________________
Insert Table 3 Here

______________________________
With respect to the other factors that could influence the audit committee, we find that
INDt is positively correlated with the CEO not exerting influence (NOINFt), but negatively
correlated with the CEO’s investment of corporate stock (CEOSHRt) and the existence of a large
outside blockholder (BLOCKt) on the audit committee (p-values of 0.0001, 0.0634 and 0.0002,
respectively). FINLITt is positively correlated with NOINFt and negatively with BLOCKt (p-
values of 0.0014 and 0.0155, respectively). Further, MEETt is positively correlated, and
NOBUSYt is negatively correlated with NOINFt (p-values of 0.0001), but not correlated with
any other factor. Finally, each of the other factors is negatively correlated with each other,
suggesting another interdependence in our measures. The correlations for the transparency
sample are almost identical.
5. Results
5.1 TESTS OF THE INFORMATIVENESS HYPOTHESES
Table 4 presents the results from our ERC tests. Panel A presents the regression
coefficients and Panel B presents the differences in coefficients across the audit committee
characteristics. We present a baseline (Base Model) that excludes audit committee
characteristics as well as separate and joint tests of independence and effectiveness model.
23
______________________________
Insert Table 4 Here

______________________________
We find the coefficient on UEt is 5.489 in the Base Model (p-value < 0.05), representing
the average ERC for all sample firms. The coefficient on UEt drops to 4.789 (p-value < 0.05)
when we separately test independence (Independence Model) and 4.159 (p-value < 0.10) when
we separately test of effectiveness (Effectiveness Model). The respective coefficients on UEt in
the Independence and Effectiveness Models represent the ERC for firms that are classified as
either not independent (some related party serves on the audit committee) or not effective (the
directors are not financially literate, do not meet regularly and serve on more than three other
boards), respectively. When we include interactions for both independence and effectiveness
jointly (Joint Model), the coefficient on UEt (ERC) is not significantly different from zero. This
is interpreted as the ERC not being significantly different from zero for those firms where the
audit committee is classified as neither independent nor effective.17
When we interact UEt with INDt in the Independence Model, we find that the coefficient
on the interaction term is 2.226 (p-value < 0.05), providing evidence that firms with independent
audit committees exhibit higher ERCs and lending support for H1a. In the Effectiveness Model,
we find that the coefficient on the interaction of UEt with FINLITt is 2.118 (p-value < 0.05),
providing evidence that firms with financially literate audit committees exhibit higher ERCs and
17
With respect to the other factors and controls we included in our informativeness tests, we find the coefficients
on the interactions of UEt with CEOSHRt GROWTHt, PRECISIONt and PERSt are significantly different from
zero. This suggests that CEO stock ownership, growth, analyst precision and earnings persistence influence
earnings informativeness beyond the other controls included and those audit committee characteristics
recommended by the BRC. Additionally, we find the coefficient on the control variable for nonlinearity is
significantly negative.
24
lending support for H2a. Neither coefficient on the interactions of UEt with MEETt or
NOBUSYt, however, is significantly different from zero.
In the Joint Model, when we regress our independence and effectiveness measures
simultaneously, the coefficients on the interaction terms reflect the extent that firms with
independent and effective audit committees exhibit higher ERCs, incremental to that exhibited
by independent or effective audit committees, alone (i.e., incremental ERC). We find that the
coefficient on the interaction of UEt with INDt is 2.106 (p-value < 0.05) and UEt with FINLITt is
1.962 (p-value < 0.05), suggesting that independence and financial literacy incrementally
improve informativeness. As with the individual tests, we find in the Joint Model that neither
coefficient on the interactions of UEt with MEETt or with NOBUSYt, is significantly different
from zero.
In testing whether the BRC’s claim that independence above all improves earnings
quality (see Table 4, Panel B), we find that the coefficient on the interaction of UEt with INDt is
significantly greater than coefficients on the interactions of UEt with MEETt and with NOBUSYt
(p-values < 0.05) lending support, in part, for H3a. We find no difference, however, in the
coefficients on the interactions of UEt with INDt and UEt with FINLITt. Overall, our results
suggest that independent and financially literate audit committees enhance earnings quality by
improving the informativeness of reported earnings.
5.2 TESTS OF THE TRANSPARENCY HYPOTHESES
Table 5 presents the results from our tests of mispricing. Panel A presents the regression
coefficients and Panel B presents the differences in coefficients across the audit committee
characteristics. In this set of tests, the coefficient on NIt represents the mispricing of accruals
25
and serves as our test metric. A negative coefficient on NIt, therefore implies that the market
overprices accruals.
______________________________
Insert Table 5 Here

______________________________
Consistent with the prior literature we find the coefficient on NIt is –3.295 for the Base
Model (p-value < 0.01), representing the average overpricing of accruals for all sample firms.
For the Independence Model, the negative coefficient on NIt increases to –3.532 (p-value <
0.01), representing the overpricing of accruals for firms whose audit committees classified as not
independent. For the Effectiveness Model, the negative coefficient on NIt increases even further
to –4.121 (p-value < 0.01), representing the overpricing of accruals for firms whose audit
committees classified as not effective. In the Joint Model, the coefficient on NIt is –4.523 (p-
value < 0.01), and represents those firms whose audit committees are classified as not
independent and not effective. Together, these results suggest that the overpricing of accruals
decreases as the level of corporate governance increases.18
When we interact NIt with INDt in the Independence Model, we find that the coefficient
on the interaction term is 1.185 (p-value < 0.05), providing evidence that the overpricing of
accruals is lower for firms with independent audit committees, thus lending support for H1b. In
the Effectiveness Model, we find that the coefficient on the interaction of NIt with MEETt is
1.182 (p-value < 0.05), providing evidence that the overpricing of accruals is lower for firms
18
With respect to the other factors and controls we include in our transparency tests, we find the coefficients on
the interactions of NIt with BETAt, GROWTHt and PRECISIONt are significantly different from zero,
consistent with the prior literature’s findings that mispricing is related to risk, growth and analyst precision. We
find no evidence that mispricing is related to CEO influence, CEO stock ownership or large outside
blockholders serving on the audit committee.
26
whose audit committees meet regularly, lending support for H2b. We also find that the
coefficient on the interaction of NIt with INDt is 1.412 (p-value < 0.01) and NIt with MEETt is
1.357 (p-value < 0.01), suggesting that independence and regular meetings incrementally
improve transparency. For both the Effectiveness and Joint Models, however, neither coefficient
on the interactions of NIt with FINLITt or with NOBUSYt, differs significantly from zero.
In Panel B of Table 5, the results from testing whether the overpricing of accruals is
reduced more when the audit committee is independent over all other characteristics. We find
that the coefficient on the interaction of NIt with INDt is significantly greater than coefficients on
the interactions of NIt with FINLITt or with NOBUSYt (p-values < 0.05 and < 0.10, respectively)
lending support, in part, for H3b. We find no difference, however, in the coefficients on the
interactions of NIt with INDt and NIt with MEETt. Overall, our results suggest that independent
audit committees that meet regularly enhance earnings quality by improving the transparency of
reported earnings.
5.3 ADDITIONAL YEAR-BY-YEAR TESTS
The results we report for our hypothesis tests are based on pooling the data across our
five-year sample period (1996-2000). Estimating our test regressions by year and averaging the
five year-by-year coefficients yield results substantially the same as those presented above, again
supporting our hypotheses. We pool the five year-by-year standard deviations in order to
calculate the t-statistics that determine the level of significance.
For our informativeness tests, when we retest the Independence Model by year, we find
that the average of the coefficients on the interaction of UEt with INDt across the five years is
3.861 (p-value < 0.01). Retesting the Effectiveness Model by year yields an average coefficient
on the interaction of UEt with FINLITt of 2.984 (p-value < 0.01), but the interactions of UEt with
27
MEETt and UEt with NOBUSYt are again, not significantly different from zero. We find similar
results to those reported in Tables 4 for the Joint Model.
For the transparency tests, when we retest the Independence Model by year, we find that
the average of the coefficients on the interaction of NIt with INDt across the five years is 1.184
(p-value < 0.05). Retesting the Effectiveness Model by year yields an average coefficient on the
interaction of UEt with MEETt of 1.181 (p-value < 0.05), but the interactions of UEt with
FINLITt and UEt with NOBUSYt are again, not significantly different from zero. We again find
similar results to those reported in Table 5 for the Joint Model.
In sum, the results from our study suggest that independent audit committees enhance
earnings quality by increasing both earnings informativeness and earnings transparency. In
contrast, we find that financially literate audit committee members enhance earnings quality only
through increasing earnings informativeness and regular audit committee meetings enhance
earnings quality only by increasing earnings transparency.
6. Conclusion
We provide evidence that audit committees comprised of independent and financially
literate directors that meet regularly enhance earnings informativeness and transparency. This
evidence suggests that financial reporting oversight by an independent and effective audit
committee improves the quality of reported earnings. These results are robust to the inclusion of
other factors identified by the literature that could influence the audit committee’s role in
financial reporting oversight, such as CEO influence, CEO shareholdings and large outside
blockholders serving on the audit committee, as well as other factors that may influence ERCs or
mispricing. Overall, these results lend supporting evidence to The Blue Ribbon Committee on
28
Improving the Effectiveness of Corporate Audit Committees’ recommendations for enhancing
audit committee oversight of the financial reporting process and suggests elements of the
Sarbanes-Oxley Act of 2002 is likely to achieve its goal of enhancing earnings quality.
One limitation of our study is that our sample is restricted to large Fortune 500 firms.
While our sample is comparable to other audit committee studies, such as Klein (2002a, 2002b),
our results may not be generalizable to smaller firms. Future research may be able to identify
whether the high level of public scrutiny common with larger firms affects the influence of
independent and effective audit committees on earnings quality.
29
REFERENCE
Abbott, L. J. Park, and S. Parker. 2000. The effects of audit committee activity and independence
on corporate fraud. Managerial Finance 26(26): 55-67.
________, S. Parker, and G. Peters. 2002. Audit committee characteristics and financial
misstatement: a study of the efficacy of certain Blue Ribbon Committee recommendations,
Working paper.
Aboody, D., J. Hughes, and J. Liu. 2002. Measuring value relevance in a (possibly) inefficient
market. Journal of Accounting Research 40(4): 965-986.
Ali, A., L. Hwang, and M. Trombley. 2003. Residual-income-based valuation predicts future
stock returns: Evidence on mispricing vs. risk explanations. The Accounting Review 78(2):
377-396.
Ali, A., and P. Zarowin. 1992. The role of earnings levels in annual earnings-returns studies.
Journal of Accounting Research 30(2): 286-296.
Beasley, M. 1996. An empirical analysis of the relation between the board of director
composition and financial statement fraud. The Accounting Review 71(4): 443-465.
Blue Ribbon Committee on Improving the Effectiveness of Corporate Audit Committees. 1999.
Report and Recommendations of the Blue Ribbon Committee on Improving the effectiveness
of Corporate Audit Committee. New York: New York Stock Exchange.
Brown, S., K. Lo, and T. Lys. 1999. Use of R2 in accounting research: measuring changes in
value relevance over the last four decades. Journal of Accounting and Economics 28 (2):83-
115.
Collins, D., and S.P. Kothari. 1989. An analysis of intertemporal and cross-sectional
determinants of earnings response coefficients. Journal of Accounting and Economics
11(2): 143-181.
Core, J., R. Holthausen, and D. Larcker. 1999. Corporate governance, chief executive officer
compensation, and firm performance. Journal of Financial economics 51(3): 371-406.
Dechow, P., R. Sloan, and A. Sweeney. 1996. Causes and consequences of earnings
manipulation: An analysis of firms subject to enforcement actions by the SEC.
Contemporary Accounting Research 13(1): 1-36.
DeZoort, F., D. Hermanson, D Archanbeau and S. Reed. 2002. Audit committee effectiveness: A
synthesis of the empirical audit committee literature. Journal of Accounting Literature 21:
38-75.
Fairfield, P., J. Whisenant, and T Yohn. 2003. Accrued earnings and growth: Implications for
future profitability and market mispricing. The Accounting Review 78(1): 353-371.
30
Freeman, R., and S. Tse. 1992. A nonlinear model of security price responses to unexpected
earnings. Journal of Accounting Research 30 (2): 185-209.
Jennings, R. 1990. A note on interpreting “incremental information content.” The Accounting

Review 65(1): 925-932.
Klein, A. 2002a. Audit committee, board of director characteristics, and earnings management.
Journal of Accounting and Economics 33(3): 375-400.
_______. 2002b. Economics determinants of audit committee independence. The Accounting

Review 77(2): 435-452.
Kothari, S. 2001. Capital markets research in accounting. Journal of Accounting and Economics
31(1-3):105-232.
Levitt, A. 1998. “The Numbers Game”, Remarks delivered at the NYU Center for Law and
Business, September 28.
Lundholm, R. and L. Myers. 2002. Bringing the future forward: The effect of disclosure on the
returns-earnings relation. Journal of Accounting Research 40(3): 809-839.
Manry, D., S. Tiras, and C. Wheatley. 2003. The influence of interim auditor reviews on the
association of returns with earnings. The Accounting Review 78(1): 251-274.
Menon, K., and J. Williams. 1994. The use of audit committees for monitoring. Journal of
Accounting and Public Policy 13(2): 121-139.
Neter, J., W. Wasserman, and M. Kutner. 1989. Applied Linear Regression Models. 2nd edition.
Homewood, IL: Richard D. Irwin.
Ohlson J., and P. Shroff. 1992. Changes versus levels in earnings as explanatory variables for
returns: Some theoretical considerations. Journal of Accounting Research 30(2): 210-226.
Sarbanes, P. and M. Oxley. 2002. Sarbanes-Oxley Act of 2002, The Public Company Accounting
Reform and Investor Protection Act. Washington D.C.: U.S. Congress.
Securities and Exchange Commission (SEC). 1999. Final Rule: Audit Committee Disclosure, 17
CFR, Parts 210, 228, 229, and 240. SEC Release No. 34-42266.
Shivdasani, A., and D. Yermack. 1999. CEO involvement in the selection of new board
members: an empirical analysis. The Journal of Finance 54 (5): 1829-1853.
Sloan, R. 1996. Do stock prices fully reflect information in accruals and cash flows about future
earnings? The Accounting Review 71(3): 289-315.
Turner, L. 1999. “Initiatives for improving the quality of financial reporting.” Remarks
delivered at “The SEC speaks in 1999” sponsored by the Practicing Law Institute,
Washington, D.C., February 10.
31
Turner, L. 2001. “Quality, transparency, and accountability.” Remarks delivered at the Financial
Executives Institutes-American Institute of CPA’s “Benchmarking the quality of earnings”
conference in New York, April 26.
Teoh, S. H. and T. Wong. 1993. Perceived auditor quality and the earnings response coefficient.
The Accounting Review 68(2): 346-366.
Vafeas, N. 2000. Board structure and the informativeness of earnings. Journal of Accounting and
Public Policy 19 (2): 139-160
Warfield, T., J. Wild, and K. Wild. 1995. Managerial ownership, accounting choices, and
informativeness of earnings. Journal of Accounting and Economics 20 (1): 61-91.
White, H. 1980. A heteroskedasticity-consistent covariance matrix estimator and a direct test for
heteroskedasticity. Econometrica 48 (4): 817-838.
Wild, J. 1996. The audit committee and earnings quality. Journal of Accounting, Auditing &
Finance 11(2): 247-276.
Xie, H. 2001. The mispricing of abnormal accruals. The Accounting Review 76(3): 357-373.
32
TABLE 1
Sample Selection and Screening Procedures
Informativeness Transparency
Observations Observations
Initial Sample of Fortune 500 firms, 1996 500 500
Possible firm-years, 1996-2000 2,500 2,500

Less firm-years:
from banking industry 205 205
from utility industry 460 460
Subtotal 1,835 1,835
Less firm-years:
Missing data on board or audit committee 343 343
Missing data on earnings announcement dates 12 12
Missing data on IBES 9 7
Missing data on Compustat 155 34
Missing data on CRSP 22 139
Subtotal 1,294 1,300
Less firm-years:
With influential observations 3 5
Total firm-years in final sample 1,291 1,295
33
TABLE 2
Sample Descriptive Statistics and Tests for Differences across Firms with 100% Independent versus <100% Independent
Audit Committees and Firms with >66% Effective versus <66% Effective Audit Committees
Panel A: Descriptive Statistics for Full Sample and Firms Partitioned by Audit Committee (AC) Independence
100% Independent <100% Independent
Full Sample Committees Committees Differences in
Std. Std. Std. Subsample
Variable Mean Median Dev. Mean Median Dev. Mean Median Dev. Means
Informativeness Sample (Full Sample, n=1291; 100% Independent group, n=811; <100% Independent group, n=480)
Cumulative Abnormal Returns -0.025 -0.064 0.359 -0.020 -0.066 0.386 -0.035 -0.063 0.306 0.0149
Unexpected Earnings -0.005 -0.001 0.023 -0.005 0.000 0.024 -0.005 -0.001 0.021 -0.0007
% Independent 0.893 1.000 0.166 1.000 1.000 0.000 0.712 0.750 0.148 0.2876 ***
% Financially Literate 0.823 0.833 0.188 0.847 1.000 0.185 0.784 0.800 0.186 0.0623 ***
Number of AC Meetings 3.884 4.000 1.571 3.888 4.000 1.570 3.877 4.000 1.573 0.0105
% Serving on < 3 Boards 0.653 0.667 0.248 0.670 0.667 0.252 0.623 0.600 0.238 0.0476 ***
% Non-Influential CEOs 0.748 1.000 0.435 0.793 1.000 0.406 0.672 1.000 0.470 0.1213 ***
CEO Stock Ownership % 0.024 0.005 0.070 0.021 0.005 0.057 0.029 0.005 0.088 -0.0075 *
% AC with Blockholders 0.036 0.000 0.185 0.021 0.000 0.143 0.060 0.000 0.238 -0.0393 ***
Market Beta 0.904 0.860 0.385 0.913 0.849 0.409 0.889 0.871 0.342 0.0240
Growth forecast 0.128 0.120 0.043 0.130 0.120 0.044 0.126 0.120 0.041 0.0042 *
Analyst Forecast Precision 0.071 0.059 0.048 0.075 0.063 0.051 0.065 0.056 0.041 0.0098 ***
Earnings Persistence 0.443 0.439 0.538 0.429 0.412 0.552 0.466 0.461 0.514 -0.0376
% Reporting Losses 0.091 0.000 0.287 0.089 0.000 0.285 0.094 0.000 0.292 -0.0048
TABLE 2: CONTINUED
Panel A: Continued
100% Independent <100% Independent
Full Sample Committees Committees Differences in
Std. Std. Std. Subsample
Variable Mean Median Dev. Mean Median Dev. Mean Median Dev. Means
Transparency Sample (Full Sample, n=1295; 100% Independent group, n=814; <100% Independent group, n=481)
Mispricing (Vt - Pt) 0.051 -0.021 0.525 0.072 -0.013 0.561 0.016 -0.027 0.456 0.0555
Reported Earnings 0.052 0.054 0.057 0.051 0.052 0.056 0.053 0.056 0.058 -0.0014
Operating Cash Flows 0.122 0.099 0.111 0.117 0.096 0.106 0.131 0.101 0.119 -0.0144 **
% Independent 0.892 1.000 0.166 1.000 1.000 0.000 0.709 0.750 0.143 0.2912 ***
% Financially Literate 0.824 0.833 0.191 0.850 1.000 0.186 0.781 0.800 0.192 0.0695 ***
Number of AC Meetings 3.874 4.000 1.589 3.903 4.000 1.597 3.825 4.000 1.577 0.0776
% Serving on < 3 Boards 0.651 0.667 0.247 0.664 0.667 0.252 0.631 0.625 0.238 0.0325 **
% Non-Influential CEOs 0.742 1.000 0.438 0.781 1.000 0.414 0.676 1.000 0.469 0.1057 ***
CEO Stock Ownership % 0.026 0.005 0.075 0.023 0.005 0.062 0.032 0.006 0.092 -0.0093 **
% AC with Blockholders 0.033 0.000 0.179 0.023 0.000 0.151 0.050 0.000 0.218 -0.0266 **
Market Beta 0.913 0.868 0.395 0.921 0.863 0.421 0.899 0.883 0.347 0.0217
Growth forecast 0.132 0.120 0.046 0.134 0.125 0.047 0.129 0.120 0.044 0.0056 **
Analyst Forecast Precision 0.069 0.059 0.041 0.072 0.059 0.043 0.064 0.056 0.035 0.0080 ***
TABLE 2: CONTINUED
Panel B: Descriptive Statistics for Firms Partitioned by Audit Committee (AC) Effectiveness
>66% Effective <66% Effective
Committees Committees Differences in
Std. Std. Subsample
Variable Mean Median Dev. Mean Median Dev. Means
Informativeness Sample (>66% Effective group, n=436; <66% Effective group, n=855)
Cumulative Abnormal Returns -0.023 -0.070 0.408 -0.027 -0.064 0.331 0.0044
Unexpected Earnings -0.007 -0.001 0.028 -0.004 -0.001 0.020 -0.0033 **
% Independent 0.918 1.000 0.149 0.880 1.000 0.172 0.0382 ***
% Financially Literate 0.947 1.000 0.136 0.760 0.750 0.180 0.1862 ***
Number of AC Meetings 4.689 4.000 1.634 3.473 3.000 1.366 1.2163 ***
% Serving on < 3 Boards 0.726 0.750 0.282 0.615 0.600 0.219 0.1116 ***
% Non-Influential CEOs 0.767 1.000 0.423 0.738 1.000 0.440 0.0286
CEO Stock Ownership % 0.026 0.006 0.066 0.023 0.005 0.073 0.0032
% AC with Blockholders 0.030 0.000 0.170 0.039 0.000 0.193 -0.0089
Market Beta 0.935 0.865 0.447 0.889 0.858 0.349 0.0463
Growth forecast 0.136 0.128 0.049 0.125 0.120 0.040 0.0114 ***
Analyst Forecast Precision 0.414 0.407 0.572 0.457 0.460 0.520 -0.0434
Earnings Persistence 0.079 0.063 0.058 0.067 0.056 0.041 0.0125 ***
% Reporting Losses 0.130 0.000 0.337 0.070 0.000 0.256 0.0603 ***
TABLE 2: CONTINUED
Panel B: Continued
>66% Effective <66% Effective
Committees Committees Differences in
Std. Std. Subsample
Variable Mean Median Dev. Mean Median Dev. Means
Transparency Sample (>66% Effective group, n=437; <66% Effective group, n=858)
Mispricing (Vt - Pt) 0.057 -0.024 0.597 0.048 -0.018 0.484 0.0085
Reported Earnings 0.048 0.047 0.062 0.054 0.056 0.054 -0.0063
Operating Cash Flows 0.112 0.092 0.105 0.128 0.101 0.114 -0.0162 **
% Independent 0.917 1.000 0.153 0.879 1.000 0.170 0.0375 ***
% Financially Literate 0.949 1.000 0.139 0.761 0.750 0.183 0.1882 ***
Number of AC Meetings 4.728 4.000 1.640 3.439 3.000 1.373 1.2883 ***
% Serving on < 3 Boards 0.723 0.750 0.276 0.615 0.667 0.223 0.1087 ***
% Non-Influential CEOs 0.767 1.000 0.423 0.730 1.000 0.444 0.0370
CEO Stock Ownership % 0.029 0.007 0.075 0.024 0.005 0.074 0.0051
% AC with Blockholders 0.025 0.000 0.157 0.037 0.000 0.190 -0.0121
Market Beta 0.930 0.865 0.446 0.904 0.870 0.367 0.0260
Growth forecast 0.141 0.130 0.052 0.128 0.120 0.042 0.0130 ***
Analyst Forecast Precision 0.075 0.063 0.049 0.066 0.056 0.035 0.0097 ***
TABLE 2: Continued
Notes:
*, ** and *** signify two-tailed significance at the 0.10, 0.05 and 0.01 level.
We define the variables as follows:
Cumulative = holding period returns cumulated from the day after the prior year’s
Abnormal earnings announcement date for year t-1 through the day current
Returns earnings announcement date for year t, less market holding period
returns for the same period;
Mispricing = intrinsic value per share (Vt) estimated as of the day after the year
t+1 earnings announcement date less market price per share (Pt) as
of the day after the year t earnings announcement date, standardized
by market price set as of the day after the year t-1 earnings
announcement date;
Unexpected = unexpected earnings reported for year t, proxied as the actual
Earnings reported earnings (from I/B/E/S) less the consensus analyst forecasts
of annual earnings measured as the first forecast after the prior
year’s earnings announcement date, standardized by market price set
as of the day after the year t-1 earnings announcement date;
Reported = earnings per share before extraordinary items reported for year t,
Earnings standardized by market price set as of the day after the year t-1
earnings announcement date;
Operating = operating cash flows per share for year t, calculated from the
Cash Flows Statement of Cash Flows, standardized by market price set as of the
day after the year t-1 earnings announcement date;
% Independent = percentage of audit committee members with no direct or indirect
relations to the firm or its employees;
% Financially = percentage of audit committee identified as financially literate;
Literate
Number of = number of audit committee meetings within the fiscal year;
AC Meetings
% Serving on = percentage of audit committee members that serve on no more than
< 3 Boards three boards;
% Non- = percentage of firms without influential CEOs, which is an indicator
Influential variable set to one if the CEO does not serve as both the Chairman
CEOs of the Board and as a member on the nominating committee,
otherwise zero;
38
TABLE 2: Continued
Notes: Continued
CEO Stock = percentage of a firm’s common stock owned by the CEO, including
Ownership % those stock options ‘in the money’ exercisable within 60 days;
% AC with = percentage of firms with large outside blockholders (five-percent or
Blockholders more) on audit committee;
Market Beta = market beta calculated using the market model with year t-1 daily returns;
Growth = analyst forecasted long-term growth from I/B/E/S measured the month
Forecast after prior year’s earnings announcement;
Analyst = one divided by the number of analyst forecasters measured the month after
Forecast prior year’s earnings announcement;
Precision
Earnings = earnings persistence at the beginning of the year measured over an eight
Persistence year period;
% Reporting = percentage of firms reporting a net loss for year t.

Losses
39
TABLE 3
Correlation Matrix of Audit Committee Characteristics and Other Factors that may
Influence the Independence and Effectivenss of Audit Committees
Panel A: Informativeness Sample

Variable INDt FINLITt MEETt NOBUSYt NOINFt CEOSHRt BLOCKt
INDt 1.000
FINLITt 0.196 1.000

0.0001
MEETt -0.019 0.047 1.000
0.4993 0.0886
NOBUSYt 0.107 -0.010 -0.060 1.000
0.0001 0.7195 0.0311
NOINFt 0.136 0.089 0.125 -0.149 1.000
0.0001 0.0014 0.0001 0.0001
CEOSHRt -0.052 -0.029 -0.039 0.034 -0.325 1.000
0.0634 0.2899 0.1637 0.2205 0.0001
BLOCKt -0.103 -0.067 -0.015 -0.005 -0.061 -0.047 1.000
0.0002 0.0155 0.5808 0.8593 0.0274 0.0923
Panel B: Transparency Sample

Variable INDt FINLITt MEETt NOBUSYt NOINFt CEOSHRt BLOCKt
INDt 1.000
FINLITt 0.205 1.000

0.0001
MEETt 0.002 0.057 1.000
0.9292 0.0409
NOBUSYt 0.074 -0.012 -0.067 1.000
0.0076 0.6710 0.0159
NOINFt 0.117 0.082 0.117 -0.128 1.000
0.0001 0.0032 0.0001 0.0001
CEOSHRt -0.061 -0.014 -0.045 0.054 -0.326 1.000
0.0293 0.6036 0.1076 0.0504 0.0001
BLOCKt -0.072 -0.062 -0.026 0.004 -0.078 -0.047 1.000
0.0099 0.0248 0.3442 0.8872 0.0050 0.0888
40
TABLE 3: Continued
Notes:
Each cell reports the Pearson Correlation and its associated p-value.

INDt = indicator variable set to one if all audit committee members have no direct or
indirect relations to the firm or its employees, otherwise zero;
FINLITt = indicator variable set to one if all audit committee members identified as
financially literate, otherwise zero;
MEETt = indicator variable set to one if the audit committee meets at least four times
(the minimum threshold) within the fiscal year, otherwise zero;
NOBUSYt = indicator variable set to one if all audit committee members serve on no more
than three boards, otherwise zero;
NOINFt = indicator variable set to one if the CEO does not serve as both the Chairman
of the Board and as a member on the nominating committee, otherwise zero;
CEOSHRt = indicator variable set to one if the percentage of a firm’s common stock
owned by the CEO is in the upper half of the sample distribution, otherwise
zero;
BLOCKt = indicator variable set to one if large outside blockholders (five-percent or
more) serve on audit committee, otherwise zero.
41
TABLE 4
Results from Regressing Market Adjusted Returns on Unexpected Earnings - Interacted
with Indicator Variables for Independence, Effectiveness and Other Controls
Panel A: Regression Results

Base Independence Effectiveness Joint
sign Model Model Model Model
Adj. R-SQR 0.1900 0.1933 0.1936 0.1962
-1.639 -2.117 -2.352 -2.777
Intercept ? (-1.79) * (-2.20) ** (-2.48) ** (-2.79) ***
5.489 4.789 4.159 3.552
UEt + (2.49) ** (2.16) ** (1.81) * (1.53)
0.858 0.810
INDt ? (2.08) ** (1.89) *
2.266 2.106
UEt*INDt + (2.22) ## (2.02) ##
0.565 0.422
FINLITt ? (1.43) (1.05)
2.118 1.962
UEt*FINLITt + (2.27) ## (2.10) ##
-0.388 -0.250
MEETt ? (-0.98) (-0.62)
-1.370 -1.158
UEt*MEETt + (-1.28) (-1.08)
-0.870 -0.926
NOBUSYt ? (-2.16) ** (-2.29) **
-1.382 -1.658
UEt*NOBUSYt + (-1.14) (-1.35)
1.024 0.990 1.163 1.106
NOINFt ? (2.61) *** (2.52) ** (2.79) *** (2.65) ***
-1.459 -1.572 -1.036 -1.314
UEt*NOINFt ? (-1.38) (-1.49) (-0.94) (-1.18)
0.801 0.530 0.862 0.629
CEOSHRt ? (1.67) * (1.08) (1.79) * (1.28)
3.381 3.226 3.646 3.514
UEt*CEOSHRt ? (3.32) *** (3.17) *** (3.57) *** (3.43) ***
-1.273 -1.197 -1.407 -1.447
BLOCKt ? (-1.53) (-1.44) (-1.63) (-1.67) *
0.456 1.275 0.675 1.303
UEt*BLOCKt ? (0.08) (0.23) (0.12) (0.24)
42
TABLE 4: CONTINUED
Panel A: Continued
-0.465 -0.365 -0.374 -0.270
BETAt ? (-0.94) (-0.74) (-0.75) (-0.54)
-1.325 -1.339 -1.231 -1.235

UEt*BETAt ? (-1.20) (-1.22) (-1.11) (-1.11)
0.115 0.114 0.164 0.160

GROWTHt ? (2.75) *** (2.72) *** (3.65) *** (3.58) ***
0.794 0.751 0.865 0.824
UEt*GROWTHt ? (6.29) *** (5.88) *** (6.64) *** (6.27) ***
-0.059 -0.056 -0.041 -0.038
PRECISIONt ? (-2.39) ** (-2.21) ** (-1.59) (-1.45)
-0.397 -0.403 -0.376 -0.372

UEt*PRECISIONt ? (-3.76) *** (-3.82) *** (-3.45) *** (-3.42) ***
1.170 1.076 0.930 0.881
PERSt ? (3.04) *** (2.79) *** (2.33) ** (2.20) **
2.588 2.576 2.413 2.482
UEt*PERSt ? (2.27) ** (2.26) ** (2.09) ** (2.14) **
0.353 0.344 0.179 0.225
LOSSt ? (0.70) (0.69) (0.33) (0.42)
1.604 1.902 0.889 1.193

UEt*LOSSt ? (1.29) (1.52) (0.69) (0.92)
UE t UE t -0.989 -1.030 -0.952 -0.990

? (-5.43) (-5.62) (-5.14) (-5.33)
*** *** *** ***
Panel B: Tests of Differences in Interaction Terms on Independence and Effectiveness

(UEt*INDt) − 0.145
+
(UEt*FINLITj,t) (0.10)
(UEt*INDt) − 3.265
+ (2.28) ##
ΜΕΕΤj,t)
(UEt*ΜΕΕΤ
(UEt*INDt) − 3.765
+ (2.17) ##
(UEt*NOBUSYj,t)
43
TABLE 4: Continued
Notes:
The regression equation is:
CAR t = α 0 + α 1 UE t + α 2 IND t + α 3 UE t * IND t + α 4 FINLITt + α 5 UE t * FINLITt + α 6 MEETt

+ α 7 UE t * MEETt + α 8 NOBUSYt + α 9 UE t * NOBUSYt + α 10 NOINFt + α 11 UE t * NOINFt
+ α12 CEOSHR t + α13 UE t * CEOSHR t + α14 BLOCK t + α 15 UE t * BLOCK t + α16 BETA t
+ α17 UE t * BETA t + α18 GROWTH t + α19 UE t * GROWTH t + α 20 PRECISION t
+ α 21 UE t * PRECISION t + α 22 PERS t + α 23 UE t * PERS t + α 24 LOSS t
+ α 25 UE t * LOSS t + α 26 UE t UE t + ν t
Each cell in Panel A reports the OLS coefficient (or differences in coefficients) and the White’s
(1980) adjusted t-statistic. *, ** and *** signify two-tailed significance at the 0.10, 0.05 and 0.01
level. #, ## and ### signify one-tailed significance at the 0.10, 0.05 and 0.01 level.

CARt = holding period returns cumulated from the day after the prior year’s
earnings announcement date for year t-1 through the day current earnings
announcement date for year t, less market holding period returns for the
same period;
UEt = unexpected earnings reported for year t, proxied as the actual reported
earnings (from I/B/E/S) less the consensus analyst forecasts of annual
earnings measured as the first forecast after the prior year’s earnings
announcement date, standardized by market price set as of the day after the
year t-1 earnings announcement date;
INDt = indicator variable set to one if all audit committee members have no direct
or indirect relations to the firm or its employees, otherwise zero;
FINLITt = indicator variable set to one if all audit committee members identified as
financially literate, otherwise zero;
NOBUSYt = indicator variable set to one if all audit committee members that serve on
no more than three boards, otherwise zero;
NOINFt = indicator variable set to one if the CEO does not serve as both the
Chairman of the Board and as a member on the nominating committee,
otherwise zero;
owned by the CEO is in the upper half of the sample distribution,
otherwise zero;
44
TABLE 4: Continued
Notes (continued):

more) serve on audit committee, otherwise zero;
BETAt = market beta calculated using the market model with year t-1 daily returns;
GROWTHt = analyst forecasted long-term growth from I/B/E/S measured the month after
prior year’s earnings announcement;
PRECISIONt = one divided by the number of analyst forecasters measured the month after
prior year’s earnings announcement;
PERSt = earnings persistence at the beginning of the year measured over an eight year
period;
LOSSt = percentage of firms reporting a net loss for year t;
UEt UEt = unexpected earnings for year t times the absolute value of unexpected
earnings.
45
TABLE 5
Results from Regressing Differences in Intrinsic Values and Market Values on Operating
Cash Flows and Earnings - Interacted with Indicator Variables for Independence,
Effectiveness, and Other Controls
Panel A: Regression Results

Adj. R-SQR 0.0502 0.0544 0.0578 0.0629
3.078 2.503 2.922 2.554
Intercept ? (1.59) (1.23) (1.46) (1.22)
0.575 0.618 0.565 0.593

OCFt + (3.10) *** (3.32) *** (3.01) *** (3.16) ***
-3.295 -3.532 -4.121 -4.523
NIt - (-2.69) *** (-2.87) *** (-3.24) *** (-3.51) ***
0.623 0.398
INDt ? (0.69) (0.43)
1.185 1.412
NIt*INDt + ##
(2.21) (2.54) ###
1.353 1.390
FINLITt ? (1.77) * (1.80) *
0.182 -0.136
NIt*FINLITt + (0.34) (-0.25)
1.173 1.166
MEETt ? (1.45) (1.43)
1.182 1.357
NIt*MEETt + ##
(2.26) (2.57) ###
0.483 0.461
NOBUSYt ? (0.56) (0.54)
0.132 0.026
NIt*NOBUSYt + (0.20) (0.04)
-2.263 -2.207 -2.864 -2.765

NOINFt ? (-2.61) *** (-2.53) ** (-3.19) *** (-3.07) ***
1.019 0.720 0.940 0.600
NIt*NOINFt ? (1.64) * (1.13) (1.51) (0.94)
46
TABLE 5: CONTINUED
Panel A: Continued
1.550 1.371 1.320 1.200
CEOSHRt ? (1.60) (1.36) (1.35) (1.17)
0.633 0.473 0.679 0.514

NIt*CEOSHRt ? (1.19) (0.87) (1.26) (0.94)
-3.336 -3.410 -3.238 -3.357

BLOCKt ? (-2.14) ** (-2.19) ** (-1.99) ** (-2.06) **
1.507 1.780 1.670 1.973
NIt*BLOCKt ?
(0.81) (0.96) (0.90) (1.07)
-2.548 -2.484 -2.441 -2.378
BETAt ? (-2.52) ** (-2.46) ** (-2.40) ** (-2.33) **
-0.953 -0.973 -0.908 -0.931
NIt*BETAt ? (-1.37) (-1.40) (-1.30) (-1.34)
0.038 0.039 0.017 0.014

GROWTHt ? (0.46) (0.48) (0.20) (0.17)
0.323 0.313 0.322 0.322

NIt*GROWTHt ? (4.63) *** (4.49) *** (4.53) *** (4.53) ***
-0.119 -0.116 -0.147 -0.150
PRECISIONt ? (-2.00) ** (-1.87) * (-2.26) ** (-2.24) **
-0.113 -0.125 -0.108 -0.116
NIt*PRECISIONt ? (-1.95) * (-2.13) ** (-1.71) * (-1.83) *
Panel B: Tests of Differences in Interaction Terms on Independence and Effectiveness

(NIt*INDt) − 1.548
+ (1.79) ##
(NIt*FINLITj,t)
(NIt*INDt) − 0.056
+
ΜΕΕΤj,t)
(NIt*ΜΕΕΤ (0.10)
(NIt*INDt) − 0.952
+ (1.57) #
(NIt*NOBUSYj,t)
47
TABLE 5: Continued
Notes:
The regression equation is:
Vt − Pt = γ 0 + γ1 OCFt + γ 2 NI t + γ3 IND t + γ 4 NI t * IND t + γ5 FINLITt

+ γ 6 NI t * FINLITt + γ 7 MEETt + γ 8 NI t * MEETt + γ 9 NOBUSYt + γ 10 NI t * NOBUSYt
+ γ11 NOINFt + γ 12 NI t * NOINFt + γ13 CEOSHR t + γ 14 NI t * CEOSHR t + γ15 BLOCK t
+ γ 16 NI t * BLOCK t + γ17 BETA t + γ 18 NI t * BETA t + γ19 GROWTH t + γ 20 NI t * GROWTH t
+ γ 21 PRECISION t + γ 22 NI t * PRECISION t + ξ t
Each cell in Panel A reports the OLS coefficient (or differences in coefficients) and the White’s
(1980) adjusted t-statistic. *, ** and *** signify two-tailed significance at the 0.10, 0.05 and 0.01
level. #, ## and ### signify one-tailed significance at the 0.10, 0.05 and 0.01 level.

Vt = intrinsic value per share per share estimated as of the day after the year t+1
earnings announcement date, standardized by market price set as of the day
after the year t-1 earnings announcement date;
Pt = market price per share as of the day after the year t earnings announcement
date standardized by market price set as of the day after the year t-1
earnings announcement date;
OCFt = operating cash flows per share for year t, calculated from the Statement of
Cash Flows, standardized by market price set as of the day after the year t-
1 earnings announcement date;
NIt = earnings per share before extraordinary items reported for fiscal year t,
standardized by market price set as of the day after the year t-1 earnings
announcement date;
INDt = indicator variable set to one if all audit committee members have no direct
or indirect relations to the firm or its employees, otherwise zero;
FINLITt = indicator variable set to one if all audit committee identified as financially
literate, otherwise zero;
NOBUSYt = indicator variable set to one if all audit committee members that serve on
no more than three boards, otherwise zero;
NOINFt = indicator variable set to one if the CEO does not serve as both the
Chairman of the Board and as a member on the nominating committee,
otherwise zero;
48
TABLE 5: Continued
Notes (continued):
owned by the CEO is in the upper half of the sample distribution,
otherwise zero;
more) serve on audit committee, otherwise zero.
BETAt = market beta calculated using the market model with daily returns for year
t-1;
GROWTHt = analyst forecasted long-term growth from I/B/E/S measured the month
after prior year’s earnings announcement;
PRECISIONt = one divided by the number of analyst forecasters measured the month after
prior year’s earnings announcement.
49

The Influence of Independent and Effective Audit Committees On Earnings Quality

Uploaded by

Copyright:

Available Formats

The Influence of Independent and Effective Audit Committees On Earnings Quality

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

The Influence of Independent and Effective Audit Committees On Earnings Quality

Uploaded by

Copyright:

Available Formats

THE INFLUENCE OF INDEPENDENT AND EFFECTIVE AUDIT

COMMITTEES ON EARNINGS QUALITY

Department of Accounting and Law, School of Management

*Corresponding author: Sam Tiras, stiras@buffalo.edu

We wish to acknowledge the helpful comments of the seminar participants at University of

independence and effectiveness of audit committees.

corporate governance by improving the independence and effectiveness of audit committees.

Levitt expressed his position as follows:

“[Q]ualified, committed, independent and tough-minded audit committees represent the

Effectiveness of Corporate Audit Committees forwarded ten recommendations on improving

awareness of certain attributes expected of audit committee members as particularly important,

specifically, recognition of the significance of the audit committee’s responsibility, time

be more informed (better) decisions by users.

By measuring earnings informativeness as the ERC, we find that audit committees

comprised of independent and financially literate directors improve earnings informativeness.

efficient financial markets (2001).”

market systematically misprices securities by overvaluing the accruals component of reported

reduce this mispricing.

improving the quality of reported earnings.

in section five, and section six concludes our study.

2.1 THE RELATION BETWEEN THE EARNINGS-RETURNS RELATION AND BOARDS

audit committees by the BRC and the Sarbanes-Oxley Act of 2002.

attempts to better characterize the nature of this relation.

2.2 THE INDEPENDENCE OF THE CORPORATE BOARD AND AUDIT COMMITTEES

fraud and earnings restatements.

2.3 THE EFFECTIVENESS OF THE CORPORATE BOARD AND AUDIT COMMITTEES

2.4 OTHER FACTORS POTENTIALLY INFLUENCING AUDIT COMMITTEES

serving on the audit committee.

financial reporting process.

CEO Stock Ownership

and the role of the audit committee in overseeing this process.

be influenced by the presence of large outside blockholders on the committee.

3. Model and Hypothesis Development

whether ERCs are higher and mispricing lower, respectively.

We estimate the ERC (our informativeness metric) by regressing cumulative abnormal

returns on unexpected earnings, as follows:3

1989; Kothari 2001).4

respectively, otherwise zero.

committee is a large outside blockholder (BLOCKt), otherwise zero.6

expectation of future growth (GROWTHt), analyst precision (PRECISIONt), earnings persistence

CAR t = α 0 + α1 UE t + α 2 IND t + α 3 UE t * IND t

(α OTHER j, t + α 11+ 2 j UE t * OTHER j, t )

αi = regression coefficients, i ∈ [0, 26];

direct comparisons across each governance characteristic.

first two hypotheses (in alternative form) below:

H1a: The association of unexpected earnings with cumulative abnormal returns

H2a: The association of unexpected earnings with cumulative abnormal returns

respectively, to identify the influence of each characteristic on earnings informativeness. We

committee characteristics. This priority placed on independence would be confirmed if α3 is

H3a: The association of unexpected earnings with cumulative abnormal returns

3.2 EARNINGS TRANSPARENCY TESTS

induced, conditional expected return (CRSP size-based decile return), as follows:8

Vt = intrinsic value per share;

rt*+1 = CRSP sized-based decile return for year t+1.

βi = regression coefficients, i ∈ [0, 2];

alone (Jennings 1990).10

variable as Vt − Pt, as follows:11

γi = regression coefficients, i ∈ [0, 2];

(our test metric).