Audit Committee, Board Characteristics, and

Auditor Switch Decisions by Andersen’s Clients*

KEN Y. CHEN, National Taiwan University

JIAN ZHOU, SUNY at Binghamton

1. Introduction
On November 14, 2001, Birner Dental Management Services, Inc. dismissed
Arthur Andersen LLP as the Company’s principal accountant. The decision to
change the accountant was recommended by the Audit Committee of the
Board of Directors of the Company and approved by the Board of
Directors. … The Company engaged Hein Associates LLP as its new princi-
pal independent accountant as of November 14, 2001. [From the 8-K filed by
Birner Dental Management Services, Inc. with the Securities and Exchange
Commission (SEC) on November 14, 2001.]

On July 29, 2002, eMagin Corporation was notified by the Securities and
Exchange Commission that Arthur Andersen LLP had notified it that Arthur
Andersen was unable to perform future audit services for eMagin, effectively
terminating eMagin’s relationship with Arthur Andersen. Consequently, as of
July 29, 2002, Arthur Andersen will no longer serve as eMagin’s independent
auditor. eMagin’s Audit committee will recommend a replacement firm to the
Board of Directors as soon as it completes its search and interview process for
a new independent auditor. [From the 8-K filed by eMagin Corporation with
the SEC on July 29, 2002.]

The difference between the 8-Ks filed with the SEC by these former clients of
Arthur Andersen LLP is striking. eMagin Corporation severed its relationship with

* Accepted by Peter Clarkson. We thank two annoymous reviewers, Susan Albring, Jan Barton, Lou
Braiotta, Kevin C. W. Chen, Shijun Cheng, Peter Clarkson (associate editor), Randy Elder, Carol
Ann Frost, Ross Fuerman, Jackie Hammersley, Henry Huang, Ryan LaFond, Bin Ke, Ling Lei,
Shu-Hsing Li, Gerry Lobo, Linda Myers, Kannan Raghunandan, Sara Reiter, Steven Schwartz,
Kevin Sun, Taychang Wang, Weimin Wang, Jay Wellman, Wei Zhang, X. Frank Zhang, Ping
Zhou, Jerry Zimmerman, Jayanthi Krishnan (discussant) and participants at the 2004 American
Accounting Association annual meeting and 2005 European Accounting Association annual
congress, Yi-Tsung Lee (discussant) and participants at the 2004 Accounting Theory and Practice
Conference at National Cheng-Chi University, and workshop participants at National Taiwan Uni-
versity and National Cheng Kung University for their helpful comments. We are also very grateful
to Jan Barton for providing us with the media coverage data. Ken Y. Chen gratefully acknowledges
the financial support of the National Science Council (project no. NSC 93-2416-H-006-045). Part
of the work was done while Jian Zhou was visiting Syracuse University. Jian Zhou gratefully
acknowledges the research support of Syracuse University. We thank Chih-Hui Cheng and Feixue
Yan for their research support.

Contemporary Accounting Research Vol. 24 No. 4 (Winter 2007) pp. 1085–117 © CAAA
doi:10.1506/car.24.4.2
1086 Contemporary Accounting Research

Andersen more than eight months later than Birner Dental Management Services.
In addition, eMagin did not initiate the auditor change. Rather, it terminated its
relationship with Andersen only after receiving a notice from the SEC that Ander-
sen would be unable to provide future audit services to eMagin. These differences
lead one to wonder about the causes of the difference in the dismissal decisions
with respect to the timing and the choice of successor auditors. Given that Ander-
sen’s reputation was clearly tarnished by the Enron scandal and that some of its
clients suffered significant negative stock price reactions around key Andersen-
Enron-related events (Chaney and Philipich 2002), it is surprising that some clients
(such as eMagin) remained with Andersen for more than nine months after Enron
announced the $1.01 billion unexpected charge to its earnings on October 16,
2001, and more than four months after Andersen was indicted for obstruction of
justice on March 14, 2002.
In this paper, we attempt to shed light on the differences in the timing of auditor
dismissal and in the choice of successor auditor made by former Andersen clients
by examining the role of audit committees. We focus on audit committees because
the audit committee has the “ultimate authority and responsibility to select, evalu-
ate, and where appropriate, dismiss the outside auditor” (Blue Ribbon Committee
Report [BRC] 1999, 14). We also examine the role of the board of directors
because the board of directors must approve the auditor change. Prior studies suggest
that audit committee size (e.g., Pincus, Rusbarsky, and Wong 1989), independence
(e.g., Lee, Mande, and Ortman 2004), number of meetings (e.g., Andersen, Mansi,
and Reeb 2004), and financial expertise (e.g., DeFond, Hann, and Hu 2005) are
some characteristics of audit committees that influence their effectiveness. Prior
literature also provides evidence that board size (e.g., Andersen et al. 2004) and
board independence (e.g., Beasley and Petroni 2001) are some of the board
attributes that influence their effectiveness. We therefore investigate the relation-
ship between audit committee characteristics, board characteristics, and auditor
dismissal decisions (i.e., timing and successor auditor choice) by Andersen clients.
We posit that more effective audit committees and more effective boards were
more likely to carefully monitor the Andersen situation and respond quickly to any
perceived threat to Andersen’s credibility. More specifically, we hypothesize that
clients with larger, more independent, and more active audit committees, and with
audit committees with greater financial expertise were more likely to dismiss
Andersen earlier and to hire a Big 4 successor auditor. We also expect that clients
with larger and more independent boards were more likely to dismiss Andersen
earlier and to choose a Big 4 successor auditor.
Using hand-collected data on the audit committee and board characteristics of
821 clients that dismissed Andersen during the sample period between October 15,
2001 and August 31, 2002, we find that clients with more independent audit com-
mittees and audit committees with greater financial expertise, as well as firms with
larger boards, dismissed Andersen sooner.1 Moreover, clients with larger and more
active audit committees were more likely to choose a Big 4 successor auditor.
Finally, we find that clients with more independent boards dismissed Andersen
sooner and were more likely to choose a Big 4 successor auditor. In sum, our

CAR Vol. 24 No. 4 (Winter 2007)

Audit Committee, Board Characteristics, and Auditor Switch Decisions 1087

results suggest that firms with more effective audit committees and boards of
directors responded quickly to the Andersen-Enron situation and were more likely
to demand higher reputation auditors.
Our study makes the following contributions to literature. We contribute to the
literature that investigates Andersen clients’ auditor choice by focusing on the roles
of the audit committees and boards of directors. Chang, Chi, and Liu (2003) find
that clients with higher sales growth rates and higher probability of financial dis-
tress delayed their dismissal of Andersen, while Barton (2005) finds that clients
more visible in the capital markets dismissed Andersen sooner and subsequently
engaged mainly Big 4 successor auditors. We focus on the impact of audit commit-
tees and boards of directors on the auditor switch decisions of Andersen clients
because the audit committee is responsible for auditor change decisions and
because the board must approve any change in auditor. We provide some evidence,
albeit mixed, that audit committee and board characteristics influenced Andersen
clients’ auditor selection decisions.
This study also resonates with recent changes in the regulation of audit com-
mittees. For example, section 301 of the Sarbanes-Oxley Act (SOX 2002) requires
audit committee members to be independent of the company’s management,2
while section 407 has requirements on audit committee financial expertise. We
provide support for these SOX requirements on audit committee independence and
financial expertise by showing that such audit committee characteristics were sig-
nificantly related to the timing of Andersen’s dismissal.
The remainder of this paper is structured as follows. Section 2 describes the
demand for auditor reputation and develops the hypotheses. Section 3 discusses
the sample selection and research design. Section 4 discusses the empirical results,
and section 5 concludes.

2. Demand for auditor reputation and hypotheses development

Demand for auditor reputation
Only a few studies have examined the consequences of a decline in an auditor’s
reputation. For example, Wilson and Grimlund (1990) examine the effect of disci-
plinary actions by the SEC and find that audit firms involved in disciplinary actions
tend to have a significant negative market reaction by losing market share relative
to their competitors and tend to be less able to retain clients than other certified
public accountant (CPA) firms. Reed, Trombley, and Dhaliwal (2000) find that
Laventhol and Horwath (LH) clients who were more highly leveraged, had less
management ownership, and issued more securities in the year after selecting the
new auditor were more likely to choose Big 6 auditors when LH declared bank-
ruptcy in November 1990. Chaney and Philipich (2002) find that Andersen’s cli-
ents experienced a statistically negative market reaction during the three days
following Andersen’s admission that it had shredded documents, while Krishna-
murthy, Zhou, and Zhou (2006) find that Andersen clients experienced negative
market reactions around the indictment date, suggesting that investors downgraded
the quality of the audits performed by Andersen.

CAR Vol. 24 No. 4 (Winter 2007)

1088 Contemporary Accounting Research

Barton (2005) finds that clients dismissing Andersen sooner and choosing
another Big 5 auditor were the companies with greater analyst and press coverage,
larger institutional ownership and share turnover, and greater amounts of cash
raised in recent security issues. Chang, Chi, and Liu (2003) find that clients with
higher sales growth and at greater risk of financial distress delayed their auditor-
switching announcements.

Audit committees and auditor switch decisions by Andersen’s clients

The audit committee plays an important monitoring role in assuring the quality of
financial reporting and corporate accountability. As a liaison between the external
auditor and the board, the audit committee bridges the information asymmetry
between them, facilitates the monitoring process (Klein 1998), and enhances the
independence of the auditor from the management (Mautz and Neumann 1977).
Thus, a properly functioning audit committee is critical in enhancing the effective
oversight of the financial reporting process and ensuring high-quality financial
reporting. Lennox and Park (2007) claim that the audit committee is the most
important governance mechanism with respect to audit firm appointments because
the audit committee is responsible for hiring the external auditor and overseeing
audit quality.
After the SEC became concerned that audit committees were failing to do
their jobs properly (Lublin and MacDonald 1999), the Blue Ribbon Committee
(1999) issued recommendations addressing the importance of audit committee
composition and operational characteristics such as size, independence, and financial
literacy.3 More specifically, the BRC report recommended that the audit commit-
tees of listed companies comprise at least three directors. This recommendation
reflects the assumption that size is a very important attribute of audit committees.
Kalbers and Fogarty (1993) propose that audit committee effectiveness is per-
ceived as a function of audit committee power. Larger audit committees with
increased organizational status and power delegated by boards of directors are thus
more likely to be recognized as authoritative bodies by management, external audi-
tors, and internal auditors. Pincus et al. (1989) suggest that audit committees are an
expensive monitoring mechanism and that firms with larger audit committees are
willing to spend more resources on this important mechanism and are likely to
improve the function of audit committees. Anderson et al. (2004) find that the
larger the audit committee, the lower the cost of debt financing. We thus expect
that a larger audit committee is likely to be a more effective mechanism, because
members in a larger audit committee have increased responsibilities and organiza-
tional status and power for their monitoring roles. Larger audit committees are also
more likely to care about auditor reputation, so we propose the following hypothesis:

HYPOTHESIS 1. Clients with larger audit committees were more likely to dismiss
Andersen earlier and hire a Big 4 auditor.

Jemison and Oakley (1983, 519) argue that an active audit committee composed
entirely of outside directors is a key element of effective corporate governance.

CAR Vol. 24 No. 4 (Winter 2007)

Audit Committee, Board Characteristics, and Auditor Switch Decisions 1089

Carcello and Neal (2003) find that audit committees with greater independence are
less likely to dismiss the audit firm following the issuance of new going-concern
reports. Abbott, Parker, and Peters (2004) document a significant negative associa-
tion between audit committee independence and financial restatements. Klein
(2002a) finds that there is a negative relation between audit committee indepen-
dence and abnormal accruals. She also finds that reductions in audit committee
independence are accompanied by large increases in abnormal accruals. These
studies suggest that audit committee independence serves to mitigate potential
sources of manager – shareholder conflicts such as opinion shopping, financial
restatements, and earnings management.
With regard to Andersen’s dismissal, managers were likely to delay the deci-
sion to switch auditors for the following reasons: (a) Although Andersen’s reputation
was clearly damaged by the Enron audit, it was possible that Andersen would sur-
vive the indictment.4 Given the uncertainty about Andersen’s indictment and the
difficulty of the transition to a new auditor (U.S. General Accounting Office
[GAO] 2003),5 managers may have wanted to wait until they received more infor-
mation concerning Andersen before deciding to dismiss it. (b) Managers may not
have liked the close scrutiny from Andersen’s successor auditors, because auditor
conservatism may have been prevalent in the post-Enron era. In fact, Cahan and
Zhang (2006) find that ex-Andersen clients had lower levels of abnormal accruals
and more negative decreases in abnormal accruals in 2002. This is consistent with
auditor conservatism and suggests the successor auditors viewed a former Ander-
sen client as a unique source of litigation risk. (c) Managers may have already built
a comfortable relationship with the audit partners from Andersen. Their loyalty to
Andersen’s partners may have prevented them from dismissing Andersen in a
timely fashion. Whittington and Pany (2006, 20) observe that “many clients select
a particular accounting firm because they have come to know and respect one of
the firm’s partners”.
On the other hand, independent audit committee members had the following
incentives to push for a new auditor: (a) Shareholders lost value as a result of
Andersen-Enron events (e.g., Chaney and Philipich 2002; Krishnamurthy et al.
2006).6 Independent directors, acting in shareholders’ interests, may have pushed
for a decisive move to dismiss Andersen. (b) Independent audit committee mem-
bers are not as concerned as the management about scrutiny by new auditors, as
documented in Cahan and Zhang 2006. Even though managers have had incentives
to delay Andersen’s dismissal to avoid the scrutiny of new auditors, independent
directors may not have shared the same concern. (c) Lee et al. (2004) find that
more independent audit committees demand higher auditor reputation. Even
though managers may have wanted to remain with Andersen if it survived, inde-
pendent directors were more likely to seek the dismissal of Andersen because they
demanded higher auditor reputation.
On the basis of these rationales, we propose the following hypothesis:

HYPOTHESIS 2. Clients with more independent audit committees were more

likely to dismiss Andersen earlier and hire a Big 4 auditor.

CAR Vol. 24 No. 4 (Winter 2007)

1090 Contemporary Accounting Research

An active audit committee is more likely to influence management or board

decisions. McMullen and Raghunandan (1996) find that the audit committees of
firms with SEC enforcement actions or earnings restatements are less likely to
have frequent meetings. Abbott and Parker (2000) find that audit committees that
are independent and meet more than twice a year are associated with the selection
of industry-specialist auditors. However, Lee et al. (2004) find no significant rela-
tion between the number of audit committee meetings and both auditor resignation
and the selection of a high-quality successor auditor. Hymowitz and Lublin (2003)
report that “many audit committees are spending far more time than they used to
reviewing financial statements and overseeing auditors, meeting 10 or 11 times a
year, up from three or four times”. Lennox (2002) finds that there is a significant
increase in audit committee meetings during the year of auditor dismissal. Ander-
son et al. (2004) find that the cost of debt financing is significantly negatively
related to the number of audit committee meetings. These studies suggest that
audit committee meetings are related to audit committee effectiveness. Therefore,
we propose the following hypothesis:

HYPOTHESIS 3. Clients with more active audit committees were more likely to
dismiss Andersen earlier and hire a Big 4 auditor.

The BRC recommends that every audit committee have at least one financial
expert.7 Audit committee members with financial expertise can perform their
oversight roles in the financial reporting process more effectively, particularly in
exercising internal controls or detecting material misstatements (Scarbrough,
Rama, and Raghunandan 1998; Raghunandan, Read, and Rama 2001; Krishnan
2005). Abbott et al. (2004) find a significantly negative association between an
audit committee having at least one member with financial expertise and the inci-
dence of financial restatement. Bédard, Chtourou, and Courtean (2004) find that
the presence of at least one financial expert on the audit committee is associated
with a lower likelihood of aggressive earnings management. DeFond et al. (2005)
find significantly positive abnormal returns around the appointment of accounting
financial experts to the audit committees, suggesting that audit committee mem-
bers with accounting financial expertise improve the audit committee’s ability to
ensure high-quality financial reporting. These studies suggest that financial experts
on audit committees may have the ability to improve the quality of financial report-
ing, an outcome consistent with the argument that financial expertise is essential to
ensure that audit committees fulfill their primary responsibilities of overseeing the
financial reporting process and enhancing financial reporting quality (SEC 2003;
PricewaterhouseCoopers 1999).
Because financial experts have invested a significant amount of effort in
developing their financial expertise, they have a strong incentive to maintain their
reputation in performing their monitoring role as audit committee members. This
is consistent with observations in prior studies that directors’ concerns about their
reputation in the external labor market causes them to be effective monitors (e.g.,
Fama and Jensen 1983). Andersen’s reputation was clearly damaged as a result of

CAR Vol. 24 No. 4 (Winter 2007)

Audit Committee, Board Characteristics, and Auditor Switch Decisions 1091

the Enron audit and other audit failures such as Baptist Foundation of Arizona,
Sunbeam, and Waste Management.8 One way of maintaining their reputation as a
financial expert was to dismiss the reputation-tarnished Andersen earlier and
choose a Big 4 firm as the successor auditor. The nonfinancial experts on the audit
committee were likely to rely on the financial experts’ opinion about Andersen.
Financial experts were also likely to be more sensitive to the value loss around
Andersen’s major events (e.g., Chaney and Philipich 2002; Krishnamurthy et al.
2006). Identifying the eligible successor auditors, reviewing potential auditors’
proposals, and selecting a new auditor are time-consuming and costly activities
(GAO 2003). Financial experts on the audit committee are likely to contribute to
the process because they are likely to have better knowledge about the candidate
audit firms and are able to quickly identify a suitable new audit firm, thus helping
to smooth the transition to a new auditor.
Given their expertise in financial reporting and reputational concerns in the
external labor market, audit committees with greater financial expertise were more
likely to dismiss Andersen sooner and hire a Big 4 successor auditor:

HYPOTHESIS 4. Clients with greater financial expertise on the audit committee

were more likely to dismiss Andersen earlier and hire a Big 4 auditor.

Boards of directors and auditor switch decisions by Andersen’s clients

Agency theory suggests that a separation of ownership and control leads to a diver-
gence in managerial and owners’ interests (Jensen and Meckling 1976). The board,
the common apex of the decision control system in public corporations, is a market-
induced, low-cost mechanism for monitoring management (Fama 1980; Fama and
Jensen 1983). Shareholders delegate their decision control rights to boards as a
more efficient way of ratifying and monitoring managerial decisions and, thus,
monitoring managerial decisions becomes essential for a board of directors to
ensure that shareholders’ interests are protected (Fama and Jensen 1983).
Prior research in finance has shown a negative relation between board size and
financial performance (e.g., Yermack 1996; Eisenberg, Sundgren, and Wells 1998).
This suggests that smaller boards are more effective monitors than large boards,
because free-riding problems among directors increase with board size (Jensen
1993).
However, Zahra and Pearce (1989, 309) argue that larger boards are more
effective monitors than small boards because they are not susceptible to manage-
rial domination, are more likely to be heterogeneous in members’ educational and
technical backgrounds, and are more likely to resist managerial domination to pro-
tect shareholders’ interests. Moreover, Chaganti, Mahajan, and Sharma (1985,
405) argue that larger boards may be valuable for the breadth of their services.
Monks and Minow (1995) argue that larger boards can commit more time and
effort to overseeing management, because larger boards from firms with several
committees allow for fewer committee assignments per director.
Furthermore, Anderson et al. (2004) find that board size is negatively associ-
ated with the cost of debt, suggesting that larger boards provide better monitoring

CAR Vol. 24 No. 4 (Winter 2007)

1092 Contemporary Accounting Research

of the financial reporting process. Although the audit committee is responsible for
the selection, evaluation, and dismissal of the auditor (e.g., BRC 1999), such deci-
sions need the approval of the board of directors. The board of directors may not
fully delegate the auditor selection decision to the audit committee, and the board
of directors has an overriding authority on the audit committee to make the final
decision. Because the board of directors is important in a company’s auditor choice
decision (Jensen 1982; Leddy 1982) and larger boards are more effective monitors
of financial reporting quality, we propose the following hypothesis:

HYPOTHESIS 5. Clients with larger boards were more likely to dismiss Andersen
earlier and hire a Big 4 auditor.

Outside or independent directors perform an important monitoring function in

public corporations because they have greater incentives than inside directors to be
effective monitors of management in order to preserve their reputational capital
(Fama and Jensen 1983), and thus they could mitigate agency conflicts between
shareholders and management. Moreover, outside directors are more vigilant than
inside directors because they focus on financial performance, which is a central
component of monitoring (Fama and Jensen 1983; Johnson, Hoskisson, and Hitt
1993). Independent directors are more likely to draw on their broader experience
and expertise in management oversight and to perform better as board members
(Kosnik 1987).
Furthermore, Beasley (1996) and Dechow, Sloan, and Sweeney (1996) find
that the proportion of independent directors on boards is negatively associated with
the likelihood of financial statement fraud, suggesting that independent directors
enhance a board’s ability to properly execute its oversight function. Anderson et al.
(2004) find that firms with large independent boards are associated with a lower
cost of debt financing, thus confirming that board independence is an important
element of the financial reporting process. Beasley and Petroni (2001) find that
boards with a higher percentage of outside directors are more likely to select a spe-
cialist Big 6 auditor. Accordingly, we propose the following hypothesis:

HYPOTHESIS 6. Clients with more independent boards were more likely to

dismiss Andersen earlier and hire a Big 4 auditor.

3. Sample selection and research design

Sample selection
As in Barton 2005, our sample period begins on October 15, 2001, the day before
Enron announced its unexpected $1.01 billion charge to earnings, and ends on
August 31, 2002, the day Andersen ceased auditing public companies. Table 1 pro-
vides details about the sample selection process. We begin with all firms listed on
COMPUSTAT as Andersen clients in 2001. COMPUSTAT lists 1,269 firms as
Andersen clients with auditor code 1. We exclude 115 firms that did not have audi-
tor change information or dismissed Andersen before October 15, 2001.9 We then
eliminate 199 firms that do not have any information about their audit committees.

CAR Vol. 24 No. 4 (Winter 2007)

Audit Committee, Board Characteristics, and Auditor Switch Decisions 1093

After deleting 3 firms without audit committee meeting information, we eliminate

89 firms that do not have institutional ownership information. We also exclude 8
firms without information on insider ownership and 25 firms without information
for computing their market value. Finally, we delete 9 firms that experienced
Andersen resignation. This leaves us with a final sample of 821 firms. The auditor
dismissal date is obtained from 8-Ks filed with the SEC. The audit committee and
board data are manually collected from 10-K reports or proxy statements filed with
the SEC. The institutional ownership data are collected from the Compact Disclo-
sure Database.

Models
We use the Cox proportional hazard model to test our hypotheses of the association of
audit committee and board characteristics with the timing of Andersen’s dismissal.

hi(t) h0(t)exp(1AC 2ACIND 3ACMTG 4ACFE 5BD

6BDIND 7ARTICLES 8ANALYSTS 9INSTI
10 NEW_FIN 11MGT_OWN 12LEVER 13LGMKV
14 MTB 15FYR ) (1),

TABLE 1
Sample selection criteria

Number
of firms
Andersen clients in 2001 from COMPUSTAT 1,269
Exclude firms without auditor change information or switched auditors
before October 15, 2001 (115)
1,154
Exclude firms without any information about audit committee (199)
955
Exclude firms without audit committee meeting information (3)
952
Exclude firms without information about institutional ownership (89)
863
Exclude firms without insider ownership information (8)
855
Exclude firms without information to calculate market value (25)
830
Exclude firms that experienced Andersen resignation (9)
821

CAR Vol. 24 No. 4 (Winter 2007)

1094 Contemporary Accounting Research

where

AC the number of audit committee members;

ACIND the percentage of outside directors on an audit committee;10
ACMTG the number of audit committee meetings in the auditor change
year;
ACFE the percentage of audit committee members who are accounting
financial experts or nonaccounting financial experts, as defined in
DeFond et al. 2005;11
BD the number of board members;
BDIND the percentage of outside directors on the board;
ARTICLES the number of articles published about the firm during the year
ending on October 15, 2001, on the basis of the Factiva data base
of major U.S. news and business publications; to be counted, an
article must include the firm name in the headline or lead paragraph
and cannot be a republished article or a recurring pricing/market
data news item;
ANALYSTS the number of analysts included in the most recent consensus fore-
cast of annual earnings in the I / B / E / S data base before October
2001;
INSTI the percentage of common shares owned by institutional share-
holders;
NEW_FIN if the firm has positive net external financing in 2001 and 2002,
and 0 otherwise; net external financing is calculated as in Richard-
son and Sloan (2003, 50);
MGT_OWN the percentage of common shares owned beneficially by managers
and directors;
LEVER the ratio of long-term debt over total assets;
LGMKV the natural log of market value;
MTB the market to book ratio;
FYR 1 if a firm’s fiscal year ends on December 31, and 0 otherwise.

The dependent variable, hi(t), is the hazard rate for firm i at time t (that is, the
probability of dismissing Andersen at time t, conditional on the firm’s not having
dismissed Andersen earlier); h0 (t) is the “baseline” hazard rate shared by all
Andersen clients at time t. A large hazard rate indicates that a firm is expected to
dismiss Andersen earlier. AC, ACIND, ACMTG, and ACFE capture audit commit-
tee characteristics, while BD and BDIND capture board characteristics. Similar to

CAR Vol. 24 No. 4 (Winter 2007)

Audit Committee, Board Characteristics, and Auditor Switch Decisions 1095

Barton 2005, we control for the impact of a firm’s visibility in the capital markets
(that is, articles, analysts, institutional ownership, and new financing) and agency
conflicts (that is, managerial ownership and leverage) on firms’ decision to dismiss
Andersen. Finally, as discussed below, we control for other factors that are likely to
be associated with the timing of Andersen’s dismissal.
We also run the following logit regression model to test the relationship
between audit committee and board characteristics and the successor auditor
choices by Andersen’s clients:

NEWAUD 1AC 2ACIND 3ACMTG 4ACFE 5BD

6 BDIND 7ARTICLES 8ANALYSTS 9INSTI
10 NEW_FIN 11MGT_OWN 12LEVER 13LGMKV
14MTB 15FYR (2),

where

NEWAUD 1 if the successor auditor is a Big 4 audit firm, and 0 otherwise.

Test variables
Hypotheses 1 to 4 predict that firms with larger, more independent, more active
audit committees, as well as audit committees with greater financial expertise, dis-
missed Andersen earlier and were more likely to replace Andersen with a Big 4
auditor. Therefore, we expect the coefficients on audit committee size (AC), inde-
pendence (ACIND), number of meetings (ACMTG), and financial expertise
(ACFE) to be positively associated with the probability of early dismissal and the
choice of a Big 4 auditor. Hypotheses 5 and 6 predict that firms with larger and
more independent boards were more likely to dismiss Andersen earlier and replace
Andersen with a Big 4 auditor, so we also expect the coefficients on board size
(BD) and board independence (BDIND) to be positive in both models.

Control variables
The control variables used in this study are divided into two categories. Variables
in the first category are similar to the client visibility variables used in Barton
2005: the level of media coverage (ARTICLES), share turnover (TURNOVER),12
analyst following (ANALYSTS), institutional ownership (INSTI), and new financing
(NEW_FIN). Barton (2005) finds that firms more visible in the capital markets dis-
missed Andersen earlier, usually to choose a Big 4 auditor and, therefore, we
expect the coefficients on these variables to be positive. We also use managerial
ownership (MGT_OWN) and leverage (LEVER) to control for potential agency
conflicts. Prior studies hypothesize that managers with small ownership stakes
have weaker incentives to act in the best interest of shareholders, a tendency that
increases the demand for higher-quality monitoring by an independent auditor
(Francis and Wilson 1988; DeFond 1992; Lennox 2005). However, the relationship
between management ownership and agency cost is not linear due to the possible
entrenchment effect associated with management ownership (Lennox 2005).

CAR Vol. 24 No. 4 (Winter 2007)

1096 Contemporary Accounting Research

Therefore, we make no prediction as to the coefficient on MGT_OWN.13 We expect

the coefficient on LEVER to be positive, because debt contracts include restrictive
covenants, thus creating a demand for higher-quality auditing to monitor compliance
with such covenants (Jensen and Meckling 1976; Watts and Zimmerman 1986).
Variables in the second category control for other potential factors likely to be
associated with the timing of Andersen’s dismissal and choice of successor auditor.14
Barton (2005) and Chang et al. (2003) find that larger firms dismissed Andersen
earlier and were more likely to choose a Big 4 successor. We use the natural loga-
rithm of market value (LGMKV) to control for the effect of firm size, and expect
the coefficient on LGMKV to be positive.15 Barton (2005) also finds that firms with
higher market-to-book ratios (MTB) dismissed Andersen sooner. We include the
market-to-book ratio in our model and expect the coefficient on MTB to be posi-
tive. It is costly for a firm to switch its auditor in the middle of the annual audit.
Firms that had not filed their 10-Ks were likely to incur the highest cost of an auditor
switch, especially if Andersen had already started the 10-K audit. These firms were
more likely to have December fiscal year-ends. Therefore, we expect that firms
with a December fiscal year-end (FYR) were more likely to dismiss Andersen later.

4. Results
Descriptive statistics
Table 2 reports the descriptive statistics for all the variables used in this study.
DURATION measures the number of days from October 15, 2001 to the auditor
dismissal date reported in the 8-Ks. Andersen clients, on average, took 223 days to
dismiss Andersen after October 15, 2001. In our final sample, 89 percent of Ander-
sen clients chose a Big 4 firm as the successor auditor. The mean and median
numbers of audit committee members were 3.34 and 3, respectively. Eighty-four
percent of audit committee members were independent. Audit committees met
approximately four times every year. About 57 percent of the audit committee
members were financial experts. The mean and median numbers of board members
were 7.77 and 7, respectively. About 73 percent of board members were independent.
On average, Andersen clients had 19 articles published about them in major
U.S. business publications during the year preceding October 15, 2001. The mean
number of analysts covering a client was 4.75. The average share turnover was
0.48. Institutional shareholders owned 42.35 percent of these firms. About 12 per-
cent of Andersen clients had positive net external financing in 2001 and 2002.
Management and directors held 17.04 percent of the shares of these firms. The
average ratio of long-term debt to assets was 0.20. The mean log market value was
5.25, while the mean market value was $1,678.45 million. The median three-day
cumulative abnormal return around Andersen’s document-shredding disclosure
date was 1 percent. The mean market-to-book ratio was 2.79, indicating the
existence of a high level of unrecorded intangible assets. Of the sample firms, 69
percent had a December 31 fiscal year-end.
Table 3 presents Pearson correlations between the dependent and independent
variables. The correlations support several hypotheses; however, the formal tests

CAR Vol. 24 No. 4 (Winter 2007)

Audit Committee, Board Characteristics, and Auditor Switch Decisions 1097

are based on multivariate regression analysis.16 The significantly negative relation-

ship between DURATION and BIG4 shows that firms that dismissed Andersen
later were less likely to choose a Big 4 successor auditor, consistent with the find-
ings in Chang et al. 2003. The significantly positive correlation between board size
(BD) and firm size (LGMKV) is 0.50, and the correlation between board size (BD)
and audit committee size (AC) is 0.48, which are somewhat higher than the corre-
sponding values of 0.37 and 0.39 in Anderson et al. 2004. The correlations
between board size (BD) and audit committee size (AC), independence (ACIND),
and meetings (ACMTG) are significantly positive, consistent with the finding in
Klein (2002b) that larger boards are associated with more effective audit committees.

Regression results
Audit committee, board characteristics, and the timing of
auditor switch by Andersen’s clients
Table 4 reports the regression results using the Cox proportional hazard model in
(1). Model 1 includes only the audit committee variables for the full sample (821

TABLE 2
Descriptive statistics for dependent and independent variables
(n 821 unless otherwise specified)

Standard Lower Upper

Mean deviation quartile Median quartile

DURATION 222.69 39.29 199.00 221.00 249.00

BIG4 0.89 0.32 1.00 1.00 1.00
AC 3.34 0.81 3.00 3.00 4.00
ACIND 0.84 0.30 0.75 1.00 1.00
ACMTG 3.92 1.81 3.00 4.00 5.00
ACFE 0.57 0.31 0.33 0.67 0.75
BD 7.77 2.39 6.00 7.00 9.00
BDIND 0.73 0.14 0.67 0.75 0.83
ARTICLES 19.06 22.81 8.00 12.00 21.00
ANALYSTS 4.75 5.46 1.00 2.00 7.00
TURNOVER 0.48 0.26 0.26 0.45 0.69
INSTI 42.35 29.13 16.72 40.07 66.71
NEW_FIN 0.12 0.33 0.00 0.00 0.00
MGT_OWN 17.04 21.33 1.69 8.47 25.88
LEVER 0.20 0.22 0.00 0.15 0.33
LGMKV 5.25 2.14 3.75 5.26 6.68
MKV 1,678.45 7,284.11 42.65 191.77 797.55
SHRED 0.004 0.07 0.03 0.01 0.02
MTB 2.79 4.86 0.89 1.67 3.36
FYR 0.69 0.46 0.00 1.00 1.00

(The table is continued on the next page.)

CAR Vol. 24 No. 4 (Winter 2007)

1098 Contemporary Accounting Research

TABLE 2 (Continued)

Notes:
DURATION the number of days spanned from October 15, 2001 to the Andersen
dismissal date from 8-Ks filed to the SEC.
BIG4 1 if the successor auditor is a Big 4 audit firm, 0 otherwise.
AC the number of audit committee members.
ACIND the percentage of outside directors on an audit committee.
ACMTG the number of audit committee meetings in the dismissal year.
ACFE the percentage of audit committee members who are accounting financial
experts or nonaccounting financial experts, as defined in DeFond, Hann,
and Hu 2005.
BD the number of board members.
BDIND the percentage of outside directors on the board.
ARTICLES the number of articles published about the firm during the year ending on
October 15, 2001, based on the Factiva data base of major U.S. news and
business publications; to be counted, an article must include the firm name
in the headline or lead paragraph and cannot be a republished article or a
recurring pricing/market data news item.
TURNOVER share turnover over the 150-day period ending on October 15, 2001 (754
observations).
ANALYSTS the number of analysts included in the most recent consensus forecast of
annual earnings in the I/B/E/S data base before October 2001.
INSTI the percentage of common shares owned by institutional shareholders.
NEW_FIN if the firm has positive net external financing in 2001 and 2002, and 0
otherwise; net external financing is calculated as in Richardson and Sloan
(2003, 50).
MGT_OWN the percentage of common shares owned beneficially by managers and
directors.
LEVER the ratio of long-term debt to total assets.
LGMKV the natural log of market value.
MKV market value, in millions of dollars.
SHRED cumulative abnormal return for day (1, 1) around January 10, 2002,
Andersen’s document-shredding disclosure date (750 observations).
MTB the market-to-book ratio.
FYR 1 if a firm’s fiscal year ends on December 31, and 0 otherwise.

CAR Vol. 24 No. 4 (Winter 2007)

TABLE 3
Correlation matrix for dependent and independent variables (p-value in parenthesis, n 821)

BIG4
AC
ACIND
ACMTG
ACFE
BD
BDIND
ARTICLES
ANALYSTS
INSTI
NEW_FIN
MGT_OWN
LEVER
LGMKV
MTB
FYR

DURATION 0.13 0.18 0.14 0.09 0.05 0.28 0.12 0.24 0.30 0.27 0.07 0.09 0.02 0.36 0.06 0.05
(0.00) (0.00) (0.00) (0.01) (0.13) (0.00) (0.00) (0.00) (0.00) (0.00) (0.03) (0.01) (0.66) (0.00) (0.09) (0.14)
BIG4 0.18 0.06 0.17 0.01 0.18 0.12 0.14 0.23 0.35 0.02 0.10 0.04 0.42 0.05 0.08
(0.00) (0.11) (0.00) (0.75) (0.00) (0.00) (0.00) (0.00) (0.00) (0.58) (0.01) (0.26) (0.00) (0.19) (0.02)
AC 0.08 0.08 0.03 0.48 0.23 0.34 0.30 0.21 0.11 0.21 0.10 0.38 0.00 0.06
(0.02) (0.02) (0.38) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.94) (0.07)
ACIND 0.04 0.05 0.13 0.13 0.08 0.08 0.12 0.01 0.04 0.06 0.15 0.03 0.00
(0.28) (0.20) (0.00) (0.00) (0.02) (0.02) (0.00) (0.76) (0.24) (0.07) (0.00) (0.40) (0.96)
ACMTG 0.05 0.15 0.09 0.15 0.20 0.16 0.03 0.05 0.00 0.17 0.01 0.13
(0.18) (0.00) (0.01) (0.00) (0.00) (0.00) (0.36) (0.20) (0.98) (0.00) (0.69) (0.00)
ACFE 0.05 0.09 0.00 0.00 0.01 0.02 0.02 0.02 0.02 0.00 0.08
(0.14) (0.01) (0.99) (0.96) (0.85) (0.58) (0.51) (0.67) (0.68) (0.97) (0.01)
BD 0.16 0.37 0.38 0.28 0.04 0.11 0.20 0.50 0.02 0.08
(0.00) (0.00) (0.00) (0.00) (0.20) (0.00) (0.00) (0.00) (0.58) (0.02)
BDIND 0.13 0.12 0.16 0.01 0.07 0.11 0.16 0.03 0.07
(0.00) (0.00) (0.00) (0.83) (0.05) (0.00) (0.00) (0.38) (0.05)
ARTICLES 0.55 0.25 0.06 0.15 0.08 0.53 0.05 0.05
(0.00) (0.00) (0.08) (0.00) (0.03) (0.00) (0.17) (0.12)
Audit Committee, Board Characteristics, and Auditor Switch Decisions

ANALYSTS 0.50 0.07 0.20 0.13 0.72 0.06 0.11

(0.00) (0.06) (0.00) (0.00) (0.00) (0.07) (0.00)

(The table is continued on the next page.)

CAR Vol. 24 No. 4 (Winter 2007)

1099
TABLE 3 (Continued)
1100

BIG4
AC
ACIND
ACMTG
ACFE
BD
BDIND
ARTICLES
ANALYSTS
INSTI
NEW_FIN
MGT_OWN
LEVER
LGMKV
MTB
FYR

INSTI 0.02 0.24 0.13 0.69 0.02 0.03

(0.65) (0.00) (0.00) (0.00) (0.43) (0.44)
NEW_FIN 0.05 0.08 0.08 0.01 0.03
(0.16) (0.02) (0.02) (0.71) (0.40)

CAR Vol. 24 No. 4 (Winter 2007)

MGT_OWN 0.08 0.18 0.01 0.05
(0.02) (0.00) (0.80) (0.12)
LEVER 0.10 0.13 0.09
(0.00) (0.00) (0.01)
LGMKV 0.16 0.09
(0.00) (0.01)
Contemporary Accounting Research

MTB 0.01
(0.78)

Note:
The variables are as defined in Table 2.
Audit Committee, Board Characteristics, and Auditor Switch Decisions 1101

firms), whereas model 2 includes board variables (BD and BDIND) to examine
their effects on the timing of auditor switch by Andersen’s clients. Model 3 adds
two additional control variables, share turnover (TURNOVER) and stock market
reaction measured around the document-shredding disclosure date January 10,
2002 (SHRED) for the reduced sample (737 firms).
The significantly positive coefficients on AC, ACIND, and ACFE in model 1
suggest that firms with larger and more independent audit committees, as well as
audit committees with greater financial expertise, dismissed Andersen sooner.
When board size and board independence are added in model 2, the coefficients on
ACIND and ACFE remain significantly positive. The significantly positive coeffi-
cients on BD and BDIND suggest that firms with larger and more independent
boards dismissed Andersen earlier. When TURNOVER and SHRED are added in
model 3, the results are qualitatively similar to those in model 2, except for board
independence, which is no longer significant.
In sum, firms with more independent audit committees and with audit commit-
tees with greater financial expertise demand better auditor reputation, as do firms
with larger boards. There is also some evidence that firms with more independent
boards demand better auditor reputation.
Consistent with Barton 2005, we find that firms with more media coverage
(ARTICLES), more analysts following (ANALYSTS), and higher institutional hold-
ings (INSTI) were more likely to dismiss Andersen earlier.17 This shows that clients
more visible in the capital markets dismissed Andersen earlier. In addition, the
coefficients on NEW_FIN and LGMKV are significantly positive in all three models,
suggesting that firms that engaged in external financing as well as larger firms dis-
missed Andersen earlier. We also find that firms with December fiscal year-ends
were less likely to dismiss Andersen earlier, probably because these firms did not
want to switch auditors in the middle of the annual audit.
Among the independent variables in (1), LGMKV consistently has the strongest
effects on the rate of dismissal in the full sample: a one standard deviation increase
in the natural logarithm of market value increases the rate of dismissal by 32.08
percent in model 2.

Audit committee, board characteristics, and the choice of successor auditors

Table 5 provides the logit regression results using the auditor choice model in (2).
Model 1 shows that audit committee size (AC) and audit committee meetings
(ACMTG) are significantly positively related to the choice of a Big 4 successor
auditor. The coefficients on AC and ACMTG remain significantly positive when we
introduce BD and BDIND in model 2 and when we further introduce TURNOVER
and SHRED in model 3. BDIND is consistently significant when it is included as a
variable of interest, demonstrating that firms with more independent boards were
more likely to choose a Big 4 successor auditor. Overall, the results from these
three models in Table 5 suggest that firms with larger and more active audit com-
mittees, as well as firms with more independent boards, demand better auditor
reputation.

CAR Vol. 24 No. 4 (Winter 2007)

TABLE 4
Regression results for Cox proportional hazard model of client defections
1102

Model 1 Model 2 Model 3

Expected Coefficient % change Coefficient % change Coefficient % change
Variables sign (p-value) in hazard (p-value) in hazard (p-value) in hazard
AC 0.09 0.02 0.05
(0.03)† 7.56 (0.36) 1.63 (0.17) 4.13
ACIND 0.28 0.24 0.20
(0.01)* 8.76 (0.02)† 7.47 (0.05)† 6.18

CAR Vol. 24 No. 4 (Winter 2007)

ACMTG 0.01 0.01 0.001
(0.27) 1.83 (0.39) 1.83 (0.48) 0.18
ACFE 0.24 0.22 0.25
(0.02)† 7.72 (0.03)† 7.06 (0.02)† 8.06
BD 0.06 0.05
Contemporary Accounting Research

(0.00)* 15.42 (0.01)* 12.69

BDIND 0.45 0.30
(0.04)† 6.50 (0.14) 4.29
ARTICLES 0.003 0.002 0.003
(0.04)† 7.08 (0.10)‡ 4.67 (0.07)‡ 7.08
ANALYSTS 0.01 0.01 0.01
(0.10)‡ 5.61 (0.08)‡ 5.61 (0.10)‡ 5.61
TURNOVER 0.05
(0.39) 1.29
INSTI 0.003 0.004 0.002
(0.04)† 9.13 (0.03)† 12.35 (0.11) 6.00

(The table is continued on the next page.)

TABLE 4 (Continued)

Model 1 Model 2 Model 3

Expected Coefficient % change Coefficient % change Coefficient % change
Variables sign (p-value) in hazard (p-value) in hazard (p-value) in hazard
NEW_FIN 0.25 0.26 0.25
(0.01)* 8.60 (0.01)* 8.96 (0.01)* 8.60
MGT_OWN ? 0.002 0.002 0.002
(0.17) 4.36 (0.19) 4.36 (0.42) 4.36
LEVER 0.14 0.26 0.20
(0.20) 3.03 (0.07) 5.56 (0.16) 4.30
LGMKV 0.15 0.13 0.15
(0.00)* 37.85 (0.00)* 32.08 (0.00)* 34.93
SHRED 0.44
(0.22) 3.13
MTB 0.01 0.01 0.01
(0.14) 4.98 (0.09)‡ 4.98 (0.15) 4.98
FYR 0.19 0.21 0.26
(0.01)* 8.37 (0.01)* 9.21 (0.00)* 11.27
No. of sample firms 821 821 737

Notes:
The variables are as defined in Table 2.
* Significant at the 1 percent level; one-tail test where appropriate.
Audit Committee, Board Characteristics, and Auditor Switch Decisions

† Significant at the 5 percent level; one-tail test where appropriate.

‡ Significant at the 10 percent level; one-tail test where appropriate.

CAR Vol. 24 No. 4 (Winter 2007)

1103
TABLE 5
Logit regressions modeling the probability of choosing Big 4 as a successor auditor
1104

Model 1 Model 2 Model 3

Expected Coefficient % change Coefficient % change Coefficient % change
Variables sign (p-value) in odds (p-value) in odds (p-value) in odds
Intercept ? 3.15 3.71 4.62
(0.00)* (0.00)* (0.00)*
AC 0.42 0.45 0.47
(0.05)† 140.30 (0.05)† 143.80 (0.09)‡ 146.40

CAR Vol. 24 No. 4 (Winter 2007)

ACIND 0.14 0.19 0.01
(0.37) 96.00 (0.33) 94.60 (0.50) 99.70
ACMTG 0.25 0.25 0.25
(0.00)* 157.30 (0.00)* 156.90 (0.01)* 155.30
ACFE 0.13 0.08 0.26
(0.38) 104.10 (0.42) 102.60 (0.28) 108.50
Contemporary Accounting Research

BD 0.08 0.09
(0.17) 83.50 (0.17) 80.70
BDIND 1.35 1.78
(0.08)‡ 121.40 (0.05)† 128.70
ARTICLES 0.01 0.01 0.01
(0.29) 84.80 (0.29) 84.50 (0.21) 74.90
ANALYSTS 0.02 0.02 0.01
(0.37) 114.00 (0.36) 114.40 (0.44) 107.50
TURNOVER 0.38
(0.30) 110.50

(The table is continued on the next page.)

TABLE 5 (Continued)

Model 1 Model 2 Model 3

Expected Coefficient % change Coefficient % change Coefficient % change
Variables sign (p-value) in odds (p-value) in odds (p-value) in odds
INSTI 0.01 0.01 0.01
(0.04)† 154.00 (0.05)† 147.50 (0.10)‡ 142.60
NEW_FIN 0.09 0.06 0.52
(0.42) 97.00 (0.44) 98.00 (0.13) 84.00
MGT_OWN ? 0.01 0.01 0.01
(0.16) 83.70 (0.14) 82.80 (0.11) 79.40
LEVER 0.02 0.03 0.74
(0.49) 99.60 (0.48) 100.70 (0.15) 85.50
LGMKV 0.71 0.75 0.91
(0.00)* 454.80 (0.00)* 498.60 (0.00)* 604.40
SHRED 2.67
(0.05)† 82.90
MTB 0.06 0.06 0.07
(0.01) 75.90 (0.01) 76.20 (0.00) 71.00
FYR 0.24 0.24 0.21
(0.19) 112.20 (0.19) 112.00 (0.25) 110.40
Pseudo R 2 0.42 0.42 0.42
No. of sample firms 821 821 737

Notes:
Audit Committee, Board Characteristics, and Auditor Switch Decisions

The variables are as defined in Table 2.

* Significant at the 1 percent level; one-tail test where appropriate.
† Significant at the 5 percent level; one-tail test where appropriate.

CAR Vol. 24 No. 4 (Winter 2007)

1105

‡ Significant at the 10 percent level; one-tail test where appropriate.

1106 Contemporary Accounting Research

In terms of the control variables, firms with higher institutional ownership

(INSTI) and larger market value (LGMKV) were more likely to choose a Big 4 suc-
cessor auditor. In model 3, the coefficient on the stock market reaction variable
(SHRED) is significantly negative, suggesting that firms that suffered a more nega-
tive stock market reaction around the shredding disclosure date were more likely to
choose a Big 4 firm as the successor auditor, probably due to their greater demand
for assurance offered by a Big 4 firm.
Among the independent variables in (2), LGMKV consistently has the strongest
effect on the odds of selecting another Big 4 auditor in the full sample: a one stan-
dard deviation increase in the natural logarithm of market value increases the odds
of selecting another Big 4 auditor by 498.60 percent in model 2.
In summary, our results so far suggest that firms with more independent audit
committees, audit committees with greater financial expertise, and larger and more
independent boards dismissed Andersen earlier. Firms with larger and more active
audit committees as well as firms with more independent boards were more likely
to choose a Big 4 successor auditor. Thus, our results suggest that firms with more
effective audit committees and boards demand better auditor reputation.

Alternative explanations of the main results

Our main argument is that more effective audit committees and boards were more
likely to dismiss Andersen earlier and choose a Big 4 successor auditor, ceteris
paribus, a finding consistent with the view that larger, more independent, more
active audit committees, and audit committees with greater financial expertise, as
well as larger and more independent boards, are more effective monitors.
Another possible explanation for our results is that early dismissal can be a
signal of a high-quality reporting system.18 More specifically, auditors could be
attracted to strong, viable companies that have good financial reporting systems,
because dismissal decisions are driven, at least in part, by competition in the audit
market. This argument states that the stronger clients would be the first to attract
the attention of other auditors, and other auditors would compete for these clients.
It would be much easier for stronger clients to find successor auditors.
We address this alternative (but not mutually exclusive) explanation to the
monitoring hypothesis in two ways.
First, we use four variables to control for the possibility that auditors will be
attracted to strong, viable companies that have good financial reporting systems.
We use Altman’s Z-score (ZSCORE)19 and return on assets (ROA) to measure the
strength and viability of a firm, because a firm’s financial strength and performance
could induce other auditors to compete for its business. Such firms would be more
likely to change auditors earlier than the weaker, less attractive clients; therefore,
we expect the coefficients on ZSCORE and ROA to be positive. Following Engel
2005, we introduce two more variables of financial reporting quality to measure a
firm’s attractiveness: (1) earnings restatements identified by the GAO (RESTA) and
(2) discretionary accruals estimated from the modified Jones model (MDA). Because
an earnings restatement indicates lower financial reporting quality (Palmrose, Rich-
ardson, and Scholz 2004; Agrawal and Chadha 2005), we expect that Big 4 audit

CAR Vol. 24 No. 4 (Winter 2007)

Audit Committee, Board Characteristics, and Auditor Switch Decisions 1107

firms were more likely to stay away from such clients and that such clients could take
longer to find a successor auditor. Big 4 auditors may find that firms with higher
levels of discretionary accruals are not attractive due to the increased risk associated
with more earnings management. Therefore, we expect that firms with higher dis-
cretionary accruals were less likely to select a Big 4 auditor and more likely to
dismiss Andersen later.
After controlling for these four variables, untabulated results for the timing
regression indicate that the coefficients on ACIND, ACFE, BD, and BDIND are still
significantly positive, consistent with those reported in model 2 of Table 4. The
four variables measuring the attractiveness of a client are not significant in explain-
ing the timing of Andersen’s dismissal.20 For the choice of a new auditor, ACMTG
and BDIND are still significantly positive after controlling for these four variables.
The coefficient on ZSCORE is significantly positive, suggesting that firms with
higher Altman’s Z-score were more likely to choose a Big 4 successor auditor. The
coefficient on RESTA is significantly negative, indicating that firms with earnings
restatements were less likely to choose a Big 4 successor auditor.21
Second, we use factor analysis to identify the common factor underlying Altman’s
Z-score, return on assets, restatement, and discretionary accruals. To facilitate the
interpretation of the common factor, we multiply restatement and discretionary
accruals by –1 because of their expected negative coefficients. The results (not tabu-
lated) indicate that the common factor is not related to the timing of Andersen’s
dismissal; however, it is strongly related to the choice of a Big 4 firm as the successor
auditor. This shows that Big 4 firms were attracted to strong, viable Andersen clients
with good financial reporting systems. However, the results on audit committee variables
and board variables are still consistent with those reported in Table 4 and Table 5.
In summary, we find that our main results remain robust after controlling for
the alternative (but not mutually exclusive) signaling explanation. We also find
some evidence that strong, viable Andersen clients with good financial reporting
systems were more likely to hire a Big 4 successor auditor.
Additional analyses and robustness tests
Examining December 31 fiscal year-end firms only
We conduct an additional analysis by restricting the sample to the 566 firms with a
December 31 fiscal year-end in Table 6. Chaney and Philipich (2002) find that
Andersen clients suffered a significant loss of market value around Andersen’s dis-
closure of document shredding in January 2002, suggesting the market had serious
doubts about the quality of Andersen’s audits. At the time, Andersen was well into
its annual audits of clients with December fiscal year-ends. Therefore, these clients
may have found it difficult to change auditors in the middle of an annual audit,
choosing to defer the auditor switch until after the completion of annual audit
(Barton 2005). So we conduct this additional analysis to see whether our empirical
results are robust when we only examine firms with December fiscal year-ends. We
link audit committee and board characteristics with the timing of Andersen’s dis-
missal in model 1 and model 2, and we link these internal governance mechanisms
with the choice of successor auditors in model 3.

CAR Vol. 24 No. 4 (Winter 2007)

TABLE 6
Additional analyses: December 31 fiscal year-end firms only
1108

Model 1 Model 2 Model 3

Cox proportional hazard model Cox proportional hazard model Logit model
Expected Coefficient % change Coefficient % change Coefficient % change
Variables sign (p-value) in hazard (p-value) in hazard (p-value) in odds
Intercept 5.73
(0.00)*
AC 0.07 0.01 1.52

CAR Vol. 24 No. 4 (Winter 2007)

(0.11) 6.06 (0.44) 1.69 (0.01)* 355.90
ACIND 0.18 0.13 0.45
(0.11) 4.60 (0.19) 3.05 (0.23) 87.40
ACMTG 0.01 0.01 0.31
(0.31) 1.84 (0.41) 1.84 (0.00)* 174.50
Contemporary Accounting Research

ACFE 0.14 0.08 0.77

(0.15) 5.09 (0.29) 3.47 (0.10)‡ 126.60
BD 0.04 0.13
(0.05)† 7.63 (0.10) 72.40
BDIND 0.81 0.34
(0.01)* 11.07 (0.40) 95.40
ARTICLES 0.002 0.002 0.02
(0.13) 5.00 (0.25) 2.47 (0.09) 63.40
ANALYSTS 0.02 0.02 0.001
(0.03)† 11.87 (0.02)† 11.87 (0.50) 100.3

(The table is continued on the next page.)

TABLE 6 (Continued)

Model 1 Model 2 Model 3

Cox proportional hazard model Cox proportional hazard model Logit model
Expected Coefficient % change Coefficient % change Coefficient % change
Variables sign (p-value) in hazard (p-value) in hazard (p-value) in odds

INSTI 0.005 0.004 0.01

(0.02)† 12.13 (0.02)† 12.13 (0.11) 153.00
NEW_FIN 0.26 0.27 0.34
(0.02)† 9.61 (0.02)† 9.61 (0.28) 89.10
MGT_OWN ? 0.004 0.003 0.002
(0.09)‡ 6.90 (0.10)‡ 6.90 (0.78) 105.10
LEVER 0.15 0.23 0.35
(0.21) 3.17 (0.17) 4.72 (0.38) 105.30
LGMKV 0.14 0.13 0.82
(0.00)* 35.50 (0.00)* 35.50 (0.00)* 579.70
MTB 0.01 0.01 0.09
(0.14) 2.52 (0.12) 2.01 (0.00) 65.10
Pseudo R 2 0.45
No. of sample firms 566 566 566

Notes:
The variables are as defined in Table 2.
* Significant at the 1 percent level; one-tail test where appropriate.
Audit Committee, Board Characteristics, and Auditor Switch Decisions

† Significant at the 5 percent level; one-tail test where appropriate.

‡ Significant at the 10 percent level; one-tail test where appropriate.

CAR Vol. 24 No. 4 (Winter 2007)

1109
1110 Contemporary Accounting Research

Following model 1 of Table 4, we only consider audit committee characteris-

tics in model 1 of Table 6. The significant coefficients on AC, ACIND, and ACFE in
model 1 of Table 4 become insignificant when we restrict the sample to only
December fiscal year-end firms. The significance levels for these three variables
are respectively 0.11, 0.11, and 0.15 for model 1 in Table 6. When we introduce
board variables in model 2, we find that board size and board independence are sig-
nificantly positively related to the hazard function, consistent with the full sample
results in model 2 of Table 4. For the December 31 year-end sample, it appears that
the timing decision was seen as more than an auditing issue because it attracted the
attention of the entire board.
The results for the successor auditor choice regression in model 3 are consis-
tent with the results for the full sample of 821 firms used in Table 5. That is, we
find that firms with larger and more active audit committees were more likely to
choose a Big 4 successor auditor for these December 31 fiscal year-end firms.
Interestingly, audit committee financial expertise also becomes significant. How-
ever, board independence is no longer significant in the auditor choice model in
Table 6, whereas it is significant in model 2 and model 3 in Table 5.

Alternative definition of financial expert

In our primary analyses in Table 4 and Table 5, we find that audit committees with
financial expertise (including both accounting and nonaccounting) were more
likely to dismiss Andersen earlier, as reported in Table 4. When we reclassify
financial experts as accounting financial experts only, the regression results (not
tabulated) are qualitatively similar to those reported in Table 4. When we further
reclassify financial experts as nonaccounting financial experts only, the regression
results (not tabulated) are also qualitatively similar to those reported in Table 4.
Also, we do not detect a relationship between audit committee financial expertise
and the choice of a Big 4 successor auditor for Andersen, consistent with the
results reported in Table 5.

Examining Big 4 sample firms only

We also test the timing of Andersen’s dismissal based only on the 727 firms that
chose a Big 4 successor. The regression results indicate that the coefficients on
ACFE, BD, and BDIND are 0.21, 0.06, and 0.51 with p-values of 0.05, 0.00, and
0.04, respectively, suggesting that financial expertise, board size, and board inde-
pendence influenced the timing of Andersen’s dismissal, consistent with the results
reported in Table 4. However, the significant coefficient on audit committee inde-
pendence (ACIND) in Table 4 is no longer significant when we restrict the sample
to the firms that selected a Big 4 auditor as a successor auditor.

Examining the existence of an audit committee

Among the 1,154 firms that terminated their relationship with Andersen after Octo-
ber 15, 2001, 199 firms did not have audit committees. The untabulated univariate
regression using the 1,154 firms shows a significant positive relationship between
the existence of an audit committee and the timing of Andersen’s dismissal and the

CAR Vol. 24 No. 4 (Winter 2007)

Audit Committee, Board Characteristics, and Auditor Switch Decisions 1111

choice of Big 4 auditors. When we run the multivariate regression using the 892
observations satisfying all the data requirements, the significant relationship between
the existence of an audit committee and the timing of Andersen’s dismissal remains,
whereas the significant relationship as to the choice of Big 4 auditors disappears.

5. Conclusions
The collapse of Andersen provides a unique setting to measure the effectiveness of
audit committees and boards of directors and to investigate the auditor switch deci-
sions by its clients. Using a unique sample of hand-collected data, we examine the
role of audit committees and boards of directors on the auditor switch decisions by
Andersen’s clients. Specifically, we investigate the impact of (a) audit committee size,
independence, activity, and financial expertise; and (b) board size and board independ-
ence on the timing of dismissal and successor auditor choice by Andersen’s clients.
We find that firms with more independent audit committees, audit committees
with greater financial expertise, and audit committees with larger boards dismissed
Andersen earlier. On the other hand, firms with larger and more active audit com-
mittees were more likely to choose a Big 4 firm as the successor auditor. We also
provide evidence that firms with more independent boards were more likely to dis-
miss Andersen earlier and to choose a Big 4 successor auditor. Our paper contributes
to the literature on auditor choice by Andersen clients by documenting the impor-
tance of audit committees and boards of directors, while controlling for important
factors identified in prior studies.
Our research also responds to the call for further research on audit committee
size and activity in DeFond and Francis 2005. Our findings have implications for
stakeholders and policymakers in that we find that the number of audit committee
meetings is an important mechanism of corporate governance,22 while the current
regulations of the New York Stock Exchange, the NASDAQ, and SOX generally
leave this area to the discretion of boards of directors and audit committees.
This study is subject to a number of limitations. First, our results only provide
evidence of a relation between audit committee, board characteristics, and auditor
switch decisions by Andersen’s clients through publicly disclosed information; we
have very little understanding of the process through which audit committee
members interact with each other and influence the auditor switch decisions.
Second, although we find that audit committee independence and financial exper-
tise were important in the timing of Andersen clients’ auditor dismissal decisions,
and that these results are robust across all three models in Table 4, only board vari-
ables remain significant in explaining the timing of Andersen’s dismissal when we
examine firms with December 31 fiscal year-ends. Finally, we treat each audit
committee meeting as equally important, an approach that does not fully reflect
the reality, because a 15-minute meeting and an hour-long meeting may vary in
importance.
Future research might find a way to examine the processes through which
audit committee members influence auditor dismissal and selection decisions. Sub-
sequent studies may also rely on survey data to more accurately measure the
importance of audit committee meetings.

CAR Vol. 24 No. 4 (Winter 2007)

1112 Contemporary Accounting Research

Endnotes
1. Caution should be exercised in interpreting the relationship between audit committee
characteristics and the timing of Andersen’s dismissal, because our results concerning
this relationship are sensitive to the exclusion of non-December 31 firms. Please refer
to the discussion of Table 6 for further details.
2. The SEC may make exemptions to the independence rule on a case-by-case basis
(Lennox and Park 2006).
3. The Sarbanes-Oxley Act (2002) also has specific requirements (e.g., sections 301, 302,
and 407) concerning the composition and financial expertise of audit committees.
4. In fact, the Supreme Court unanimously overturned Andersen’s conviction on May 31,
2005. This is purely a symbolic victory for Andersen because it had ceased to practice
auditing on August 31, 2002 and had lost all of its clients.
5. Ninety-two percent of respondents to a GAO survey indicated that transitioning to a
new auditor is difficult (GAO 2003).
6. Chaney and Philipich (2002) document that Andersen clients experienced market value
loss around the disclosure of the document-shredding date, while Krishnamurthy et al.
(2006) find that Andersen clients experienced negative market reaction around the
indictment date.
7. The BRC’s recommendation related to audit committee competence states: “At least
one member of the audit committee should have accounting or financial management
expertise (defined as past employment or professional certification in accounting or
finance, or comparable experience including service as a corporate officer with
financial oversight responsibility)” (BRC 1999).
8. Andersen was indicted for obstruction of justice on March 14, 2002. About two weeks
earlier, on March 1, 2002, Andersen agreed to pay $217 million to settle class-action
lawsuits related to the audit of Baptist Foundation of Arizona. On November 7, 2001,
Waste Management settled its class-action lawsuits for $457 million. As part of a
malpractice settlement, Andersen agreed to pay Waste Management $20 million. In
June 2001, Andersen agreed to pay a $7 million civil fine after the SEC accused it of
“knowingly or recklessly” issuing false and misleading audit reports for Waste
Management for the years 1993 through 1996 that inflated the company’s earnings by
more than $1 billion.
9. The initial sample in Barton 2005 is 1,299, and 123 firms either had no auditor change
information or made an auditor switch before October 15, 2001. Our sample is thus
comparable to Barton’s.
10. An outside director is defined as a director whose only relationship with the firm is to
serve on the audit committee. “Gray” directors are considered to be insiders. As in
Shivdasani and Yermack 1999, gray directors include retired employees, relatives of
the chief executive officer (CEO), and persons with disclosed conflicts of interest such
as outside business dealings with the company or interlocking director relationships
with the CEO.
11. DeFond et al. (2005) define a financial expert as belonging to one of the following two
categories: (a) accounting financial expert — all directors with experience as a public
accountant, an auditor, a principal or chief financial officer, a controller, or a principal
or chief accounting officer. These categories are chosen on the basis of suggestions

CAR Vol. 24 No. 4 (Winter 2007)

Audit Committee, Board Characteristics, and Auditor Switch Decisions 1113

included in the initial version of the Sarbanes-Oxley Act proposed by the SEC; and
(b) nonaccounting financial expert — all directors with experience as a CEO or
president of a for-profit corporation.
12. We exclude share turnover in the main regression to reduce the loss of observations due
to missing data. The results are generally robust to the inclusion of share turnover. See
Table 4 and Table 5 for regression results with and without share turnover.
13. Lennox (2005) finds a significantly negative relation between (low and high) regions of
management ownership and the appointment of high-quality auditors, whereas DeFond
(1992) finds a weak negative association (for some model specifications only) and
Francis and Wilson (1988) find no such significant association.
14. We did not control for the statistically insignificant control variables in the multivariate
proportional hazard model in Barton 2005. These variables include business-line
diversification, geographic diversification, volatility, auditor tenure, and unfavorable
audit opinion.
15. We also use the (1, 1) cumulative abnormal return around the document-shredding
date (SHRED) to control for the firm-specific measure of the effect of Andersen’s
reputation loss, because Asthana, Balsam, and Krishnan (2004) find some evidence
that the negative market reaction to the disclosure of Andersen’s document shredding
was related to the timing of Andersen’s dismissal. However, we exclude this variable
from our sample selection criteria to maximize the sample size. Our results, reported in
Table 4 and Table 5, are generally robust to the inclusion or exclusion of this variable.
16. Audit committee size (AC), audit committee independence (ACIND), number of audit
committee meetings (ACMTG), board size (BD), and board independence (BDIND) are
significantly negatively correlated with DURATION, suggesting that such firms were
more likely to dismiss Andersen earlier. The significantly positive correlations between
audit committee size (AC), number of audit committee meetings (ACMTG), board size
(BD), board independence (BDIND), and successor auditor choice (BIG4) show that
firms with larger and more active audit committees and firms with larger and more
independent boards were more likely to choose a Big 4 successor auditor.
17. The coefficients on ARTICLES and on ANALYSTS are significant across all three
models. The coefficient on INSTI is significant for both model 1 and model 2, and is at
11 percent significance level for model 3.
18. Cahan and Zhang (2006) examine whether successor auditors after Andersen’s demise
required more conservative accounting for their ex-Andersen clients and find that
successor auditors viewed an Andersen audit as a unique source of litigation risk. Their
study can be thought of as a signaling explanation in that successor auditor
conservatism could be a signal of financial reporting quality.
19. Altman’s 1968 Z-score is defined as Z 1.2w1 1.4w 2 3.3w 3 0.6w4 1.0w5,
where w1 working capital/total assets, w 2 retained earnings/total assets,
w3 earnings before interest and taxes/total assets, w4 market value of equity/book
value of total debt, and w 5 sales/total assets.
20. In the timing regression, the coefficient on ROA (MDA) is positive (negative), as
expected. Contrary to our expectation, the coefficient on ZSCORE (RESTA) is negative
(positive). All four coefficients are not significant.

CAR Vol. 24 No. 4 (Winter 2007)

1114 Contemporary Accounting Research

21. In the auditor choice regression, the coefficient on ROA is positive as expected. The
coefficient on MDA is positive, contrary to our initial expectation. Both coefficients are
not significant.
22. The importance of audit committee meetings is also evident from Levitt’s 1998
observation: “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.”

References
Abbott, L., and S. Parker. 2000. Auditor selection and audit committee characteristics.
Auditing: A Journal of Practice & Theory 19 (2): 47–66.
Abbott, L., S. Parker, and G. Peters. 2004. Audit committee characteristics and restatements.
Auditing: A Journal of Practice & Theory 23 (1): 69–87.
Agrawal, A., and S. Chadha. 2005. Corporate governance and accounting scandals. Journal
of Law and Economics 48 (2): 371–406.
Altman, E. 1968. Financial ratios, discriminant analysis, and the prediction of corporate
bankruptcy. Journal of Finance 23 (4): 589–609.
Anderson, R., S. Mansi, and D. Reeb. 2004. Board characteristics, accounting report
integrity, and the cost of debt. Journal of Accounting and Economics 37 (1): 315–42.
Asthana, S., S. Balsam, and J. Krishnan. 2004. Audit firm reputation, auditor switches, and
client stock price reactions: The Andersen experience. Working paper, Temple University.
Barton, J. 2005. Who cares about auditor reputation? Contemporary Accounting Research
22 (3): 549–86.
Beasley, M. 1996. An empirical analysis of the relation between the board of director
composition and financial statement fraud. The Accounting Review 71 (4): 443–65.
Beasley, M., and K. Petroni. 2001. Board independence and audit-firm type. Auditing: A
Journal of Practice & Theory 20 (1): 97–114.
Bédard, J., S. M. Chtourou, and L. Courtean. 2004. The effect of audit committee expertise,
independence, and activity on aggressive earnings management. Auditing: A Journal of
Practice & Theory 23 (2): 13–35.
Blue Ribbon Committee (BRC). 1999. Report and recommendations of the Blue Ribbon
Committee on Improving the Effectiveness of Corporate Audit Committees. Stamford,
CT: BRC.
Cahan, S., and W. Zhang. 2006. After Enron: Auditor conservatism and ex-Andersen clients.
The Accounting Review 81 (1): 49–82.
Carcello, J., and T. Neal. 2003. Audit committee characteristics and auditor dismissals
following “new” going-concern reports. The Accounting Review 78 (1): 95–117.
Chaganti, R., V. Mahajan, and S. Sharma. 1985. Corporate board size, composition and
corporate failures in retailing industry. Journal of Management Studies 22 (2): 400–17.
Chaney, P., and K. Philipich. 2002. Shredded reputation: The cost of audit failure. Journal of
Accounting Research 40 (4): 1221–45.

CAR Vol. 24 No. 4 (Winter 2007)

Audit Committee, Board Characteristics, and Auditor Switch Decisions 1115

Chang, C., W. Chi, and C. Liu. 2003. Client characteristics and auditor switch in an audit
failure. Working paper, National Cheng-Chi University.
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.
DeFond, M. 1992. The association between changes in client from agency costs and auditor
switching. Auditing: A Journal of Practice & Theory 1 (1): 16–31.
DeFond, M., and J. Francis. 2005. Auditing research after Sarbanes-Oxley. Auditing:
A Journal of Practice & Theory 24 (Supplement): 5–30.
DeFond, M., R. Hann, and X. Hu. 2005. Does the market value financial expertise on audit
committees of boards of directors? Journal of Accounting Research 43 (2): 154–94.
Eisenberg, T., S. Sundgren, and M. Wells. 1998. Larger board size and decreasing firm value
in small firms. Journal of Financial Economics 48 (1): 35–54.
Engel, E. 2005. Discussion of “Does the market value financial expertise on audit
committees of boards of directors?”. Journal of Accounting Research 43 (2): 195–204.
Fama, E. 1980. Agency problems and the theory of the firm. Journal of Political Economy
88 (2): 288–307.
Fama, E., and M. Jensen. 1983. Separation of ownership and control. Journal of Law and
Economics 26 (2): 301–25.
Francis, J. and E. Wilson. 1988. Auditor changes: A test of theories relating to agency costs
and auditor differentiation. The Accounting Review 63 (4): 663–82.
Hymowitz, C., and J. Lublin. 2003. Corporate reform: The first year: Boardrooms under
renovation; scandals prompted changes, but critics say more are needed to prevent
another Enron. Wall Street Journal (eastern ed.), July 22, B1.
Jemison, D., and P. Oakley. 1983. Corporate governance in mutual insurance companies.
Journal of Business Research 11 (4): 501–22.
Jensen, R. 1982. The competition in auditing: An assessment — A discussion. In Fourth
Symposium on Auditing Research, 452–66. Urbana, IL: University of Illinois at
Urbana-Champaign.
Jensen, M. 1993. The modern industrial revolution, exit, and the failure of internal control
system. The Journal of Finance 48 (3): 831–80.
Jensen, M., and W. Meckling. 1976. Theory of the firm: Management behavior, agency
costs, and ownership structure. Journal of Financial Economics 3 (4): 305–60.
Johnson, R., R. Hoskisson, and M. Hitt. 1993. Board of director involvement in
restructuring: The effects of board versus managerial controls and characteristics.
Strategic Management Journal 14 (Special Summer Issue): 33–50.
Kalbers, L., and T. Fogarty. 1993. Audit committee effectiveness: An empirical investigation
of the contribution of power. Auditing: A Journal of Practice & Theory 12 (1): 24–49.
Klein, A. 1998. Firm performance and board committee structure. The Journal of Law &
Economics 41 (1): 275–303.
Klein, A. 2002a. Audit committee, board of director characteristics, and earnings
management. Journal of Accounting and Economics 33 (3): 375–400.
Klein, A. 2002b. Economic determinants of audit committee independence. The Accounting
Review 77 (2): 435–52.

CAR Vol. 24 No. 4 (Winter 2007)

1116 Contemporary Accounting Research

Kosnik, R. 1987. Greenmail: A study of board performance in corporate governance.

Administrative Science Quarterly 32 (2): 163–85.
Krishnamurthy S., J. Zhou, and N. Zhou. 2006. Auditor reputation, auditor independence
and the stock market impact of Andersen’s indictments on its client firms.
Contemporary Accounting Research 23 (2): 465–90.
Krishnan, J. 2005. Audit committee quality and internal control: An empirical analysis. The
Accounting Review 80 (2):649–75.
Leddy, T. 1982. The competition in auditing: An assessment — A discussion. In Fourth
Symposium on Auditing Research, 467–82. Urbana, IL: University of Illinois at
Urbana-Champaign.
Lee, H., V. Mande, and R. Ortman. 2004. The effect of audit committee and board of
director independence on auditor resignation. Auditing: A Journal of Practice &
Theory 23 (2): 131–46.
Lennox, C. 2002. Opinion shopping, audit firm dismissals, and audit committees. Working
paper, Hong Kong University of Science and Technology.
Lennox, C. 2005. Management ownership and audit firm size. Contemporary Accounting
Research 22 (1): 205–27.
Lennox, C. S., and C. W. Park. 2007. Audit firm appointments, audit firm alumni, and audit
committee independence. Contemporary Accounting Research 24 (1): 235–58.
Levitt, A. 1998. The numbers game. Remarks delivered at the NYU Center for Law and
Business, September 28.
Lublin, J., and E. MacDonald. 1999. Panel to propose boosting independence and power of
board audit committee. Wall Street Journal (eastern ed.), February 8, A1.
Mautz, R., and F. Neumann. 1977. Corporate audit committees: Policies and practices.
Cleveland, OH: Ernst and Ernst.
McMullen, D., and K. Raghunandan. 1996. Enhancing audit committee effectiveness.
Journal of Accountancy 182 (2): 79–81.
Monks, R., and N. Minow. 1995. Corporate governance. Oxford, UK: Basil Blackwell.
Palmrose, Z., V. Richardson, and S. Scholz. 2004. Determinants of market reactions to
restatement announcements. Journal of Accounting and Economics 37 (1): 59–89.
Pincus, K., M. Rusbarsky, and J. Wong. 1989. Voluntary formation of corporate audit
committees among NASDAQ firms. Journal of Accounting and Public Policy 8 (4):
239–65.
PricewaterhouseCoopers. 1999. Audit committee: Best practices for protecting
shareholders’ interest. New York: PricewaterhouseCoopers.
Raghunandan, K., W. Read, and D. Rama. 2001. Audit committee composition, “gray
directors”, and interaction with internal auditing. Accounting Horizons 15 (2): 105–18.
Reed, B., M. Trombley, and D. Dhaliwal. 2000. Demand for audit quality: The case of
Laventhol and Horwath’s auditees. Journal of Accounting, Auditing and Finance
15 (2): 183 – 98.
Richardson, S., and R. Sloan. 2003. External financing, capital investment and future stock
returns. Working paper, University of Pennsylvania and University of Michigan.
Sarbanes-Oxley Act of 2002. Pub. L. no. 107-204, 116 Stat. 745.

CAR Vol. 24 No. 4 (Winter 2007)

Audit Committee, Board Characteristics, and Auditor Switch Decisions 1117

Scarbrough, D., D. Rama, and K. Raghunandan. 1998. Audit committee composition and
interaction with internal auditing: Canadian evidence. Accounting Horizons 12 (1):
51–62.
Securities and Exchange Commission (SEC). 2003. Final rules: Standards relating to listed
company audit committees. Washington, DC: SEC.
Shivdasani, A., and D. Yermack. 1999. CEO involvement in the selection of new board
members: An empirical analysis. The Journal of Finance 54 (5): 1829–53.
U.S. General Accounting Office (GAO). 2003. Accounting firm consolidation: Selected
large public company views on audit fees, quality, independence, and choice. GAO-03-
1158. Washington, DC: GAO.
Watts, R., and J. Zimmerman. 1986. Positive accounting theory. Englewood Cliffs, NJ:
Prentice Hall.
Whittington, R., and K. Pany. 2006. Principles of auditing and other assurance services.
New York: McGraw-Hill/Irwin.
Wilson, T., Jr. and R. Grimlund. 1990. An examination of the importance of an auditor’s
reputation. Auditing: A Journal of Practice & Theory 9 (2): 43–59.
Yermack, D. 1996. Higher market valuation of companies with a small board of directors.
Journal of Financial Economics 40 (2): 185–211.
Zahra, S., and J. Pearce II. 1989. Boards of directors and corporate performance: A review
and integrative model. Journal of Management 15 (2): 291–334.

CAR Vol. 24 No. 4 (Winter 2007)