Do social ties between external auditors and audit committee

members affect audit quality?*

Xianjie He†, Jeffrey Pittman‡, Oliver M. Rui§, Donghui Wu**

Abstract
We examine whether social ties between engagement auditors and audit committee members
shape audit outcomes. Although these social ties can facilitate information transfer and help
auditors alleviate management pressure to waive correction of detected misstatements, cozy
interpersonal relations can undermine auditors’ monitoring of the financial reporting process.
We measure social ties by alma mater connections, professor-student bonding, and employment
affiliation and audit quality by the propensity to render modified audit opinions, financial
reporting irregularities, and firm valuation. Our evidence implies that social ties between
engagement auditors and audit committee members impair audit quality. In additional results
consistent with expectations, we generally find that this relation is concentrated where social
ties are more salient, or firm governance is relatively poor and agency conflicts are more severe.
Implying reciprocity stemming from social networks, we also report some suggestive evidence
that audit fees are higher in the presence of social ties between an engagement auditor and the
audit committee. Collectively, our analysis lends support to the narrative that the negative
implications namely, worse audit quality and higher audit fees of these social ties may
outweigh the benefits.
Key words: Social ties; External auditors; Audit committee
JEL Codes: A14; G30; M42

* We thank Eddy Cardinaels, Ferdi Gul, and Xijia Su for their insights on an earlier version of this paper.
Our research has also benefited from comments from participants at the 2014 International Symposium
on Audit Research and the 2014 American Accounting Association Annual Meeting, and workshops at
Shanghai University of Finance and Economics, the University of Hong Kong, and Xiamen University.
Shuwei Sun and Ruiguang Zhao provided excellent research assistance and the China Institute of
Certified Public Accountants generously supplied audit field work day data for this research project.
Xianjie He acknowledges financial support from the National Natural Science Foundation of China (No.
71202003; No. 71472113) and the MOE Project of Key Research Institute of Humanities and Social Sciences
at Universities (No. 14JJD630005). Jeffrey Pittman appreciates financial support from Canada’s Social
Sciences and Humanities Research Council as well as Memorial University’s Chair in Corporate
Governance and Transparency. Oliver Rui acknowledges financial support of a CEIBS research grant and
a grant from National Science Fund Committee of China (No. 71372203). Donghui Wu acknowledges
Rega Financial Research Project funding support.
† School of Accountancy, Shanghai University of Finance and Economics. E-mail: he.xianjie@mail.shufe.
edu.cn.
‡ Faculty of Business Administration, Memorial University of Newfoundland. E-mail: jpittman@mun.ca.

§ China Europe International Business School (CEIBS). E-mail: oliver@ceibs.edu.

** School of Accountancy and Center for Institutions and Governance, the Chinese University of Hong

Kong. E-mail: donghui.wu@cuhk.edu.hk.

1. Introduction

A burgeoning literature documents that social networks influence economic agents’

behavior. In this paper, we examine the importance of social ties between an engagement
auditor and AC members to audit outcomes in China. 1 As illustrated in Figure 1, firm
management can affect external auditing by exerting influence over connected external auditors
via linkage A or AC members via linkage B. A number of papers have investigated how the
interpersonal links between firm management and external auditors (linkage A) affect audit
outcomes. Although some evidence suggests that manager-auditor ties impair audit and
earnings quality (e.g., Menon and Williams 2004; Guan et al. 2016), other research implies that
such ties do not threaten auditor independence and may even improve audit quality (e.g.,
Geiger et al. 2008). Partly motivated by the regulatory concerns surrounding AC
independence from management, considerable research implies that AC members’ formal ties,
with management (linkage B) can undermine AC oversight quality. In particular,
financial/familial ties or informal social ties between firm management and AC members can
reduce audit or financial reporting quality (e.g., Carcello and Neal 2000, 2003; Klein 2002;
Bédard et al. 2004; Krishnan 2005; Hwang and Kim 2012; Bruynseels and Cardinaels 2014).

However, there is hardly any research on whether economic outcomes vary systematically
with how external auditors interact with AC members via linkage C. This issue is interesting
as both external auditing and the AC are important governance structures intended to protect
the integrity of firms’ accounting numbers, and there are extensive interactions between the two
parties. Besides that the AC is responsible for overseeing the auditing process, the external
auditor may need to obtain information relevant to the audit from the AC (PCAOB 2007, AS5,
Para. 17). After observing that prior studies seldom examine demand-side factors, DeFond
and Zhang (2014) call for more evidence on the role that auditor-AC interactions play in audit
quality. Similarly, in calling for more research on the impact of AC characteristics on audit
quality, they stress the importance of analyzing the personal and social characteristics of

1 We use the term “engagement auditors” rather than “partners” given that China diverges from most
other countries in that the signatory auditors need not be partners. Engagement auditors there sign the
audit report in order to clarify who is responsible for the audit (MOF 1995a). Individual auditors are
explicitly accountable for audit reports issued in their names; e.g., regulatory sanctions for audit failures
are imposed on the individual signatory auditors involved (Chen et al. 2010; Gul et al. 2013).

1
committee members.

Social ties between AC members and auditors could affect audit quality in one of two
competing ways. On one hand, social ties could impede AC members and engagement
auditors from exercising due diligence in their mutual monitoring during the external audit
process, translating into lower quality audits. Intimacy between engagement auditors and AC
members can also lead to auditors’ unwarranted trust in information provided by ACs, injecting
bias into auditors’ judgment and decision making. On the other hand, by enhancing the trust
between AC members and engagement auditors, social ties can facilitate information transfer
between the two parties and reduce errors in auditor judgment. Additionally, in the presence
of a supportive AC, the auditor is in a better position to resist client pressure to waive
corrections when controversial issues in financial reporting arise, leading to better audits.

It follows that the question of whether audit quality would be enhanced or diminished by
social ties between AC members and external auditors has to be tackled empirically. For
several reasons, the Chinese market provides an opportune setting for analyzing this issue.
First, set against a guanxi (relationship oriented) culture pervading socioeconomic life in China,
social ties among individuals play a major role in economic activities there (Xin and Pearce
1996). Second, engagement auditors are required to sign the audit report in China and are
held accountable for the audit. Extensive recent research implies that signing auditors’
incentives and characteristics shape audit outcomes in China (Chen et al. 2010; Gul et al. 2013;
He et al. 2015). Third, in China, demographic data is available for identifying social ties
between AC members and engagement auditors. Consequently, Chinese public firms provide
a fertile testing ground for isolating whether social ties influence individual auditors’ practices.

Measuring social ties between engagement auditors and AC members with their alma mater
connections, teacher-student bonding, or employer affiliation, we find that such ties reduce the
auditors’ propensity to issue modified audit opinions (MAOs). Although MAOs are a direct
outcome of the audit process, we confront concerns surrounding the validity of this audit
quality measure by triangulating our analysis with the incidence of ex post financial reporting
irregularities, which reflects whether the auditor issued a clean opinion on materially deficient
financial statements. Moreover, we consider whether the capital markets reward in the form
of higher valuations firms that refrain from hiring auditors with social ties to the AC.
Evidence from these analyses corroborates our MAO results. We also generally find that the

2
importance of these social ties to audit quality is concentrated in firms with more salient links
evident in cases where a connected AC member chairs the committee, has an accounting
background, or has less reputation capital to protect; or the connected auditor is the more senior
auditor on the engagement or shares local ties with AC members. Consistent with
expectations, another set of results suggest that the detrimental impact of AC-auditor social ties
on audit quality subsides for firms with better governance structures and milder agency
conflicts. Finally, providing some support for reciprocity prevailing in these social networks,
we find that audit fees are slightly higher in the presence of AC-auditor social ties.

This study advances our understanding of the determinants of AC oversight effectiveness.

In some ways, the AC serves as a bridge between firm management and the external auditors.
On one side of this bridge, there is ample evidence on the links between the AC and
management. In contrast, we are not aware of any archival research exploring how social
networks between auditors and AC members operate on the other side of this bridge. Naiker
and Sharma (2009) find that the presence of former audit partners on the AC is associated with
fewer internal control weaknesses. Our research differs from theirs in two main respects.
First, we evaluate whether the connection between a current engagement auditor and the AC
impacts audit quality whereas they focus on potential connections between former audit
partners and current audit teams. Second, our focus on a broader specification of connections
is constructive for isolating whether social links between external auditors and AC members
shape audit quality.

To the best of our knowledge, China is the only country for which all of the data necessary
for examining our research questions is available. However, our focus on China naturally
raises concerns about the external validity of our evidence, particularly given the major cultural,
legal, and institutional differences between China and developed economies such as the U.S.
Apart from stressing that our research should be interpreted with this caution in mind, we
exploit the fairly unique institutions at work in China to deepen our analysis on the role that
AC-auditor social ties play in shaping audit quality. For example, although ACs in China are
similar to those in the U.S. in some ways; e.g. at least one member must have accounting
expertise, they diverge in others; e.g., public firms in China are not required to have an AC.
We capitalize on this institutional structure by examining whether the audit quality
implications of having an AC are materially sensitive to the presence of AC-auditor social ties.

3
Reinforcing our main evidence, these results imply that ACs lead to higher audit quality,
although the presence of AC-auditor social connections erodes this impact.

We also confront external validity concerns with additional analysis that exploits
cross-sectional variation in the data.2 For example, we would be on less solid ground in
interpreting our research as relevant to other jurisdictions if the evidence was concentrated in
clients of low-quality audit firms given that, in sharp contrast to the U.S., the Big Four audit
firms only hold a small fraction of the Chinese audit market. However, we find no perceptible
evidence that the impact of AC-auditor social ties on audit quality varies systematically
between clients of small and large audit firms. In another difference with U.S. institutions
governing external auditing, two engagement auditors sign the audit report in China. To
facilitate comparisons to other countries, we analyze whether social ties matter more to audit
quality when the connection comes through the more senior auditor, who has more
responsibility for the engagement. We find that social ties involving the senior auditor are
more damaging to audit quality, providing some support for the perspective that our research
has implications for other countries. Reinforcing the relevance of our research to other
jurisdictions, additional analysis implies that the role that AC-auditor social ties play in shaping
audit quality is more evident when firm governance is weaker and AC members have less
reputation capital at stake. Finally, our finding that ACs improve audit quality only in the
absence of AC-auditor social ties is consistent with Carcello et al.’s (2011) evidence that AC
effectiveness is compromised when management has more influence in the director
nomination/selection process.

Considerable regulatory attention has been devoted to the importance of AC independence

from firm management. In comparison, the independence of the AC from the outside auditor
has largely been taken for granted in public policy circles. In examining both audit quality and
fees, our large-sample evidence suggests that the costs of social links between the AC and
engagement auditors outweigh their benefits. Reflecting its first-order economic materiality,
the presence of AC-auditor social ties translates into the likelihood of receiving an MAO falling
by about 40% according to our coefficient estimates. These results are particularly relevant to
auditing standard setters when considering external auditors’ reliance on information provided
by ACs. Our evidence also suggests that, apart from formal financial ties or familial

2 We thank the Editor for suggesting additional analysis to help mitigate external validity concerns.

4
affiliations, disclosure of informal ties between engagement auditors and AC members would
benefit outside users in gauging audit quality. From a policy standpoint, our analysis also
coarsely informs the debate on whether compelling audit firms to divulge the identity of the
engagement auditor would improve accounting transparency.

Our analyses extend individual-level auditing research (e.g., Chi et al. 2009; Zerni 2012;
Carcello and Li 2013; Gul et al. 2013; Knechel et al. 2013a, 2013b) to include the role of social ties
between engagement auditors and ACs, helping to empirically clarify whether audit quality
hinges on individual auditors’ “socioeconomic characteristics” (DeFond and Zhang 2014: 304).
Our evidence may benefit audit firms eager to improve their performance by, for example,
reallocating their quality control resources to ensure that auditors with social ties to AC
members are reassigned to other clients (Francis and Michas 2013; Knechel et al. 2013b).

In Section 2, we develop the motivation for our testable predictions. Section 3 describes
the research methods. Section 4 covers the sample and data used in the analysis. We report
our main evidence in Section 5 and additional analysis in Section 6. Section 7 concludes.

2. Hypothesis development

2.1 Social networks and economic efficiency

The embeddedness theory developed by Granovetter (1985) emphasizes the importance of

social relations in analyzing economic activities in modern industrial society. Since much
information about economic activities can be subtle and difficult to verify, concrete social
relations generate trust between people and encourage cooperation (Granovetter 2005).
Consistent with socially connected individuals enjoying better information flow, Cohen et al.
(2008, 2010a) document that Wall Street money managers and financial analysts benefit from
their social ties with managers of public firms. Similarly, Engelberg et al. (2012) show that the
presence of interpersonal links between firm and bank managers improves monitoring by
facilitating the exchange of information between lenders and borrowers, while Schmidt (2015)
finds that stronger social connections between CEOs and independent board directors are
associated with higher bidder announcement returns during mergers and acquisitions.
Although social networks facilitate information sharing between economic agents, board
directors could also exploit their information advantage to benefit managers at the expense of

5
shareholders, resulting in inferior corporate governance. For example, socially connected
boards are associated with excessive CEO compensation and lower pay-/turnover-performance
sensitivity (Subrahmanyam 2008; Nguyen 2012), worse earnings management (Krishnan et al.
2011; Hwang and Kim 2012; Bruynseels and Cardinaels 2014), more frequent financial
misconduct (Chidambaran et al. 2012), lower firm valuation and more value-destroying
investments (Fracassi and Tate 2012).

2.2 Social networks between engagement auditors and audit committee members

Prior research on social networks implies that connections between engagement auditors
and AC members can have both positive and negative implications for audit quality. This
tension underlying whether these social ties are, on balance, better or worse for audit quality
helps motivate our research.

In one direction, social ties enhance trust and facilitate information flows between
individuals. To the extent that the AC possesses additional information concerning the
company’s accounting system, internal control environment, or transactions, better
communication with the AC will help the auditor improve their performance. However, such
information could be sensitive and difficult to transmit. Prior research suggests that social
networks enable people to discuss sensitive issues that would otherwise not be shared (Sias and
Cahill 1998; Gibbons 2004), which also applies to financial reporting context (Bruynseels and
Cardinaels 2014). In the presence of personal connections, AC members are likely to be candid
with engagement auditors. Rennie et al. (2010) find that management’s openness of
communication is essential to the efficient and effective conduct of an audit. Similarly, the
auditor’s trust in AC members who are socially connected with them would enable the auditor
to obtain the AC’s insights on important transactions and managers’ incentives, leading to
higher audit quality.

Another channel through which social networks may positively affect audit quality is the
AC’s support of the external auditor when controversial issues arise. Previous studies
generally suggest that more independent ACs are associated with better audit and financial
reporting quality. One explanation is that an independent AC could help alleviate
management pressure by supporting the auditor’s position in proposing audit adjustments or
issuing a going-concern report (Carcello and Neal 2000; DeZoort et al. 2008; Cohen et al. 2010b).

6
A more independent AC also shields auditors from dismissal after making audit decisions that
are unfavorable to management (Carcello and Neal 2003). Conceivably, interpersonal links
between AC members and auditors strengthen this positive impact. Mapping into DeAngelo’s
(1981) theory characterizing audit quality as the joint probability that the auditor will detect and
report material accounting omissions or misstatements, AC-auditor social connections may
improve audit quality by stimulating information flows and protecting auditor independence.

In the other direction, cozy interpersonal relations can be detrimental to monitoring and
governance according to some prior social network research. Extending this argument to the
audit setting, AC-auditor social ties may lead to a substandard audit for two reasons. First, the
AC serves as the firm’s primary interface with the external auditor. It is the committee’s duty
to supervise the firm’s internal audit system, to review its internal control system and financial
reports, and to oversee the external audit. A socially-connected auditor could be
reluctant whether consciously or unconsciously to challenge internal audit or control
procedures or financial reporting policies since a negative opinion on these matters is likely to
damage their social links. Likewise, social pressures could also undermine the AC’s external
audit oversight. Although the negative effects of concessions will unfold gradually, the
consequences of “not being nice” are likely to be immediate. Psychology studies find that
most people are far more concerned about immediate effects of their actions (e.g., Loewenstein
1996), suggesting that the auditor or AC are likely to refrain from challenging each other when
they are closely connected. Indeed, Nelson (2006: 31) stresses: “From a social perspective,
auditors may feel pressure not to disappoint or harm clients (who may be friends or former
colleagues)…”

Second, when sharing a common social background, auditors and AC members are more
likely to develop homophily, i.e., an affinity for similar others. While homophily fosters
mutual understanding and communications between parties (Bhowmik and Rogers 1971), it
also engenders favoritism bias (Uzzi 1996). This bias can lower the auditor’s professional
skepticism, resulting in them overly relying on audit evidence collected from familiar
individuals. Behavioral auditing research suggests that trust affects the auditor’s judgment,
although this may leave them vulnerable to becoming overly sympathetic toward clients
(Thompson 1995; Moore et al. 2006). For example, auditor-subjects choose a less rigorous audit
strategy when manager-subjects exhibit more trust-attracting behavior (King 2002). Similarly,

7
auditors who are more trusting of others are less likely to pay attention to audit evidence
suggesting aggressive reporting (Rose 2007). In the audit process, the auditor obtains
information from the AC that could be critical to their judgment and opinion formation. The
favoritism bias arising from social ties can lead to auditors’ unwarranted trust in such
information, injecting unintentional bias into auditors’ judgment and reporting decisions.
Consequently, social ties with the AC could impede the auditor from applying a more rigorous
audit process or exercising due professional care.

2.3 Audit committees in China

It would be hard to accept that AC-auditor social ties can materially affect audit outcomes in
China unless the audit committees under study actively monitor the financial reporting process.
Chinese listed firms largely began setting up ACs in 2002 when the China Securities Regulatory
Commission (CSRC, the equivalent of the U.S. SEC) released the Code of Corporate Governance for
Listed Companies, which recommends that corporate boards form ACs to oversee audit-related
matters.3 A recent KPMG (2015) survey suggests that ACs in China, compared with the global
average, devote significantly more time to issues pertinent to financial reporting. This
reconciles with recent archival research implying that regulatory and market forces discipline
director behavior in China (Xin et al. 2013; Jiang et al. 2016). Importantly, although early
survey evidence suggested that ACs in China largely play a ceremonial role (e.g., Lin et al. 2008),
extensive recent research reinforces that ACs there improve governance and accounting
transparency (e.g., Chen and Zhang 2010; Lo et al. 2010; Chang et al. 2013; Lin and Lee 2013).
To provide additional insight on how ACs operate in practice, we interviewed several AC
members of Chinese public firms, which led to the following observations. First, the ACs
perform a fiduciary duty in overseeing financial reporting. Second, given their responsibility
for reviewing many internal documents to ratify corporate decisions on such issues as
restructuring, mergers and acquisitions, and related-party transactions, AC members may
acquire information that is not readily available to the external auditors. Third, ACs and the
external auditors meet regularly to discuss such issues as the audit opinion formulation, and

3 According to the Code, the main duties of the AC include: (i) proposing the appointment of an external
audit firm; (ii) overseeing the company’s internal audit system; (iii) serving as the liaison between the
company’s internal and external auditors; (iv) reviewing the company’s financial information and its
disclosure; and (v) reviewing the company’s internal control system.

8
accounting judgments and estimates.

2.4 Hypotheses

It remains an empirical question whether the positive or negative aspects of social

connections between engagement auditors and the AC dominate in shaping audit quality.
Accordingly, we state our first hypothesis in the null form as follows:

H1: The presence of social ties between engagement auditors and audit committee
members has no impact on audit quality.
In a standard approach for improving identification (e.g., Rajan and Zingales 1998), we next
consider whether the effects of AC-auditor social links on audit quality hinge on the importance
of these ties. Since an AC chair likely plays a more important role than other members in
monitoring the financial reporting process, we predict that the impact of social ties will be
concentrated in cases where connections are forged through AC chairs. A number of studies,
including DeFond et al. (2005), Krishnan (2005), and Dhaliwal et al. (2010), document that better
monitoring accompanies accounting expertise on the AC. Accordingly, we expect that the
audit quality impact of social ties will be isolated in AC members with accounting expertise.
Prior research suggests that auditor seniority matters to audit quality. Bonner and Pennington
(1991) find that domain-related knowledge is gradually acquired with experience. Bonner and
Lewis (1990) stress that experience provides an opportunity for the development of expertise.
We expect that social ties play a larger role in shaping audit quality when the connection comes
through the more senior auditor. Xin and Pearce (1996) suggest that guanxi in China depends
on shared identification with school, place of work, and hometown. We posit that bonding
tends to be stronger when people come from the same geographic region. It follows that the
impact of social ties on audit quality is stronger if AC members and engagement auditors work
in the same province. Fama and Jensen (1983) argue that reputation capital motivates outside
directors to closely monitor the firm. Fich and Shivdasani (2007) find that subsequent board
appointments fall for directors of firms suspected or charged with fraud, which they attribute to
reputational damage. In another way to examine cross-sectional variation in the importance of
social ties, we expect that the impact of AC-auditor connections will depend on the reputational
capital that directors have at stake. The above analysis leads to our second hypothesis:

H2: The impact of social ties between engagement auditors and AC members on audit

9
 quality will vary with: whether a connected AC member chairs the committee, has an
accounting background, or comes from the same province as an engagement auditor;
and the connected AC member’s reputational capital.
In firms with relatively poor governance structures and more severe agency problems,
AC-auditor social ties likely play a larger role in shaping audit quality. On one hand, social
ties encourage information sharing between ACs and auditors and improve auditor
independence, translating into audits becoming more effective in alleviating agency problems.
On the other hand, by fostering favoritism and reducing auditor skepticism, AC-auditor social
links can exacerbate agency conflicts. We expect governance to be weaker and agency
problems to be worse when ownership is heavily concentrated (La Porta et al. 2000; Fan and
Wong 2002), institutional or foreign investors are absent (Hartzell and Starks 2003; Parrino et al.
2003; Khanna and Palepu 2000), or analyst coverage is thin (Irani and Oesch 2013; Chen et al.
2015). Similarly, corporate governance suffers when firms have powerful entrenched CEOs,
such as those who are internally promoted (Adam et al. 2005; May 1995) or have longer tenure
(Hermalin and Weisbach 1988, 1998). Additionally, Doidge et al. (2007) show that when
country-level institutions are weak, economic agents may not have sufficient incentives to
develop reputational capital since the payoffs to good reputation are inherently low. 4
Consequently, we expect that the importance of AC-auditor social ties to audit outcomes is
sensitive to regional institutional infrastructure quality. This discussion provides motivation
for our third prediction:

H3: The impact of social ties between engagement auditors and AC members on audit
quality will vary with the firm’s governance quality, which is evident in the extent of
corporate ownership concentration, the equity stake held by institutional investors, the
presence of foreign investors, the extent of analyst coverage, whether the firm is run by
a powerful CEO, and the institutional infrastructure quality in the region in which the
firm is located.

4 Our choice and specification of governance measures is also motivated by prior research on corporate
governance in China. Specifically, high ownership concentration of Chinese firms is associated with
more tunneling activities (Liu and Tian 2012); ownership by mutual funds translates into better firm
performance (Yuan et al. 2008); the presence of foreign investors reduces expropriation by controlling
shareholders (Huang and Zhu 2015); greater analyst coverage improves firms’ information environments
(Li et al. 2016); key managers having longer tenure is associated with higher fraud incidence (Chen et al.
2006); and Chinese firms operating in regions characterized by weaker institutions are less likely to
commit to transparent financial reporting (Wang et al. 2008).

10
3. Empirical methods

3.1 Social ties and auditor reporting behavior

We operationalize the construct of social networks with (i) common educational experience,
(ii) teacher-student bonding, and (iii) employer affiliation.5 Education is one of the strongest
forces, next only to ethnicity, age, and religion, that shape people’s social worlds (McPherson et
al. 2001). People that select the same schools are likely to share common interests or
backgrounds, while educational experiences may engender some common beliefs. They also
have opportunities to interact with each other through alumni activities. It follows that a
common educational background enables individuals to develop social networks and form
homophily. In China, the bonding between professor and student is important, which is
evident in academics routinely accepting appointments to corporate ACs. 6 We identified
whether an AC member is currently a dean, associate dean, department chair, or a full professor
at the school from which the engagement auditor graduated. These academics are usually in
charge of alumni affairs at their institutions. Alumni events provide an opportune setting for
academics to develop closer connections with alumni. Moreover, audit partners, who are
naturally considered excellent alumni, are frequently invited to sit on different committees at
their alma mater. This provides another channel through which partners can form close
connections with academics on the ACs. In addition, the mobility of university faculties in
China was minimal during our sample period; i.e., senior academics typically taught at the
same school from which they had graduated. This reinforces that AC members and

5 Our empirical strategy involves comprehensively examining our research questions with a broad
specification of AC-external social ties. However, this may come at the expense of undermining external
validity since the professor-student bonding in channel (ii) may be fairly unique to China. Later in the
paper, we explain that our core evidence generally holds when we separately analyze each channel. We
also attempted to follow recent research by analyzing AC-auditor social ties forged through
business/country clubs and charities (e.g., Bruynseels and Cardinaels 2014). However, we could
pinpoint hardly any social connections coming through these channels in China; i.e., data constraints
make reliable identification exceedingly difficult, even when relying on extensive hand collection.
Measurement error of unknown severity stemming from this potential under-identification of actual
AC-auditor social ties admits noise into the data, although this would bias our tests against rejecting the
null hypotheses.
6 In his classic paper, Jacobs (1979: 243) stresses that teacher-student relationships are grounded in guanxi:
“... a base for a guanxi depends upon two or more persons having a commonality of shared
identification. …For example, the teacher and student in a teacher-student guanxi share identification
with an experience important to both of them.” In empirical work, the teacher-student relationship is
often used by scholars to reflect the linkage between individuals in China (e.g., Farh et al. 1998).

11
engagement auditors, if they are connected as defined above, likely overlapped when the
auditors attended that school. Finally, considering that employment affiliation between senior
managers and auditors matters to audit quality in the U.S. (Menon and Williams, 2004; Baber et
al., 2014), we also include AC-auditor employment affiliation as another form of social ties.

To test the importance of social ties between AC members and engagement auditors to audit
quality, we examine auditors’ propensity to issue modified audit opinions to their clients.
Compared with measures based on attributes of audited financial statements (e.g., abnormal
accruals or earnings response coefficients), the audit opinion represents a more direct measure
of audit quality since it is directly under the auditor’s control (DeFond and Zhang 2014). Prior
research suggests that MAOs have reasonable power to capture variation in audit quality in
China (Chen et al. 2010; Chan and Wu 2011). China’s Independent Auditing Standards specify
four types of audit opinions namely, unqualified, qualified, disclaimer, and adverse and
stipulate that explanatory notes can be used with unqualified opinions when necessary.7 We
classify auditors’ reports into two categories: (i) clean reports; and (ii) MAOs that comprise
unqualified opinions with explanatory notes, qualified opinions, and disclaimers.

Our MAO measure differs from the going-concern opinions (GCOs) typically used in the
U.S. auditing literature. Focusing on financially distressed clients, a GCO-based analysis
examines how auditors evaluate whether there is doubt surrounding their clients’ ability to
continue as a going-concern. Since an audit modification due to a GAAP violation is
permissible in China, MAOs could be issued to highly profitable firms that have questionable
accounting practices (e.g., Chen et al. 2001). Consequently, audit opinions in China reflect not
only auditors’ evaluation of clients’ financial condition, but also their judgment on the fairness
of the financial statements. It follows that measuring audit quality with MAOs suits our
research setting.

Audit quality jointly stems from an auditor’s competence and independence (DeAngelo
1981). In our testing ground, issuing an MAO implies that the auditor is able to discover and
report material accounting misstatements. As stressed earlier, AC-auditor social ties may

7 No firm in our sample has received an adverse opinion. Although the CICPA interprets unqualified
opinions with explanatory notes in a manner similar to the “emphasis of matter” in U.S. GAAS, this type
of audit report is often issued in lieu of a qualified opinion in China; previous China-based studies all
treat it as a form of audit opinion modification (e.g., DeFond et al. 2000; Chen et al, 2010; Chan and Wu,
2011).

12
influence audit outcomes in one of two competing ways. On one hand, the closer relationship
suggests that AC members are more forthcoming in sharing information with auditors. More
efficient information transmission should enable auditors to detect irregularities, leading to a
higher likelihood of issuing MAOs. Alternatively, auditors are better able to withstand client
pressure to render a clean opinion when they are backed by a socially-connected AC. On the
other hand, intimacy between engagement auditors and AC members may deter them from
challenging each other’s work, reducing the prospect of discovering irregularities.
Connections can also foster auditors’ unwarranted trust in information provided by the AC,
biasing auditors’ judgment and decision making. As a result, an acquiescent auditor could
issue a clean opinion when an MAO is warranted. In short, the net impact of AC-auditor
social connections on audit quality remains an empirical question. We examine which effect
dominates by estimating the following logistic regression model (client firm and time subscripts
are omitted for simplicity):

MAO = + 1ACT + 2MGT_Chair/CEO + 3MGT_Other + Ci + t + k + , (1)

where MAO is equal to one for a modified opinion, and zero otherwise; ACT is assigned the
value one if at least of one of the AC members has a school tie with one of the engagement
auditors and their age gap is less three years, or they have teacher-student bonding, or they
used to work at the same employer, and zero otherwise; MGT_Chair/CEO is equal to one if the
CEO or board chair has social ties namely, alma mater and employer affiliation
connections with one of the engagement auditors;8 MGT_Other is set to one if at least one of
the top managers (including board directors other than CEO, board chair, or AC members) has
social ties developed through educational experience, teacher-student bonding, or employer
affiliation with one of the engagement auditors, and zero otherwise; Ci is a vector of client and
auditor characteristics that may affect audit quality; t and k are year and industry fixed effects,
respectively; and is the regression error term.

In all estimations, we control for the social ties between engagement auditors and firms’ top
managers and other board directors given that Kwon and Yi (2012) and Guan et al. (2016) report

8The teacher-student bonding is irrelevant in measuring CEO/Chair-auditor ties. Reflecting that the
CEO and Board Chair roles are typically full-time jobs, we observe no cases in the data where a CEO or
Board Chair takes an academic position at an educational institution.

13
that school ties affect audit quality. In our data, ACT is positively correlated at the 1% level
with both MGT_Chair/CEO ( = 0.114) and MGT_Other ( = 0.122), likely reflecting the influence
of social networking among firm managers, AC members, and auditors over the choice of
auditors or formation of ACs (Lennox and Park 2007).

The observed AC-auditor social links are largely pre-determined in that the formation of
interpersonal connections typically predates the events being studied by a long time (Cohen et
al. 2008, 2010a; Engelberg et al. 2012). Consistent with this prior research on social networks
and governance, the ACT variable is specified as exogenous in Model (1). Nevertheless, one
concern is potential endogeneity stemming from the non-randomness in the match between
ACs and engagement auditors via social ties. Although such endogeneity does not necessarily
bias for our findings, we apply propensity score matching (PSM) to confront this issue. As we
show in Section 5.3, the implementation of PSM well aligns treatment observations (i.e., those
with ACT equal to one) with matched controls in the observable characteristics; in the PSM
analysis, all of our core results hold. Also, the similarity between treatment and control firms
suggests that endogeneity is not likely to pose a material threat to our inferences (Atanasov and
Black 2015).

We follow prior Chinese auditing research by controlling for this set of client and audit
characteristics (e.g., DeFond et al. 2000; Chen et al. 2001): Big4 (an indicator variable for clients
of an international Big 4 auditor); Top10 (an indicator variable for clients of a domestic top 10
auditor, ranked by the number of listed clients in the year under study); LTA (the natural
logarithm of client’s total assets); Leverage (the ratio of total liabilities to total assets); Loss (an
indicator variable for negative net income); ROA (net income divided by total assets ); Liquidity
(the ratio of current assets to current liabilities); REC (net accounts receivables divided by total
assets); and INV (inventories divided by total assets). Since large auditors are usually more
competent or more independent (DeAngelo 1981; DeFond et al. 2000; Wang et al. 2008; Chan
and Wu 2011), we expect the coefficients on Big4 and Top10 to be positive. Since auditors are
more likely to issue MAOs when clients have higher audit risks, we expect the coefficients on
Leverage, Loss, REC, and INV to be positive and those on LTA, ROA, and Liquidity to be negative.

3.2 Social ties and ex post audit outcomes

A lower propensity to render MAOs is typically interpreted as evidence of low quality

14
auditing on the grounds that an auditor commits a Type II error by issuing a clean opinion
when an MAO is warranted. This conclusion, however, is confounded by the Type I error, i.e.,
an MAO is issued when a clean one is appropriate (Deng et al. 2012; Fogel-Yaari and Zhang
2013). In our setting, social ties with AC members may enhance auditors’ trust of clients’
financial reporting policies. As a result, auditors are less likely to disagree with management
over the application of accounting principles or estimates. To the extent that management
choice is appropriate, enhanced trust reduces Type I errors. Alternatively, supported by the
AC, socially tied auditors are in a better position to persuade management to make audit
adjustments in order to comply with GAAP in financial reporting. In either event, social
connections reduce the need for an auditor to issue an MAO. This suggests that the negative
link between AC-auditor social ties and the propensity to issue MAOs may not necessarily
reflect compromised audit quality.

The ex ante disagreements between auditors and client management over accounting issues
or audit adjustments are unobservable. However, we can analyze two observable ex post
outcomes: clients’ earnings restatements and regulatory sanctions in the event of accounting
irregularities. Unambiguously reflecting accounting problems and thus low-quality audits,
both outcome variables are associated with auditor attributes (Lennox and Pittman 2010;
Francis et al. 2013; Gul et al. 2013). Importantly, this analysis will help clarify whether the
observed association between AC-auditor social ties and the propensity to issue MAOs reflects
errors in auditor judgment and decision making. Specifically, there are six possible scenarios
as follows:

Ex ante association between social ties and MAOs

Positive Negative
(1) Correct decision (2) Type II errors
Positive
Ex post association (fewer Type II errors)
(3) Type I errors (4) Correct decision
between social ties No
(fewer Type I errors)
and irregularities (5) Type I errors (6) Correct decision
Negative
(fewer Type I errors)
In both scenarios (1) and (6), the auditors reduce reporting errors since the propensity to
issue MAOs is consistent with the ex post occurrence of irregularities. In scenario (4), where
the incidence of irregularities is not higher for audits conducted by connected auditors,
rendering fewer MAOs reduces the auditors’ Type I errors and thus the costs to both clients and

15
auditors. We are particularly interested in scenario (2): if socially connected auditors less
frequently issue MAOs and, ex post, there are more irregularities for their clients, then the
evidence suggests Type II errors and supports impaired audit quality due to AC-auditor social
ties. Although scenarios (3) and (5) are theoretically possible, it is far less likely to occur since
auditors will likely lose the engagement when they over-issue MAOs.

We employ the following logistic model to test whether accounting irregularities are related
to the AC-auditor social connections:

Irregularity = + 1ACT + 2MGT_Chair/CEO + 3MGT_Other + Ci + t + k + , (2)

where Irregularity is an indicator variable that is equal to one if earnings in the current year are
restated in subsequent years or the firm is subsequently sanctioned by the regulators due to
financial misrepresentation in the current year, and zero otherwise;9 Ci includes the following
client and auditor characteristics that may affect the incidence of accounting irregularities: Big4,
Top10, LTA, Leverage, ROA, Loss, and MAO. All other variables are defined as before.

We triangulate the above ex post audit outcomes with a valuation analysis. To the extent
that AC-auditor social ties undermine audit quality, it follows that the worse information
asymmetry accompanying these links will translate into a lower valuation (Fan and Wong 2005;
Guedhami et al. 2014). In contrast, if AC-auditor social ties sufficiently enhance information
sharing between AC members and engagement auditors according to investor perceptions (i.e.,
information asymmetry falls with these connections), then these ties should be positively
reflected in the firm’s valuation. In exploring whether AC-auditor ties affect firm value, we
measure firm value with Tobin’s Q after Fan and Wong (2005) in this model:

Tobin’s Q = + 1ACT + 2MGT_Chair/CEO + 3MGT_Other + Ci + t + k + , (3)

where Tobin’s Q is the market value of common equity plus the book value of debt divided by
total assets at the end of the year; and Ci includes these factors that may affect firm valuation:
Big4, Top10, LTA, Leverage, ROA, and Liquidity. After recent research (e.g., Jiang et al. 2010;
Guedhami et al. 2014), we also control for Growth (the annual sales growth rate) and CAPEXP

9 Since both restatements and regulatory sanctions capture the same construct of accounting
irregularities from an ex post perspective, we combine these two measures into a single variable. Our
core findings are materially insensitive to examining restatement and regulatory sanctions separately.

16
(the ratio of capital expenditures over total assets). All other variables are defined as before.

3.3 Models for testing H2 and H3

We analyze whether the impact of social ties matters more in different situations under the
predictions in H2 and H3 with the following model:

y = + 11ACT_X + 12ACT_X’ + 2MGT_Chair/CEO + 3MGT_Other + Ci + t + k + , (4)

where y represents the dependent variable, MAO, Irregularity, or Tobin’s Q. In this model, we
replace ACT with a pair of test variables that bisect the subsample where ACT is equal to one.
Specifically, ACT_X denotes the cases where ACT = 1 and the observation possesses an attribute
of interest such that the negative effects of AC-auditor social ties are expected to be stronger,
and ACT_X’ denotes the cases where ACT = 1 and the observation does not have that attribute.
We specify the bisecting variables in Section 5.4.

4. Sample and data

Our sample period covers 2004 to 2010. The sample period starts in 2004 because firms’
disclosure of the background information on their directors, including the AC members, is quite
sparse before 2004. Our primary data source is the China Stock Market and Accounting
Research Data Base (CSMAR). From the 11,143 A-share firm-year observations available in
CSMAR,10 we exclude: (i) 4,052 observations that do not have ACs; and (ii) 93 observations that
have missing values for variables used in the multivariate regression analysis. After imposing
these screens, our final sample consists of 6,998 firm-year observations.

We downloaded the profiles of firms’ executives and board directors, including AC

members, from CSMAR. From these profiles, we manually collected their information for their
junior college, undergraduate, and postgraduate education as well as employment experience.
As stipulated by China’s Independent Auditing Standards No. 7 – Audit Report and related practice

10 A-shares in the Shanghai and Shenzhen stock exchanges refer to those issued to domestic investors.
A total of 114 firms since 1992 have been authorized to issue B-shares to foreign investors. The B-share
firms are not included in the sample because their regulatory environments are considerably different
from those of A-share firms. For example, until 2007, B-share firms were required to prepare an
additional set of financial statements for foreign investors according to International Accounting
Standards, and to be audited by an international audit firm.

17
guidelines issued in 1995 (MOF 1995a and 1995b), engagement auditors must sign the audit
reports in order to clarify the responsibility for the audits performed. Typically, an audit
report is signed by two auditors: one mainly administers the fieldwork, and the other mainly
performs the review work. We collect signatory auditors’ education information from
http://cmispub.cicpa.org.cn, an on-line enquiry system developed by the Chinese Institute of
Certified Public Accountants (CICPA). From the CICPA, we also obtain the audit field work
day data to measure audit effort. Data for individual auditors’ working experience were
publicly available from the CSRC at http://assdata.csrc.gov.cn/.

We retrieve data on enforcement actions taken by the CSRC or MOF (the Ministry of
Finance) for corporate misconduct from CSMAR.11 Consistent with Hung et al. (2015), we only
retain enforcements related to financial misrepresentation.12 Consequently, the nature of the
frauds under study is similar to that of those identified from the Accounting and Auditing
Enforcement Releases (AAERs) in the U.S.-based literature (e.g., Lennox and Pittman 2010).
We hand collect restatement data from firms’ annual reports. Specifically, we reviewed the
“Material Accounting Errors” section of footnotes to the financial statements and identified
corrections to earnings or shareholders’ equity in prior years. After Wang and Wu (2012) and
Gul et al. (2013), we excluded restatements arising from changes in accounting standards or
government tax rules, or mergers and acquisitions since such restatements are clearly not
caused by intentional errors.

In Table 1, Panel A shows the time-series distribution of our sample observations,

partitioned into two groups by ACT. In almost 15% of the sample firm-years, AC members are
socially connected to engagement auditors. In the early years under study, there are relatively
more observations where AC members are socially connected to engagement auditors. This
likely reflects that the director market was less developed when the CSRC initially encouraged
listed firms to establish ACs. In a fledgling market, firms are more apt to turn to their auditors

11Both the CSRC and the MOF have the authority to sanction firms for problematic accounting practices.
The MOF is in charge of all accounting affairs in China, while the CSRC is empowered to handle
accounting and disclosure issues related to publicly traded companies.
12 Specifically, we keep the following categories of corporate misconduct compiled by CSMAR: (1)
fabricating profits; (2) overstating assets; (3) delaying disclosures; (4) false statements; (5) material
omissions; (6) expropriation of corporate assets by the large shareholders; and (7) fraudulent IPOs.
Misconduct such as market manipulation, insider trading, or violations of laws and regulations that are
not directly related to financial presentation are excluded from our analysis.

18
for advice on AC composition, leading to the appointment of AC members from the auditors’
circle of acquaintances. This practice gradually becomes less prevalent over time as the
director market matures. In Panel B, we report the distribution of sample observations by
industry. The relative frequencies of observations are fairly evenly spread between the two
groups of observations partitioned by ACT, suggesting that there is no major clustering of social
ties between engagement auditors and AC members in certain industries.

The descriptive statistics for the regression variables are reported in Table 2. To mitigate
the undue influence of outlying observations and potential coding errors, we winsorize the
continuous variables at the 1st and 99th percentiles. Only about 0.6% of the observations are
audited by engagement auditors who have social ties with the firm’s CEO/Chair, evident in the
mean of MGT_Chair/CEO, while the percentage of observations in which engagement auditors
are socially tied to other top executives is much higher at 7.1%, as shown by the mean of
MGT_Other. The means of MAO and Irregularity reveal that about 7.6% of the observations
have received MAOs from their auditors and the incidence rate of financial reporting
irregularities is about 8.1%. Turning to the independent variables, the means of Big4 and
Top10 indicate that about 6.3% and 43.4% of sample firms are audited by the international Big N
and domestic Top 10 auditors, respectively. The statistics for the other independent variables
suggest that they are reasonably distributed with some degrees of variation.

5. Empirical results

5.1 Social ties and audit outcomes

Table 3 reports multivariate analysis based on two audit outcome variables, MAO and
Irregularity. To correct for potential serial correlation and heteroskedasticity in the pooled
cross-sectional data, we use the Z/t-statistics based on the Huber/White/sandwich estimator
(clustered) for variance throughout the paper (Wooldridge 2002). For the MAO regression in
Column (1), we find that the coefficients on ACT and MGT_Chair/CEO are significantly negative
at the 1% and 5% levels, respectively, while the coefficient on MGT_Other fails to load. In
other words, these results imply that a common social background between the engagement
auditors and AC members or the CEO/Chair reduces the likelihood of MAOs, although social
ties between engagement auditors and other top managers has no perceptible impact on the

19
auditors’ reporting behavior. For the control variables, the coefficients all have the predicted
signs except that the coefficient on Liquidity does not load and the coefficient on INV is
significantly negative. 13 To evaluate economic importance, we consider the odds ratio
estimates for changing the independent variables by one unit. For ACT, the odds ratio
estimate is 0.391, suggesting that an observation with AC-auditor social ties is only 40% as likely
to receive an MAO as an observation without such ties. From an economic standpoint, this
impact is highly material.

Although the MAO regression results are more consistent with the impaired auditor
independence narrative underlying the prediction in H1, less frequent MAOs could also stem
from lower rates of Type I errors due to the enhanced trust between firm management and
auditors. To explore this alternative explanation, we turn to the analysis of ex post audit
quality measures based on accounting irregularities. As shown in Column (2), the coefficient
on ACT (MGT_Chair/CEO) is positive and statistically significant at the 1% (10%) level.
Reflecting its first-order economic importance, the odds ratio of 1.363 for ACT indicates that the
odds of accounting irregularities for observations where AC-auditor social ties are present are
0.363 times larger than those for observations without such ties.

Our evidence suggests that AC-auditor social ties are associated with an ex ante lower
probability of MAO issuance and an ex post higher incidence of accounting irregularities.
Jointly, these results imply that personal links introduce more Type II errors rather than reduce
Type I errors in audit reporting decisions. As the MAO and irregularity analyses are
performed separately, it is important to ascertain whether clients that practice problematic
accounting indeed have received clean audit opinions from connected auditors. We therefore
re-estimate the MAO regression after including the main effect of Irregularity and its interaction
term with ACT to evaluate whether the impact of ACT on MAO issuance is conditional on the
actual occurrence of financial misrepresentations. Results presented in Column (3) show that
the coefficient on Irregularity is positive and significant, lending support to the intuition that
auditors are more likely to issue MAOs in the presence of clients’ problematic accounting
practices. Importantly, such a tendency declines when AC members are socially connected
with engagement auditors, which is evident in the negative and statistically significant (at the 1%

13Consistent with Wang et al. (2008) and Gul et al. (2013), one plausible explanation is that inventory
build-up may convey positive signals to auditors if this is driven by an increase in customer demand.

20
level) coefficient on ACT*Irregularity.14 Accordingly, our results imply that social ties reduce
the likelihood that auditors render MAOs to firms with financial reporting irregularities,
suggesting higher Type II error rates from an ex post standpoint.

5.2 Social ties and firm value

We explore whether AC-auditor social ties reduce firm value in Table 4. The regression
results for Model (3) are reported in Column (1), where we find that the coefficient on ACT is
significantly negative at the 1% level. The coefficient estimate of ACT is -0.154, implying that
social ties between the engagement auditors and AC members are associated with a discount in
Tobin’s Q amounting to 5.99% relative to the sample mean value (see Table 2). Although the
coefficient on MGT_Other is significantly negative, the one on MGT_Chair/CEO fails to load.
Control variables other than CAPEXP (capital expenditures) all have significant coefficients in
the predicted directions. In light of the finding that ACT impairs audit quality (Table 3), a
natural question is whether the negative impact of ACT on valuation is stronger among cases
that are characterized by weak monitoring due to the AC-auditor social networks. To analyze
this issue, we bisect ACT as follows: we code ACT_PQWM equal to 1 for firms that suffer from
poor financial reporting quality, although they are not issued MAOs (i.e. ACT_PQWM equals 1
if ACT = 1, Irregularity = 1 and MAO = 0), and 0 otherwise; ACT_NPQWM equals 1 if ACT = 1
and ACT_PQWM = 0. We then re-estimate the valuation regression after replacing ACT with
ACT_PQWM and ACT_NPQWM, with the results reported in Column (2) of Table 4. We find
that the coefficient on ACT_NPQWM is negative and statistically significant, suggesting that, in
the presence of AC-auditor social ties, the market discounts firms without an ex post indication
of compromised auditor monitoring. This is not surprising given that the partition of sample
observations by ACT_PQWM and ACT_NPQWM is based on an ex post observations (i.e., the
existence of reporting irregularities) whereas the market’s pricing is based on ex ante assessment
of the potential impact of ACT on monitoring effectiveness. Nevertheless, the coefficient on
ACT_PQWM is also significantly negative at the 1% level and its magnitude is about 4.5 times as
large as that of ACT_NPQWM. An untabulated F-test suggests that the difference between

14 Ai and Norton (2003) show that coefficient estimates of interaction terms in a logistic regression are
often biased when estimated by conventional methods. Accordingly, we report the estimates based on
the Norton et al. (2004) method, which corrects for the bias, for the interaction terms. Thus, the
coefficients on the interaction terms are not comparable to the coefficients on other variables.

21
ACT_PQWM and ACT_NPQWM coefficients are statistically significant (p < 0.01). In other
words, the negative effect of ACT on firm valuation is more pronounced among cases where
firms have poor financial reporting quality but have received clean audit opinions from their
auditors.

5.3 Robustness checks

To more fully capture the social ties construct, we consider three channels in defining social
ties between ACs and engagement auditors. The first and third channels, namely ties formed
via schooling and working experience, respectively, are common in the social tie literature.
However, our second channel, the bonding between teacher and student, may be fairly unique
to China, engendering concerns that the resulting evidence may suffer from relatively poor
external validity. Accordingly, we consider in Panel A of Table 5 whether our core evidence
holds when we separately examine social ties coming through each channel.15 For brevity, we
only report estimates of our experimental variable ACT, although all of the control variables are
included in fitting regressions. We find that all three types of social ties are associated with a
higher incidence of irregularities, while professor-student and employment ties have a
significantly negative impact on MAO issuance. Collectively, these results imply that social
ties are associated with Type II errors, i.e., the subsequent irregularity incidence suggests that
connected auditors should modify their opinions more often, yet they fail to do so. We also
find that alma mater ties and professor-student ties are associated with a lower valuation.
However, the impact of employment ties on valuation is not significant (t = -1.57), implying that
the market is less sensitive to these ties than alma mater or professor-student ties. Extensive
prior research implies that education and workplace ties matter to corporate governance in the
U.S. and elsewhere. Accordingly, dispelling the concern that our main results are driven by

15 Focusing on the educational ties in the first channel also provides the opportunity to sharpen the
analysis by exploring cross-sectional variation in the data. Prior research implies that attending the
same academic department within a university or the same elite university engenders closer alma mater
networks (Nguyen, 2012). Consistent with expectations, we find strong, robust evidence that
AC-auditor educational ties have a larger impact on audit outcomes when the connected audit committee
member graduated from the same academic department or elite university as an engagement auditor (we
relax the three-year age gap restriction in order to generate sufficient power). We follow He and Ke
(2014) by treating the Project 211 universities although representing only 6% of all universities in China,
they account for 96% of the country’s key laboratories and 70% of its scientific research funding as elite
institutions.

22
teacher-student bonding provides some support for the perspective that our research has
implications for countries beyond China.

The AC-auditor social ties under study can be forged through the appointment of a new AC
member who is socially connected with the incumbent engagement auditors, or through a new
engagement auditor being appointed to a client with a socially-connected AC member. To
analyze the audit quality implications of social ties formed at each stage, we re-estimate the
regressions after re-specifying the ACT variable strictly by each type of tie formation. The
results are reported in Panel B of Table 5. We find that connections formed via AC member
appointment are associated with worse outcomes (at the 5% level or better) for all three
regressions. Similarly, although ties formed through auditor appointment are irrelevant to
Tobin’s Q, they are reliably associated with worse audit quality in terms of MAO issuance and
Irregularity incidence. Overall, the evidence suggests that, from a policy standpoint, the
AC-auditor social ties should be considered at both AC member and external auditor
appointment.

There are relatively more observations with AC-auditor ties in early years of the 2004-2010
sample period. To test whether our results are stable over time, we run Fama-MacBeth (1973)
regressions. In this analysis, statistical inferences are based on the distribution of yearly
regression coefficients (Fama and MacBeth 1973; Fama and French 2001). As shown in the
Panel C of Table 5, the coefficients for ACT are in the expected directions and significant at
conventional levels (for d.f. = 6) for all three outcomes. The coefficients in the MAO and
Irregularity regressions are negative and positive, respectively, in all seven years, and the
binomial test rejects the randomness of the coefficient signs at the 1% level (untabulated).
With signs being negative in six out of the seven years, the ACT coefficients in the Tobin’s Q
regressions are also significantly negative at a p-value of 0.055 in the binomial test. The
Fama-MacBeth analysis corroborates our findings from the pooled regressions and suggests
that our results are quite stable over time.

In our main analysis, we follow extant research on social connections by controlling for
covariates that may influence the outcome variables. This nevertheless hinges on the linearity
assumption regarding the relations between these covariates and the outcome variables.
Consequently, we implement a propensity score matching (PSM) approach to rigorously control
for variation in firm characteristics between firms with AC-auditor ties and those without.

23
Aligning two groups of firms in observable characteristics relaxes the linearity assumption.
Specifically, we regress ACT on relevant covariates by the logistic model to generate a
propensity score. For each treatment observation (i.e., ACT = 1), we select a matching control
(i.e., ACT = 0) from the same year that has the closest propensity score. Given that our three
main regression models differ in their covariates and that the sample size is smaller in the
Tobin’s Q regression due to missing values in additional variables, we implement the PSM for
these three models separately. Reassuringly, untabulated results indicate that the matching is
successful in that there are no significant differences at the 10% level between the treatment and
control groups in the relevant covariates for each of the matched samples. To a certain extent,
the similarity between the treated and controls in observable covariates also provides assurance
that they do not differ on unobservable characteristics (Atanasov and Black 2015). Regression
results based on the matched samples are reported in Panel D of Table 5, where we find that our
core results persist at the 1% level in all three audit quality regressions. This implies that
differences in the characteristics between the treatment and control firms, if any, do not bias
toward our findings.

5.4 Variation in the effects of AC-auditor social ties

Next, we analyze the prediction in H2 that the impact of social ties on audit quality varies
with their importance or incentives. As specified in Model (4), we use ACT_X and ACT_X’ to
bisect observations with AC-auditor ties according to AC member and/or auditor
characteristics. To conserve space, we only report these two variables in Table 6, although the
regressions include all of the control variables. The fraction of observations within each group
indicated by ACT_X and ACT_X’ are shown in brackets beneath these variables.

In Panel A, we define the bisecting variables by whether the connected AC member chairs
the committee. We find that the coefficient on ACT_X is significant at the 1% level with
expected signs for all three outcome variables, namely MAO, Irregularity, and Tobin’s Q. In
contrast, none of the coefficients on ACT_X’ is statistically significant at conventional levels. In
short, the impact of social ties tends to be concentrated where engagement auditors are
connected with AC chairs, consistent with the notion that the AC chair plays a vital role in
monitoring the financial reporting process. In Panel B, we sort connected cases by whether the
AC member has an accounting background. We estimate statistically significant (at the 1%

24
level) coefficients, with predicted signs, on ACT_X across the three regressions. In contrast,
ACT_X’ is marginally significant only when the dependent variable is Tobins’ Q. Our evidence
lends support to the intuition that the detrimental impact of social ties on audit quality is
concentrated in AC members with an accounting background. In Panel C, we analyze whether
AC-auditor connections play a more prominent role when the connection is formed via the
senior engagement auditor.16 The results indicate that only ACT_X enters significantly (at the
1% level), with coefficient signs suggesting the connections built via the senior auditor impair
audit quality. These results suggest that the social tie effects are primarily driven by the senior
auditor, who usually plays a more influential role in the audit engagement. This finding also
implies that our evidence is relevant to other jurisdictions where only one engagement auditor
signs or is identified in the audit report. Turning to the hometown tie effects in Panel D, we
split the sample according to whether the connected AC members and engagement auditors
work in the same province or province-level municipality.17 Although all of the coefficients on
ACT_X and ACT_X’ have the predicted signs, the one on ACT_X’ is insignificant in the Tobin’s
Q regression, and the effect of ACT_X is economically stronger than that of ACT_X’ in the MAO
regressions. This analysis provides some support that tighter social connections evident in the
presence of hometown ties exert a larger impact on audit quality. To examine whether the
social tie effects are sensitive to AC members’ incentives, we explore the impact of director
reputation. In Panel E of Table 6, we turn on ACT_X if the number of directorships held by the
connected AC members is below the sample top quartile value. The evidence is consistent
with the intuition that the impact of AC-auditor social ties on audit consequences is attenuated
for the AC members with more reputation capital to protect: the coefficients on ACT_X are
significant at the 1% level with signs suggesting worse audit quality for all three regressions,
whereas those on ACT_X’ are statistically indistinguishable from zero.

Next, we examine the prediction in H3 that the role that social ties play in shaping audit

16 We primarily identify the senior auditor as the one with more experience (measured as the number of
years since the auditor first signs an audit report for a public firm). However, if the signatory auditors
have identical length of experience, we then consider the one with a larger client base (defined as total
assets audited by the individual in all previous years) as the more experienced engagement auditor. In
rare cases in which the signatory auditors have the same length of experience as well as size of client base,
we use the first signatory auditor, reading from left to right or top to bottom, on the audit report.
17 In this definition, we exclude Beijing and Shanghai because it is well known that residents of these two

municipalities often come from other places, implying the absence of hometown ties measured at this
level.

25
quality varies with corporate governance. The results for the variables of interest are
presented in Table 7.18 We first partition the observations with ACT equal to one by the
indicators of ACT_X and ACT_X’ according to ownership concentration, institutional
shareholding, foreign shareholding, and analyst coverage in Panels A, B, C, and D
respectively. 19 For all three audit quality proxies, none of the ACT_X’ coefficients are
statistically significant in the predicted direction. Thus, we find no evidence of impaired audit
quality in the presence of AC-auditor ties for firms with a more dispersed ownership structure,
higher institutional shareholding, foreign investors, or greater analyst coverage. In contrast,
we find supportive evidence at the 1% level for the worse governance cases denoted by the
ACT_X indicators; i.e., firms with a shareholder holding a large equity stake, lower institutional
shareholdings, no foreign investors, and lesser analyst coverage. Turning to another
dimension of corporate governance, CEO power, we group observations by whether the CEO is
internally promoted and CEO tenure in Panels E and F, respectively. Consistent with
expectations, we find supportive evidence in all three audit outcome regressions for the
observations with stronger CEO power stemming from internal promotion or longer tenure.
In contrast, for observations where CEOs are not internally promoted or have shorter tenure,
the AC-auditor ties only significantly reduce MAO issuance with a weaker economic impact.
At the region level, we specify the marketization index to measure the quality of a province’s
formal institutions.20 As shown in Panel G, for ACT_X’, the indicator for connected cases
where the firm is headquartered in top-quartile provinces in terms of marketization, its
coefficient is significant only in the MAO regression at the 10% level, while the coefficients on

18 In unreported analysis, we also consider the moderating role that board independence plays given
prior research implying that more independent boards improve corporate governance in China (e.g.,
Chen et al. 2006; Jiang et al. 2016). Reinforcing our other results on the prediction in H3, we find
evidence at the 1% level for two of the three outcomes that audit quality suffers when the proportion of
independent directors is below the sample top quartile. In contrast, we find no evidence that
AC-auditor social ties undermine audit quality when firms protect shareholders’ interests by appointing
relatively more independent boards.
19 Our sample does not include the B- or H-share firms, i.e., firms that issued shares to foreign investors
directly. Accordingly, foreign investments in our sample firms are made through the Qualified Foreign
Institutional Investor program.
20 The marketization index is compiled by Fan et al. (2011) to measure the province-level development of
the capital market, private sector, product market, factor market, legal institutions, and contract
enforcement. It has been routinely applied in Chinese accounting/auditing research (e.g., Wang et al.
2008; Jian and Wong 2010). The index is available up to 2009. As it is quite stable over time, we
extrapolate the 2009 index to subsequent years as necessary.

26
ACT_X all load at the 1% level in the predicted directions. This evidence supports the
prediction in H3 that the impact of AC-auditor social ties on audit outcomes is larger for firms
located in regions with poorer institutional infrastructure.

Altogether, the evidence in Table 7 lends support to the prediction in H3. The effects of
AC-auditor social connections on audit quality tend to be more concentrated among the firms
with more severe agency problems due to their inferior corporate governance.

6. Additional analysis

6.1 The interactive effect between AC-auditor and CEO/Board Chair-auditor social ties

To provide insight into whether the role that AC-auditor social connections play in shaping
audit quality varies according to whether the CEO/Board Chair also has social ties with the
auditor, we introduce the interaction term, ACT*MGT_Chair/CEO, to the regressions. In Table
8, we report some evidence implying that the impact is more pronounced in the presence of
social ties between the CEO/Board Chair and the auditor: ACT*MGT_Chair/CEO is positive and
statistically significant at the 1% level in the Irregularity regression in Column (2), although this
variable fails to load in Columns (1) and (3) (MAO and Tobin’s Q, respectively). Importantly,
after adding the interaction term, our key results persist with the coefficients on the main effect
of ACT remaining statistically significant in the predicted directions, suggesting that
AC-auditor social ties are associated with worse audit quality. Therefore, our findings
concerning the impact of AC-auditor social ties are largely independent of the social networks
between engagement auditors and the CEO/Board Chair. Interestingly, the main effect of
MGT_Chair/CEO loads negatively in the Irregularity regression. Given the significantly
positive coefficient on the interaction term of ACT*MGT_Chair/CEO, the CEO/Board
Chair-auditor social ties increase the incidence of accounting irregularities only when AC
members are socially connected with the engagement auditors.21 This finding reinforces our
argument that the AC, as a bridge between firm management and the external auditors, plays
an important role in the auditing process.

21Note that we report the estimates based on the Norton et al. (2004) method for ACT*MGT_Chair/CEO.
As such that the coefficient on ACT*MGT_Chair/CEO is not directly comparable to that on
MGT_Chair/CEO. The conventional coefficient estimate of ACT*MGT_Chair/CEO is 15.179 (p < 0.01).

27
6.2 Audit firm type and the effects of AC-auditor social ties

We next investigate whether the social tie effects differ between engagement auditors from
large versus small audit firms. This analysis provides some insight into whether Chinese
evidence can be generalized to other markets that are dominated by large audit firms. Similar
to their U.S. counterparts, large audit firms in China provide better audit services (DeFond et al.
2000; Wang et al. 2008; Chan and Wu 2011). To the extent that large audit firms implement, for
example, better quality control procedures, the social-tie effects could be smaller in this
situation. Conversely, due to their large scale and operational complexity, large audit firms
may not closely monitor their engagement auditors, allowing them more leeway in
decision-making.

Consistent with DeFond et al. (2000) and Wang et al. (2008), we specify large audit firms as
the Big 4 or the Top 10 domestic audit firms according to their share of the total number of
listed clients. We estimate Models (1) to (3) separately for the large and small audit firm
sub-samples. In untabulated results, we find that the ACT variable is significantly associated
with worse audit consequences for both sets of clients, except that its coefficient is insignificant
in the Irregularity regression estimated on the large audit firm sub-sample. Importantly, we
find that the coefficients on ACT are statistically indistinguishable between the two sub-samples,
potentially reflecting the tension over how audit firm size affects engagement auditor practices.
Clarifying that our findings are not specific to small audit firms lends some support to the
perspective that our research is relevant to countries besides China.22

6.3 Can audit quality still be improved in the presence of AC-auditor social ties?

Since we focus on social connections between ACs and engagement auditors, our sample
does not include observations without ACs. Given that ACs are known to improve financial
reporting and our evidence that AC-auditor social ties undermine monitoring, we next examine

22 To examine whether the Big 4 firms differ from domestic firms, we also partition the sample into three
client groups: Big 4 audit firms, large domestic audit firms, and small audit firms. For the Big 4
subsample, the coefficient on MAO fails to load and that on Tobin’s Q is positive and statistically
significant. In the Irregularity regression, the coefficient on ACT is not estimable due to a lack of
variation in this variable when ACT is equal to one. Thus, there are some results implying that social
ties do not undermine audit quality supplied by the Big 4, which have more valuable reputations at stake
according to extensive prior research. However, we interpret this evidence as only suggestive given that
only 6% of the sample firms appoint a Big 4 auditor.

28
whether, relative to firms without ACs, audit quality is higher in firms with AC-auditor social
ties. We begin this analysis by comparing the firms without ACs to firms with ACs that are
free from AC-auditor social ties (i.e., ACT = 0). The regression models are specified the same
as before, except that we replace the ACT variable with ACOMMIT, an indicator for observations
that have ACs. We report these results in the first three columns of Table 9. We find that
audit outcomes are consistently better for firms with such ACs: the MAO issuance likelihood is
higher, the firms are less likely to experience an Irregularity, and they enjoy a higher Tobin’s Q.
Next, we compare the firms without ACs to the firms with ACs that are socially connected with
engagement auditors (i.e., ACT = 1), with the results reported in the last three columns of Table
9. The significantly negative coefficient on ACOMMIT in Column (4) suggests that, relative to
firms without an AC, engagement auditors are less likely to issue MAO in the presence of social
ties with AC members. However, we find no significant difference in the incidence of
Irregularity between the two groups of firms in Column (5), while firms with connected ACs are
associated with higher Tobin’s Q in Column (6). Accordingly, we find no consistent evidence
that audit quality differs between firms without ACs and firms that have ACs involving
AC-auditor ties. Altogether, the evidence in Table 9 suggests that while ACs help improve
audit quality, these benefits are eroded when these ACs contain socially connected AC
members.

6.4 AC-auditor social ties and audit pricing

In our final set of tests, we analyze whether social networks between ACs and an
engagement auditor influence the auditor-client contracting process evident in audit fees. As
discussed earlier, an upside of AC-auditor social ties is better information sharing between the
two parties, which may translate into a more efficient audit and lower audit fees. While our
overall evidence points to the detrimental effects of social ties in terms of audit quality, clients
could still benefit from cheaper audits. We empirically clarify this issue by examining whether
audit fees are systematically lower in the presence of AC-auditor social ties in this model:

Fee = + 1ACT + 2MGT_Chair/CEO + 2MGT_Other + Ci + t + k + , (5)

where Fee is the natural logarithm of annual audit fees; 23 Ci includes these potential

23 The Fee variable does not include non-audit fees. In China, clients seldom purchase non-audit
services from their auditors. During our sample period, only 2.83% of firms report non-audit fees. All

29
determinants of audit fees: Big4, Top10, MAO, LTA, ROA, Liquidity, Leverage, REC, INV, and Loss.
All other variables are defined as before. We expect the coefficients on Big4 and Top10 to be
positive stemming from the superior reputation or expertise of large auditors. We also expect
LTA to enter positively because large firms are usually more complex, requiring auditors to
exert more effort on these engagements. MAO, Leverage, REC, INV, and Loss are expected to
have positive coefficients since larger values in these variables indicate higher audit risks,
whereas Liquidity and ROA are expected to load with negative coefficients since they mitigate
audit risks.

Results for the audit fee regression are reported in Column (1) of Table 10. We observe
that the coefficient on ACT is positive and significant at the 5% level, suggesting that audit fees
are higher when engagement auditors are socially connected with ACs. Although the
economic impact is fairly negligible, the above evidence does not support that social ties benefit
clients through lower fees.24 Still, it remains plausible that such a fee premium spuriously
captures the additional audit effort in the presence of poor financial reporting quality, as our
preceding analysis shows. To address this issue, we control for audit effort in the fee
regression. We measure audit effort in two alternative ways: as the natural logarithm of the
number of calendar days between the fiscal year end and the audit report date scaled by 365
(e.g., Ghosh and Tang 2015); and the natural logarithm of the number of audit field work days
according to data compiled by the CICPA. In Table 10, the results based on these two audit
effort measures are reported in Columns (2) and (3). Consistent with the theory that audit
costs and, in turn, fees rise when auditors expend more effort, we find that both audit effort
measures are positively correlated at the 1% level with audit fees. More importantly, the
coefficients on ACT remain significantly positive (at the 5% level) in both regressions. To
tackle the concern that social ties may directly influence audit effort, we next regress the audit
effort measures on social tie variables and those that proxy for audit risk and/or complexity in
successive estimations. As shown in Columns (4) and (5), the coefficients on ACT are
statistically indistinguishable from zero. In short, we find no evidence suggesting that the

of our core evidence reported persists after replacing audit fees with total fees.
24 According to our coefficient estimates, fees are only 2.74% (= e0.027 – 1) higher in the presence of

AC-auditor ties. Since the sample mean audit fees amount to RMB 519,177 yuan (equivalent to
US$68,174 at the June 30, 2007 midpoint of our sample period), this translates into a premium of about
RMB 14,209 yuan (US$1,866).

30
presence of AC-auditor social ties has a perceptible impact on audit effort.

Collectively, the results in Table 10 suggest a fee premium stemming from AC-auditor
social networks, providing some evidence consistent with the social norm of reciprocity (e.g.,
Sanchez et al. 2007). In our setting, socially connected AC members may approach firm
management to negotiate a more favorable audit fee arrangement. The reciprocity theory
predicts that a small favor can produce a sense of obligation to provide a larger return favor
(Becker 1990). The more favorable audit contractual terms may thus lead auditors to feel
obligated to act positively toward their clients. As such, social ties with AC members lead to
unconscious errors in auditors’ judgment and decision making, reducing their ability to detect
misstatements. However, this interpretation has to be set against the presence of AC-auditor
social ties only having a small economic impact on audit fees.

7. Conclusions

As the liaison between the company’s management and its external auditor, the AC plays a
major role in overseeing the financial reporting process. In examining the importance of social
connections, extant research has predominantly focused on whether AC members’ formal or
informal ties with firm management influence financial reporting quality. We investigate
whether AC-auditor social connections shape audit outcomes in China, where the biographic
data necessary for analyzing our research questions are available and the guanxi (networking)
culture prevails in socioeconomic activities.

AC-auditors social ties could affect audit quality in one of two competing ways. In one
direction, social ties facilitate information flows between the AC and auditor and protect
auditor independence from management, leading to higher quality audits. In the other
direction, AC-auditor social ties could impede the auditor from exercising due care and even
inject unintentional bias into auditors’ judgment and decision making, resulting in lower
quality audits.

Our evidence implies that the costs of AC-auditor social connections outweigh their
potential benefits. Specifically, we estimate that the likelihood of receiving a modified audit
opinion (MAO) for observations where AC-auditor social ties are present is about 40% smaller
than that for observations where such ties are absent, and such ties increase the odds of clients’

31
financial reporting irregularities by 0.363 times. Moreover, social ties reduce the likelihood
that auditors render MAOs to firms with ex post financial reporting irregularities. These
results lend support to the intuition that the lower propensity to issue non-clean opinions
represents higher Type II error rates in connected auditors’ decision-making. This evidence is
corroborated by a valuation analysis suggesting that investors perceive that audit quality
suffers when auditors are socially linked to their client’s ACs. In results supporting another
prediction, we find that the audit quality impact is concentrated where social ties are more
salient, which is evident in a connected audit committee member chairing the committee,
having an accounting background, sharing local ties with an engagement auditor, or having
lesser reputational capital at stake. Also, we report evidence implying that the impact of social
ties between AC members and engagement auditors on audit quality is more pronounced
where the firm has more severe agency conflicts evident in relatively poor governance
structures or the entrenchment of powerful CEOs. This cross-sectional variation pattern
reinforces the inference that social networks between AC members and engagement auditors
undermine audit quality. Finally, consistent with the social norm of reciprocity, we document
that engagement auditors who are socially connected with the AC attract an audit fee premium,
although the amounts involved are economically small.

Our research has several policy and practical implications. First, our findings suggest that
financial report users would benefit from the disclosure of the identity of the engagement
auditor. This information would enable users to spot the interpersonal links between the
engagement auditor and AC members, putting them in a better position to gauge the audit
quality impact. Our valuation analysis suggests that Chinese investors price such information.
Second, whether AC members and the outside auditors are independent from each other has
largely been overlooked in regulatory circles. Our analysis indicates that the cozy
interpersonal relations between these two parties can jeopardize audit quality to the extent that
having a connected AC is almost equivalent to not establishing an AC at all. It would be
constructive for regulators to consider the role that informal personal ties play when setting
policies surrounding the independence of ACs and external auditors. Finally, for practitioners
eager to strengthen corporate governance, they should also be wary of the detrimental effect of
social networks in the context of external auditing when developing firm-level governance
structures. For public accounting firms attempting to improve their performance, our research

32
suggests that social ties between engagement auditors and client AC members should be
considered in the allocation of their scarce quality control resources and the assignment of
individuals to audit engagements.

33
References

Adams, R.B., H. Almeida, and D. Ferreira. 2005. Powerful CEOs and their impact on corporate
performance. Review of Financial Studies 18 (4): 1403–1432.
Ai, C., and E. Norton. 2003. Interaction terms in logit and probit models. Economics Letters 80 (1):
123–129.
Atanasov, V. and B. Black. 2015. Shock-based causal inference in corporate finance and accounting
research. Critical Finance Review: forthcoming.
Baber W.R., J. Krishnan, and Y. Zhang. 2014. Investor perceptions of the earnings quality
consequences of hiring an affiliated auditor. Review of Accounting Studies 19 (1): 69–102.
Becker, L.C. 1990. Reciprocity. Chicago: University of Chicago Press.
Bédard, J., S.M., Chtourou and L. Courteau. 2004. The effect of audit committee expertise,
independence, and activity on aggressive earnings management. Auditing: A journal of Practice &
Theory 23 (2): 13–35.
Bhowmik, D. and E. Rogers. 1971. Homophily-heterophily: Relational concepts for communication
research. Public Opinion Quarterly 34 (4): 523–538.
Bonner, S.E. and B.L. Lewis. 1990. Determinants of auditor expertise. Journal of Accounting Research
28 (Supplement): 1–20.
Bonner, S.E. and N. Pennington. 1991. Cognitive processes and knowledge as determinants of
auditor expertise. Journal of Accounting Literature 10: 1–50.
Bruynseels, L. and E. Cardinaels. 2014. The audit committee: Management watchdog or personal
friend of the CEO. The Accounting Review 89 (1): 113–145.
Carcello, J.V. and C. Li. 2013. Costs and benefits of requiring an engagement partner signature:
Recent experience in the United Kingdom. The Accounting Review 88 (5): 1511–1546.
Carcello, J.V. and T. Neal. 2000. Audit committee composition and auditor reporting. The Accounting
Review 75 (4): 453–467.
Carcello, J.V., and T. Neal. 2003. Audit committee characteristics and auditor dismissals following
“new” going-concern reports. The Accounting Review 78 (1): 95–117.
Carcello, J.V., T. Neal., Z.V. Palmrose, and S. Scholz. 2011. CEO involvement in selecting board
members, audit committee effectiveness, and restatements. Contemporary Accounting Research 28
(2): 396–430.
Chan, K.H. and D. Wu. 2011. Aggregate quasi rents and auditor independence: Evidence from audit
firm mergers in China. Contemporary Accounting Research 28 (1): 175–213.
Chang, H., X. Chen, and N. Zhou. 2013. Determinants and consequences of audit committee
effectiveness: Evidence from China. Available at: http://mitsloan.mit.edu/events/2013-asia-
conference-in-accounting/pdf/Determinants_and_Consequences_of_Audit_Committee_Effect.
pdf.
Chen, C., S. Chen, and X. Su. 2001. Profitability regulation, earnings management, and modified
audit opinions: Evidence from China. Auditing: A Journal of Practice & Theory 20 (2): 9–30.
Chen, G., M. Firth, N. Gao, and M. Rui. 2006. Ownership structure, corporate governance, and fraud:
Evidence from China. Journal of Corporate Finance 12 (3): 424-448.
Chen, J. and H. Zhang. 2010. The impact of regulatory enforcement and audit upon IFRS compliance
– Evidence from China. European Accounting Review 19 (4): 665–692.

34
Chen, S., S. Sun, and D. Wu. 2010. Client importance, institutional improvements, and audit quality
in China: An office and individual auditor level analysis. The Accounting Review 85 (1): 127–158.
Chen, T., J. Harford, and C. Lin. 2015. Do analysts matter for governance? Evidence from natural
experiments. Journal of Financial Economics 115 (2): 383–410.
Chi, W., H. Huang, Y. Liao, and H. Xie. 2009. Mandatory audit partner rotation, audit quality, and
market perception: Evidence from Taiwan. Contemporary Accounting Research 26 (2): 359–391.
Chidambaran, N.K., S. Kedia, and N.R. Prabhala. 2012. CEO-director connections and corporate
fraud. Working paper. Available at: http://dx.doi.org/10.2139/ssrn.2023030.
Church, B., S. Davis, and S. McCraken. 2008. The auditor’s reporting model: A literature overview
and research synthesis. Accounting Horizons 22 (1): 69–90.
Cohen, L., A. Frazzini, and C. Malloy. 2008. The small world of investing: Board connections and
mutual fund returns. Journal of Political Economy 116 (5): 51–979.
Cohen, L., A. Frazzini, and C. Malloy. 2010a. Sell-side school ties. Journal of Finance 65 (4): 1409–1437.
Cohen, J., G. Krishnamoorthy, and A. Wright. 2010b. Corporate governance in the
post-Sarbanes-Oxley era: Auditors’ experiences. Contemporary Accounting Research 27 (3): 751–
786.
DeAngelo, L. 1981. Auditor size and audit quality. Journal of Accounting and Economics 3 (3): 183–199.
DeFond, M. and J. Zhang. 2014. A review of archival auditing research. Journal of Accounting and
Economics 58 (2–3): 275–326.
DeFond, M. and J.R. Francis. 2005. Audit 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): 153–193.
DeFond, M., T.J. Wong, and S. Li. 2000. The impact of improved auditor independence on audit
market concentration in China. Journal of Accounting and Economics 28 (3): 269–305.
Deng, M., N. Melumad, and T. Shibano. 2012. Auditors’ liability, investments, and capital markets:
A potential unintended consequence of the Sarbanes-Oxley Act. Journal of Accounting Research 50
(5): 1179–1215.
DeZoort, T., P. Harrison, and M. Taylor. 2006. Accountability and auditors materiality judgments:
The effects of differential pressure strength on conservatism, variability, and effort. Accounting,
Organizations and Society 31 (4): 373–390.
DeZoort, T., D. Hermanson, and R. Houston. 2008. Audit committee member support for proposed
audit adjustments: Pre-SOX versus post-SOX judgments. Auditing: A Journal of Practice & Theory
27 (1): 85–104.
Dhaliwal, D., V. Naiker, and F. Navissi. 2010. The association between accruals quality and the
characteristics of accounting experts and mix of expertise on audit committees. Contemporary
Accounting Research 27 (3): 787–827.
Doidge, C., G.A. Karolyi, and R.M. Stulz. 2007. Why do countries matter so much for corporate
governance? Journal of Financial Economics 86 (1): 1–39.
Engelberg, E., P. Gao, and C.A. Parsons. 2012. Friends with money. Journal of Financial Economics 103
(1): 169–188.
Fama E. and K. French. 2001. Disappearing dividends: Changing firm characteristics or lower
propensity to pay. Journal of Financial Economics 60 (1): 3–43.

35
Fama, E. and M. Jensen. 1983. The separation of ownership and control. Journal of Law and Economics
26 (2): 301–325.
Fama E. and J. MacBeth. 1973. Risk, return, and equilibrium: Empirical tests. Journal of Political
Economy 81 (3): 607–636.
Fan, G., X. Wang, and H. Zhu. 2011. The 2011 Report on the Relative Process of Marketization of Each
Region in China. Beijing: The Economic Science Press. In Chinese.
Fan, J. and T.J. Wong. 2002 Corporate ownership structure and the informativeness of accounting
earnings in East Asia. Journal of Accounting and Economics 33 (3): 401–425.
Fan, J. and T.J. Wong. 2005. Do external auditors perform a corporate governance role in emerging
markets? Evidence from East Asia. Journal of Accounting Research 43 (1): 35–72.
Farh, J.L., A.S. Tsui, K.R. Xin, and B.S. Cheng. 1998. The influence of relational-demography and
guanxi: The Chinese case. Organization Science 9 (4): 471–488.
Fich, E.M. and A. Shivadasani. 2007. Financial fraud, director reputation, and shareholder wealth.
Journal of Financial Economics 86 (2): 306–336.
Fogel-Yaari, H. and P. Zhang. 2013. Is an auditor’s propensity to issue going concern opinions
related to audit quality? Working paper. Available at: http://dx.doi.org/10.2139/ssrn.2197781.
Fracassi, C. and G.A. Tate. 2012. External networking and internal firm governance. Journal of Finance
67 (1): 153–194.
Francis, J.R., P. Michas, and M. Yu. 2013. Office size of Big 4 auditors and client restatements.
Contemporary Accounting Research 30 (4): 1626–1661.
Francis, J.R. and P. Michas. 2013. The contagion effect of low-quality audits. The Accounting Review
88 (2): 521–552.
Geiger, M.A., C. S. Lennox, and D. S. North. 2008. The hiring of accounting and finance officers from
audit firms: How did the market react? Review of Accounting Studies 13 (1): 119–143.
Ghosh, A. and C.Y. Tang. 2015. Assessing financial reporting quality of family firms: The auditors’
perspective. Journal of Accounting and Economics 60 (1): 95–116.
Gibbons, D.E. 2004. Friendship and advice networks in the context of changing professional values.
Administrative Science Quarterly 49 (2): 238–262.
Granovetter, M. 1985. Economic action and social structure: The problem of embeddedness.
American Journal of Sociology 91 (3): 481–510.
Granovetter, M. 2005. The impact of social structure on economic outcomes. Journal of Economic
Perspectives 19 (1): 33–50.
Guan, Y., L. Su, D. Wu, and Z. Yang. 2016. Do school ties between auditors and client executives
influence audit outcomes? Journal of Accounting and Economics 61 (2–3): 506–525.
Guedhami, O., J.A. Pittman, and W. Saffar. 2014. Auditor choice in politically connected firms.
Journal of Accounting Research 52 (1): 107–162.
Gul, F.A., D. Wu, and Z. Yang. 2013. Do individual auditors affect audit quality? Evidence from
archival data. The Accounting Review 88 (6): 1993–2023.
Hartzell, J.C. and L.T. Starks. 2003. Institutional investors and executive compensation. Journal of
Finance 58 (6): 2351–2374.
He, X. and B. Ke. 2014. What determines individual auditors’ career outcomes? Working paper,
Shanghai University of Finance and Economics and Nanyang Technological University.

36
He, X., J.A. Pittman, and O.M. Rui. 2015. Reputational implications for partners after a major audit
failure: Evidence from China. Journal of Business Ethics: forthcoming.
Hermalin, B.E. and M.S. Weisbach. 1988. The determinants of board composition. Rand Journal of
Economics 19 (4): 589–606.
Hermalin, B.E. and M.S. Weisbach. 1998. Endogenously chosen boards of directors and their
monitoring of the CEO. American Economic Review88 (1): 96–118.
Huang, W. and T. Zhu. 2015. Foreign institutional investors and corporate governance in emerging
markets: Evidence of a split-share structure reform in China. Journal of Corporate Finance 32: 312–
326.
Hung, M.Y., T.J. Wong, and F. Zhang. 2015. The value of political ties and market credibility:
Evidence from corporate scandals in China. Contemporary Accounting Research: forthcoming.
Hwang, B-H. and S. Kim. 2012. Social ties and earnings management. Working paper. Available at:
http://dx.doi.org/10.2139/ssrn.1215962.
Irani, R.M. and D. Oesch. 2013. Monitoring and corporate disclosure: Evidence from a natural
experiment. Journal of Financial Economics 109 (2): 398–418.
Jacobs, J.B. 1979. A preliminary model of particularistic ties in Chinese political alliance: Kan-ch'ing
and kuan-hsi in a rural Taiwanese township. China Quarterly 78: 237–273.
Jian, M. and T.J. Wong. 2010. Propping through related party transactions. Review of Accounting
Studies 15 (1): 70–105.
Jiang, G.H., Charles M.C. Lee, and H. Yue. 2010. Tunneling through intercorporate loans: The China
experience. Journal of Financial Economics 98 (1): 1–20.
Jiang, W., H. Wan, and S. Zhao. 2016. Reputation concerns of independent directors: Evidence from
individual director voting. Review of Financial Studies 29 (3): 655–696.
Khanna, T. and K.G. Palepu 2000. Emerging market business groups, foreign investors, and
corporate governance. In Concentrated Corporate Ownership, edited by R. Morck, 265–294.
University of Chicago Press.
King, R. 2002. An experimental investigation of self-serving biases in an auditing trust game: The
effect of group affiliation. The Accounting Review 77 (2): 265–284.
Klein, A. 2002. Audit committee, board of director characteristics, and earnings management. Journal
of Accounting and Economics 33 (3): 375–400.
Knechel, W.R., L. Niemi, and M. Zerni. 2013a. Empirical evidence on the implicit determinants of
compensation in Big 4 audit partnerships. Journal of Accounting Research 51 (2): 349–387.
Knechel, W.R., G. Krishnan, M. Pevzner, L. Shefchik, and U. Velury. 2013b. Audit quality: Insights
from the academic literature. Auditing: A Journal of Practice & Theory 32 (Supplement 1): 385–421.
KPMG’s Audit Committee Institute. 2015. 2015 Global audit committee survey. Available at:
https://www.kpmg-institutes.com/content/dam/kpmg/auditcommitteeinstitute/pdf/2015/2
015-global-audit-committee-survey.pdf.
Krishnan, G.V., K.K., Raman, K. Yang, and W. Yu. 2011. CFO/CEO-board social ties, Sarbanes-Oxley,
and earnings management. Accounting Horizons 25 (3): 537–557.
Krishnan, J. 2005. Audit committee quality and internal control: An empirical analysis. The
Accounting Review 80 (2): 649–675.
Kwon, S.Y. and H. Yi. 2012. Do social ties between CEOs and engagement audit partners affect audit
fees and audit quality? Working paper, Korea University.

37
La Porta, R., F. Lopez-de-Silanes, A. Shleifer, and R.W. Vishny. 2000. Investor protection and
corporate governance. Journal of Financial Economics 58 (1–2): 3–27.
Lennox, C.S. and C. Park. 2007. Audit firm appointments, audit firm alumni, and audit committee
independence. Contemporary Accounting Research 24 (1): 235–258.
Lennox, C.S. and J.A. Pittman. 2010. Big five audits and accounting fraud. Contemporary Accounting
Research 27 (1): 209–247.
Li, Z., T.J. Wong, and G. Yu. 2016. The dyadic ties of managers and financial analysts and their
externality on a firm’s information environment. Harvard Business School working paper #16–
135.
Lin, C.H. and P.-C. Lee. 2013. Comparison of effectiveness between two supervisory systems in
China: With and without audit committees. Journal of Business and Economics 4 (1): 928–947.
Lin, Z., J. Xiao, and Q. Tang. 2008. The roles, responsibilities and characteristics of audit committee
in China. Accounting, Auditing & Accountability Journal 21 (5): 721–751.
Liu, Q. and G. Tian. 2012. Controlling shareholder, expropriations and firm’s leverage decision:
Evidence from Chinese non-tradable share reform. Journal of Corporate Finance 18 (4): 782–803.
Lo, A.W., Wong, R.M., and M. Firth. 2010. Can corporate governance deter management from
manipulating earnings? Evidence from related-party sales transactions in China. Journal of
Corporate Finance 16 (2): 225–235.
Loewenstein, G. 1996. Behavioral decision theory and business ethics: Skewed trade-offs between
self and other. In D.M. Messick and A.E. Tenbrunsel, eds., Codes of Conduct. New York: Russell
Sage Foundation.
May, D.O. 1995. Do managerial motives influence firm risk reduction strategies? Journal of Finance 50
(4): 1291–1308.
McPherson, M., L. Smith-Lovin, and J.M. Cook. 2001. Birds of a feather: Homophily in social
networks. Annual Review of Sociology 27: 415–444.
Menon, K. and D. Williams. 2004. Former audit partners and abnormal accruals. The Accounting
Review 79 (4): 1095–1118.
MOF (the Ministry of Finance of the People’s Republic of China). 1995a. Independent Auditing
Standards No. 7: Audit Report. Beijing, China. In Chinese.
MOF (the Ministry of Finance of the People’s Republic of China). 1995b. Practice Guidelines of Certified
Public Accountants in China No. 1: Annual Report Audit. Beijing, China. In Chinese.
Moore, D.A., P.E. Tetlock, L. Tanlu, and M.H. Bazerman. 2006. Conflicts of interest and the case of
auditor independence: Moral seduction and strategic issue recycling. Academy of Management
Review 31 (1): 10–29.
Naiker, V., and D.S. Sharma. 2009. Former audit partners on the audit committee and internal
control committee. The Accounting Review 84 (2): 559–587.
Nelson, M. 2006. Ameliorating conflicts of interest in auditing: Effects of recent reforms on auditors
and their clients. Academy of Management Review 31 (1): 30–42.
Nguyen, B.D. 2012. Does the Rolodex matter? Corporate elite's small world and the effectiveness of
boards of directors. Management Science 58 (2): 236–252.
Norton, E., H. Wang, and C. Ai. 2004. Computing interaction effects and standard errors in logit and
probit models. Stata Journal 4 (2): 154–167.
Parrino, R., R.W. Sias, and L.T. Starks. 2003. Voting with their feet: Institutional ownership changes

38
 around forced CEO turnover. Journal of Financial Economics 68 (1): 3–46.
PCAOB (the Public Company Accounting Oversight Board). 2007. Auditing Standard No. 5: An Audit
of Internal Control Over Financial Reporting That Is Integrated with An Audit of Financial Statements.
Washington, D.C.
Rajan, R.G., and L. Zingales. 1998. Financial dependence and growth. American Economic Review 88
(3): 559–586.
Rennie, M.D., L.S., Kopp and W.M. Lemon. 2010. Exploring trust and the auditor-client relationship:
Factors influencing the auditor’s trust of a client representative. Auditing: A Journal of Practice &
Theory 29 (1): 279–293.
Rose, J.M. 2007. Attention to evidence of aggressive financial reporting and intentional misstatement
judgments: Effects of experience and trust. Behavioral Research in Accounting 19 (1): 215–229.
Sanchez, M.H., C.P. Agoglia, and R.C. Hatfield. 2007. The effect of auditors’ use of a
reciprocity-based strategy on auditor-client negotiations. The Accounting Review 82 (1): 205–240.
Schmidt, B. 2015. Costs and benefits of friendly boards during mergers and acquisitions. Journal of
Financial Economics 117 (2): 424–447.
Sias, P.M., and D.J. Cahill. 1998. From coworkers to friends: The development of peer friendships in
the workplace. Western Journal of Communication 62 (3): 273–299.
Subrahmanyam, A. 2008. Social networks and corporate governance. European Financial Management
14 (4): 633–662.
Thompson, L. 1995. “They saw a negotiation”: Partisanship and involvement. Journal of Personality
and Social Psychology 68 (5): 839–853.
Uzzi, B. 1996. The sources and consequences of embeddedness for the economic performance of
organizations: The network effect. American Sociological Review 61 (4): 674–698.
Wang, Q., T.J. Wong, and L. Xia. 2008. State ownership, the institutional environment, and auditor
choice: Evidence from China. Journal of Accounting and Economics 46 (1): 112–134.
Wang, X. and M. Wu. 2012. The quality of financial reporting in China: An examination from an
accounting restatement perspective. China Journal of Accounting Research 4 (4): 167–196.
Wooldridge, J.M. 2002. Econometric Analysis of Cross Section and Panel Data. Cambridge: MIT Press.
Xin, K.R. and J.L. Pearce. 1996. Guanxi: Connections as substitutes for formal institutional support.
Academy of Management Journal 39 (6): 1641–1658.
Xin, Q., M. Huang, and H.Y. Yick. 2013. Misstatements by public firms and regulatory sanctions
against independent directors. Management World (2013–5): 131–175. (In Chinese).
Yuan, R., J. Xiao, and H. Zou. 2008. Mutual funds’ ownership and firm performance: Evidence from
China. Journal of Banking and Finance 32 (8): 1552–1565.
Zerni, M. 2012. Audit partner specialization and audit fees: Some evidence from Sweden.
Contemporary Accounting Research 29 (1): 312–340.

39
Figure 1 The relationship among firm management, external auditor, and audit committee

C
Audit committee External auditor

B A

Management

40
 Table 1 Sample distribution
Panel A. Distribution by year
Year ACT = 1 ACT = 0 Total Percentage
2004 73 332 405 18.02%
2005 87 479 566 15.37%
2006 101 570 671 15.05%
2007 146 840 986 14.81%
2008 193 1,097 1,290 14.96%
2009 195 1,227 1,422 13.71%
2010 222 1,436 1,658 13.39%
Total 1,017 5,981 6,998 14.53%
Industry N % ACT = 1 ACT = 0
Farming, Forestry, Animal Husbandry, and Fishing 154 2.20% 1.28% 2.36%
Mining 132 1.89% 1.28% 1.99%
Food and Beverage 272 3.89% 3.54% 3.95%
Textile, Apparel, Fur and Leather 289 4.13% 4.82% 4.01%
Furniture Manufacturing 25 0.36% 0.79% 0.28%
Paper and Allied Products; Printing 132 1.89% 2.06% 1.86%
Petroleum, Chemical, Plastics, and Rubber Products 737 10.53% 12.98% 10.12%
Electronics 296 4.23% 4.42% 4.20%
Metal and Non-metal 620 8.86% 8.85% 8.86%
Machinery, Equipment, and Instrument Manufacturing 1,146 16.38% 15.04% 16.60%
Medicine and Biological Products 475 6.79% 5.60% 6.99%
Other Manufacturing 60 0.86% 0.39% 0.94%
Utilities 317 4.53% 3.83% 4.65%
Construction 146 2.09% 1.38% 2.21%
Transportation and Warehousing 295 4.22% 4.42% 4.18%
Information Technology 449 6.42% 4.03% 6.82%
Wholesale and Retail Trades 463 6.62% 10.42% 5.97%
Banking and Financial Institutions 8 0.11% 0.00% 0.13%
Real Estate 373 5.33% 5.70% 5.27%
Public Facilities and Other Services 197 2.82% 2.75% 2.83%
Communication and Cultural Industries 54 0.77% 0.10% 0.89%
Conglomerates 358 5.12% 6.29% 4.92%
Total 6,998 100.00% 100.00% 100.00%
The sample includes all A-share companies that have established audit committees, after excluding
observations with missing values in variables necessary for multivariate regression analysis. ACT is
equal to one if at least one of audit committee member has social ties with one of the engagement
auditors, and zero otherwise.

41
 Table 2 Descriptive statistics
Variable Mean Median Min. Max. Std. Dev. N
ACT 0.145 0.000 0.000 1.000 0.352 6,998
MGT_Chair/CEO 0.006 0.000 0.000 1.000 0.075 6,998
MGT_Other 0.071 0.000 0.000 1.000 0.257 6,998
MAO 0.076 0.000 0.000 1.000 0.265 6,998
Irregularity 0.081 0.000 0.000 1.000 0.273 6,998
Big4 0.063 0.000 0.000 1.000 0.242 6,998
Top10 0.434 0.000 0.000 1.000 0.496 6,998
LTA 21.509 21.400 18.493 25.057 1.218 6,998
Leverage 0.540 0.519 0.068 2.529 0.313 6,998
Loss 0.113 0.000 0.000 1.000 0.317 6,998
ROA 0.027 0.030 -0.382 0.225 0.076 6,998
Liquidity 1.646 1.217 0.113 10.948 1.628 6,998
REC 0.095 0.070 0.000 0.416 0.090 6,998
INV 0.174 0.138 0.000 0.747 0.155 6,998
Tobin’s Q 2.571 1.935 0.883 13.612 2.030 6,760
Growth 0.245 0.146 -0.771 5.162 0.676 6,760
CAPEXP 0.053 0.037 -0.074 0.273 0.060 6,760
ACOMMIT 0.637 1.000 0.000 1.000 0.481 10,983
Fee 13.160 13.122 11.983 15.009 0.559 5,881
Audit effort 1 -1.463 -1.411 -3.820 -0.718 0.334 5,881
Audit effort 2 3.166 3.219 1.609 5.663 0.800 5,589
Variable definitions:
ACT = 1 if at least one of AC members has social ties, formed by educational experience,
teacher-student bonding, or employer affiliation, with one of the engagement
auditors, and 0 otherwise.
MGT_Chair/CEO = 1 if the CEO or board chair has social ties, including alma mater and employer
affiliation connections, with one of the engagement auditors, and 0 otherwise.
MGT_Other = 1 if at least one of the top managers (including board directors other than CEO,
board chair, or audit committee members) has social ties, developed through
educational experience, teacher-student bonding, or employer affiliation, with one
of the engagement auditors, and 0 otherwise.
MAO = 1 for observations where auditors issue modified audit opinions (including
unqualified opinions with explanatory notes, qualified, and disclaimed), and 0
otherwise.
Irregularity = 1 if earnings in the current year are restated in a subsequent year, or the
observation is subsequently sanctioned by the CSRC, MOF, or stock exchanges
due to accounting frauds or accounting irregularities in the current year, and 0
otherwise.
Big4 = 1 for observations that are audited by the international Big 4 auditors, and 0
otherwise.
Top10 = 1 for observations that are audited by the domestic top 10 auditors, ranked by the
number of listed clients in the year of interest, and 0 otherwise.
(The table continues on the next page.)

42
 Table 2 (Cont.)
LTA = The natural logarithm of client’s total assets at the fiscal year end.
Leverage = Ratio of total liabilities to total assets.
Loss = 1 for observations with negative net income, and 0 otherwise.
ROA = Net income divided by total assets at the fiscal year end.
Liquidity = Ratio of current assets to current liabilities at the fiscal year end.
REC = Ratio of net accounts receivables to total assets at the fiscal year end.
INV = Ratio of inventories to total assets at the fiscal year end.
Tobin’s Q = Market value of common equity plus book value of debt divided by total assets at
the fiscal year end.
Growth = The difference between sales of the current and the prior year divided by prior
year sales.
CAPEXP = The ratio of capital expenditures over total assets at the fiscal year end.
ACOMMIT = 1 if there is an audit committee, and 0 otherwise.
Fee = The natural logarithm of annual audit fees.
Auditor Effort 1 = The natural logarithm of number of days between the audit report date and the
fiscal year end scaled by 365.
Auditor Effort 2 = The natural logarithm of the number of audit field work days.
All the continuous variables are winsorized at the 1st and 99th percentile.

43
 Table 3 Analysis of audit quality
(1) (2) (3)
Variables y = MAO y = Irregularity y = MAO
ACT -0.940*** 0.310*** -0.759***
(-4.05) (2.55) (-3.08)
Irregularity 1.033***
(6.23)
ACT * Irregularity -0.077***
(-2.66)
MGT_Chair/CEO -1.254** 0.734* -2.070***
(-2.09) (1.70) (-3.25)
MGT_Chair/CEO * Irregularity 0.021
(0.14)
MGT_Other 0.059 -0.278 -0.089
(0.23) (-1.48) (-0.30)
MGT_Other * Irregularity 0.216
(1.42)
Big4 1.265*** -1.290*** 1.351***
(3.54) (-3.95) (3.80)
Top10 0.295** -0.066 0.284**
(2.31) (-0.69) (2.20)
LTA -0.857*** 0.061 -0.870***
(-11.07) (1.32) (-11.12)
Leverage 3.906*** 0.254* 3.892***
(9.89) (1.91) (9.64)
ROA -3.767*** -0.736 -3.902***
(-3.80) (-1.01) (-3.91)
Liquidity 0.013 0.017
(0.19) (0.24)
REC 1.151 0.785
(1.60) (1.08)
INV -2.553*** -2.546***
(-3.79) (-3.88)
Loss 1.032*** 0.566*** 0.930***
(5.06) (3.09) (4.50)
MAO 0.932***
(5.52)
Constant 14.694*** -2.114** 14.572***
(8.94) (-2.05) (8.79)
Pseudo R2 0.455 0.105 0.467
N 6,998 6,998 6,998
All variables are defined in Table 2. The numbers in parentheses are Z-statistics corrected for serial
correlation and heteroscedasticity with the Huber/White/sandwich estimator (clustered) for variance.
The estimates for the interaction terms in Column (3) are based on the Norton et al. (2004) method. ***,
**, and * suggest two-tailed statistical significance at the 1%, 5%, and 10% levels, respectively.

44
 Table 4 Analysis of firm value
(1) (2)
Variables y = Tobin’s Q y = Tobin’s Q
ACT -0.154***
(-3.16)
ACT_PQWM -0.534***
(-4.19)
ACT_ NPQWM -0.119**
(-2.32)
MGT_Chair/CEO 0.400 0.444
(1.11) (1.24)
MGT_Other -0.136* -0.129*
(-1.89) (-1.78)
Big4 0.569*** 0.563***
(7.66) (7.59)
Top10 0.235*** 0.232***
(5.40) (5.34)
LTA -0.744*** -0.743***
(-28.65) (-28.65)
Leverage 2.461*** 2.465***
(13.79) (13.82)
ROA 6.363*** 6.357***
(10.98) (10.97)
Liquidity 0.314*** 0.315***
(12.53) (12.54)
GROWTH 0.063* 0.063*
(1.73) (1.73)
CAPEXP -0.119 -0.117
(-0.31) (-0.31)
Constant 16.463*** 16.455***
(29.57) (29.56)
Adj. R2 0.339 0.339
N 6,760 6,760
ACT_PQWM equal to 1 if ACT = 1, Irregularity = 1 and MAO = 0, and 0 otherwise; ACT_NPQWM
equals 1 if ACT = 1 and ACT_PQWM = 0. All other variables are defined in Table 2. The numbers in
parentheses are t-statistics corrected for serial correlation and heteroscedasticity with the
Huber/White/sandwich estimator (clustered) for variance. ***, **, and * suggest two-tailed statistical
significance at the 1%, 5%, and 10% levels, respectively.

45
 Table 5 Robustness checks
(1) (2) (3)
y = MAO y = Irregularity y = Tobin’s Q
Panel A. Social tie channels
Channel (i) only -0.600 0.553*** -0.326***
(-1.46) (2.73) (-3.14)
Channels (ii) only -0.943*** 0.383*** -0.131**
(-3.62) (2.89) (-2.50)
Channels (iii) only -1.054** 0.665*** -0.195
(-2.03) (2.72) (-1.57)
Panel B. Social tie formation types
Through the appointment of a new AC -0.828*** 0.293** -0.262***
member (-3.14) (2.12) (-4.77)
Through the appointment of a new -1.292*** 0.504** 0.142
engagement auditor (-2.93) (2.22) (1.57)
Panel C. The Fama-MacBeth (1973) test results
ACT -1.244** 0.361*** -0.102*
(-3.87) (4.80) (-2.38)
Panel D. Regression results under the PSM approach
PSM -1.002*** 0.691*** -0.170***
(-2.92) (3.64) (-2.59)
This table reports regression estimates of the ACT variable. All of the control variables are included in the
regression estimation, although their estimates are not tabulated for brevity.
In Panel A, we restrict the analysis to cases where AC-auditor social ties come through channel (i), (ii), and
(iii) respectively. Here, channels (i), (ii), and (iii) refer to ties formed via common schooling experience,
professor-student bonding, and common working experience, respectively.
In Panel B, we analyze two forms of AC-auditor social tie formation: (i) ties that are formed through the
appointment of a new AC member who is socially connected with the incumbent engagement auditors,
and (ii) ties are formed through a new engagement auditor being appointed to a firm with a
socially-connected AC member.
In Panel C, we present Fama-MacBeth (1973) regression results, which are based on the distribution of
ACT coefficients from yearly regressions.
In Panel D, we estimate the regressions on samples matched by covariates that may influence audit
outcome variables. To construct the matched samples, we first regress ACT on relevant covariates by
the logistic model to generate a propensity score. Then, for each treatment observation (i.e., ACT = 1),
we select a matching control (i.e., ACT = 0) from the same year that has the closest propensity score.
Except Panel C, the numbers in parentheses are t-statistics corrected for serial correlation and
heteroscedasticity with the Huber/White/sandwich estimator (clustered) for variance. ***, **, and * suggest
two-tailed statistical significance at the 1%, 5%, and 10% levels, respectively.

46
 Table 6 Regression results by audit committee member and/or auditor characteristics
(1) (2) (3)
Variables y = MAO y = Irregularity y = Tobin’s Q
Panel A. ACT_X = 1 if ACT = 1 and AC chair has social ties with one of the engagement auditors
ACT_X -1.362*** 0.331** -0.166***
[69.22%] (-4.05) (2.34) (-2.96)
ACT_X’ -0.395 0.263 -0.127
[30.78%] (-1.21) (1.24) (-1.49)
Panel B. ACT_X = 1 if ACT = 1 and at least one of the connected AC members has an accounting
background
ACT_X -1.284*** 0.322** -0.154***
[77.78%] (-4.21) (2.38) (-2.80)
ACT_X’ -0.188 0.268 -0.156*
[22.22%] (-0.58) (1.11) (-1.69)
Panel C. ACT_X= 1 if ACT = 1 and at least one of the AC members have social ties with the senior
signing auditor
ACT_X -1.648*** 0.347** -0.177***
[69.91%] (-4.98) (2.45) (-3.07)
ACT_X’ -0.070 0.228 -0.103
[30.09%] (-0.23) (1.09) (-1.29)
Panel D. ACT_X = 1 if ACT = 1 and the connected AC members and engagement auditors work in
the same region other than Beijing and Shanghai
ACT_X -1.845*** 0.305* -0.377***
[38.05%] (-3.46) (1.76) (-5.27)
ACT_X’ -0.602 ** 0.314 * -0.018
[61.95%] (-2.42) (1.96) (-0.30)
Panel E. ACT_X = 1 if ACT = 1 and the number of directorships held by the connected AC
member is below the sample top quartile
ACT_X -1.031*** 0.412*** -0.184***
[81.32%] (-3.95) (3.23) (-3.40)
ACT_X’ -0.542 -0.335 -0.024
[18.68%] (-1.14) (-0.96) (-0.25)
ACT_X is defined at the heading of each panel. ACT_X’ denotes the cases where ACT = 1 and the
observation does not has a specific attribute as defined in ACT_X. The numbers in the brackets are
the percentages of observations within each group. In the interest of brevity, only estimates of
variables of interest are reported, although all of the control variables are included in the regression
estimation. The numbers in parentheses are t-statistics corrected for serial correlation and
heteroscedasticity with the Huber/White/sandwich estimator (clustered) for variance. ***, **, and *
suggest two-tailed statistical significance at the 1%, 5%, and 10% levels, respectively.

47
 Table 7 Regression results by corporate governance
(1) (2) (3)
Variables y = MAO y = Irregularity y = Tobin’s Q
Panel A. ACT_X = 1 if ACT = 1 and the percentage of shareholdings by the largest shareholders is
above the bottom quartile
ACT_X -1.016*** 0.403*** -0.176***
[75.81%] (-3.86) (3.02) (-3.32)
ACT_X’ -0.691 -0.031 -0.086
[24.19%] (-1.56) (-0.12) (-0.84)
Panel B. ACT_X = 1 if ACT = 1 and the percentage of shareholdings by institutional shareholders
is below the sample top quartile
ACT_X -0.949*** 0.445*** -0.351***
[73.55%] (-3.79) (3.42) (-6.60)
ACT_X’ -0.884 -0.282 0.394***
[26.45%] (-1.55) (-0.92) (3.98)
Panel C. ACT_X = 1 if ACT = 1 and the firms’ shares are held by foreign investors via the
Qualified Foreign Institutional Investor (QFII) program
ACT_X -0.894*** 0.403*** -0.427***
[71.98%] (-3.72) (3.06) (-8.01)
ACT_X’ -1.528 -0.116 0.542***
[28.02%] (-1.51) (-0.38) (5.93)
Panel D. ACT_X = 1 if ACT = 1 and the number of analyst following is below the sample top
quartile
ACT_X -0.894*** 0.403*** -0.427***
[71.98%] (-3.72) (3.06) (-8.01)
ACT_X’ -1.528 -0.116 0.542***
[28.02%] (-1.51) (-0.38) (5.93)
Panel E. ACT_X = 1 if ACT = 1 and the CEO is promoted from the inside
ACT_X -1.064*** 0.557*** -0.201***
[56.05%] (-2.98) (3.85) (-3.16)
ACT_X’ -0.819 *** -0.095 -0.094
[43.95%] (-2.85) (-0.46) (-1.32)
Panel F. ACT_X = 1 if ACT = 1 and the CEO tenure length is above the sample bottom quartile
ACT_X -1.007*** 0.406*** -0.217***
[70.11%] (-3.33) (2.99) (-4.06)
ACT_X’ -0.821 ** 0.001 0.005
[29.89%] (-2.43) (0.01) (0.05)
Panel G. ACT_X = 1 if ACT = 1 and the marketization index of the province in which the client
firm is headquartered is below the sample top quartile
ACT_X -0.950*** 0.514*** -0.217***
[77.58%] (-3.70) (4.10) (-3.96)
ACT_X’ -0.899* -0.598 0.064
[22.42%] (-1.83) (-1.61) (0.67)
(The table continues on the next page.)

48
 Table 7 (Cont.)
ACT_X is defined at the heading of each panel. ACT_X’ denotes the cases where ACT = 1 and the
observation does not has a specific attribute as defined in ACT_X. The numbers in the brackets are
the percentages of observations within each group. In the interest of brevity, only estimates of
variables of interest are reported, although all of the control variables are included in the regression
estimation. The numbers in parentheses are t-statistics corrected for serial correlation and
heteroscedasticity with the Huber/White/sandwich estimator (clustered) for variance. ***, **, and *
suggest two-tailed statistical significance at the 1%, 5%, and 10% levels, respectively.

49
 Table 8 Analysis of the effect on the interaction between ACT and MGT_Chair/CEO
(1) (2) (3)
Variables y = MAO y = Irregularity y = Tobin’s Q
ACT -0.941*** 0.286** -0.150***
(-4.02) (2.30) (-3.07)
MGT_Chair/CEO -1.402** -14.080*** 0.759
(-2.25) (-39.48) (1.49)
ACT*MGT_Chair/CEO 0.019 0.206*** -0.529
(0.64) (2.37) (-0.76)
MGT_Other 0.059 -0.289 -0.134*
(0.23) (-1.52) (-1.86)
Big4 1.264*** -1.291*** 0.570***
(3.54) (-3.96) (7.67)
Top10 0.295** -0.062 0.234***
(2.31) (-0.65) (5.39)
LTA -0.857*** 0.062 -0.744***
(-11.07) (1.33) (-28.65)
Leverage 3.907*** 0.254* 2.462***
(9.89) (1.90) (13.80)
ROA -3.766*** -0.737 6.363***
(-3.79) (-1.01) (10.98)
Liquidity 0.013 0.314***
(0.19) (12.53)
REC 1.152
(1.60)
INV -2.553***
(-3.79)
Loss 1.032*** 0.561***
(5.06) (3.05)
MAO 0.933***
(5.53)
GROWTH 0.064*
(1.74)
CAPEXP -0.128
(-0.34)
Constant 14.694*** -2.118** 16.463***
(8.94) (-2.06) (29.56)
Pseudo/Adj. R2 0.455 0.106 0.339
N 6,998 6,998 6,760
All variables are defined in Table 2. The numbers in parentheses are t-statistics corrected for serial
correlation and heteroscedasticity with the Huber/White/sandwich estimator (clustered) for variance.
The estimates for the interaction terms are based on the Norton et al. (2004) method. ***, **, and *
suggest two-tailed statistical significance at the 1%, 5%, and 10% levels, respectively.

50
 Table 9 The effect of audit committees on audit outcomes
Firms without audit committees vs. firms with ACT = 0 Firms without audit committees vs. firms with ACT = 1
(1) (2) (3) (4) (5) (6)
Variables y = MAO y = Irregularity y = Tobin’s Q y = MAO y = Irregularity y = Tobin’s Q
ACOMMIT 0.225** -0.172** 0.602*** -0.718*** 0.103 0.358***
(2.07) (-2.26) (17.47) (-3.00) (0.80) (6.98)
Big4 0.582** -0.874*** 0.564*** 0.309 -1.060*** 0.674***
(2.01) (-3.67) (8.75) (0.71) (-3.14) (6.78)
Top10 0.033 -0.189** 0.244*** -0.094 -0.302*** 0.277***
(0.32) (-2.41) (6.80) (-0.62) (-2.82) (5.99)
LTA -0.642*** 0.010 -0.682*** -0.424*** 0.048 -0.539***
(-11.03) (0.26) (-32.33) (-5.02) (0.91) (-18.33)
Leverage 3.514*** 0.078 2.214*** 3.736*** 0.251 2.168***
(11.70) (0.74) (17.67) (7.74) (1.52) (12.77)
ROA -4.198*** -1.184** 5.637*** -4.583*** -1.381* 5.661***
(-5.30) (-2.11) (13.57) (-3.72) (-1.68) (10.60)
Liquidity -0.028 0.276*** -0.047 0.242***
(-0.56) (14.70) (-0.64) (9.85)
Loss 1.191*** 0.695*** 1.248*** 0.826***
(7.32) (4.82) (5.08) (4.21)
REC 1.119** 1.040
(2.12) (1.37)
INV -3.152*** -4.062***
(-5.56) (-4.58)
MAO 0.824*** 0.527***
(6.28) (2.67)
GROWTH 0.062* 0.031
(1.89) (0.74)
CAPEXP 0.302 0.489
(0.98) (1.24)
Constant 10.154*** -1.665** 14.573*** 4.491** -2.392** 11.762***
(8.12) (-2.03) (32.52) (2.44) (-2.08) (18.80)
Pseudo/Adj. R2 0.448 0.072 0.331 0.475 0.076 0.289
N 9,966 9,966 8,999 5,002 5,002 4,217
All variables are defined in Table 2. The numbers in parentheses are t-statistics corrected for serial correlation and heteroscedasticity with the
Huber/White/sandwich estimator (clustered) for variance. ***, **, and * suggest two-tailed statistical significance at the 1%, 5%, and 10% levels,
respectively.

51
 Table 10 Analysis of audit fees and audit effort
(1) (2) (3) (4) (5)
Variables y = Fee y = Fee y = Fee y = Audit effort1 y = Audit effort2
ACT 0.027** 0.026** 0.031** 0.005 -0.002
(2.04) (2.01) (2.31) (0.39) (-0.06)
MGT_Chair/CEO 0.096* 0.088* 0.084 0.087 -0.039
(1.78) (1.67) (1.53) (1.35) (-0.32)
MGT_Other 0.007 0.007 0.009 -0.004 -0.009
(0.36) (0.38) (0.48) (-0.22) (-0.23)
Big4 0.692*** 0.688*** 0.680*** 0.045*** 0.202***
(22.92) (22.78) (22.47) (2.92) (4.59)
Top10 0.114*** 0.113*** 0.124*** 0.011 -0.048**
(11.28) (11.23) (12.11) (1.18) (-2.36)
MAO 0.132*** 0.124*** 0.133*** 0.084*** 0.071
(5.72) (5.42) (5.66) (3.84) (1.58)
LTA 0.296*** 0.295*** 0.283*** 0.014*** 0.183***
(52.85) (52.81) (48.27) (3.16) (18.21)
ROA 0.080 0.136 0.112 -0.582*** -0.248
(0.82) (1.38) (1.12) (-6.11) (-1.36)
Liquidity -0.009** -0.010*** -0.009** 0.007** -0.014**
(-2.46) (-2.64) (-2.32) (1.97) (-2.06)
Leverage 0.106*** 0.107*** 0.104*** -0.008 -0.001
(5.20) (5.28) (5.02) (-0.36) (-0.02)
REC 0.071 0.053 0.051 0.186*** 0.376***
(1.06) (0.80) (0.74) (3.32) (2.77)
INV -0.122*** -0.122*** -0.115*** -0.008 0.008
(-3.04) (-3.05) (-2.84) (-0.21) (0.10)
Loss 0.017 0.013 0.020 0.048*** -0.045
(0.81) (0.60) (0.91) (2.62) (-1.03)
Audit Effort 1/Audit Effort 2 0.096*** 0.060***
(6.78) (8.54)
Constant 6.685*** 6.857*** 6.765*** -1.798*** -0.569**
(53.21) (54.08) (52.58) (-17.80) (-2.49)
Adj. R2 0.556 0.559 0.561 0.051 0.180
N 5,881 5,881 5,589 5,881 5,589
All variables are defined in Table 2. The numbers in parentheses are t-statistics corrected for serial correlation and heteroscedasticity with the
Huber/White/sandwich estimator (clustered) for variance. ***, **, and * suggest two-tailed statistical significance at the 1%, 5%, and 10% levels,
respectively.

52