Corporate Governance

China Journal of Accounting Research 5 (2012) 321–342
Contents lists available at SciVerse ScienceDirect
China Journal of Accounting Research

journal homepage: www.elsevier.com/locate/cjar
Corporate governance and audit fees: Evidence from

companies listed on the Shanghai Stock Exchange
Xingze Wu
Research Institute of Accounting and Finance, Nanjing University, China
A R T I C L E I N F O A B S T R A C T
Article history: This study uses data from companies listed on the Shanghai Stock Exchange to
Received 28 February 2012 investigate the relationship between corporate governance and audit fees. Full
Accepted 10 October 2012 sample results reveal a significant negative relationship between corporate gov-
Available online 12 November ernance and audit fees, and subsample results further show that corporate gov-
2012 ernance’s influence on audit fees is affected by corporate growth. The negative
relationship between corporate governance and audit fees is economically and
JEL classification: statistically significant in sample companies that grew moderately during the
G38 sample period, and mixed or insignificant in companies that experienced overly
M42 fast or negative growth.
Ó 2012 China Journal of Accounting Research. Founded by Sun Yat-sen
Keywords: University and City University of Hong Kong. Production and hosting by
Corporate governance Elsevier B.V. All rights reserved.
Audit fees
SSE Corporate Governance
Sector
Corporate growth
1. Introduction
China’s special audit market has important theoretical and empirical implications for the determinants of
audit fees (Zhu and Yu, 2004). Of the various determinants thus far proposed in the literature, corporate gov-
ernance constitutes a relatively new research topic (Larcker and Richardson, 2004; Cai, 2007). Since the
demise of Enron and WorldCom, however, the internal corporate governance of listed companies has become
E-mail address: [email protected]

q
This study was supported by a Nanjing University IAPHD Project and sponsored by the Qing Lan Project of Jiangsu Province.
Production and hosting by Elsevier
1755-3091/$ - see front matter Ó 2012 China Journal of Accounting Research. Founded by Sun Yat-sen University and City University of
Hong Kong. Production and hosting by Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.cjar.2012.10.001
322 X. Wu / China Journal of Accounting Research 5 (2012) 321–342
a topic of considerable research interest (Liu and Hu, 2006). Auditors themselves have also begun to attach
greater importance to evaluations of internal corporate governance. On 15 February 2006, China’s Ministry of
Finance announced the issuance of new auditing standards. These standards implement a risk-oriented audit
approach that attaches importance to the risks associated with a firm’s governance structure. Auditing Stan-
dard No. 1211 states clearly that auditors must pay attention to the governance structure of the audited entity.
However, it remains unclear whether corporate governance has any effect on audit fees and, if it does have
such an effect, how it influences audit fees.
There are at least two arguments concerning the relationship between corporate governance and audit fees.
The first is informed by substitution theory and the second by signaling theory, and the two lead to different
conclusions. Substitution theory posits that the more perfect the internal corporate governance structure of a
firm, and hence the lower the agency costs, the fewer risks the audit firm and auditor will encounter and thus
the lower the audit fee that will be charged. In other words, an audit is seen as a form of external governance
for which effective internal corporate governance may substitute to some degree. Signaling theory argues that
managers signal high-level corporate governance to external stakeholders1 by inviting a more rigorous exter-
nal audit, which inevitably leads to higher audit fees, i.e., companies with strong corporate governance pay
higher audit fees to accounting firms. The mixed empirical evidence reported to date leaves unanswered the
question of which theory better explains corporate practice.
Most of the literature on the relationship between corporate governance and audit fees concentrates on one
or more aspects of corporate governance, such as ownership, board of director or management characteristics,
as proxy variables for corporate governance (Pan, 2008). Although the use of such proxies renders it easy to
collect and treat data, it has a number of disadvantages. For example, it introduces the possibility of omitted
variables in the models because all corporate governance characteristics are not included. In addition, different
characteristics may interact with one another in a manner too complex to identify, thus producing possibly
biased results. Finally, as the influence of single characteristics on the level of corporate governance is uncer-
tain, it is doubtful whether a proper corporate governance proxy exists. For example, some scholars believe
that CEO duality impairs corporate governance, whereas others take the opposite view. It is thus clear that
identifying the relationship between audit fees and corporate governance on the basis of such a proxy is prob-
lematic, although a more comprehensive corporate governance variable would mitigate or eliminate such
problems to a considerable extent.
The Shanghai Stock Exchange (SSE) introduced the SSE Corporate Governance Sector in 2007, thus offer-
ing a good opportunity for a comprehensive investigation of the relationship between corporate governance
and audit fees. The listed companies within this sector are subject to greater public scrutiny of their corporate
governance structures. After preliminary examination of listed companies’ application qualifications, the
appraisal working group of the Corporate Governance Sector publishes the application materials of those that
qualify on its official website for public appraisal. The overall aim is to involve public investors in the appraisal
process and encourage all market participants to pay greater attention to the issue of corporate governance.
The SSE also invites professional research institutions to appraise the SSE Corporate Governance Sector and
to judge the governance structures of the companies submitting applications. These research institutions
include CITIC Securities Co., Ltd., Guotai Junan Securities Co., Ltd., Shenyin & Wanguo Securities Co.,
Ltd. and Haitong Securities Co., Ltd., among others. Experts and scholars have also been invited to form
an Expert Consultative Committee for Appraisal of the Corporate Governance Sector, which meets regularly
to discuss the method, process and results of the appraisal process, thus ensuring its objectivity and standard-
ization. This rigorous appraisal process ensures that listed companies undergo comprehensive assessment of
their corporate governance level prior to inclusion in the SSE Corporate Governance Sector. As noted, it also
makes possible a comprehensive investigation of the relationship between corporate governance and audit
fees.
This study uses inclusion in the SSE Corporate Governance Sector as a proxy for corporate governance to
empirically investigate the relationship between corporate governance and audit fees after controlling for the
1
It is obvious that companies will not pay higher audit fees to convey a signal to the market merely for signaling purposes. Rather, such
motives as obtaining financing from the market, boosting firm value or reducing financing costs generally explain signaling behavior.
X. Wu / China Journal of Accounting Research 5 (2012) 321–342 323
other main factors associated with audit fees. Compared to the proxies used in most of the recent literature,
the proxy used here is more comprehensive, authoritative and easily understood, and it is also easily collected.
If a relationship between corporate governance and audit fees is confirmed, listed firms may use such confir-
mation in the future to negotiate audit fees with accounting firms, which is one of the main innovations and
contributions of this study.
The focus on risk under the risk-oriented audit approach is likely to lead to interactions between corporate
growth and internal governance in listed companies. Companies experiencing overly fast or negative growth
are characterized by greater risk (Lang et al., 1996) and their internal corporate governance may suffer an
adverse change in stability, thus providing management with the motivation to manage reported earnings.
Companies that grow steadily and moderately, in contrast, are often in the maturity stage.2 They thus expe-
rience a lower degree of risk and their internal corporate governance is relatively stable. Cui et al. (2007) exam-
ine the relationship between corporate growth and financial risk and find the probability that a company
experiences financial crisis increases dramatically when its growth rate exceeds what the authors call a reason-
able growth rate. They also report a significant positive relationship between the probability of financial crisis
and excessive growth rates and an insignificant relationship between the probability of financial crisis and the
real growth rate of non-excessively growing companies (Cui et al., 2007). In reality, many companies appear to
collapse suddenly. Enron and WorldCom in the United States and the Giant Group and Qinchi Alcohol in
China are representative examples. In line with the foregoing discussion, this study examines subsamples
grouped by corporate growth in addition to the full sample. The full sample reveals a significant negative rela-
tionship between corporate governance and audit fees, and the subsample results also show that corporate
governance’s influence on audit fees is affected by corporate growth. The negative relationship between cor-
porate governance and audit fees is economically and statistically significant in sample companies that grew
moderately during the sample period, whereas the relationship is mixed or insignificant for companies that
experienced overly fast or negative growth.
The remainder of this paper is organized as follows. Section 2 reviews the literature on audit fees and cor-
porate governance. Section 3 develops the research hypotheses, which are grounded in theoretical analysis.
Section 4 describes the data and variables. The full sample and subsample regression results are provided
in Sections 5 and 6, respectively. Section 7 reports the result of a sensitivity test and Section 8 concludes
the paper.
2. Overview of prior research
2.1. Factors associated with audit fees
Audit fees have been a subject of interest in the auditing literature since the pioneering research of Simunic
(1980). Simunic (1980) posits that audit fees are determined by the loss exposure of the auditee, the apportion-
ment rate of loss between the audit firm and the auditee, and the production function and characteristics of the
audit firm. He provides empirical evidence to show that the scale of the auditee is the main factor influencing
audit fees, although the number of consolidated subsidiaries included in the auditee’s financial statements,
number of industries in which the auditee operates, ratio of the auditee’s assets abroad to total assets at
year-end, ratio of receivables to total assets at year-end, ratio of inventory to total assets at year-end and
whether an auditee incurred a loss in the most recent 3 years or received a “subject to” qualified opinion also
have a significant influence. Simunic finds the ratio of net income to total assets at year-end, auditor tenure
and audit firm scale to have no significant influence on audit fees. Francis (1984) investigates the Australian
audit market using a modified Simunic model and also finds the scale of listed companies’ assets and a variable
reflecting the complexity of business transactions or events (the number of consolidated subsidiaries) to be sig-
nificantly related to audit fees. However, contrary to Simunic (1980), Francis also finds the scale of the audit
firm to be significantly related to audit fees. Francis and Stokes (1986) investigate the 96 largest and 96 small-
2
The typical lifecycle of an enterprise comprises four stages, i.e., start-up, growth, maturity and decline. Although a low rate of growth
is a common characteristic of the start-up and maturity stages, this study considers it to be associated with the maturity stage alone, as
Chinese legal regulations prohibit firms in the start-up stage from listing on the A-share market.
est publicly traded non-finance companies in the Australian Graduate School of Management Annual Report
Data Files and find that Big 8 price premiums are observed for small auditees but not for large auditees. Gul
(2001) takes the opinion that audit fees can be considered simply as a function of firm size, complexity and
audit risk.
In December 2001, the China Securities Regulatory Commission promulgated “Standards Concerning the
Contents and Formats of Information Disclosure by Companies Offering Securities to the Public No. 2-Con-
tents and Formats of Annual Reports (Revised in 2001)” and “Question and Answer Document Concerning
the Standards of Information Disclosure by Companies Offering Securities to the Public No. 6-Payments to
Accounting Firms and Disclosure.” These documents state that listed companies are required to disclose their
audit fees in their annual reports from 2001 onwards. The new regulations prompted a number of Chinese
scholars to carry out empirical studies of audit fees using data from Chinese listed companies. Most of these
studies adopt the model developed by Simunic (1980) and use financial variables (Liu and Hu, 2006). Wang
(2002) was one of the first in China to investigate audit fees empirically. He reports the scale of the auditee and
audit firm, audit complexity and audit risk, the industry in which the auditee operates and whether the auditee
receives a qualified opinion to have an effect on audit fees. Wu (2003) cites auditee scale, whether an auditee
has been audited by one of the Big 5, audit opinion, ratio of accounts receivable to total assets and the ratio of
inventory to total assets as the main factors influencing audit fees. Han and Zhou (2003) find the auditee’s
total assets, audit opinion, number of consolidated subsidiaries and debt ratio to be significantly related to
audit fees. Liu et al. (2003) analyze 590 companies and conclude that the scale and location of a listed com-
pany and the complexity of its business transactions are the main factors influencing audit fees, whereas there
are no significant relationships with the ratio of inventory to total assets, ratio of long-term debt to total
assets, loss occurrence, audit tenure and audit firm scale. Zhu and Guo (2006) investigate the issues surround-
ing audit fee increases in companies with no changes in accounting firms and find company expansion and an
increase in the debt ratio to be the two main explanatory factors. In addition, they also find a change in the
ratio of cash to current debt, intention to opinion shop and earnings management to be significantly related to
an audit fee increase, although a change in return on equity (ROE) and changes in the ratios of accounts
receivable and inventory to total assets exhibit no relationship. Most of the empirical studies to date find audi-
tee scale and complexity and whether a firm has been audited by one of the “Big N” firms to have a signifi-
cantly positive influence on audit fees (Simunic, 19803; Francis and Stokes, 1986; Gul, 2001; Wu, 2003; Han
and Zhou, 2003; Liu and Hu, 2006). With regard to the ratios of inventory to total assets and accounts receiv-
able to total assets, domestic and overseas findings differ, with studies carried out overseas usually reporting a
positive relationship between these ratios and audit fees (e.g., Simunic, 1980) and domestic studies finding no
such relationship.
2.2. Corporate governance and audit fees
Although many studies have examined the factors influencing audit fees, the relationship between corporate
governance and audit fees is only now beginning to receive extensive research attention. The preliminary evi-
dence is inconsistent (Cai, 2007). Overseas studies generally begin with the hypothesis that audits are a form of
external governance and investigate the influence of agency costs and board of director characteristics on audit
pricing. For example, Gul et al. (1998) examine the association between the magnitude of earnings/accruals (as
a proxy for agency costs) and audit pricing and find a positive relationship. They also find audit prices to be
lower for family companies than other kinds of companies and report the number of independent directors to
be negatively related to audit fees. Gul and Tsui (2001) testify to the influence of agency costs on audit pricing
in the Australian audit market. Carcello et al. (2002) investigate the association between board of director
characteristics and external audit fees using Fortune 1000 data, and find a significant positive relationship
between audit fees and board independence, expertise and diligence. Hay et al. (2004) believe that the promul-
gation of the Sarbanes–Oxley Act, Section 404 of which demands that listed companies disclose internal con-
trol information, will increase opportunities to investigate the association between corporate governance and
3
Simunic (1980) does not investigate the influence of the Big N on audit fees.
audit fees directly, although our review of the overseas literature indicates no such increase. Chinese research-
ers, in contrast, have paid increased attention to the issue in recent years. Drawing on the ownership perspec-
tive, Zhang and Zhang (2005) find the audit fees of state-owned listed companies to be low relative to those of
other types of firms and Gao and Gao (2008) report the stockholding ratio of managers to be significantly
associated with audit fees. In contrast, Zhang and Xu (2005) show there to be no significant relationship
between audit fees and the proportion of state-owned shares. Li and Wang (2006) examine the role played
by board of director characteristics and find the audit fee rate to be significantly and negatively related to
the number of independent directors on the board, but insignificantly related to the number of board meetings
and the existence of an audit committee. Using a framework of internal corporate governance and data on A-
share listed companies from 2001 to 2003, Liu and Hu (2006) analyze the relationship between audit pricing
and agency costs, and find that a number of the corporate governance factors that may influence agency costs
(i.e., the proportion of independent directors on the board, the stockholding ratio of senior managers and
president–CEO duality) also have a significant influence on audit fees, subject to the existence of other vari-
ables. Cai (2007) investigates the influence of corporate governance structure on audit fees from the perspec-
tive of the audit service provider and provides evidence to show that accounting firms charge companies with a
larger board of directors higher audit fees than they do non-state-owned companies featuring CEO duality or
a moderate managerial share ratio.
The aforementioned research tests the relationship between corporate governance and audit fees empiri-
cally from different perspectives, although the theoretical basis of most is substitution theory, with signaling
theory receiving little attention to date. Most of this research also considers corporate governance character-
istics such as shareholdings, board of director and management variables as proxies for corporate governance
(Pan, 2008). As noted in the introduction, there are several limitations to the use of such proxies. To address
these limitations, this paper analyzes the relationship between corporate governance and audit fees from the
perspectives of substitution theory and signaling theory, and uses inclusion in the SSE Corporate Governance
Sector to proxy for corporate governance.
3. Theory and hypotheses
As a form of external governance, independent auditing can mitigate agency conflicts among stakeholders
and reduce agency costs (Jensen and Meckling, 1976; Watts and Zimmerman, 1983; Fan and Wong, 2005).
Companies with serious agency problems thus have an incentive to hire auditors with a strong reputation
to send a signal to the market that they are attempting to reduce agency costs to improve firm value (Wang
and Zhou, 2006; Wang, 2009). However, if a company suffers no serious agency problems, it is unnecessary for
it to hire high-profile auditors. Analysis from the audit supplier’s perspective using the equation, audit
risk = material misstatement risk detection risk, suggests that the greater the material misstatement risk
assessed, the greater the likelihood of misstating a financial report, the lower the level of detection risk, the
larger the amount of audit work and the higher the audit cost. Carcello et al. (2002) find better internal firm
governance to result in less audit risk. Auditors assign a lower level of inherent risk and control risk4 to com-
panies characterized by such governance. Hence, audit effort and audit costs decline as a result of lower audit
fees. In contrast, auditors assess companies with poor internal governance as having higher levels of inherent
risk and control risk. For these firms, auditors need to spend more time, perform more audit work and bear
greater audit risk, and, accordingly, they collect higher audit fees. This discussion leads to the following
hypothesis.
Hypothesis 1a. Audit fees are lower for companies with high-level corporate governance.
At the same time, the information economics perspective suggests the existence of information asymmetry
between firms and external investors. Owing to the absence of a mechanism for imparting information, “bad
money drives out good” is the prevailing sentiment in the market. Signaling provides the best way to mitigate
4
There are two risk-oriented audit approaches, traditional and modern. Inherent risk and control risk in the traditional audit risk
approach have been replaced with material misstatement risk in the modern approach.
information asymmetry (Spence, 1973). The two basic methods of conveying a signal in the audit market are
to choose reputable information intermediaries voluntarily to assure outside investors of the credibility of
accounting information (Fan and Wong, 2005) and to purchase more audit services (Carcello et al., 2002).
Both methods result in higher audit costs and fees. It is obvious that the only companies with the incentive
to adopt these methods are those with better corporate governance. Such companies prefer the strict test of
an external audit to signal their governance level to the market and improve firm value. Therefore, companies
with high-level corporate governance may also experience higher audit fees, which leads to the following alter-
native hypothesis.
Hypothesis 1b. Audit fees are higher for companies with high-level corporate governance.
4. Data and variables
4.1. Data and sample
Considering that the SSE Corporate Governance Sector was introduced in 2007, with its constituents finally
confirmed at the end of that year, this study’s preliminary sample comprises all A-share companies listed on
the SSE for the 2007–2008 period. The following selection procedure was executed. First, in line with similar
studies (Liu and Hu, 2006; Cai, 2007), we removed observations of financial enterprises. Second, we removed
observations with incomplete data. Third, we removed observations listed on or after November 2, 2007,
which is the expiration date for voluntary applications from listed companies, according to the “Appraisal
Measures of SSE Corporate Governance Sector.”5 Finally, to alleviate the influence of outliers, we removed
all observations whose Tobin’s Q value falls outside the range of the mean minus two times the standard devi-
ation and the mean plus two times the standard deviation. The final sample contains 602 observations for 2007
(149 in the SSE Corporate Governance Sector) and 678 for 2008 (184 in the SSE Corporate Governance Sec-
tor). Table 1 summarizes the sample selection procedure.
Our primary data source was Beijing University’s China Center for Economic Research (CCER) database.
Some data, including the components of the SSE Corporate Governance Sector, H-share issuance, number of
a company’s subsidiaries and the number of industries in which a company operates, were collected manually
from the Sina Finance website (www.finance.sina.com.cn), Juchao website (www.cninfo.com.cn) and the
annual financial reports of the sample firms.
4.2. Model and variables
We modify and extend the Simunic (1980) model according to the Chinese institutional environment and
construct the following multiple linear regression model.
Lnfee ¼ b0 þ b1 Gov þ b2 TobinQ þ bBig4 þ b4 LnAssets þ b5 H S tock þ b6 Loss þ b7 Recint þ b8 Invint
þ b9 Segment þ b10 Subs rt þ e: ð1Þ
The explained variable in Model 1 is Lnfee, which is defined as the natural logarithm of the current year’s
external audit fees. The explanatory variable is Gov, which represents corporate governance. Previous studies
have adopted two types of variables to proxy for corporate governance: one or more aspects or characteristics
of corporate governance and a variable encompassing the comprehensive aspects of such governance. For
example, Larcker and Richardson (2004) use the structure of the board of directors, Carcello et al. (2002)
the characteristics of the board of directors and Liu and Hu (2006) the type of final controller, ownership con-
centration, board independence, CEO duality and managerial shareholdings. All of these proxies are examples
of the first type of variable. Studies using proxies of the second type are primarily concerned with the effec-
tiveness of corporate governance, e.g., Beiner et al. (2003) and Drobetz et al. (2004). Pan (2008) is the only
study of the relationship between audit fees and corporate governance to use the corporate governance index
5
Some observations belong to two or three of the elimination categories. For example, a company listed after November 2, 2007 is also a
company with incomplete data.
Table 1
Summary of sample selection criteria.
Selection procedure 2007 2008
A-share Companies included in SSE A-share Companies included in SSE
companies listed Corporate Governance Sector companies Corporate Governance Sector
on SSE listed on SSE
Total 851 199 864 231
Less: financial enterprises 19 10 20 13
Companies with incomplete data 219 40 144 30
Companies listed on or after 5 – 10 2
November 2, 2007
Outliers 6 – 12 2
Final sample observations 602 149 678 184
developed by the Nankai University Research Center of Corporate Governance as a proxy for such gover-
nance. The current study also adopts a more comprehensive proxy of corporate governance, Gov. Different
from Pan (2008), however, and for the reasons stated in the introduction, this study uses a dummy variable,
i.e., inclusion in the SSE Corporate Governance Sector. Gov takes the value of 1 if a company is included in
the sector, and otherwise 0. The control variables are as follows.
Previous research shows that firm size is a very important factor influencing audit fees (Simunic, 1980;
Wang, 2002; Chen et al., 2005). Theoretically, the larger a company is, the greater its business and accounting
activities and hence the greater the audit adjustment needed. In China, the administrative rules and regulations
on audit fees issued by the Chinese Institute of Certified Public Accountants or local administrative depart-
ments state that accounting firms should charge audit fees that are based on the customer’s assets (i.e., firm
size). In line with existing analysis and usual practice (e.g., Simunic, 1980; Larcker and Richardson, 2004;
Liu et al., 2003), we include LnAssets as a variable representing the natural logarithm of total assets at
year-end to proxy for firm size. We expect a positive relationship between firm size and audit fees.
The two main measures of firm complexity used in previous research are the number of consolidated
subsidiaries (Subs_rt) and the number of industries in which a company is involved in (Segment). Both are
used to measure firm complexity in this study, with a square root transformation to the number of consoli-
dated subsidiaries performed according to the procedure used by Chen and Zhou (2006), Liu and Hu
(2006) and Li and Wang (2006). To ensure data comparability, we include only those subsidiaries directly
established and held by the sample companies in counting the number of subsidiaries. The number of indus-
tries in which a firm is involved in is determined by the types of business (classified by industry) disclosed in its
annual financial report. We consider such data to be missing if no corresponding data is disclosed in the
annual report, and assign a 1 to Segment if only the main business data classified by product is disclosed. Posi-
tive relationships are expected between these variables and audit fees.
In line with Simunic (1980) and Larcker and Richardson (2004), we use the ratios of accounts receivable to
total assets (Recint) and inventory to total assets (Invint) at the fiscal year-end to proxy for a company’s asset
risk. We also use a dummy variable (Loss) to indicate whether a company has suffered a loss in the most recent
3 years. This variable takes the value of 1 if a loss occurs, and 0 otherwise. We expect the coefficients of all
three variables to be positive.
Some A-share companies are also listed overseas, e.g., on the New York Stock Exchange or Hong Kong
Stock Exchange. Because the annual reports of these companies need to be audited by both domestic and
overseas auditors, they pay both foreign and domestic audit fees, although many fail to disclose them sepa-
rately. We thus include a dummy variable (H_stock) to control for this factor. We assign it a 1 if the company
is listed on the Hong Kong Stock Exchange, and otherwise 0.
Most researchers to date have ignored corporate growth, so whether it is related to audit fees or not
remains unknown. We argue here that both audit costs and risk vary with corporate growth, and, accordingly,
audit fees also vary with growth. A high growth rate is generally accompanied by an increase in total assets,
inventory and/or divisions, which results in greater audit effort and higher audit costs. In addition, a high
growth rate also presents a challenge for management, which may struggle to maintain control. There are
Table 2
Definitions of variables in Model 1.
Name Definition
Explained variable Lnfee Natural logarithm of amount of current year’s external audit fee
Explanatory variable Gov Dummy = 1 if included in SSE Corporate Governance Sector, otherwise 0
Control variables LnAssets Natural logarithm of total assets at the end of the year
Segment Number of industries in which a company is involveda
Subs_rt Square root of number of consolidated subsidiariesb
Recint Accounts receivable/total assets at the end of the year
Invint Inventory/total assets at the end of the year
Loss Dummy = 1 if auditee incurred loss in any of past three fiscal years,
otherwise 0
H_Stock Dummy = 1 if auditee is an H-share company, otherwise 0
TobinQ Value of Tobin’s Q
Big4 Dummy = 1 if audited by Big 4 accounting firm, otherwise 0
a
Collected manually from financial statements.
b
Collected manually from financial statements.
numerous examples of companies experiencing a sudden decline after years of fast-paced growth (e.g., Sanjiu
Medical & Pharmaceutical Co., Ltd., the Giant Group, and the Sanzhu Group). Such cases are often charac-
terized by out-of-control operational and financial management. Hence, a high corporate growth rate may
increase audit risk. To reduce such risk, auditors are likely to increase the number of audit tests, resulting
in higher audit costs. Although the total assets of companies experiencing negative growth may be on the
decline, their incentives to engage in earnings management may strengthen in the face of pressure to report
a profit rather than a loss to retain listing status. Such companies may also undergo frequent management
changes. Both factors increase the audit risk of companies with negative growth. Companies that enjoy steady,
moderate growth, in contrast, are characterized by a lower degree of risk. It is thus possible that the relation-
ship between audit fees and corporate growth may feature a U-shape rather than a linear shape. In the pre-
vious literature, Tobin’s Q and the price-to-book ratio (P/B) are the variables most commonly used to measure
corporate growth (Xiao and You, 2009). In this study, we use Tobin’s Q (TobinQ).
The foregoing control variables primarily represent the characteristics of the demand for audit services.
However, the characteristics of the audit service supplier are also critical influential factors in audit fee deter-
mination, as proved both theoretically and empirically. Francis (1984), Firth (1985), DeFond et al. (2000), Ire-
land and Lennox (2002) and Chen et al. (2007) find evidence of a Big N premium using stock market data from
Australia, New Zealand, Britain, Hong Kong and China, respectively. Using data on 15 countries and dis-
tricts, Choi et al. (2008) also identify a Big 4 premium after controlling for the litigation environment of
the countries/districts under study. In line with previous research, we include Big_4 in our model. We assign
it a value of 1 if the accounting firm belongs to the Big 4,6 and otherwise 0. We expect Big_4 to be positively
related to audit fees.
In addition to these control variables, some scholars argue that profit capability, debt level and industry are
also important factors influencing audit fees. Accordingly, we include return on assets (ROA) (to represent
profit capability), LEVERAGE (a proxy for debt level) and industry variables based on the China Securities
Regulatory Commission (CSRC) industry classification (with finance industry observations eliminated and
manufacturing used as the benchmark) and run a regression using data for 2007 and 2008. The results show
the coefficients of neither ROA nor LEVERAGE to be significant, which is consistent with Zhang and Xu
(2005) and Liu et al. (2003). The coefficients for all of the industry variables, with the exception of the real
estate industry (which has a significantly negative sign), are insignificant. As these additional control variables
add little explanatory power to the model (the adjusted R2 increases by less than 0.04) and exert little influence
on the initial explanatory variables, we do not include them.
Table 2 lists the type, name and definitions of the variables included in Model 1.
6
The Big 4 in this study are Ernst & Young Hua Ming, Deloitte Huayong Certified Public Accountants Co., Ltd., Pricewaterhous-
eCoopers Zhongtian and KPMG Huazhen.
Table 3
Descriptive statistics of variables in Model 1 (2007).
Variable N Mean Std. Dev Median Min Max
Continuous variables
Lnfee 602 13.31661 0.7950948 13.12236 11.91839 18.00517
TobinQ 602 2.036617 1.036366 1.72355 0.5047 7.3216
LnAssets 602 21.71051 1.197106 21.5589 18.49332 27.30113
Recint 602 0.0821234 0.08855 0.0570468 0 0.9750174
Invint 602 0.1689402 0.1492284 0.1353458 0.0001945 0.8766935
Segment 602 2.458472 1.529228 2 1 9
Subs_rt 602 2.759411 1.454904 2.645751 0 11.13553
Value = 1 Value = 0
Freq. Percentage Freq. Percentage
Dummy variables
Gov 602 0.2475083 0.4319234 149 24.75 453 75.25
H_stock 602 0.0481728 0.2143092 29 4.82 573 95.18
Loss 602 0.2292359 0.4206908 138 23.92 464 76.08
Big4 602 0.0913623 0.288363 55 9.14 547 90.86
Lnfee = natural logarithm of amount of current year’s external audit fee.
Gov = 1 if sample company is included in SSE Corporate Governance Sector, and 0 otherwise.
TobinQ = value of Tobin’s Q.
LnAssets = natural logarithm of total assets at the end of the year.
H_stock = 1 if auditee is an H-share company, and 0 otherwise.
Loss = 1 if auditee incurred a loss in any of the past three fiscal years, and 0 otherwise.
Recint = accounts receivable/total assets at the end of the year.
Invint = inventory/total assets at the end of the year.
Big4 = 1 if audited by a Big 4 accounting firm, and 0 otherwise.
Segment = number of industries in which a company is involved.
Subs_rt = square root of number of consolidated subsidiaries.
4.3. Descriptive statistics
Tables 3 and 4 present the results of the descriptive statistics for the 2007 and 2008 observations, respec-
tively. The 2007 sample includes 602 observations, 149 of which (or 24.75% of the total) are included in
the SSE Corporate Governance Sector. The 2008 sample includes 678 observations, 184 of which (or
27.14% of the total) are included in this sector. Although the number of observations included in the SSE Cor-
porate Governance Sector in 2008 increased by 35 (or 23.5%) over 2007, the ratio of observations in the sector
relative to the total sample is almost the same for the 2 years. There are 35 A- and H-share companies (4.57%)
and 51 companies audited by the Big 4 (7.52%) in the 2008 sample, an increase of two and decrease of four,
respectively, relative to the 2007 sample. The number of companies suffering a loss in the most recent 3 years
reached 162 in 2008, an increase of 24 over 2007, although the proportion remained roughly the same in the
2 years. A minor increase in the mean of the natural logarithm of audit fees can be seen in 2008, although the
mean and median are close in that year. The mean of the natural logarithm of total assets is similar. The mean
and median of Tobin’s Q in 2008 are remarkably lower than those in 2007, most likely because of the 2008
international financial crisis. In both years, the Tobin’s Q mean is much higher than the median. Closer scru-
tiny of the sample suggests that this result stems from a number of restructured companies with extraordi-
narily high Tobin’s Q values, but that also have changes in total assets, inventory and branches after
restructuring, thus we do not eliminate these observations. Descriptive statistics also show that the average
number of industries in which a company was involved in 2007 was 2.46, with a maximum of 9 and a
minimum of 1, and the average number of consolidated subsidiaries in that year was 9.73, with a maximum
of 124 and a minimum of 0.7 The figures for 2008 are almost the same.
7
These figures refer to the number of consolidated subsidiaries, whereas the figures in Tables 3 and 4 are the square roots of the number
of consolidated subsidiaries.
5. Empirical results
5.1. Univariate analysis
Table 5 presents the results of univariate analysis of the audit fees and firm characteristics of the sample
companies and the characteristics of their audit firms. This analysis compares companies included in the
SSE Corporate Governance Sector (Governance Sector hereafter) with other firms (Non-governance Sector
hereafter). It can be seen from Panel A that the mean difference in audit fees between the two groups of firms
is highly significant (p-value = 0) in 2007 and 2008, thus providing preliminary evidence that audit fees are
correlated with inclusion in the Governance Sector. However, it is possible that the difference is caused by fac-
tors other than corporate governance (e.g., the scale of total assets). Panel B presents the means of the firm
characteristics of companies in the two groups. We can see that there are significant differences (1% level,
two-tailed) between the groups in terms of firm size, listing on the Hong Kong Stock Exchange, loss occur-
rence and number of subsidiaries in both 2007 and 2008. Furthermore, the difference between the Governance
and Non-governance Sectors is positive for all characteristics other than loss occurrence. The results also show
that there is no statistically significant difference between the two groups in terms of corporate growth, the
ratio of accounts receivable to total assets, the ratio of inventory to total assets and the number of industries
in which a firm is involved. Panel C presents the means of the audit firm characteristics, from which it can be
seen that the between-group difference is highly statistically significant. We can also see that the proportion of
companies audited by one of the Big 4 is larger in the Governance than Non-governance Sector. We believe
that these results show that companies audited by a Big 4 audit firm are much more likely to be included in the
SSE Corporate Governance Sector. Our inference is as follows. If it is true that the Big 4 provide superior
audit quality and can boost the corporate governance level of an auditee, then the results in Panel C show
that related parties recognize companies with good corporate governance, although the SSE Corporate Gov-
ernance Sector is appraised and promulgated on the basis of voluntary applications.
5.2. Multiple regression analysis
5.2.1. Corporate governance and audit fees

We now examine Model 1 using the 2007 and 2008 data. The results are presented in Table 6. The maxi-
mum variance inflation factor (VIF) values are 2.06 and 2.11 in 2007 and 2008, respectively, which indicates
that multicollinearity is not a serious issue.8 The adjusted R2 is 0.7700 in 2007 and 0.7766 in 2008, which indi-
cates that the explanatory power of our model is high and in line with the level achieved in similar research
worldwide.9 The regression results show the coefficient of Gov to be 0.0772639 (significant at the 5% level,
two-sided) in 2007 and 0.0559151 (significant at the 10% level, one-sided) in 2008. These results confirm the
negative influence of corporate governance on audit fees. Further analysis using the 2008 results shows that
companies included in the Governance Sector enjoy a RMB64,46710 (or 5.44%) discount on audit fees over
their Non-governance Sector counterparts,11 thus supporting Hypothesis 1a.
5.2.2. Influence of other factors on audit fees

Our empirical results show firm size to be significantly and positively related to audit fees at the 1% level,
which is consistent with the findings of Simunic (1980), Wu (2003), Han and Zhou (2003) and Zhu and Guo
(2006). Consistent with Francis (1984) and Wu (2003), the coefficient of Big_4 is significantly positive. In
addition, both H_stock and Loss are significantly and positively associated with audit fees at the 1% level.
8
Multicollinearity is believed not to constitute a serious problem if the VIF value is less than 10. The VIF values in Table 6 are all lower
than this critical point, and hence we conclude that there is no serious multicollinearity among the variables in our model.
9
According to Zhang and Liu (2006), the explanatory power of models in domestic audit fee research is generally low (the highest is
0.49, with most in the range of 0.3–0.4), whereas that of models in similar international research is high (most reach 0.7–0.8).
10
Based on the sample firms’ mean audit fees in 2008, i.e., RMB1,185,059.
11
The same conclusion can be drawn using the data for 2007. Holding other factors constant, companies included in the Governance
Sector enjoy an RMB86,631 (or 7.44%) audit fee discount over their counterparts in the Non-governance Sector (based on the mean audit
fees for 2007, i.e., 1,164,405).
Table 4
Descriptive statistics of variables in Model 1 (2008).
Variable N Mean Std. Dev Median Min Max
Lnfee 678 13.36693 0.7904243 13.21767 11.51293 18.00517
TobinQ 678 1.26587 0.4261108 1.14915 0.2157 3.2511
LnAssets 678 21.77525 1.246619 21.60967 18.47492 27.346
Recint 678 0.0777098 0.0781033 0.0549182 0 0.5255643
Invint 678 0.1872533 0.189845 0.1420945 2.02e14 2.460644
Segment 678 2.570796 1.65401 2 1 10
Subs_rt 678 2.771297 1.504253 2.645751 0 11.61895
Value = 1 Value = 0
Freq. Percentage Freq. Percentage
Dummy variables
Gov 678 0.2713864 0.4450033 184 27.14 494 72.86
H_stock 678 0.0457227 0.2090373 31 4.57 647 95.43
Loss 678 0.2389381 0.4267497 162 23.89 516 76.11
Big4 678 0.0752212 0.2639427 51 7.52 627 92.48
Big4 = 1 if audited by Big 4 accounting firm, and 0 otherwise.
Further analysis shows the variation in the coefficients of these four variables over the 2 years to be small,
which indicates the stability of their relationship with audit fees. We also find a strong relationship between
firm complexity and audit fees, with the coefficient of Subs_rt statistically and economically significant at the
1% level. However, the regression results on the relationship between Segment and Lnfee show an inconsis-
tency between 2007 and 2008. The coefficient of Segment is insignificant in 2007 and significant at the 1% level
in 2008. Empirical results also show the ratios of accounts receivable to total assets and inventory to total
assets to be economically and statistically associated with audit fees at a significant level. The coefficient of
Recint is significant at the 10% level (two-sided) in 2007 and 2008 (one-sided), whereas that of Invint is signif-
icant at the 10% level (one-sided) in 2007 and at the 5% level (two-sided) in 2008. This evidence is inconsistent
with the results of Liu et al. (2003) and Zhu and Guo (2006) and with those of Wu (2003), who reports a sig-
nificant positive relationship between accounts receivable (and inventory) and audit fees. In our results, the
sign of Invint is negative. Common sense suggests that the larger a company’s inventory, the greater the audit
risk and the higher the audit cost. Hence, the relationship between inventory and audit fees should be positive.
The cause of the adverse result reported herein is left for future research. One possibility is that the assessed
audit risk of inventory is low, thus prompting a simplified audit procedure.
6. Subsample regressions
It is our belief that corporate growth interacts with corporate governance in such a complex manner that it
is difficult to identify the relationship between them using common methods, i.e., including an interaction term
in the regression equation. Too high or too low a rate of corporate growth will affect firm value and result in
an increase in firm risk. Moderate growth, in contrast, is sustainable growth and the governance structure of
companies experiencing moderate growth is generally more stable. Accordingly, we divide the sample compa-
nies into three subsamples, companies with negative growth, moderate growth and overly fast growth,
depending on their Tobin’s Q value. If this value is less than 1, we include it in the first group. If it is greater
332
Table 5
Comparison between governance and non-governance sectors.
2007 2008
Observations Governance Non-governance Difference p- Observations Governance Non-governance Difference p-
Governance Non-governance Sector Sector Value Governance Non- Sector Sector Value
X. Wu / China Journal of Accounting Research 5 (2012) 321–342

Sector Sector Sector governance
Sector
Panel A: Comparison of mean of natural logarithm of audit fees
Lnfee 149 453 13.61134 13.21967 0.3916634 0.000 184 494 13.68533 13.24834 0.4369928 0.000
Panel B: Comparison of mean of auditee characteristics
TobinQ 149 453 1.99375 2.050717 0.0569669 0.561 184 494 1.237726 1.276353 0.0386277 0.294
LnAssets 149 453 22.41935 21.47736 0.9419973 0.000 184 494 22.5234 21.49659 1.026809 0.000
H_stock 149 453 0.1006711 0.0309051 0.0697661 0.001 184 494 0.0923913 0.0283401 0.0640512 0.000
Loss 149 453 0.0402685 0.2913907 0.2511223 0.000 184 494 0.0326087 0.3157895 0.2831808 0.000
Recint 149 453 0.0775032 0.0836431 0.0061399 0.463 184 494 0.0703157 0.0804638 0.0101482 0.133
Invint 149 453 0.1689522 0.1689362 0.0000159 0.999 184 494 0.1705773 0.1934646 0.0228873 0.163
Segment 149 453 2.375839 2.485651 0.1098123 0.448 184 494 2.51087 2.593117 0.0822478 0.565
Subs_rt 149 453 3.031421 2.669941 0.3614794 0.008 184 494 3.212619 2.606918 0.6057007 0.000
Panel C: Comparison of mean of auditor characteristics
Big4 149 453 0.1744966 0.0640177 0.110479 0.000 184 494 0.1467391 0.0981561 0.096844 0.000
Table 6
Multiple regression results of corporate governance and audit fees in Model 1. Explained variable: Lnfee.
Variable Expected 2007 VIF 2008 VIF
Sign
Intercept ? 5.366365 5.792809 –
(12.97)*** (14.98)***
Gov ? 0.0772639 1.19 0.0559151 1.24
(1.97)** (1.52)
TobinQ ? 0.0804095 1.21 0.1232368 1.24
(4.87)*** (3.21)***
LnAssets + 0.3398957 2.06 0.3210999 2.11
(18.23)*** (18.77)***
H_stock + 1.250754 1.47 1.270307 1.47
(14.21)*** (14.95)***
Loss + 0.1324712 1.17 0.1008036 1.20
(3.31)*** (2.67)***
Recint + 0.3269621 1.05 0.3221629 1.09
(1.81)* (1.64)
Invint + 0.1542988 1.05 0.1875702 1.02
(1.45) (2.40)**
Big4 + 0.6413022 1.56 0.6060236 1.59
(9.52)*** (8.66)***
Segment + 0.0132502 1.10 0.0249864 1.11
(1.24) (2.68)***
Subs_rt + 0.0883677 1.22 0.0935901 1.25
(7.49)*** (8.58)***
N 602 Average 1.31 678 Average 1.33
F 202.19*** 223.48***
R2 0.7738 0.7701
Adj R2 0.7700 0.7667
*
Two-tailed significance at the 0.10 level.
**
***
than the mean12 of the total sample, we include it in the third group, and if it is greater than or equal to 1 and
less than or equal to the mean of the total sample, we include it in the second group. The descriptive statistics
for the three subsamples in 2007 and 2008 are presented in Tables 7 and 8.
To examine the relationship between corporate governance and audit fees in these subsamples, we eliminate
the variable TobinQ in Model 1 and construct Model 2. The definitions of the variables in Model 2 are the
same as those in Model 1.
Lnfee ¼ b0 þ b1 Gov þ b2 Big4 þ b3 LnAssets þ b4 H S tock þ b5 Loss þ b6 Recint þ b7 Invint þ b8 Segment

þ b9 Subs rt þ e: ð2Þ
12
The regression results are similar when we use a threshold other than the mean, such as the median.
Tables 9 and 10 present the Model 2 regression results using the subsamples for 2007 and 2008, respec-
tively.13 The results show a significant negative relationship between corporate governance and audit fees
in the moderate growth sample for both 2007 and 2008. The coefficients for both years are near 10%. This
relationship is insignificant in the negative growth and overly fast growth subsamples. The sign of Gov is
inconsistent between the two samples, possibly because corporate governance and corporate growth have
opposing effects on audit fees. In other words, audit risk’s positive effect on audit fees in negative and overly
fast growth firms offsets the negative effect of corporate governance on audit fees to some extent. For this rea-
son, we observe a mixed and insignificant result.
7. Sensitivity test
Because a company’s inclusion in the SSE Corporate Governance Sector is the result of a self-selection (vol-
untary application) process,14 it is possible that some companies in the Non-governance Sector have good cor-
porate governance, but simply have not submitted an application. If this is the case, the results will be biased.
To determine whether our conclusions are robust, we perform a sensitivity test using the two-stage procedure
developed by Heckman (1976).
In the first stage, we estimate a Probit choice equation and obtain inverse Mills ratios. In the second stage,
we include the inverse Mills ratios as an explanatory variable in the primary model to control for the potential
endogeneity induced by self-selection. A number of strict constraints are necessary in implementing the Heck-
man (1976) procedure successfully. For example, at least one exogenous independent variable that has no
direct effect on the dependent variable in the second-stage regression should be included in the first-stage
choice model. Lennox et al. (2012) find that many accounting studies fail to select proper variables when using
selection models and thus obtain inconsistent results.15 Hence, we carefully select the explanatory variables in
the first stage and include 11 factors considered to have an effect on corporate governance, such as Auditcomm
(establishment of an audit committee), Dual (CEO duality), DirScale (board of director scale), Fnctl (final con-
troller type) and M_Stockholder (frequency of stockholder meetings) in the choice model. Of these factors, at
least Fnctl and M_Stockholder have no significant effect on audit fees16 and can thus play the role of an exog-
enous independent variable excluded in the second-stage regression. The first-stage choice equation is as fol-
lows (Model 3).
ProbitðGov ¼ 1Þ
¼ b0 þ b1 Auditcomm þ b2 Dual þ b3 First þ b4 Second þ b5 Auditopinion þ b6 M Dir
þ b7 M Supervisor þ b8 M Stockholder þ b9 DirScale þ b1 0IndDirb11 Fnctl þ b12 TobinQ
þ b13 Big4 þ b14 LnAssets þ b15 H stock þ b16 Loss þ b17 Recint þ b18 nvint þ b19 Segment
þ b20 Subs rt þ e: ð3Þ
13
We also use the P/B ratio as a criterion to regress Model 2. The results are consistent with those reported.
14
The Expert Consultative Committee for Corporate Governance Sector Appraisal was officially founded in September 2007. In the same
month, the “Appraisal Measures of the SSE Corporate Governance Sector (Draft Version)” was published to solicit the opinions of all
parties. The “Appraisal Measures of the SSE Corporate Governance Sector” were officially released on October 9, after which listed
companies could voluntarily submit applications. By November 2, 2007, the SSE had received valid application materials from 255
companies, published these materials and solicited public comment. Then, on the basis of the appraisal results of the Expert Consultative
Committee, the final Sector list included 199 companies.
15
Lennox et al. (2012) report the results of selection models to be sensitive to model shape and note that an absence of exclusion
restrictions can lead to severe multicollinearity problems.
16
Few researchers report the type of final controller or frequency of stockholder meetings to have a significant influence on audit fees. We
obtain results that are consistent with previous research using data for 2006, 2007 and 2008.
Table 7
Descriptive statistics of variables in subsample (2007).
Variable Negative growth (N = 20) Moderate growth (N = 372) Overly fast growth (N = 210)

Mean Std. Dev Median Min Max Mean Std. Dev Median Min Max Mean Std. Dev Median Min Max
Lnfee 13.72293 1.023866 13.49333 12.42922 15.95558 13.37225 0.875986 13.21767 11.91839 18.00517 13.17936 0.565867 13.12236 11.91839 15.75137
LnAssets 22.74855 1.175557 22.72143 20.73494 24.98712 21.96001 1.181509 21.71716 19.91666 27.30113 21.16967 1.004719 21.11371 18.49332 24.15856
Recint 0.047737 0.061352 0.021253 0.000511 0.213918 0.082786 0.091054 0.060145 0 0.975017 0.084225 0.085850 0.053721 0.000025 0.480798
Invint 0.141426 0.139168 0.118763 0.002116 0.528862 0.173637 0.151263 0.139192 0.000195 0.876694 0.163240 0.146684 0.128984 0.000937 0.812561
Segment 1.9 1.165287 1.5 1 5 2.424731 1.469265 2 1 9 2.571429 1.650618 2 1 8
Subs_rt 2.654131 1.300035 2.645751 0 5.91608 2.797157 1.448523 2.64575 0 9.539392 2.702572 1.483857 2.44949 0 11.13553
Value = 1 Value = 0 Value = 1 Value = 0 Value = 1 Value = 0
Freq. Percentage Feeq. Percentage Feeq. Percentage Freq. Percentage Freq. Percentage Freq. Percentage
Dummy variables
Gov 0.15 0.366348 0 3 15 17 85 0.266129 0.442528 0 99 27 273 73 0.223810 0.417792 0 47 22 163 78
H_stock 0.25 0.444262 0 5 25 15 75 0.053763 0.225854 0 20 5 352 95 0.019048 0.137019 0 4 2 206 98
Loss 0.2 0.410391 0 4 20 16 80 0.193548 0.395611 0 72 19 300 81 0.295238 0.457240 0 62 30 148 70
Big4 0.25 0.444262 0 5 25 15 75 0.104839 0.306758 0 39 10 333 90 0.052381 0.223326 0 11 5 199 95

335
336
Table 8
Descriptive statistics of variables in subsample (2008).

Variable Negative Growth (N = 157) Moderate Growth (N = 274) Overly Fast Growth (N = 247)
Mean Std. Dev Median Min Max Mean Std. Dev Median Min Max Mean Std. Dev Median Min Max
Lnfee 13.75278 1.102427 13.45884 12.25486 18.00517 13.32945 0.682444 13.21767 12.20607 16.70588 13.16326 0.541855 13.12236 11.51293 15.73243
LnAssets 22.45753 1.382253 22.13158 19.87016 27.346 21.9469 1.109534 21.7087 18.73775 25.14214 21.15117 0.989002 21.10941 18.47492 24.85722
Recint 0.059195 0.060625 0.039716 0 .3320666 0.077571 0.080662 .0585952 0 0.525564 0.089632 0.082916 0.067365 0 0.399786
Invint 0.168962 0.166717 0.126629 2.02e14 .7681355 0.199915 0.191007 .1442456 3.36e14 0.940148 0.184833 0.201679 0.147006 7.73e14 2.460644
Segment 2.624204 1.718728 2 1 10 2.49635 1.635791 2 1 9 2.619433 1.635762 2 1 8
Subs_rt 2.829213 1.532347 2.645751 0 9.69536 2.858444 1.576183 2.645751 0 9.486833 2.637812 1.397386 2.44949 0 11.61895
Value = 1 Value = 0 Value = 1 Value = 0 Value = 1 Value = 0
Freq. Percentage Freq. Percentage Freq. Percentage Freq. Percentage Freq. Percentage Freq. Percentage
Dummy variables
Gov 0.286624 0.453631 0 45 29 112 71 0.277372 0.448521 0 76 28 198 72 0.226721 0.419561 0 56 23 191 77
H_stock 0.165605 0.372915 0 26 17 131 83 0.018248 0.134093 0 5 2 269 98 0 0 0 0 0 247 100
Loss 0.152866 0.361010 0 24 15 133 85 0.229927 0.421556 0 63 23 211 77 0.303644 0.460764 0 75 30 172 70
Big4 0.165605 0.372915 0 26 17 131 98 0.065693 0.248199 0 18 7 256 93 0.028340 0.166280 0 7 3 240 97

Table 9
Multiple regression results of corporate governance and audit fees in Model 2 (2007). Explained variable: Lnfee.
Variable Expected sign Negative growth VIF Moderate growth VIF Overly fast growth VIF
Intercept ? 2.138893 – 4.86583 – 7.902495 –
(0.37) (9.86)*** (12.34)***
Gov ? 0.1673906 1.92 0.0980085 1.19 0.0383723 1.21
(0.41) (1.98)** (0.61)
LnAssets + 0.4615434 7.48 0.3709851 1.82 0.2307721 1.65
(1.83)* (16.20) *** (7.55)***
H_stock + 1.45432 2.49 1.36174 1.45 0.9150677 1.19
(3.78)*** (12.76)*** (4.81)***
Loss + 0.572797 2.92 0.150768 1.13 0.033121 1.26
(1.27) (2.80)*** (0.56)
Recint + 1.33218 2.86 0.2718718 1.07 0.1819705 1.06
(0.45) (1.19) (0.63)
Invint + 1.362542 1.66 0.2010797 1.06 0.2201791 1.04
(1.36) (1.48) (1.32)
Big4 + dropped – 0.6569943 1.60 0.539355 1.18
(7.97)*** (4.64)***
Segment + 0.1629168 1.86 0.0144943 1.12 0.0198448 1.13
(1.28) (1.00) (1.29)
Subs_rt + 0.0058437 1.83 0.0936485 1.19 0.1162623 1.45
(0.05) (6.20) *** (5.98)***
N Total 602 20 Average 2.88 372 Average 1.29 210 Average 1.24
F 9.80*** 172.33*** 39.91***
R2 0.8770 0.8108 0.6424
Adj R2 0.7875 0.8061 0.6263
*
**
***
The definitions of the variables in Model 3 can be found in Table 11. It should be noted that because the
SSE Corporate Governance Sector was launched in 2007, the data used in the choice equation is for 2006. We
obtain 725 sample observations17 and the regression results are presented in Table 12.
The second step is to include the inverse Mills ratios (Lambda) in Model 1 and construct Model 4. The
regression results are reported in Table 13.
Lnfee ¼ b0 þ b1 Gov þ b2 TobinQ þ b3 Big4 þ b4 nAssets þ b5 H Stock þ b6 Loss þ b7 Recint þ b8 Invint

þ b9 Segment þ b10 Subs rt þ b11 Lambda þ e: ð4Þ
17
A total of 835 companies were listed on the SSE in 2006. According to the “Appraisal Measures of the SSE Corporate Governance
Sector,” necessary conditions for inclusion in the SSE Corporate Governance Sector are having been listed on the SSE for 12 months and
no special treatment status. Hence, we eliminate 74 sample observations marked ST or ST and 5 sample observations listed after October
9, 2006. In addition, we also eliminate 15 sample companies in the finance industry and 16 with incomplete data. The final sample thus
includes 725 observations.
Table 10
Multiple regression results of corporate governance and audit fees in Model 2 (2008). Explained variable: Lnfee.
Variable Expected sign Negative growth VIF Moderate growth VIF Overly fast growth VIF
Intercept ? 4.15681 – 5.789669 – 6.978095 –
(4.85)*** (11.11)*** (11.22)***
Gov ? 0.1201422 1.29 0.1018529 1.21 0.036208 1.26
(1.41) (1.98)** (0.59)
LnAssets + 0.3973288 2.30 0.3294044 1.60 0.2736142 1.61
(10.21)*** (13.75)*** (9.22)***
H_stock + 1.321769 1.85 1.2595 1.22 Dropped –
(10.21)*** (7.28)***
Loss + 0.1286156 1.14 0.0343346 1.18 0.1343597 1.33
(1.22) (0.64) (2.33)**
Recint + 2.2625 1.22 0.063834 1.11 0.05694 1.01
(3.51)*** (0.23) (0.20)
Invint + 0.3300925 1.07 0.2394448 1.05 0.0421028 1.01
(1.50) (2.13)** (0.36)
Big4 + 0.3631917 2.04 0.650547 1.35 0.8505657 1.15
(2.67)*** (6.62)*** (5.71)***
Segment + 0.0408373 1.13 0.0125083 1.13 0.0318546 1.12
(1.86)* (0.92) (2.13)**
Subs_rt + 0.0808408 1.24 0.0968029 1.33 0.0925125 1.30
(3.14)*** (6.31)*** (4.90)***
N Total 678 157 Average 1.48 274 Average 1.24 247 Average 1.23
F 91.00*** 87.99*** 38.95***
R2 0.8478 0.7500 0.5670
Adj R2 0.8385 0.7415 0.5524
Segment = number of industries a company involved in.
*
**
***
As can be seen from Table 13, the coefficient and sign of Gov are larger than and consistent with the benchmark
results, respectively. The coefficient in 2008 is approximately 70% greater than that of the benchmark. Although
Gov’s significance level declines slightly in 2007, it remains significant at the 10% level (two-sided). Its significance
level in 2008 reaches the 5% level (two-sided). The coefficients of most of the control variables vary within 10%,
with the exception of those of Recint and Invint. The significance level of the control variables is the same in the
2 years, except for Recint in 2007 (which changes from significant at the 10% level to insignificant) and Segment in
2008 (from the 1% level to the 5% level). The t-statistics for Lambda are 0.51 and 1.28 in 2007 and 2008,
respectively. The maximum VIF value is 2.44, which indicates that the model suffers no serious multicollinearity
problems. Thus, the sensitivity test results demonstrate that our findings are robust to self-selection.
8. Conclusions and limitations
This paper reports the results of an empirical investigation of the relationship between corporate gover-
nance and audit fees using data disclosed in the annual financial reports of companies listed on the Shanghai
Table 11
Definitions of variables in Model 3.
Name Definition
Explained Gov Dummy = 1 if included in SSE Corporate Governance Sector, otherwise 0
variable
Explanatory Auditcomm Dummy = 1 if audit committee is set up, otherwise 0
variables Dual Dummy = 1 if president and CEO are the same person, otherwise 0
First Shares held by first major shareholder/total shares at year-end
Second Shares held by second major shareholder/total shares at year-end
Auditopinion Numerical variable: 1 if a clean opinion, 2 if an unqualified opinion with emphasis of matter paragraph, 3
if a qualified opinion, 4 if a disclaimer of opinion
M_Dir Frequency of board of director meetings held in a fiscal year
M_Supervisor Frequency of board of supervisor meetings held in a fiscal year
M_Stockholder Frequency of stockholder meetings held in a fiscal year
DirScale Number of members of board of directors disclosed in annual report
IndDir Number of independent directors on the board of directors disclosed in annual report
Fnctl Dummy = 1 if owned by the state, otherwise 0
TobinQ Value of Tobin’s Q
Big4 Dummy = 1 if audited by Big 4 accounting firm, otherwise 0
LnAssets Natural logarithm of total assets at the end of the year
H_Stock Dummy = 1 if auditee is an H-share company, otherwise 0
Loss Dummy = 1 if auditee incurred a loss in any of the past three fiscal years, otherwise 0
Recint Accounts receivable/total assets at the end of the year
Invint Inventory/total assets at the end of the year
Segment Number of industries in which a company is involved
Subs_rt Square root of number of consolidated subsidiaries
Table 12
Probit regression results. Explained variable: Gov.
Variable Auditcomm Dual First Second Auditopinion M_Dir M_Supervisor M_Stock
holder
Result 0.1164814 0.270078 0.5186868 0.025993 0.506515 0.013953 0.0543591 0.012199
(1.04) (1.40) (1.25) (0.03) (1.95)** (0.75) (1.48) (0.25)
Variable DirScale IndDir Fnctl TobinQ Big4 LnAssets H_Stock Loss
Result 0.0181251 0.1075773 0.0592087 0.3857623 0.1731616 0.2789833 0.3502073 1.072429
(0.49) (1.41) (0.44) (2.55)** (0.73) (3.95)*** (0.84) (4.93)***
Variable Recint Invint Segment Subs_rt Intercept N LR chi2 Pseudo R2
Result 0.536 0.085569 0.114073 0.040388 7.272783 725 109.26*** 0.1696
(0.96) (0.31) (2.76)*** (1.05) (4.85)***
Figures in parentheses are Z-values.

**
***
Stock Exchange in the first and second years after the introduction of the SSE Corporate Governance Sector,
i.e., 2007 and 2008. The results based on the full sample show this relationship to be significant and negative.
In general, the audit fees of companies included in this sector are 5.44–7.44%18 lower than those of their Non-
governance Sector counterparts. These results suggest that substitution theory provides a better explanation of
the relationship between corporate governance and audit fees than signaling theory. Subsample data also
shows corporate governance’s influence on audit fees is affected by corporate growth. The negative relation-
18
The percentages are 5.44% in 2008 and 7.44% in 2007.
Table 13
Regression results of Model 4. Explained variable: Lnfee.
2007 2008
Benchmark Two-stage regression VIF Benchmark Two-stage regression VIF
Intercept 5.366365 5.684087 – 5.792809 6.164501 –
(12.97)*** (13.88)*** (14.98)*** (16.05)***
Gov 0.0772639 0.0903212 2.19 0.0559151 0.0930575 2.00
(1.97)** (1.72)* (1.52) (2.01)**
TobinQ 0.0804095 0.073934 1.22 0.1232368 0.1145037 1.25
(4.87)*** (4.58)*** (3.21)*** (2.98)***
LnAssets 0.3398957 0.3246977 1.98 0.3210999 0.3032975 2.05
(18.23)*** (17.55)*** (18.77)*** (17.79)***
H_stock 1.250754 1.124668 1.40 1.270307 1.094173 1.42
(14.21)*** (12.51)*** (14.95)*** (12.32)***
Loss 0.1324712 0.141715 1.32 0.1008036 0.1104056 1.28
(3.31)*** (3.41)*** (2.67)*** (2.89)***
Recint 0.3269621 0.2719119 1.06 0.3221629 0.2791031 1.09
(1.81)* (1.54) (1.64) (1.45)
Invint 0.1542988 0.1257606 1.04 0.1875702 0.1685315 1.02
(1.45) (1.19) (2.40)** (2.18)**
Big4 0.6413022 0.6755913 1.51 0.6060236 0.6677469 1.58
(9.52)*** (10.15)*** (8.66)*** (9.43)***
Segment 0.0132502 0.0100982 1.10 0.0249864 0.022193 1.11
(1.24) (0.96) (2.68)*** (2.40)**
Subs_rt 0.0883677 0.0935087 1.22 0.0935901 0.0961795 1.27
(7.49)*** (8.03)*** (8.58)*** (8.79)***
Lambda 0.085819 2.44 0.0626612 2.09
(0.51) (1.28)
N 602 592 Average 1.50 678 662 Average 1.47
F 202.19*** 159.50*** 223.48*** 171.28***
R2 0.7738 0.7516 0.7701 0.7435
Adj R2 0.7700 0.7468 0.7667 0.7392
Lambda = inverse Mills ratio.
Note: The difference in the number of observations between the two-stage regression (Model 4) and basic regression (Model 1) for 2007 is
due to the observations in the latter including companies listed after 2006. The procedure used to calculate Lambda means the Lambda
values for these observations are missing, which is why the difference occurs in 2008.
*
**
***
ship between corporate governance and audit fees is found to be economically and statistically significant in
sample firms that experienced moderate growth during the sample period, relative to those that experienced
overly fast or negative growth, for which the relationship is mixed and insignificant.
The SSE Corporate Governance Sector was introduced near the end of 2007. Although we find corporate
governance to have an economically significant influence on audit fees, the degree of statistical significance is
relatively low (10% level, one-sided in the full-sample regression for 2008). There are two main explanations
for this finding in addition to the effect of corporate growth. First, audit fees are characterized by inertia.
When audit firms initially negotiate their fees with clients prior to provision of the first audit service, they per-
form a comprehensive evaluation of the company, determine the audit risk level, estimate the audit costs and
finally determine the charging criteria. Although regulations require audit firms to perform such routine work
as evaluating the audit risk level and determining the audit procedure and test scope every year, in practice
they may keep audit fees fixed for many years, thus demonstrating inertia. Of the 536 sample companies in
2008 that exhibited comparability19 to those in 2007, 283 companies (or 48.29%) saw no change in audit fees.
Second, it takes time for stakeholders to comprehend the signal conveyed by corporate governance. As noted
in the introduction to this paper, there are two competing explanations concerning the relationship between
corporate governance and audit fees, one informed by substitution theory and the other by signaling theory. If
listed companies are rational economic beings, then they will prefer substitution theory to signaling theory, as
its logic suggests that audit fees will decrease and firm value increase. Acceptance of signaling theory is more
complicated. Signaling high-level corporate governance through a high-quality audit requires a large expen-
diture on auditing. Hence, a company’s acceptance of signaling theory depends on the tradeoff between expen-
diture and the expected return.20 The situation is the opposite for audit firms. They tend to prefer signaling
theory, as it allows them to charge higher fees with no increase in audit risk, whereas the logic of substitution
theory requires that they balance a decrease in fees and an increase in audit risk with a reduction in the number
of audit tests. Both auditees and auditors clearly need time to consider the economic consequences of signaling
good corporate governance and adopt audit plans that favor themselves when negotiating audit fees. Reaching
consensus may take a considerable amount of time. Our empirical evidence is largely in accord with the first
explanation, i.e., that audit fees are characterized by inertia, although its validity requires testing with data for
and beyond 2009.
This study suffers two limitations. The first lies in the sample data. Because audit fees may be calculated in a
variety of ways (e.g., they may or may not be inclusive of travel expenses and the fees for interim reports), we
cannot infer whether the data on companies that do not disclose detailed audit fees are consistent with those of
other companies. The second limitation lies in self-selection. Although we perform a sensitivity test using the
two-stage procedure developed by Heckman (1976), we cannot completely rule out the influence of self-selec-
tion owing to the complexity of dealing with such a problem.
References
Beiner, S., Drobetz, W., Schmid, F., Zimmermann, H., 2003. Is board size an independent corporate governance mechanism? Kyklos 57,
327–356.
Cai, J., 2007. Corporate governance, audit risk and audit fees. Auditing Research (in Chinese) 3, 65–71.
Carcello, J.V., Hermanson, D.R., Neal, T.L., Riley Jr., R.R., 2002. Board characteristics and audit fees. Contemporary Accounting
Research 19 (3), 365–384.
Chen, D.H., Zhou, C.Q., 2006. The effect of self-selection system on audit fees: empirical evidence from Chinese listed companies. Journal
of Finance and Economics (in Chinese) 3, 44–55.
Chen, C.J.P., Su, X.J., Wu, X., 2005. Abnormal audit fees and the improvement of unfavourable audit outcomes. China Accounting and
Finance Review 7 (4), 1–54.
Chen, C.J.P., Su, X.J., Wu, X., 2007. Market competitiveness and Big 5 pricing: evidence from China’s binary market. The International
Journal of Accounting 42 (1), 1–24.
Choi, J.H., Kim, J.B., Liu, X., Simunic, D.A., 2008. Audit pricing, legal liability regimes, and Big 4 premiums: theory and cross-country
evidence. Contemporary Accounting Research 25 (1), 55–99.
Cui, X.G., Wang, L.Y., Xu, H., 2007. High-speed growth, financial crisis and risk forecasting. Accounting Research (in Chinese) 12, 55–
62.
DeFond, M.L., Francis, F.R., Wong, T.J., 2000. Auditor industry specialization and market segmentation: evidence from Hong Kong.
Auditing: A Journal of Practice and Theory 19 (Spring), 49–66.
Drobetz, W., Schillhofer, A., Zimmermann, H., 2004. Corporate governance and expected stock returns: evidence from Germany.
European Financial Management 10, 267–293.
19
Here, comparability refers to the company disclosing its audit fees in both its 2007 and 2008 annual reports.
20
The return may be an increase in firm value owing to an improvement in corporate governance. As consideration of this issue is not the
main research objective of the study, we provide no evidence of the relationship between corporate governance and firm value.
Fan, J.P.H., Wong, T.J., 2005. Do external auditors perform a corporate governance role in emerging markets? Evidence from East Asia.
Journal of Accounting Research 43, 35–72.
Firth, M., 1985. An analysis of audit fees and their determinants in New Zealand. Auditing: A Journal of Practice and Theory 4 (Spring),
23–37.
Francis, J., 1984. The effect of audit firm size on audit prices: a study of the Australian market. Journal of Accounting and Economics 6,
133–151.
Francis, J.R., Stokes, D.J., 1986. Audit prices, product differentiation, and scale economies: further evidence from the Australian market.
Journal of Accounting Research 24, 383–393.
Gao, M.M., Gao, Y., 2008. Corporate governance and audit pricing: empirical evidence from Chinese listed companies. Productivity
Research 9 (147–148), 155.
Gul, F.A., 2001. Hong Kong Auditing: Economic Theory and Practice. City University of Hong Kong Press.
Gul, F.A., Tsui, J.S.L., 2001. Free cash flow, debt monitoring, and audit pricing: further evidence on the role of director equity ownership.
Auditing: A Journal of Practice and Theory 20, 72–84.
Gul, F.A., Tsui, J.S.L., Chen, C.J.P., 1998. Agency Costs and Audit Pricing: Evidence on Discretionary Accruals (Working Paper).
Han, H.J., Zhou, S.C., 2003. Research on accountants’ payments in the Chinese stock market: empirical analysis of Chinese listed
companies’ data. Management World 2, 15–22.
Hay, D., Knechel, W.R., Wong, N., 2004. Audit Fee: A Meta-analysis of The Effect of Supply and Demand Attributes, University of
Auckland (Working Paper).
Heckman, J.J., 1976. The common structure of statistical models of truncation, sample selection and limited dependent variables and a
simple estimator for such models. Annals of Economic and Social Measurement 5 (4), 475–492.
Ireland, J.C., Lennox, C.S., 2002. The large audit firm fee premium: a case of selectivity bias? Journal of Accounting, Auditing and
Finance 17, 73–91.
Jensen, M., Meckling, W., 1976. Theory of the firm: managerial behavior, agency costs and ownership structure. Journal of Financial
Economics 3, 305–360.
Lang, L., Ofek, E., Stulz, R.M., 1996. Leverage, investment, and firm growth. Journal of Financial Economics 40 (1), 3–29.
Larcker, D.F., Richardson, S.A., 2004. Fees paid to audit firms, accrual choices, and corporate governance. Journal of Accounting
Research 42 (3), 625–658.
Lennox, C., Francis, J., Wang, Z., 2012. Selection models in accounting research. The Accounting Review 87 (2), 589–616.
Li, B.X., Wang, P.X., 2006. Board characteristics and ratios of audit fee to asset. China Accounting Review 1, 105–118.
Liu, M.H., Hu, B., 2006. Corporate governance, agency costs and audit pricing: an empirical analysis on Chinese A-stock market panel
data from 2001 to 2003. Research on Financial and Economic Issues 2, 72–79.
Liu, B., Ye, J.Z., Liao, Y.Y., 2003. An empirical study of the determinants of audit fees. Auditing Research (in Chinese) 1, 44–47.
Pan, K.Q., 2008. Corporate governance, audit risk and audit pricing: empirical evidence based on CCGINK. Nankai Business Review 1,
106–112.
Simunic, D., 1980. The pricing of audit services: theory and evidence. Journal of Accounting Research 18, 161–190.
Spence, M., 1973. Job market signaling. The Quarterly Journal of Economics 87 (3), 355–374.
Wang, Z.L., 2002, Determinants of Audit Fees and Audit Quality: Evidence from China’s Listed Companies. Doctoral Paper, Shanghai
University of Finance and Economics.
Wang, Y., 2009. Control chain, agency conflicts and auditor choice. Accounting Research (in Chinese) 6, 65–72.
Wang, P., Zhou, L.A., 2006. Auditor’s choice by China’s listed companies and its governance role. China Accounting Review 2, 88–98.
Watts, R.L., Zimmerman, J.L., 1983. Agency problems, auditing, and the theory of the firm: some evidence. Journal of Law and
Economics 26 (October), 613–633.
Wu, L.N., 2003. The determinants of audit fees: evidence from the Chinese stock market’s first-time disclosure of audit fees. China
Accounting Review (in Chinese) 1, 113–128.
Xiao, M., You, J., 2009. Private benefits of control, growth opportunities and investor protection. China Journal of Accounting Research
2, 123–145.
Zhang, J.X., Liu, C.L., 2006. A study review on audit fees and its inspiration. Contemporary Finance & Economics 7, 123–128.
Zhang, J.X., Xu, Y., 2005. The determinants of audit fees: evidence from China’s listed companies in 2001–2003. China Accounting
Review (in Chinese) 1, 99–116.
Zhang, M., Zhang, Q.F., 2005. Ownership structure, auditor choice and audit pricing: evidences from Chinese securities market.
Accounting and Finance, 1. <http://www.sinoss.net/qikan/2011/0130/11784.html>.
Zhu, X.P., Guo, Z.Y., 2006. Research on the information content of an increase in audit fees without auditor switching. Auditing
Research (in Chinese) 2, 64–68.
Zhu, X.P., Yu, Q., 2004. The determinants of audit fees: an empirical analysis in China. China Accounting Review 2, 393–408.

Corporate Governance

Uploaded by

Copyright:

Available Formats

Corporate Governance

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Corporate Governance

Uploaded by

Copyright:

Available Formats

China Journal of Accounting Research 5 (2012) 321–342

Contents lists available at SciVerse ScienceDirect

China Journal of Accounting Research

Corporate governance and audit fees: Evidence from

E-mail address: [email protected]

Production and hosting by Elsevier

2. Overview of prior research

2.1. Factors associated with audit fees

2.2. Corporate governance and audit fees

3. Theory and hypotheses

4. Data and variables

4.1. Data and sample

4.2. Model and variables

4.3. Descriptive statistics

5.1. Univariate analysis

5.2. Multiple regression analysis

5.2.1. Corporate governance and audit fees

5.2.2. Inﬂuence of other factors on audit fees

X. Wu / China Journal of Accounting Research 5 (2012) 321–342

Lnfee ¼ b0 þ b1 Gov þ b2 Big4 þ b3 LnAssets þ b4 H S tock þ b5 Loss þ b6 Recint þ b7 Invint þ b8 Segment

X. Wu / China Journal of Accounting Research 5 (2012) 321–342

Lnfee = natural logarithm of amount of current year’s external audit fee.

X. Wu / China Journal of Accounting Research 5 (2012) 321–342

Lnfee = natural logarithm of amount of current year’s external audit fee.

Lnfee ¼ b0 þ b1 Gov þ b2 TobinQ þ b3 Big4 þ b4 nAssets þ b5 H Stock þ b6 Loss þ b7 Recint þ b8 Invint

8. Conclusions and limitations

You might also like