lin and liu - 2009 - the determinants of auditor switching from the perspective of corporate governance in china

An effective corporate governance mechanism may inhibit the controlling shareholder from switching to a smaller auditor to exploit the minority shareholders.. Keywords: Corporate Governa

The Determinants of Auditor Switching from

the Perspective of Corporate Governance in

China

Z Jun Lin and Ming Liu*

ABSTRACT

firms and their auditor switching types in the Chinese context Two types of auditor switching – namely switching to a larger auditor or switching to a smaller auditor – are identified and examined

demonstrate that to realize opaqueness gains, firms with weaker corporate governance generally are more likely to switch

to a smaller auditor rather than to a larger one

affect a firm’s auditor switching decision An effective corporate governance mechanism may inhibit the controlling shareholder from switching to a smaller auditor to exploit the minority shareholders The emerging economies are usually featured with concentrated ownership and insufficient legal protection of minority shareholders Compared with prior studies, this paper generates findings more applicable to the emerging economies

of the structural arrangement of corporate governance of the listed firms, the independent auditing process and the credibility of financial reporting in an emerging market like China The findings also suggest that in order to bolster the confidence of the market participants, the Chinese government should promote the reform of the corporate governance system and enforce effective regulations, in particular on firms’ auditor switches

Keywords: Corporate Governance, Supervisory Board, Government Ownership, State-Owned Enterprise (SOE), Auditor Switching, China

INTRODUCTION

The purpose of this study is to investigate the association

between a firm’s internal corporate governance

mecha-nism and its auditor switching decisions in the Chinese

context An independent auditing function can detect and

disclose earnings management and other types of

miscon-duct by business managers or controlling shareholders In

general, if a firm has established a sound corporate

gover-nance mechanism, the firm’s management or its controlling

shareholders will not have a free hand in making decisions

on auditor selection, and vice versa Hence, there should be

an association between a firm’s corporate governance and its auditor selection decision This study empirically investi-gates the relationship between a firm’s internal corporate governance mechanism (proxied by ownership concentra-tion, effectiveness of the supervisory board [SB] monitoring and shared board of directors [BoD] chairman and CEO) and their auditor switching decisions in the context of cor-porate governance practice in China

This study was motivated by several factors First, as cor-porate governance has a positive impact on corcor-porate finan-cial reporting and auditing processes, a study of auditor switching with respect to the internal corporate governance mechanism may assist analysis of auditing quality and the auditor’s roles in ensuring the credibility of corporate finan-cial disclosures Second, for the controlling owners, there is

a tradeoff between hiring a high-quality auditor to lower the

*Address for correspondence: Department of Accounting and Information

Manage-ment, Faculty of Business Administration, University of Macau, Macau, China E-mail:

Mliu@umac.mo

Corporate Governance: An International Review, 2009, 17(4): 476–491

costs of raising capital and hiring a low-quality auditor to

maintain the gains from the opaqueness of corporate

gover-nance (called “opaqueness gains,” such as tunneling

behav-iors to transfer resources from a listed firm to its controlling

shareholder) The ongoing bear market in China during the

2001–2004 period provides a good opportunity to

disen-tangle these two incentives, thus allowing us to pinpoint the

association between a firm’s internal corporate governance

mechanism and its auditor switching decisions Third, the

Chinese Institute of Certified Public Accountants (CICPA)

began to rank auditors in China in the early 2000s in order to

improve the transparency of the Chinese auditing market,

thus allowing the possibility of identifying high-quality

audi-tors in the Chinese market An investigation of the

determi-nants of auditor switching from the perspective of corporate

governance should contribute to an understanding of the

necessity and utility of independent audits in China

Our regression results show that firms with larger

control-ling shareholders (a higher degree of ownership

concentra-tion) or firms where the positions of board of directors’

chairman and CEO are held by the same person are more

likely to switch to a smaller auditor rather than to a larger

one However, the supervisory board monitoring strength is

not a significant factor underlying auditor switching

deci-sions The findings suggest that firms with a weak internal

corporate governance mechanism generally tend to switch

to smaller or more pliable auditors to sustain the opaqueness

gains The finding that the supervisory board function

(proxied by its size to represent monitoring effectiveness)

does not have a significant influence on auditor switching

decisions may imply that the supervisory board’s

monitor-ing function is not consistently effective in practice

The remainder of the paper is arranged as follows The

second section reviews the relevant literature The third

section develops the hypotheses to examine the association

between the internal corporate governance mechanism and

auditor switching decisions The fourth section introduces a

regression model to test the hypotheses The fifth section

presents and discusses the empirical results Sensitivity tests

are presented in the sixth section Finally, the seventh section

provides some conclusions

LITERATURE REVIEW The Development of Auditing Profession in China

Shortly after the founding of the People’s Republic of China

in 1949, the auditing profession in China disappeared entirely

due to the public (state) ownership of all production means

An independent auditing function was virtually nonexistent

in the planned economy before the 1980s when the state

owned and ran enterprises directly But the mushrooming of

Sino-foreign joint ventures, brought about by the

govern-ment’s adoption of the “open-door” policy in the early 1980s,

led to the emergence of independent auditing Due to the

involvement of non-state equity interests in joint-ventures, it

became necessary to have independent professionals, or

certified public accountants (CPAs), to verify capital

contri-butions and audit annual financial statements and income

tax returns (Lin, Tang, & Xiao, 2003) Thus, the CICPA,

a quasi-governmental organization in charge of national

administration of CPAs and auditing firms, was established

in the early 1980s Following the business restructuring cam-paign, shareholding (stock) companies reappeared in the Chinese economy at the turn of 1990s, resulting in a further sharp increase in the demand for external audits The estab-lishment of the Shanghai and Shenzhen stock exchanges and the promulgation of new accounting and auditing standards also played an important role in this process The China Securities Regulatory Commission (CSRC) required that the annual reports of all listed firms be audited by registered Chinese CPAs

The Role of the Auditing Function

In contemporary market economies, business incorporation leads to the separation of ownership and management Professional managers, rather than owners (shareholders), are directly involved in daily business operations Due to various self-interests and information asymmetries, busi-ness managers are able to pursue their own interests at the expense of those of the owners and other stakeholders (Jensen & Meckling, 1976) One of the binding mechanisms over management operations and information disclosure is the auditing function performed by independent profession-als (Watts and Zimmerman, 1986)

Nonetheless, the utility of the auditing function depends upon quality of the audit which is determined by the inde-pendence and expertise of the auditors (DeAngelo, 1981; Watkins, Hillison, & Morecroft, 2004) Audit quality is con-sidered to be commensurate with the size of the auditors, i.e., larger auditors should have a higher degree of indepen-dence, possess more industrial expertise and resources, and bear higher reputation costs, so they can provide higher-quality auditing services (DeAngelo, 1981; Lennox, 2005) Investors perceive the accounting numbers (e.g., earnings and book values) audited by large auditors to have better information quality, and therefore attach greater market value to them (Lennox, 2005; Watkins et al., 2004)

DeFond and Subramanyam (1998) argue that there are incentives for the controlling shareholders of firms to pursue their own interests by manipulating the accounting numbers or transferring resources through “tunneling” behaviors Thus, the controlling shareholders will weigh their own self-interests when making auditor selection deci-sions (Johnson, La Porta, Lopez-de-Silanes, & Shleifer, 2000;

La Porta, Lopez-de-Silanes, Shleifer, & Vishny, 2002) On the one hand, selecting a large auditor will signal to the market that the financial statements are more reliable, thus the firms may benefit from lower capital-raising costs on the equity or debt market On the other hand, large auditors may be more stringent in detecting and reporting “tunneling behavior” and hence may deprive the controlling shareholders of their opaqueness gains (Johnson & Lys, 1990) In particular, when firms receive unfavorable audit reports, they might initiate

an auditor switch, searching for a more pliable auditor with the goal of “opinion shopping” (DeFond & Subramanyam, 1998; Johnson & Lys, 1990; Watkins et al., 2004) Auditor switching may take different forms, including switching to a smaller auditor and switching to a larger auditor Nonethe-less in the extant literature there is a general lack of research differentiating the two types of auditor switching

Corporate Governance and the Auditing Function

Corporate governance evolved with the separation of

own-ership and management underlying the modern corporation

system (Ang, Cole, & Lin, 2000) There are various types of

principal-agent relations, e.g., between owners and

manage-ment, between creditors and owners/managemanage-ment, and

between controlling shareholders and minority

sharehold-ers A primary objective of corporate governance is to

monitor the behavior of the various interested parties and to

reduce the agency costs underlying the various

principal-agent relations (Karpoff, Malatesta, & Walkling, 1996) Thus,

corporate governance can be defined as “a set of

mecha-nisms, both industrial and market-based, that induce the

self-interested parties of a company to make decisions that

maximize the value of the company to its owners” (Denis &

McConnell, 2003)

An audit provides external monitoring over a firm’s

finan-cial reporting by independent professionals (auditors) and

therefore serves a fundamental role in reinforcing

informa-tion credibility However, the true effectiveness of external

auditing is subject to the actuality and the development of the

corporate governance environment (Holm & Laursen, 2007)

A sound corporate governance mechanism should ensure

that firms appoint qualified auditors and that the auditors

exercise independent and effective monitoring over the

financial reporting process and attest to the financial

state-ments’ conformity with the Generally Accepted Accounting

Principles (GAAP) Thus, corporate governance plays a role

in enhancing the effectiveness of the audit function

Francis and Wilson (1988) examined the relationship

between a firm’s agency costs and its demand for audit

quality in the US market They find that, concentrated

own-ership, a proxy for close alignment of interest between a

controlling shareholder and the listed firm, can substitute

for good audit quality and therefore is associated with

choice of a low-quality auditor However, in China, as

own-ership is highly concentrated and legal protection of

minor-ity shareholders is insufficient, ownership concentration

may lead to severe entrenchment problems and therefore

represent poor corporate governance

The association between corporate governance and

exter-nal auditing is an important issue worthy of serious study In

particular, as the agency relationship, the business

adminis-trative system, and corporate governance practices in

emerg-ing markets such as China differ substantially from those in

the developed economies (Tam, 2000), we are interested in

finding out whether empirical results are similar in the

Chinese market as in the United States This study will not

only enrich the extant auditing literature, but will also further

promote the development of corporate governance and

inde-pendent auditing practices in the emerging economies

DEVELOPMENT OF THE HYPOTHESES

Auditing plays two major functions: internally, auditing

can be used to monitor management behavior and reduce

agency costs (Jensen & Meckling, 1976); and externally,

auditing can be used by potential creditors and investors to

evaluate the cost of capital of the company, as described by

the results of a survey of members of the American Institute

of Certified Public Accountants (AICPA) (Carpenter & Strawser, 1971): “Almost universally, the reason expressed [for the change in auditors] was that the underwriters informed the client that a ‘national known auditor’ was nec-essary to sell their offerings at the highest possible price.”

In China, on average two-thirds of the shares are held by block shareholders (Lin, Liu, & Zhang, 2007) The block shareholder has the power to appoint a management team, and through the management team has access to inside information However, because the controlling shareholders also have the power to select the auditor,1 the minority shareholders do not expect to rely on auditing to effectively monitor management Hence, the internal function of audit-ing to reduce owner-manager agency costs does not exist in China and for this function there is insufficient demand for audit quality

As there is no significant bond market in China, listed firms are primarily financed by equity rather than debt After the establishment of Shanghai Stock Exchange and the Shenzhen Stock Exchange, Chinese listed firms achieved cumulative financing of RMB 1.16 trillion (RMB 6.80 = U$1) between 1992 and 2004 In 2000 the total market capitaliza-tion hit RMB 1.61 trillion, however the bear market thereaf-ter resulted in the slumping of the market value by RMB 44 trillion (CSRC, 2005) During the weak market period of

2001 to 2004 listed firms were not enthusiastic about offering new equity securities to the public, therefore the external use

of audits was also substantially diminished

In contrast, controlling shareholders are motivated to hire low-quality auditors to seek potential opaqueness gains at all times, especially if they are operating in a weak corporate governance environment where there are only loose monitor-ing and bindmonitor-ing contracts (Felo, Krishnamurthy, & Solieri, 2003) To address the research question empirically, we set out to test the association between the firms’ internal corpo-rate governance mechanism and their auditor switching deci-sions, i.e., whether firms with a weak internal corporate governance mechanism are more likely to switch to a lower-quality auditor

If such an association does not exist, internal corporate governance mechanism may not impact the type of auditor switches Alternatively, if auditors of different size offer monitoring services with varied levels of quality, the firms’ internal corporate governance mechanism may impact the switch type, in respect of the varied monitoring functions on the firms’ opaqueness gains The association may be evi-dence of the cost-benefit based demand for audit monitor-ing: when the opaqueness gains outweigh the benefits of lowering capital raising costs, lower-quality auditors would

be preferred by Chinese firms, and especially by firms with weak internal corporate governance mechanisms, as these firms have more opaqueness gains to protect (DeFond, Wong, & Li, 2000; Lin et al., 2007) By switching to a smaller auditor, the controlling shareholder (the agent) may com-fortably exploit the wealth of the minority shareholders without being watched closely by the auditor In contrast, switching to a larger auditor leads to more rigid audit moni-toring and hence tunneling behaviors will be confined

We used three proxies to measure a firm’s internal cor-porate governance mechanism: ownership concentration (shareholding of the largest owner); the effectiveness of

supervisory board (size of the supervisory board); and

shared CEO-Chair (whether the board of directors’

chair-man and CEO positions are held by the same person)

A high ownership concentration is a distinct feature of

listed firms in China A Chinese listed firm usually has a

large controlling shareholder,2who often is the government

or the parent state-owned enterprise.3Nonetheless, the

own-ership structure affects corporate governance and corporate

value in many different ways Johnson et al (2000) argue that

more narrowly held firms may face greater agency costs

because the controlling shareholders will have a dominant

influence on corporate affairs and they can easily bypass

monitoring by other shareholders La Porta,

Lopez-de-Silanes, and Shleifer (1999) and La Porta et al (2002) show

that in the emerging transitional economies, the controlling

shareholders may expropriate the minority shareholders

through aggressive “tunneling” behaviors They further

argue that “the central agency problem in large corporations

around the world is that of restricting expropriation of

minority shareholders by controlling shareholders” (La

Porta et al., 1999) This is particularly true for Chinese listed

firms in which the controlling shareholders usually hold a

very high percentage of the equity shares

In China, the controlling shareholders have frequently

intervened in the operations of the listed firms to benefit the

parent companies, e.g., using the listed firms as guarantors

for loan applications for the parent and related companies

and therefore exposing the listed firms to extra financial and

operating risks In fact, the controlling shareholders of many

listed firms are keen to raise funds only on the stock market

They frequently engage in benefit transfers through the

mis-appropriation of funds and related-party transactions to

expropriate the interests of the minority shareholders,

which, if detected, may invite external intervention by

minority shareholders and other stakeholders (Lin et al.,

2007) The desire to maximize self-interest through

“tunnel-ing” behaviors leads the listed firms to avoid being

moni-tored by a high-quality auditor The more concentrated the

ownership structure (i.e., with a larger controlling

share-holder), the weaker the internal corporate governance

mechanism Therefore, firms with larger controlling

share-holders are expected more likely to switch to pliable

auditors to realize opaqueness gains through tunneling

behaviors or other types of misconduct, as stated below:

Hypothesis 1 (H1): All other things being equal, a Chinese firm

with a higher percentage of total shares held by its controlling

shareholder will more likely switch to a smaller auditor.

Pursuant to the Chinese Company Law, all Chinese firms

adopted a German-style dual-board governance system,

thus each listed firm has both a board of directors and a

supervisory board The supervisory board is composed of

the shareholders’ representatives (including Chinese

Com-munist Party officials) and an appropriate proportion of

employee representatives, who are nominated by the firm’s

employee union The Company Law specifically defines the

supervisory board as a monitoring mechanism to carry out a

series of responsibilities, including: (1) monitoring the

per-formance of the directors and senior managers to ensure

compliance with the laws, regulations, and the articles of

incorporation; (2) reviewing the financial affairs of the firm;

(3) requesting the directors and senior managers to alter and/or rectify their personal activities if deemed in conflict with the firm’s objectives; (4) proposing specific shareholder meetings whenever deemed necessary; (5) fulfilling any other duties that are stipulated in the articles of incorpora-tion of the firm; and (6) submitting a supervisory board report to the shareholders’ annual general meeting The Standard Code of Corporate Governance for Listed Compa-nies in China issued by the CSRC and the State Economic and Trade Commission in 2002 further requires that super-visory board members should have some professional knowledge or work experience in the areas of law and accounting (CSRC, 2002)

Based on the requirements of the Company Law, the supervisory board shall independently and effectively carry out supervision over the activities of the directors and the management as well as examine the financial affairs of the firm Such a German-style two-tiered board system with the co-existence of a board of directors and a supervisory board has become the backbone of corporate governance in most Chinese listed firms since the mid-1990s Using an event study, Dahya, Karbhari, Xiao, and Yang (2003) report that investors consider the supervisory board to be an important device of corporate governance in China Chen (2005) finds that there is a positive association between the size of the supervisory board and the level of corporate governance, suggesting that a larger supervisory board should be more effective in carrying out its legitimate monitoring responsi-bilities We use the number of supervisory board members

as a proxy for the monitoring effectiveness of the supervi-sory board and have the second hypothesis:

Hypothesis 2 (H2): All other things being equal, a Chinese firm with fewer supervisory board members will more likely switch

to a smaller auditor.

Within a sound corporate governance structure, the board

of directors must ensure that the management acts in the best interests of the shareholders The board of directors is responsible for execution of the resolutions passed by the shareholders’ meetings and for appointing, removing, and remunerating senior managers Traditionally, the sharing of the position of board of directors’ chairman and CEO has been common in the United States However, in most Euro-pean, British, and Canadian businesses, in an effort to ensure better corporate governance, these two positions are often split Combining the two positions does have its advantages, allowing the CEO multiple perspectives on the firm as a result of his/her multiple roles and empowering him/her to act with determination Nonetheless, this prac-tice results in less transparency of the CEO’s activities, and

as such his/her actions can go unmonitored, which paves the way for scandals and corruptions To the contrary, sepa-ration of the two positions allows the board of directors’ chairman, on behalf of the stockholders, to be more impar-tial in overseeing the work of the CEO and the overall per-formance of management (La Porta et al., 2002; Petra, 2006) Investors, researchers, and government officials have gradually accepted the view that the best practices of corpo-rate governance require the separation of the roles of board

of directors’ chairman and CEO Such a corporate gover-nance device has received a boost since 2003 In practice,

market regulators and professional bodies in many

devel-oped countries require that the two important positions be

separated (Jiraporn, Young, & Davidson, 2005) In 2002 the

CSRC also adopted this requirement in its Standard Code of

Corporate Governance for Listed Companies in China

Con-sistent with the association between internal corporate

gov-ernance and the auditing function, we have the third

hypothesis stated as the following:

Hypothesis 3 (H3): All other things being equal, a Chinese firm

with the board of directors’ chairman and CEO positions held

by the same person is more likely to switch to a smaller auditor.

RESEARCH METHODOLOGY

Model Specification

We intend to examine the determinants of audit switching

from the perspective of the internal corporate governance

mechanism of Chinese listed firms Our sample includes

firms that switched auditors from 2001 to 2004 We classified

all firms that switched auditors only once during the 4-year

test period into two types – those switching to a larger

auditor (upward switching, or US firms) and those

switch-ing to a smaller auditor (downward switchswitch-ing, or DS firms)

according to the ranking order of auditors in China, which

was compiled by the CICPA in terms of the CPA firms’

annual audit revenue (see Appendix) As elaborated earlier,

the size of the auditing firm is regarded as an effective

surrogate for the independence and monitoring strength of

the auditors (Copley & Douthett, 2002; DeAngelo, 1981)

Thus we construct a model to test whether the firm’s internal

corporate governance mechanism (proxied by ownership

concentration, supervisory board, and shared CEO-Chair)

is associated with the different types of auditor

switch-ing (namely, upward switchswitch-ing or downward switchswitch-ing)

Downward switching (DS) can be expressed as a function of

the three corporate governance variables in which we are

interested and the related control variables: DS = f

(owner-ship concentration, supervisory board, shared CEO-Chair, control variables, error terms)

As we classify all auditor switches into two types (upward switches or downward switches), there are only two values possible (0 or 1) for the dependent variable Hence, we run logit regression, which is used to predict the probability of

an occurrence of an event by fitting the data to a logistic curve It makes use of several predictor variables that may be either numerical or categorical The logistic function is useful because it can take as an input any value from nega-tive infinity to posinega-tive infinity, whereas the output is con-fined to values of either 0 or 1 The dependent variable represents the exposure to some set of risk factors, whereas f(dependent variable) = 1/(1 + exp (-dependent variable)) represents the probability of a particular outcome, given that set of risk factors The following logit model is used to test Hypotheses 1 to 3

+

β

14

+β ∗ β +β β∗ R

+β19 04∗ +ε (1) Please see Table 1 for a description of the variables in the model

Three groups of variables, as discussed below, help explain why certain factors trigger a certain type of auditor switching The first group consists of corporate governance variables, namely ownership concentration (LSH), super-visory board (SB), and shared CEO-Chair positions (CEOCHR) Although the controlling shareholders may seek

to influence auditor selection so as to facilitate their tunnel-ing behaviors, they are subject to the constraints of the cor-porate governance structure in place The second group consists of company-specific variables that have been tested

or are considered helpful to explain auditor switching in

TABLE 1 Description of Variables

DS = 1 if the firm switches to a smaller auditor; 0 otherwise

LSH = largest owner’s shareholding as a percentage of total shares

SB = number of members of the SB

GOV = 1 if the largest shareholder is a government agency; 0 otherwise

OPI = 1 if the firm receives an unclean auditor’s opinion for the previous year; 0 otherwise

LEV = long-term liabilities divided by total assets at the end of the previous year

MB = market to book ratio at the end of the previous year, calculated as the market value of stocks divided by

the book value

LOSS = 1 if the firm experiences a loss for the previous year; 0 otherwise

Yr02 = 1 if the auditor switching occurs in Year 2002; 0 otherwise

Yr03 = 1 if the auditor switching occurs in Year 2003; 0 otherwise

Yr04 = 1 if the auditor switching occurs in Year 2004; 0 otherwise

prior studies, including government control (GOV),

audi-tor’s opinion (OPI), size (LNASSET), financial leverage

(LEV), market-to-book ratio (MB), profitability (LOSS), and

new issues (NWISS) As there were new requirements in the

Standard Code of Corporate Governance during the test

period, we incorporate the third group of interactions

vari-ables to capture the yearly impact of the gradual adoption of

these new requirements, namely Yr02, Yr03, and Yr04 With

the interactions terms, we intend to test whether the impact

of corporate governance mechanism on auditor switching

varies with the progress of corporate governance practice

In the regression, the dependent variable is defined by the

types of auditor switching, thus it is coded 1 if a firm

switched to an auditor that was smaller than its predecessor

For the independent variables, we expect b1(for ownership

concentration) and b3 (for shared CEO-Chair) to have

posi-tive signs as firms with a high ownership concentration and

a shared board of directors’ chairman and CEO are more

likely to switch to a smaller auditor But b2(for supervisory

board) is expected to be negative as a large or strong

super-visory board may inhibit the firm from switching to a

smaller auditor

Government controlled and non-government controlled

(privately owned or privately controlled) firms may have

differing corporate governance structures and may also have

different considerations when making auditor switching

decisions In general, government agencies have a stronger

influence over government-controlled firms and therefore

can more easily access the firm’s financial information

for their decision making (Chan, Lin, & Mo, 2006)

There-fore, government-controlled firms may have less demand for

high-quality independent audits and may have a greater

propensity, compared with non-government controlled

firms, to switch to smaller auditors Hence we add the

vari-able GOV to capture the effect of government control over

the firm’s auditor switching decisions This variable is coded

as 1 if the largest owner of the firm is a government agency

and 0 otherwise It is expected to be positively associated

with a downward switch of auditors

We also control for the effects the auditor’s opinion, firm

size, financial leverage (risk), growth potential, profitability,

and whether there is new issue of equity after the auditor

switching Prior auditor switching studies focus mainly on

the auditing markets in the Western countries Nonetheless,

the basic theories and findings in prior research on auditor

switching should be applicable to this study as well because

the Chinese auditing profession in recent years has gradually

adopted international accounting and auditing standards

One very common reason cited for auditor switching is

the qualifications of the auditor’s opinions Prior research

has found that firms receiving unfavorable audit reports are

more likely to switch auditors (DeFond & Subramanyam,

1998) We expect that the auditor’s opinion (receiving an

unfavorable auditor opinion in the prior year = 1) to be

posi-tively related to a downward switch of auditors Large firms

may be less likely to switch to a smaller auditor, as financial

analysts and the financial press will more closely scrutinize

their auditor switches Following Friedlan (1994), we use the

log of the total assets to control for the size effect of the firms

and we expect it to be negative in the regression model

Reed, Trombley, and Dhaliwal (2000) find that firms

select-ing Big 6 auditors tend to be highly leveraged, whereas Titman and Trueman (1986) predict otherwise As there are opposing arguments and findings regarding the association between a firm’s leverage and its auditor switching, we do not predict the sign of the coefficient for financial leverage

We also include the market-to-book ratio to control for the propensity of growing firms to switch to less conservative auditors (DeFond & Subramanyam 1998) Moreover, Sainty, Taylor, and Williams (2002) document that profitability may affect the selection of auditors In our model, we expect a net loss in the prior year (LOSS) to be positively associated with

a downward switch of auditors

Firms may change auditors (especially from a low-quality auditor to a high-quality auditor) to increase the marketabil-ity of new securities (Carpenter & Strawser, 1971) Pae and Yoo (2001) document a negative relationship between audit quality and the cost of raising capital, i.e., a firm can reduce its costs of raising capital by hiring a quality auditor We therefore include the variable of new issues (NWISS) in our regression model, which equals 1 if the listed firm issues new equity to the public in the two years after its auditor switching and 0 otherwise The new issue is used to proxy for the firm’s intentions to issue new equity at the time of the auditor switching Although the firms in our sample gener-ally did not intend to issue new equity to the public during our test period, for accuracy and completeness we still include the variable Since 2002, new corporate governance requirements have been announced and implemented, therefore we add the year dummy variables for 2002 to 2004

to capture the potential impact of these new corporate gov-ernance devices on the firms’ auditor switching decisions during the test period

Sampling

Our sample covers A-share firms that switched auditors from the beginning of 2001 to the end of 2004.4There are two main reasons to limit the sample firms that made auditor switches to this time period The first one is the availability of the ranking of Chinese auditors, which has been compiled

by the CICPA since 2002 Thus it provides the possibility of identifying or classifying the different types of auditor switches (i.e., an upward switch or a downward switch) The second reason is that during this time period, firms had little intention to offer equity to the public, therefore the opaque-ness gains from weak corporate governance significantly outweighed the benefits from lowering the costs of raising capital Hence, the bear market period from 2001 to 2004 is appropriate to test the association between the firms’ inter-nal corporate governance mechanism and their auditor switching decisions Data were collected from the China Stock Market and Accounting Research (CSMAR) Database, the TEJ database (carrying financial information and stock

market data compiled by the Taiwan Economic Journal), and

authoritative national newspapers or magazines designated

by the CSRC to publish financial reports of listed firms, such

as China Securities Daily, Shenzhen Securities Times, and

Shanghai Securities News The collected data on the sample

firms were cross-checked and verified by different data sources to ensure their reliability A description of the data is provided in Table 2

TABLE

Table

At the end of 2004, there were 1,387 A-share firms listed

on the two stock exchanges in China, among which 316

firms (22.7 per cent) switched auditors during the four-year

period from 2001 to 2004 This implies that generally a firm

is not willing to switch auditors because of the potential high

costs associated with auditor switching, such as the costs of

searching for and renegotiating with a new auditor and the

potentially unfavorable market responses to an auditor

switch (Reed et al., 2000; Watkins et al., 2004) Panel A of

Table 2 presents the sample size for this study Financial,

transportation, and utility firms are excluded because the

nature of their operations is very different from that of other

types of firms We also delete firms that switched auditors

more than once during the four-year period The frequent

switching of auditors may indicate some serious underlying

reasons that are beyond the scope of this study

Further-more, firms that switched twice or more may have switched

to a larger auditor at one time and to a smaller auditor the

other time, making it difficult to categorize the type of

switching The final sample consists of 233 firms

EMPIRICAL RESULTS

Panel B of Table 2 presents the basic statistics on the tested

variables Among the 233 sample firms, 134 firms switched

to larger auditors and 99 firms switched to smaller auditors,

based on the rankings prepared by the CICPA On average,

the largest controlling owners held 48.94 per cent of the total

shares of the sample firms, indicating a high ownership

concentration in the Chinese listed firms About 79 per cent

of our sample firms were directly owned by the government

or governmental agencies, reflecting that most Chinese

listed firms were originally carved out from state-owned

enterprises and that various government agencies remain

the largest owners of the listed firms In the sample, 16 per

cent (37/233) of the firms received unfavorable auditor

opin-ions before their auditor switches This high percentage may

support the assertion that there is an association between

receiving unfavorable auditor reports and switching

audi-tors The average size of the supervisory board was about

4.21, with a minimum of 1 member and a maximum of 12

members In 6 per cent of the sample firms (15/233), the

CEO also held the position of board of directors’ chairman

Very few firms intended to issue equity during the bear

market, as indicated in the table that only 3 per cent (8/233)

of the sample firms issued equity in the two-year period

after their auditor switching

Panel C of Table 2 presents the correlation coefficient

matrix for the variables used in the regression model A

downward switch is significantly and positively correlated

with ownership concentration, shared CEO-Chair, and

market-to-book ratio; and it is significantly and negatively

related to firm size The correlation coefficients among the

independent variables are moderate, with no value

exceed-ing 50 (the largest is 44)

Table 3 provides the empirical results from the regression

which tests whether a firm with a weak internal corporate

governance mechanism is inclined to switch to a smaller

auditor Since there are only two values possible for the

dependent variable (1 for a downward switch and 0 for an

upward switch), we use a logit regression With a Chi-square

of 40.03, p< 01, and a pseudo R-square of 21, the regression model is satisfactory in differentiating firms switching to smaller auditors from firms switching to larger auditors at

an acceptable level of significance The probability of a down-ward switch can be expressed as f(DS) = 1/(1 + exp(-DS)), i.e., the larger the value of the DS, the higher the probability

of a downward switch

The coefficient for ownership concentration is positively significant at the 5 per cent level (Coeff = 02, Wald = 4.78, p< 05), which supports H1 If under the current situa-tion the probability of a downward switch f(DS) is exactly 50 (equal probability of an upward switch or a downward switch), then an increase of 10 per cent of the total sharehold-ing will cause f(DS) to increase to 55 This suggests that firms with a higher degree of ownership concentration (i.e., higher percentage of equity shares held by the largest shareholder) are more likely to switch to a smaller auditor Although the test results seem to be the same as Francis and Wilson 1988 findings, our interpretation is completely different The rela-tively high internal ownership in upward switches should better align the interests between managers and sharehold-ers, but may not be significant enough to expropriate the minority shareholders, and hence proxies for effective corpo-rate governance (Francis & Wilson, 1988; Sabherwal & Smith, 2008) Effective corporate governance can substitute for a quality auditor, and therefore is associated with switching to

a lower-quality auditor in the United States In contrast, in China, due to the awfully high level of concentration, the divergence between cash flow rights and control rights, and the ineffective legal protection, entrenchment problems become detrimental, hence ownership concentration proxies for poor corporate governance Poor corporate governance allows the controlling shareholders more space to hire a lower-quality auditor to realize their opaqueness gains, therefore leading to our results – a high ownership concen-tration, a proxy for poor corporate governance, is associated with switching to a lower-quality auditor

The coefficient for the size of the supervisory board is insignificant (Coeff = 03, Wald = 05, p> 10), therefore H2

is not supported, indicating that the effectiveness of super-visory board monitoring may not be related to whether a firm switches to a larger or a smaller auditor Another pos-sible interpretation is that in practice the monitoring role of the supervisory board is dubious, thus the supervisory board presently does not have a significant impact on a firm’s auditor switching decision In fact, as supervisory board members are mainly from inside the firm, whether the supervisory board can effectively play a monitoring role is controversial Some researchers contend that the supervi-sory board is mainly decorative in China (Dahya et al., 2003) and our findings seem to support this viewpoint

Consistent with H3, the coefficient for the shared CEO-Chair is positive and significant at the 5 per cent level (Coeff = 1.28, Wald = 3.85, p< 05) If the current status is the separation of the CEO and the board of directors’ chairman and the probability of a downward switch f(DS) is exactly 50, then a change to the combination of the two key roles will lead f(DS) to increase to 78 Thus, a firm is more likely

to switch to a smaller auditor if its CEO also holds the position of board of directors’ chairman, as expected

βε19

b0

b1

b2

b3

b4

b5

b6

b7

b8

b9

b10

b11

b12

b13

b14

b15

b16

b17

b18

b19