Corporate governance and audit fees: Evidence from companies listed on the Shanghai Stock Exchange

Wu, Xingze Article Corporate governance and audit fees: Evidence from companies listed on the Shanghai Stock Exchange China Journal of Accounting Research Provided in Cooperation with: Sun Yat-sen University Suggested Citation: Wu, Xingze (2012) : Corporate governance and audit fees: Evidence from companies listed on the Shanghai Stock Exchange, China Journal of Accounting Research, ISSN 1755-3091, Elsevier, Amsterdam, Vol. 5, Iss. 4, pp. 321-342, https://doi.org/10.1016/j.cjar.2012.10.001 This Version is available at: https://hdl.handle.net/10419/187577 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by-nc-nd/3.0/ 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 ARTICLE INFO Article history: Received 28 February 2012 Accepted 10 October 2012 Available online 12 November 2012 JEL classification: G38 M42 Keywords: Corporate governance Audit fees SSE Corporate Governance Sector Corporate growth ABSTRACT This study uses data from companies listed on the Shanghai Stock Exchange to investigate the relationship between corporate governance and audit fees. Full sample results reveal a significant negative relationship between corporate governance and audit fees, and subsample results further show that corporate governance’s influence on audit fees is affected by corporate growth. The negative relationship between corporate governance and audit fees is economically and statistically significant in sample companies that grew moderately during the sample period, and mixed or insignificant in companies that experienced overly fast or negative growth. Ó2012 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. 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 governance 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 1755-3091/$ - see front matter Ó2012 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 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 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 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 Standard 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 stakeholders 1 by inviting a more rigorous external 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 uncertain, 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 problematic, 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 offering 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 standardization. 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. 322 X. Wu / China Journal of Accounting Research 5 (2012) 321–342 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 confirmation 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 experience a lower degree of risk and their internal corporate governance is relatively stable. Cui et al. (2007) examine 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 reasonable 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 relationship 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 corporate 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 corporate 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 apportionment 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 significantly 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 small2 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. X. Wu / China Journal of Accounting Research 5 (2012) 321–342 323 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-Contents 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 company 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 surrounding 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 auditee scale and complexity and whether a firm has been audited by one of the “Big N”firms to have a significantly positive influence on audit fees (Simunic, 1980 3 ;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 receivable 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 evidence 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 promulgation of the Sarbanes–Oxley Act, Section 404 of which demands that listed companies disclose internal control information, will increase opportunities to investigate the association between corporate governance and 3 Simunic (1980) does not investigate the influence of the Big Non audit fees. 324 X. Wu / China Journal of Accounting Research 5 (2012) 321–342 audit fees directly, although our review of the overseas literature indicates no such increase. Chinese researchers, in contrast, have paid increased attention to the issue in recent years. Drawing on the ownership perspective, 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 Ashare 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 variables. Cai (2007) investigates the influence of corporate governance structure on audit fees from the perspective 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 empirically 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 characteristics 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 risk 4 to companies 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. X. Wu / China Journal of Accounting Research 5 (2012) 321–342 325 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 alternative 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 Qvalue falls outside the range of the mean minus two times the standard deviation 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 Sector). 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 ¼b0þb1Gov þb2TobinQ þbBig4þb4LnAssets þb5HStock þb6Loss þb7Recint þb8Invint þb9Segment þb10Subs rt þe: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 concentration, 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 effectiveness 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. 326 X. Wu / China Journal of Accounting Research 5 (2012) 321–342 developed by the Nankai University Research Center of Corporate Governance as a proxy for such governance. 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 departments 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 consolidated 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 industries 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. Positive 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 separately. 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 1 Summary of sample selection criteria. Selection procedure 2007 2008 A-share companies listed on SSE Companies included in SSE Corporate Governance Sector A-share companies listed on SSE Companies included in SSE Corporate Governance Sector 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 November 2, 2007 5 – 10 2 Outliers 6 – 12 2 Final sample observations 602 149 678 184 X. Wu / China Journal of Accounting Research 5 (2012) 321–342 327 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 characterized 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 relationship between audit fees and corporate growth may feature a U-shape rather than a linear shape. In the previous literature, Tobin’s Qand 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 determination, as proved both theoretically and empirically. Francis (1984), Firth (1985), DeFond et al. (2000), Ireland and Lennox (2002) and Chen et al. (2007) find evidence of a Big Npremium using stock market data from Australia, New Zealand, Britain, Hong Kong and China, respectively. Using data on 15 countries and districts, 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 R 2 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. 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 involved a Subs_rt Square root of number of consolidated subsidiaries b 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. 6 The Big 4 in this study are Ernst & Young Hua Ming, Deloitte Huayong Certified Public Accountants Co., Ltd., PricewaterhouseCoopers Zhongtian and KPMG Huazhen. 328 X. Wu / China Journal of Accounting Research 5 (2012) 321–342 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 Continuous variables 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 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 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. X. Wu / China Journal of Accounting Research 5 (2012) 321–342 335 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 Continuous variables 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 2.02e14 .7681355 0.199915 0.191007 .1442456 3.36e14 0.940148 0.184833 0.201679 0.147006 7.73e14 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 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 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. 336 X. Wu / China Journal of Accounting Research 5 (2012) 321–342 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 observations 17 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 ¼b0þb1Gov þb2TobinQ þb3Big4þb4nAssets þb5H Stock þb6Loss þb7Recint þb8Invint þb9Segment þb10Subs rt þb11Lambda þe:ð4Þ 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 (0.37) – 4.86583 (9.86) *** – 7.902495 (12.34) *** – Gov ? 0.1673906 (0.41) 1.92 0.0980085 (1.98) ** 1.19 0.0383723 (0.61) 1.21 LnAssets + 0.4615434 (1.83) * 7.48 0.3709851 (16.20) *** 1.82 0.2307721 (7.55) *** 1.65 H_stock + 1.45432 (3.78) *** 2.49 1.36174 (12.76) *** 1.45 0.9150677 (4.81) *** 1.19 Loss + 0.572797 (1.27) 2.92 0.150768 (2.80) *** 1.13 0.033121 (0.56) 1.26 Recint + 1.33218 (0.45) 2.86 0.2718718 (1.19) 1.07 0.1819705 (0.63) 1.06 Invint + 1.362542 (1.36) 1.66 0.2010797 (1.48) 1.06 0.2201791 (1.32) 1.04 Big4 + dropped – 0.6569943 (7.97) *** 1.60 0.539355 (4.64) *** 1.18 Segment + 0.1629168 (1.28) 1.86 0.0144943 (1.00) 1.12 0.0198448 (1.29) 1.13 Subs_rt + 0.0058437 (0.05) 1.83 0.0936485 (6.20) *** 1.19 0.1162623 (5.98) *** 1.45 NTotal 602 20 Average 2.88 372 Average 1.29 210 Average 1.24 F9.80 *** 172.33 *** 39.91 *** R 2 0.8770 0.8108 0.6424 Adj R 2 0.7875 0.8061 0.6263 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 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. * Two-tailed significance at the 0.10 level. ** Two-tailed significance at the 0.05 level. *** Two-tailed significance at the 0.01 level. 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. X. Wu / China Journal of Accounting Research 5 (2012) 321–342 337 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 governance and audit fees using data disclosed in the annual financial reports of companies listed on the Shanghai 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 (4.85) *** – 5.789669 (11.11) *** – 6.978095 (11.22) *** – Gov ?0.1201422 (1.41) 1.29 0.1018529 (1.98) ** 1.21 0.036208 (0.59) 1.26 LnAssets + 0.3973288 (10.21) *** 2.30 0.3294044 (13.75) *** 1.60 0.2736142 (9.22) *** 1.61 H_stock + 1.321769 (10.21) *** 1.85 1.2595 (7.28) *** 1.22 Dropped – Loss + 0.1286156 (1.22) 1.14 0.0343346 (0.64) 1.18 0.1343597 (2.33) ** 1.33 Recint + 2.2625 (3.51) *** 1.22 0.063834 (0.23) 1.11 0.05694 (0.20) 1.01 Invint +0.3300925 (1.50) 1.07 0.2394448 (2.13) ** 1.05 0.0421028 (0.36) 1.01 Big4 + 0.3631917 (2.67) *** 2.04 0.650547 (6.62) *** 1.35 0.8505657 (5.71) *** 1.15 Segment + 0.0408373 (1.86) * 1.13 0.0125083 (0.92) 1.13 0.0318546 (2.13) ** 1.12 Subs_rt + 0.0808408 (3.14) *** 1.24 0.0968029 (6.31) *** 1.33 0.0925125 (4.90) *** 1.30 NTotal 678 157 Average 1.48 274 Average 1.24 247 Average 1.23 F91.00 *** 87.99 *** 38.95 *** R 2 0.8478 0.7500 0.5670 Adj R 2 0.8385 0.7415 0.5524 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 Big 4 accounting firm, and 0 otherwise. Segment = number of industries a company involved in. Subs_rt = square root of number of consolidated subsidiaries. * Two-tailed significance at the 0.10 level. ** Two-tailed significance at the 0.05 level. *** Two-tailed significance at the 0.01 level. 338 X. Wu / China Journal of Accounting Research 5 (2012) 321–342 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 Nongovernance 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 relationTable 11 Definitions of variables in Model 3. Name Definition Explained variable Gov Dummy = 1 if included in SSE Corporate Governance Sector, otherwise 0 Explanatory variables Auditcomm Dummy = 1 if audit committee is set up, otherwise 0 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 (1.04) 0.270078 (1.40) 0.5186868 (1.25) 0.025993 (0.03) 0.506515 (1.95) ** 0.013953 (0.75) 0.0543591 (1.48) 0.012199 (0.25) Variable DirScale IndDir Fnctl TobinQ Big4 LnAssets H_Stock Loss Result 0.0181251 (0.49) 0.1075773 (1.41) 0.0592087 (0.44) 0.3857623 (2.55) ** 0.1731616 (0.73) 0.2789833 (3.95) *** 0.3502073 (0.84) 1.072429 (4.93) *** Variable Recint Invint Segment Subs_rt Intercept N LR chi2 Pseudo R 2 Result 0.536 (0.96) 0.085569 (0.31) 0.114073 (2.76) *** 0.040388 (1.05) 7.272783 (4.85) *** 725 109.26 *** 0.1696 Figures in parentheses are Z-values.  Two-tailed significance at the 0.10 level. ** Two-tailed significance at the 0.05 level. *** Two-tailed significance at the 0.01 level. 18 The percentages are 5.44% in 2008 and 7.44% in 2007. X. Wu / China Journal of Accounting Research 5 (2012) 321–342 339 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 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 (12.97) *** 5.684087 (13.88) *** – 5.792809 (14.98) *** 6.164501 (16.05) *** – Gov 0.0772639 (1.97) ** 0.0903212 (1.72) * 2.19 0.0559151 (1.52) 0.0930575 (2.01) ** 2.00 TobinQ 0.0804095 (4.87) *** 0.073934 (4.58) *** 1.22 0.1232368 (3.21) *** 0.1145037 (2.98) *** 1.25 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 (1.81) * 0.2719119 (1.54) 1.06 0.3221629 (1.64) 0.2791031 (1.45) 1.09 Invint 0.1542988 (1.45) 0.1257606 (1.19) 1.04 0.1875702 (2.40) ** 0.1685315 (2.18) ** 1.02 Big4 0.6413022 (9.52) *** 0.6755913 (10.15) *** 1.51 0.6060236 (8.66) *** 0.6677469 (9.43) *** 1.58 Segment 0.0132502 (1.24) 0.0100982 (0.96) 1.10 0.0249864 (2.68) *** 0.022193 (2.40) ** 1.11 Subs_rt 0.0883677 (7.49) *** 0.0935087 (8.03) *** 1.22 0.0935901 (8.58) *** 0.0961795 (8.79) *** 1.27 Lambda 0.085819 (0.51) 2.44 0.0626612 (1.28) 2.09 N602 592 Average 1.50 678 662 Average 1.47 F202.19 *** 159.50 *** 223.48 *** 171.28 *** R 2 0.7738 0.7516 0.7701 0.7435 Adj R 2 0.7700 0.7468 0.7667 0.7392 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 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. 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. * Two-tailed significance at the 0.10 level. ** Two-tailed significance at the 0.05 level. *** Two-tailed significance at the 0.01 level. 340 X. Wu / China Journal of Accounting Research 5 (2012) 321–342 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 perform 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 comparability 19 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 expenditure on auditing. Hence, a company’s acceptance of signaling theory depends on the tradeoff between expenditure 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-selection 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. X. Wu / China Journal of Accounting Research 5 (2012) 321–342 341 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 CCGI NK . 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. 342 X. Wu / China Journal of Accounting Research 5 (2012) 321–342