DeepDyve requires Javascript to function. Please enable Javascript on your browser to continue.
Competitive pressure, economies of scale, and auditor industry specialisation premium Competitive pressure, economies of scale, and auditor industry specialisation premium
Chen, Shenglan; Ma, Hui
2014-04-03 00:00:00
China Journal of Accounting Studies, 2014 Vol. 2, No. 2, 96–117, http://dx.doi.org/10.1080/21697213.2014.926197 Competitive pressure, economies of scale, and auditor industry specialisation premium Shenglan Chen* and Hui Ma School of Economics & Management, Inner Mongolia University, Hohhot 010021, China This paper examines the premium due to auditor industry specialisation in China’s audit market. Using a sample of China’s listed firms from 2001 to 2011, we find that industry specialist auditors earn significant premiums at the province level. We also find that competitive pressure has a negative effect on the industry specialisa- tion premium. Furthermore, we document that specialist auditors pass on lower cost savings from economies of scale to clients when facing greater competitive pressure. The results remain robust after controlling for the possible endogeneity issues and using different samples for sensitivity tests. Our results have important implications for future research on the auditor industry specialisation premium. Keywords: audit fee premium; auditor industry specialisation; competitive pressure; economies of scale 1. Introduction Audit firms with industry specialisation are able to develop more industry-specific knowledge and expertise, thereby enabling them to identify financial statement fraud and provide higher quality services than others (Balsam, Krishnan, & Yang, 2003; Minutti-Meza, 2013; Simunic, 1980). If audit firms develop industry specialisation as an important strategy to enhance the quality of audit services or as differentiated prod- ucts to meet client demand, they should be able to charge a relative fee premium. China’s regulators believe that developing industry specialisation is an important strategy for the development of the whole audit industry. For example, in October 2009, the general office of the State Council forwarded the document ‘Several opinions on accelerating the development of China’s CPA industry’ issued by the Ministry of Finance, and emphasised the importance of developing industry specialisation in the audit market. In September 2011, ‘China’s CPA industry development plan (2011– 2015)’ was issued by the Chinese Institute of Certified Public Accountants (CICPA), which specifically suggests enhancing auditors’ industry specialisation and developing international competitive advantages. However, the research on auditor industry special- isation in China’s audit market is scarce. Only a few studies examine whether industry specialist auditors can earn premiums and the empirical evidence has been somewhat mixed. One main reason for the mixed empirical evidence is that the studies use national-level specialisation measures. Recent studies in the American audit market have shifted attention to local-level auditor industry specialisation due to the questions *Corresponding author. Email: chen_shenglan@126.com Paper accepted by Xi Wu. © 2014 Accounting Society of China China Journal of Accounting Studies 97 about reasonableness of national-level specialisation measures (Francis, Reichelt, & Wang, 2005; Fung, Gul, & Krishnan, 2012). First, local offices of the audit firms are the primary units of decision-making and are responsible for determining fee contracts. It is therefore important and meaningful for researchers to investigate the pricing deci- sion of local-level industry specialists. Second, local offices are likely to be better units for analyses of audit outcomes than national offices. It is possible that national-level industry specialists are not local-level industry leaders or some audit firms are local industry leaders alone without also being national-specific industry leaders. The special- isation premium may be underestimated if national clienteles are used to measure audi- tor industry specialisation. However, few studies try to measure auditor industry specialisation in China’s audit market at the local level. An important feature of China’s audit market is its fierce competition. Unlike devel- oped economies in which the Big 4 auditors dominate, concentration in China’s audit market is rather low (Chen, Sun, & Wu, 2010; Lu, Wang, & Wu, 2012). Such a buyer’s market is likely to afford clients more bargaining power and impose pressure on auditors fighting for their slice of the cake. Audit firms are more likely to offer a lower price compared with that of a competitor to maintain market share and competi- tive advantage. Thus, competitive pressure can be an indispensable factor that can force audit firms to pass on some industry specialisation premiums. However, there is no research on the effects of competitive pressure on auditor industry specialisation premiums in China. In addition, economies of scale may be another indispensable factor when the industry specialisation premium is investigated (Chen & Ma, 2013; Fung et al., 2012). Intuitively, the higher the industry specialisation premiums passed on to clients due to competitive pressure, the lower the scale economies that can be passed to clients. How- ever, there is no research that investigates whether industry specialist auditors pass on lower scale economies to clients when facing greater competitive pressure. This paper examines whether industry specialists can charge a fee premium and what factors determine auditor industry specialisation premiums in China’s audit mar- ket. More specifically, we first construct a province-industry measure to investigate whether there are industry specialisation premiums using a sample of China’s listed firms over the period 2001–2011. Next, considering the importance of competitive pres- sure, we examine the effect of competitive pressure on the industry specialisation pre- mium. Intuitively, when audit firms are facing greater pressure from competitors, they are more willing to lower audit fees to attract more clients, which means competitive pressure could mitigate industry specialisation premiums. Finally, we examine the inter- action effect between competitive pressure and economies of scale on the industry spe- cialisation premium in audit pricing. Our results indicate that industry specialist auditors (e.g., province-industry leaders) earn significant premiums (22.1%, on average) compared with non-specialist auditors. However, Wu and Zhang (2012) find that there are 17.7% specialisation premiums, using national clienteles to measure auditor industry specialisation. These findings col- lectively show that the specialisation premium at the province-industry level is more pronounced than that at the national-industry level. In addition, consistent with our pre- diction, we find that specialisation premiums are moderated by competitive pressure. Finally, we find the effects of competitive pressure and economies of scale on the industry specialisation premium to be highly substitutive. The negative effect of prov- ince-industry scale on the industry specialisation premium is less significant for auditors facing greater competitive pressure. 98 Chen and Ma Our study contributes to an understanding of audit pricing by industry specialist auditors in a number of ways. First, it adds to current knowledge about the pricing effects of auditor industry specialisation. Based on the unique characteristics of China’s audit market, we find that industry specialist auditors earn significant premiums. Sec- ond, we employ province-industry specialisation measures to investigate the effect of auditor industry specialisation on audit fees. The evidence suggests that industry spe- cialisation premiums are underestimated when national-level industry specialisation measures are used. Third, to the best of our knowledge, we are the first to document evidence that the province-industry specialisation premium is mitigated by competitive pressure. A possible explanation for such mitigation is that specialist auditors have to give up some fee premiums due to external competitive pressure. Such evidence is helpful to understand the low audit pricing in China’s audit market. Fourth, our study extends the line of research on audit firms’ scale economies. Chen and Ma (2013) find that the negative effect of province-industry scale on audit fees for clients is more pronounced for specialist auditors. Our evidence suggests that the extent to which cost savings from scale economies can be passed on to clients also depends on the competi- tive pressure from the closest competitor (auditor) in the audit industry. The remainder of this paper proceeds as follows. Section 2 reviews the relevant literature. Section 3 describes the institutional background. Section 4 develops the hypotheses. Section 5 presents the research design. Section 6 discusses the descriptive statistics and empirical results and Section 7 concludes. 2. Literature review Audit firms have incentives to supply specialised services to meet unique client needs in ways that are not replicated easily by competing audit firms. Industry specialist audi- tors possess more industry-specific knowledge and expertise in identifying disclosure issues and the practice experience of auditing clients, which enables them to recognise industry-specific audit issues more effectively and give a more precise audit judgment (Krishnan, 2003; Simunic, 1980). This suggests auditor industry specialisation can be used as an effective way to satisfy the clients and to differentiate an industry-specialist from other competitors (Balsam et al., 2003; Fung et al., 2012; Minutti-Meza, 2013). Meanwhile, the demand for high-quality audit service will increase due to friction in the capital market. Clients have a strong tendency to pay high audit fees for industry- specialist auditors in order to improve accounting information quality, mitigate informa- tion asymmetry, reduce capital costs and enhance investment efficiency (Bae, Choi, & Rho, 2012; Craswell, Francis, & Taylor, 1995; Dunn & Mayhew, 2004; Jaggi, Gul, & Lau, 2012; Krishnan, 2003; Li, Xie, & Zhou, 2010). Prior studies that examine the impact of audit industry specialisation on audit pric- ing at the national level yield mixed evidence. Recent studies have shifted attention to city-level auditor industry specialisation based on the rationale that the primary audit work and decision-making involving clients occurs in the local offices of auditors (Fung et al., 2012). First, large audit firms are decentralised and operate as a network of semi-autonomous local practice offices, which means local offices are responsible for determining fee contracts (Ferguson, Francis, & Stokes, 2003; Francis, Stokes, & Anderson, 1999; Francis et al., 2005). Second, the industry specialisation premiums may be underestimated when national clienteles are used to measure auditor industry specialisation (Francis et al., 2005). China Journal of Accounting Studies 99 An important feature of China’s audit market is its fierce competition. It is therefore meaningful for audit firms to develop industry specialisation. Industry specialisation is particularly valuable because it can serve as a differentiation strategy to service a rela- tively large group of clients possessing the same basic characteristics. A few studies examine whether auditor industry specialisation affects audit pricing using national cli- enteles to measure auditor industry specialisation (Han & Chen, 2008; Wu & Zhang, 2012). Nonetheless, the specialisation premiums may be underestimated when national- level data are used to measure auditor industry specialisation, considering the important role of local offices in audit pricing. It is widely recognised that important corporate decisions are fundamentally affected by competitive pressure. This is because the increased competitive pressure reduces price-cost margins and forces firms to adopt a new strategy (Hall, 1988; Harrison, 1994). Previous literature has examined the response of firms to competitive pressure, and finds that innovation, product introduction, and investment are increased if firms are facing greater competitive pressure (Fresard & Valta, 2013; Vives, 2008). The spatial competition model is a powerful tool in examining the effect of compe- tition on product pricing. There have been considerable advances in spatial economic theory since the seminal work of Hotelling (1929), showing how two identical, single- product firms compete on price and location in a bounded linear market. The value and importance of the spatial competition model has been emphasised with the development of an uncooperative game. Chan (1995) and Chan (1999) analyse the different price decisions in the audit market using spatial competition theory, which offer some propo- sitions for empirical research. Because market activities are performed at dispersed points in space, each supplier finds only a few competitors in its immediate neighbour- hood. Moreover, greatest competitive pressure on pricing comes from the competitor who is closest to the supplier. Numan and Willekens (2012a) argue that the relative market location (by industry market share) of a Big 4 auditor to its closest Big 4 com- petitor impacts competitive pressure and therefore audit pricing. They find that the size of the audit fee premium from industry specialisation is not only affected by industry specialisation per se, but also by the degree of competitive pressure from the closest competitor in the market segment. Traditionally, the fall in unit costs associated with the rise in production scale is explained by technological factors or internal scale relationships. The supplier’s tech- nology may exhibit scale economies and pass them on to buyers through a lower unit price. Scale economies may increase price competition among firms (Cachon & Harker, 2002). Thus, the firm has an incentive to adjust for their price of product to maintain competitive advantages in the market (Hall, 1988; Klette, 1999). Likewise, for an audit firm, scale economies can arise from substantial investment in general audit technology (e.g., audit software development or hardware acquisition), human capital development (e.g., staff training) and advertisement investment. Such investments are fixed costs and are likely to be shared among all of the clients. Once these investments are in place, additional clients can be serviced at a lower marginal cost than the cost of servicing the first few clients (Carson, Simnett, Soo, & Wright, 2012; Fung et al., 2012; Junius, 1997). Audit cost is an essential factor for audit firms in determining fee contracts. Consequently, if an audit firm wants to maintain or enlarge market share, some or all of these cost savings (scale economies) are passed on to clients, which should produce a reduction in audit fees (Fields, Fraser, & Wilkins, 2004; Gist, 1994). However, Fung et al. (2012) find that the negative effect of city- industry scale on audit fees is observed only for clients of specialist auditors. The most 100 Chen and Ma important reason is that cost savings from scale economies are greater in specialist auditors, and audit firms are willing to share these cost savings with clients to maintain market share. Chen and Ma (2013) examine the effect of scale economies on industry specialisation premium, and find similar evidence in China’s audit market. However, there is no research that investigates whether industry specialist auditors pass on lower scale economies to clients when facing greater competitive pressure. 3. Institutional background Until the late 1970s, China operated a planned economic system under which an enter- prise was either state-owned or collectively owned. Both types of enterprise were run directly by the government, with little room for market mechanisms. As such, there was no need for independent accounting and auditing until the early 1980s when the economic reform and open-door policies were adopted by the government. Since the 1990s, the audit profession in China has experienced rapid growth driven by an increasing demand for independent audits. China’s audit firms were established and initially owned by government or other sponsoring bodies. This situation has caused much concern regarding auditor independence in China. A fundamental institutional change in the China’s auditing market is the audit firm disaffiliation program around 1997–1999. The disaffiliation improves auditor independence through the separation of audit firms from their sponsoring government agencies (Simunic & Wu, 2012). The audit market in China is different from that of developed countries in two ways. First, audit quality in China is still perceived as relatively poor compared with that in the US and other developed countries (Chen et al., 2010). A recent change in the market strategy of audit firms in China involves the emphasis on development of industry specialisation. China’s regulators believe that developing industry specialisa- tion is an important strategy for the development of the whole audit industry. For example, in October 2009, the general office of the State Council forwarded the ‘Several opinions on accelerating the development of China’s CPA industry’ made by the Ministry of Finance, and emphasised the importance of developing industry special- isation in an audit market. In September 2011, ‘China’s CPA industry development plan (2011–2015)’ was issued by the Chinese Institute of Certified Public Accountants (CICPA), which definitely suggests enhancing auditors’ industry specialisation and developing international competitive advantages. Second, unlike the US and other developed markets where the Big 4 auditors are the major players, the Chinese audit market for listed companies is dominated by local Chinese audit firms. The competition among audit firms is more pronounced in China due to active participation of small- and mid-sized local audit firms and the low con- centration of Big 4 auditors. In 2013, there were 8,209 audit firms in China, and the local Chinese audit firm Ruihua entered the top 4 in China’s audit market for the first time. At the end of 2013, there were 40 audit firms qualified to audit approximately 2,400 listed companies, which means that one qualified firm had fewer than 60 listed clients on average. Such a buyer’s market is likely to afford clients more bargaining power and impose pressure on auditors fighting for their share of the market. 4. Hypothesis development Accounting professionals are typically based in specific practice offices and audit cli- ents in the same geographic locale; hence, their expertise and knowledge is both office China Journal of Accounting Studies 101 and client-specific (Ferguson et al., 2003; Francis et al., 2005). Furthermore, consider- ing different fiscal and taxation industrial policies and business environments, develop- ing industry specialisation at local office level can maintain competitive advantage and satisfy clients’ demands. The Chinese auditors usually operate only in the local market due to stronger geographical influences in the selection of audit firms (Wu & Zhang, 2012). Clients, in turn, have greater knowledge of and confidence in the expertise of locally based personnel who actually perform audits. The above argument assumes that audit firms are unable to fully achieve uniform audit quality across offices, and that a certain amount of overall audit expertise is office-specific. Industry specialists at the local level can charge higher audit fees, since they have obvious advantages in audit expertise, audit equipment and human capital. As a corollary, we argued that province- specific industry leaders can earn specialisation premiums. Our first hypothesis is thus as follows: Hypothesis 1: Ceteris paribus, the industry specialist auditors earn higher premiums than non-specialist auditors at the province level. Unlike developed economies in which the Big 4 auditors audit the majority of listed companies, concentration in China’s audit market for listed companies is rather low. Such a buyer’s market is likely to afford clients more bargaining power and impose pressure on auditors fighting for their slice of the cake. Therefore, audit firms tend to adopt a low-price strategy to compete with other audit firms. Following Numan and Willekens (2012a), we assume that the competitive pres- sure from the closest competitor will have a negative effect on audit fee premiums from industry specialisation. Such a negative effect can be more significant in China, since competition is quite intense among the non-Big 4 audit firms that dominate China’s audit market. More specifically, the farther (closer) is the closest competitor’s product-space location relative to that of the incumbent auditor, the less (greater) competitive pressure will be, ceteris paribus. To pursue their own maximum interest, clients urge audit firms to reduce the audit industry specialisation premiums through their bargaining power. Audit firms have no choices but to accept their request, because the clients always threaten to purchase audit services from the least-cost supplier, which can provide similar audit services (Chan, 1999). This is because cli- ents bear a lower transaction cost of switching auditors under such circumstances (DeAngelo, 1981). As a result, competitors’ relative product-space locations also affect the industry specialisation premiums that the incumbent auditor will be able to charge, leading to lower industry specialisation premiums. This implies that the spe- cialisation premiums will be moderated by competitive pressure. This leads to our second hypothesis: Hypothesis 2: Ceteris paribus, the industry specialist auditors earn lower premiums under greater competitive pressure. The industry specialisation premium for specialist auditors is moderated not only by competitive pressure, but also by scale economies. However, one cannot draw a conclusion that the audit firms continuously reduce the audit fee. To maximise profit, any audit firm cannot tender an audit fee that is lower than its current auditing cost to the client in the long run (Chan, 1999; Johnstone, Bedard, & Ettredge, 2004). Likewise, the specialist auditors will charge a higher industry specialisation premium where 102 Chen and Ma possible. Thus, an obvious assumption is that the overall benefits that audit firms can share with clients through competitive pressure or scale economies are limited. Based on spatial competition theory, Chan (1995) argues that the equilibrium audit fee charged to a client is a function of audit cost, and is also directly related to the audit pricing of the closest competitor in the audit market. Thus, each audit firm can calculate the audit fee offered by the other audit firms, and respond with an audit fee that is attractive to the client but still leaves it with an economic profit. In advance of the audit activity a specialist auditor may set a priority of charging an industry special- isation premium. However, the closest competitor can force the specialist auditor to decrease the industry specialisation premium, especially in China’s audit market. If the industry specialisation premiums reduced by competitive pressure are still much higher than the target profits, audit firms are able to pass on higher cost savings from scale economies to clients. In contrast, specialist auditors will pass on lower cost savings from scale economies to clients if they are forced to decrease the industry specialisa- tion premium when facing greater competitive pressure. The logic behind it is that once the industry specialisation premiums reduced by competitive pressure are determined, audit firms can determine the extent to which scale economies may be passed on to clients. It means that specialist auditors pass on lower cost savings from economies of scale to clients when facing greater competitive pressure. This leads to our third hypothesis: Hypothesis 3: Ceteris paribus, there is a substitutive effect, between competitive pressure and economies of scale, on the industry specialisation premium. 5. Research design To examine the relationship between industry specialisation, competitive pressure and audit fee, we run the following equation in line with existing studies on audit fee (Choi, Kim, Liu, & Simunic, 2009; Choi, Kim, Kim, & Zang, 2010; Francis et al., 2005; Francis & Yu, 2009). We cluster standard errors for each company to address potential problems associated with the non-independence of the panel observations (Petersen, 2009). Laf ¼ b þ b Spec þ b CP þ b Scale þ b Spec CP þ b Spec Scale þ b CP 0 1 2 3 4 5 6 Scale þ b Spec CP Scale þ b Size þ b Cata þ b Quick þ b Lev 7 8 9 10 11 þ b ROI þ b SOE þ b OP þ b Loss þ b Cha þ b ProvS þ b HHI 12 13 14 15 16 17 18 þ b IPS þ b Local þ b DS þ Year fixed effect þ Industry fixed effect þ e (1) 19 20 21 The dependent variable (Laf) is the natural log of audit fees (in thousands of dollars). We obtained audit fees data, which only includes domestic audit fees from CSMAR. Considering the provincial administration in China’s audit market, we use province clienteles to measure auditor industry specialisation, auditor competitive pressure and auditor economies of scale. Following Ferguson et al. (2003), Francis et al. (2005) and Fung et al. (2012), we use audit fees (province clienteles) to construct the specialisation variables. First, we compute each audit firm’s share in each one-digit industry group (manufacture industry classification is according to the two-digit) in each province, and then rank the auditors based on their province-level industry share. The auditor with the largest share at the China Journal of Accounting Studies 103 province level is designated the industry specialist (Spec); others are non-industry specialists. The use of concentration measures assumes that all firms in an industry face the same level of competition, but this is often not the case in practice. Taking these con- siderations into account, in this study we choose not to focus on the effect of auditor concentration on pricing. Instead, drawing on spatial economics, we examine how the relative location of competing auditors in the same market segment affects audit pricing. To capture the auditor’s competitive pressure from its closest competitor (our second hypothesis), we follow Numan and Willekens (2012a) and Numan and Wille- kens (2012b) and define competitive pressure between the incumbent auditor and its closest competitor based on province-level industry share. Hence, we define competitive pressure (CP) as the absolute difference between the incumbent audit supplier‘s market share in the client’s industry and the market share of its closest competitor. To explain the empirical results more intuitively, we replace CP with –CP, which means the larger the CP, the greater will be the competitive pressure. Auditor industry specialisation measures are widely used in the literature. However, proxies for scale economies are less commonly employed. Furthermore, some relevant data are not generally available in the auditing context, and our measure of scale econ- omies is thus indirect. Following Fung et al. (2012), we use the number of clients the auditor has in each province-industry for each year (Number) to measure the auditor scale (Scale). Considering the highly skewed nature of the variable Number, we rank Number across all province-industry combinations and use its percentile transformation in the regression. Following prior research (Choi et al., 2009, 2010; Francis et al., 2005; Francis & Yu, 2009), we control for several factors that could affect audit pricing. Size is the natu- ral logarithm of the company’s total assets in that year and is positively associated with fees. Cata is the ratio of current assets to total assets and is positively associated with fees. Quick is the ratio of the current assets (excluding inventory) to current liabilities and negatively associated with fees. Lev is the ratio of long-term debt to total assets. There is an expectation that the audit fee is lower for client companies having higher leverage because the lenders carry out a monitoring role. ROI is the ratio of earnings before interest and tax to total assets and is positively associated with fees. SOE is an indicator variable that indicates the controlling shareholder is a government agency and is negatively associated with fees. OP is an indicator variable that indicates the audit report is modified and is positively associated with fees. Loss is an indicator variable that indicates a loss in a given year and is positively associated with fees. Cha is an indicator variable that indicates auditor change in a given year and is negatively associated with fees. ProvS is natural log of aggregate audit fees for all firms audited by the company’s auditor for each province and is positively associated with fees. HHI is the Herfindahl concentration index per audit market and is positively associated with fees. IPS is the audit fees an audit firm generates in an industry as a percentage of the total fees generated by an audit firm in a province and is positively associated with fees. Local is an indicator variable that indicates audit offices and client firms are located in the same region and is positively associated with fees. DS is the natural loga- rithm of number of business segments and is positively associated with fees. Table 1 shows the detailed calculation for these variables. All data are taken from the CSMAR database. 104 Chen and Ma Table 1. Variable definitions. Predicted Variables sign Definition Laf Natural log of audit fees (in thousands of dollars) Spec + Indicator variable coded 1 for auditors that are province-industry leaders, and 0 otherwise CP – Competitive pressure which is smallest absolute fee market share difference between the incumbent auditor and his closest competitor in an audit market. We replace CP with –CP in the regression Scale – Percentile rank of the province-industry number of audit clients for each audit firm (variable values range from 0.01 to 1) Size + Natural log of total assets Cata + Ratio of current assets to total assets Quick – Ratio of current assets (excluding inventories) to current liabilities Lev – Ratio of long-term debt to total assets ROI + Ratio of earnings before interest and tax to total assets SOE – Indicator variable that indicates the controlling shareholder is a government agency OP + Indicator variable that indicates the audit report is modified Loss + Indicator variable that indicates a loss in a given year Cha – Indicator variable that indicates auditor change in a given year ProvS + Natural log of aggregate audit fees for all firms audited by the company’s auditor for each province HHI + Herfindahl concentration index per audit market, HHI=1–ΣP , P is the i i market share of each audit firms in province-industry audit market IPS + Audit fees an audit firm generates in an industry as a percentage of the total fees generated by an audit firm in each province Local + Indicator variable that indicates audit offices and client firms are located in the same region DS + Natural logarithm of number of industries the client is involved 6. Descriptive statistics and empirical results Our sample covers specialist auditors between 2001 and 2011. There was very little available data before 2001. We start with all China’s A-share listed companies, audited by non-Big 4 firms, for which basic data are available from the CSMAR database dur- ing the period 2001–2011, yielding 13,624 firm-year observations. We then exclude: (1) firms for which some data are unavailable on CSMAR (1,066); and (2) firms in the financial sector (359). In addition, because both industry expertise and auditor scale are identified at the province level, we exclude: (3) observations for province-industry-year combinations for which there are fewer than two companies and cities with only one auditor (1,728). Hence, we have 10,471 observations. To mitigate the undue influence of extreme values, we winsorise all continuous variables at the bottom and top 1% per- centiles. Table 2 reports the sample distribution by year and industry. The annual trend in specialist auditors suggests that the number of firms audited by specialist audit firms increased throughout our sample period from 325 in 2001 to 542 in 2011. However, the number of firms that employ non-specialists auditors is still larger than the number of firms audited by non-specialist auditors. A careful examination of Table 2 also reveals that employing specialist auditors tended to take place in industries such as Petroleum, Chemical, Plastics and Rubber Products Manufacturing, Machinery, Equip- ment and Instrument Manufacturing, and Medicine and Biological Products. Fewer China Journal of Accounting Studies 105 Table 2. Sample distribution. 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 Total Panel A: Distribution by year Spec 325 338 360 369 335 298 310 349 390 448 542 4,064 Non-Spec 411 430 463 480 471 469 483 594 678 813 1,115 6,407 Total 736 768 823 849 806 767 793 943 1,068 1,261 1,657 10,471 Panel B: Distribution by industry Spec Non-Spec Total Farming, forestry, animal husbandry and fishing 98 94 192 Mining 61 58 119 Food and beverage 176 216 392 Textile, apparel, fur and leather 217 222 439 Paper and allied products; printing 40 39 79 Petroleum, chemical, plastics and rubber products manufacturing 500 819 1,319 Electronics 180 263 443 Metal or non-metal mineral 311 662 973 Machinery, equipment and instrument manufacturing 649 1,297 1,946 Medicine and biological products 387 305 692 Other manufacturing 61 60 121 Utilities 144 199 343 Construction 68 127 195 Transportation and warehousing 162 226 388 Information technology 252 298 750 Wholesale and retail trades 340 369 709 Real estate 136 365 501 Public facilities and other services 92 131 223 Communication and cultural industries 28 29 57 Conglomerates 244 346 590 Total 4,064 6,407 10,471 Spec firms are firms that are audited by industry specialist auditors at province level. Non-Spec firms are those that are audited by non-specialist auditors at province level. 106 Chen and Ma Table 3. Descriptive statistics. Mean Median SD Min Max Panel A: Full sample (n = 10,471) Laf 6.177 6.131 0.513 5.011 8.161 Spec 0.388 0.000 0.487 0.000 1.000 CP −0.162 −0.073 0.202 −0.817 0.000 Scale 0.526 0.326 0.264 0.251 0.996 Size 14.447 14.360 1.060 11.852 17.894 Cata 0.555 0.567 0.211 0.085 0.969 Qui 1.465 0.873 2.043 0.103 13.932 Lev 0.065 0.022 0.094 0.000 0.433 ROI 0.060 0.057 0.074 −0.232 0.313 SOE 0.612 1.000 0.487 0.000 1.000 OP 0.074 0.000 0.262 0.000 1.000 Loss 0.106 0.000 0.308 0.000 1.000 Cha 0.193 0.000 0.395 0.000 1.000 ProvS 17.275 17.252 1.064 14.657 19.461 HHI 0.377 0.347 0.175 0.116 0.868 IPS 0.370 0.235 0.328 0.025 1.000 Local 0.887 1.000 0.316 0.000 1.000 DS 0.893 1.099 0.691 0.000 2.398 Panel B: Spec (n = 4,064) Laf 6.344 6.310 0.543 5.011 8.161 CP −0.289 −0.227 0.229 −0.817 0.000 Scale 0.668 0.700 0.263 0.251 0.996 Size 14.632 14.518 1.051 11.852 17.894 Cata 0.543 0.552 0.205 0.085 0.969 Qui 1.344 0.857 1.749 0.103 13.932 Lev 0.066 0.028 0.091 0.000 0.433 ROI 0.061 0.057 0.071 −0.232 0.313 SOE 0.641 1.000 0.480 0.000 1.000 OP 0.063 0.000 0.243 0.000 1.000 Loss 0.096 0.000 0.294 0.000 1.000 Cha 0.194 0.000 0.395 0.000 1.000 ProvS 17.059 17.058 1.047 14.657 19.461 HHI 0.446 0.458 0.177 0.116 0.868 IPS 0.344 0.237 0.286 0.025 1.000 Local 0.920 1.000 0.272 0.000 1.000 DS 0.957 1.099 0.692 0.000 2.398 Panel C: Non-Spec (n=6,407) Laf 6.071 6.016 0.463 5.011 8.161 CP −0.082 −0.036 0.129 −0.817 0.000 Scale 0.436 0.295 0.222 0.251 0.982 Size 14.330 14.248 1.049 11.852 17.894 Cata 0.563 0.577 0.214 0.085 0.969 Qui 1.542 0.884 2.206 0.103 13.932 Lev 0.064 0.018 0.095 0.000 0.433 ROI 0.059 0.057 0.076 −0.232 0.313 SOE 0.594 1.000 0.491 0.000 1.000 OP 0.081 0.000 0.272 0.000 1.000 (Continued) China Journal of Accounting Studies 107 Table 3. (Continued). Mean Median SD Min Max Loss 0.113 0.000 0.316 0.000 1.000 Cha 0.193 0.000 0.395 0.000 1.000 ProvS 17.412 17.448 1.052 14.657 19.461 HHI 0.334 0.293 0.159 0.116 0.868 IPS 0.386 0.234 0.351 0.025 1.000 Local 0.866 1.000 0.340 0.000 1.000 DS 0.852 0.693 0.688 0.000 2.398 Panel D: Spec – Non-Spec Mean t-value Median z-value difference difference *** *** Laf 0.327 30.562 0.294 27.853 *** *** CP −0.209 −61.596 −0.192 −57.593 *** *** Scale 0.218 −6.897 0.405 40.141 *** ** * , , and denote statistical significance at the level of 1%, 5%, and 10%, respectively. Laf is natural log of audit fees (in thousands of dollars). Spec firms are firms that are audited by industry specialist auditors at province level. Non-Spec firms are those that are audited by non-specialist auditors at province level. Spec is an indicator variable coded 1 for auditors that are province-industry leaders and 0 otherwise. CP is competitive pressure, which is the smallest absolute fee market share difference between the incumbent auditor and his closest competitor in an audit market, which is replaced by –CP in regression. Scale is the percentile rank of the province-industry number of audit clients for each audit firm (variable val- ues range from 0.01 to 1). Size is natural logarithm of total assets. Cata is the ratio of current assets to total assets. Quick is the ratio of the current assets to current liabilities. Lev is the ratio of long-term debt to total assets. ROI is the ratio of earnings before interest and tax to total assets. SOE is an indicator variable that indicates the controlling shareholder is a government agency. OP is an indicator variable that indicates the audit report is modified. Loss is an indicator variable that indicates a loss in a given year. Cha is an indicator variable that indicates auditor change in a given year. ProvS is natural log of aggregate audit fees for all firms audited by the company’s auditor for each province. HHI is the Herfindahl concentration index per audit mar- ket. IPS is the audit fees an audit firm generates in an industry as a percentage of the total fees generated by an audit firm in a province. Local is an indicator variable that indicates audit offices and client firms are located in the same region. DS is the natural logarithm of number of business segments. The test for mean difference is Student’s t-test, and the test for median difference is Wilcoxon test. firms purchase specialist audit service in industries such as Mining, Paper and Allied Products; Printing, and Communication and Cultural Industries. In summary, this is consistent with the idea that clients in different industries have unique demands for high quality audit service. Table 3 shows the descriptive statistics of all variables. Panel A reports the full sample statistics; Panels B and C report the statistics for spec firms and non-spec firms. We also report the difference between spec firms and non-spec firms in Panel D. We see that the mean (median) Laf for the full sample is 6.177 (6.131). The mean (median) Laf for the spec firms is 6.344 (6.310), while that for the non-spec firms is 6.071 (6.016). The differences in the means and medians are –0.327 and –0.294, respectively, both of which are significant at the 1% level. The mean (median) CP for the spec firms is –0.289 (–0.227), while that for the non-spec firms is –0.082 (–0.036), the difference of 0.209 (0.192) being significant at the 1% level. The mean (median) Scale for the spec firms is 0.668 (0.700), while that for the non-spec firms is 0.436 (0.295), the dif- ference of –0.218 (–0.405) being significant at the 1% level. The above univariate tests indicate that industry specialists are significantly more likely to charge a higher audit fee and bear less competitive pressure from the closest competitor and enjoy larger 108 Chen and Ma Table 4. Pearson correlation matrix. Laf Spec Scale CP Size Cata Qui Lev ROI SOE OP Loss Cha ProvS HHI IPS Local Spec 0.259*** Scale 0.085*** 0.429*** CP −0.163*** −0.501*** −0.362*** Size 0.624*** 0.139*** 0.011 −0.073*** Cata 0.000 −0.046*** 0.033*** 0.084*** −0.075*** Qui −0.128*** −0.047*** 0.041*** 0.068*** −0.171*** 0.270*** Lev 0.165*** 0.013*** −0.086*** −0.010 0.377*** −0.299*** −0.168*** ROI 0.100**** 0.014 0.034*** −0.039*** 0.143*** 0.063*** 0.156*** −0.028*** SOE 0.023* 0.047*** −0.032*** −0.050*** 0.208*** −0.189*** −0.165*** 0.125*** −0.079*** OP −0.056*** −0.033*** −0.061*** 0.039*** −0.228*** −0.028*** −0.080*** −0.058*** −0.289*** −0.054*** Loss −0.060*** −0.027*** −0.056*** 0.030** −0.164*** −0.086*** −0.108*** 0.004 −0.632*** −0.005 0.377*** Cha 0.037*** 0.001 −0.044*** −0.002 0.035*** −0.038*** −0.073*** 0.018 −0.053*** 0.020** 0.059*** 0.068*** ProvS 0.250*** −0.161*** 0.201*** 0.173*** 0.113*** 0.172*** 0.161*** −0.056*** 0.084*** −0.183*** −0.027*** −0.044*** −0.008 HHi 0.027*** 0.310*** 0.043*** −0.701*** 0.010 −0.134*** −0.077*** 0.036*** 0.026*** 0.090*** −0.024* −0.005 0.013 −0.334*** IPS 0.027*** −0.063*** −0.289*** 0.081*** 0.036*** 0.004 −0.050*** 0.019 −0.039*** 0.013 0.063*** 0.035*** 0.101*** −0.313*** −0.018 Local 0.078*** 0.082*** 0.251*** −0.072*** 0.049*** 0.040*** 0.038*** 0.005 0.035*** 0.021** −0.027*** −0.031*** −0.024*** 0.264*** −0.011 −0.424*** DS 0.077*** 0.074*** 0.006 −0.064*** 0.077*** −0.060*** −0.177*** 0.040*** −0.042*** 0.095*** −0.025** −0.003 −0.006 −0.104*** 0.072*** 0.001 0.017 *** ** * , , and denote statistical significance at the level of 1%, 5%, and 10%, respectively. The dependent variable (Laf) is the natural log of audit fees (in thousands of dollars). Spec is indicator variable coded 1 for auditors that are province-industry leaders and 0 other- wise. CP is competitive pressure, which is smallest absolute fee market share difference between the incumbent auditor and his closest competitor in an audit market, and which is replaced by –CP in regression. Scale is percentile rank of the province-industry number of audit clients for each audit firm (variable values range from 0.01 to 1). Size is natural logarithm of total assets. Cata is the ratio of current assets to total assets. Quick is the ratio of the current assets to current liabilities. Lev is the ratio of long-term debt to total assets. ROI is the ratio of earnings before interest and tax to total assets. SOE is an indicator variable that indicates the controlling shareholder is a government agency. OP is an indicator variable that indicates the audit report is modified. Loss is an indicator variable that indicates a loss in a given year. Cha is an indicator variable that indicates auditor change in a given year. ProvS is the natural log of aggregate audit fees for all firms audited by the company’s auditor for each province. HHI is the Herfindahl concentration index per audit market. IPS is the audit fees an audit firm generates in an industry as a percentage of the total fees generated by an audit firm in a province. Local is an indicator variable that indicates audit offices and client firms are located in the same region. DS is the natural logarithm of number of business segments. China Journal of Accounting Studies 109 Table 5. Results for OLS regressions. (1) (2) (3) (4) HP (5) LP (6) Dependent variable: Laf *** *** *** *** *** Spec 0.221 0.128 0.471 0.257 0.163 (16.12) (7.57) (11.11) (6.64) (4.27) *** *** CP −0.481 0.061 0.556 (–10.80) (0.99) (4.26) *** *** Spec × CP −0.420 −1.818 (–5.63) (–9.02) *** Scale 0.064 −0.059 −0.162 (1.08) (–1.53) (–3.81) *** *** Spec × Scale −0.387 −0.179 −0.052 (–5.91) (–2.80) (–0.82) *** Scale × CP −1.721 (–5.38) *** Spec × Scale × CP 2.618 (7.02) *** *** *** *** *** *** Size 0.283 0.296 0.282 0.304 0.250 0.272 (32.44) (33.14) (32.28) (27.26) (25.44) (31.63) Cata 0.053 0.059 0.052 0.076 0.028 0.051 (1.29) (1.40) (1.28) (1.47) (0.56) (1.27) *** *** *** *** *** *** Qui −0.018 −0.017 −0.017 −0.017 −0.018 −0.017 (–6.17) (–5.72) (–6.02) (–3.98) (–5.19) (–6.05) * ** * ** Lev −0.137 −0.182 −0.146 −0.138 −0.146 −0.177 (–1.84) (–2.39) (–1.96) (–1.44) (–1.56) (–2.39) *** *** *** ** ** *** ROI 0.291 0.259 0.284 0.293 0.251 0.255 (3.14) (2.70) (3.06) (2.30) (2.12) (2.80) *** *** *** ** *** *** SOE −0.051 −0.055 −0.048 −0.044 −0.056 −0.048 (–3.40) (–3.47) (–3.19) (–2.24) (–3.17) (–3.23) *** *** *** *** *** *** OP 0.152 0.156 0.153 0.146 0.132 0.141 (6.75) (6.71) (6.72) (5.09) (4.39) (6.27) *** *** *** *** *** Loss 0.066 0.065 0.068 0.035 0.076 0.060 (Continued) 110 Chen and Ma Table 5. (Continued). (1) (2) (3) (4) HP (5) LP (6) (3.79) (3.70) (3.95) (1.50) (3.26) (3.51) * ** Cha −0.016 −0.014 −0.015 0.000 −0.027 −0.015 (–1.66) (–1.46) (–1.58) (0.01) (–2.20) (–1.57) *** *** *** *** *** *** ProvS 0.111 0.091 0.100 0.155 0.088 0.116 (11.53) (9.12) (10.42) (11.61) (7.54) (11.76) ** ** *** * HHI 0.107 −0.145 −0.064 0.216 −0.087 −0.110 (2.43) (–2.51) (–1.14) (3.88) (–1.37) (–1.91) *** *** *** *** *** *** IPS 0.152 0.140 0.153 0.148 0.136 0.130 (6.42) (5.71) (6.50) (4.26) (5.37) (5.74) Local 0.020 0.033 0.011 0.013 0.050 0.025 (0.78) (1.25) (0.45) (0.38) (1.60) (0.99) *** *** *** *** *** *** DS 0.049 0.050 0.047 0.045 0.054 0.050 (5.37) (5.32) (5.29) (3.88) (4.91) (5.72) *** *** Intercept −0.129 0.162 0.169 −1.233 0.823 0.084 (–0.64) (0.77) (0.83) (–4.74) (3.47) (0.41) Year Yes Yes Yes Yes Yes Yes Industry Yes Yes Yes Yes Yes Yes N 10,471 10,471 10,471 5,190 5,281 10,471 Adj. R 0.511 0.490 0.518 0.569 0.451 0.535 F 81.959 78.918 79.733 60.829 43.137 78.903 F-Test for difference in coefficients between the LP subsample and HP subsample: Spec × Scale: F = 5.036, p = 0.025 *** ** * Standard errors are adjusted for heteroscedasticity and clustering at firm-level and the adjusted t-statistics are given in parentheses. , , and denote statistical significance at the level of 1%, 5%, and 10%, respectively. The dependent variable (Laf) is the natural log of audit fees (in thousands of dollars). Spec is the indicator variable coded 1 for auditors that are province-industry leaders and 0 otherwise. CP is competitive pressure, which is the smallest absolute fee market share difference between the incumbent auditor and his closest competitor in an audit market, and which is replaced by –CP in regression. Scale is percentile rank of the province-industry number of audit clients for each audit firm (variable values range from 0.01 to 1). Size is natural logarithm of total assets. Cata is the ratio of current assets to total assets. Quick is the ratio of the current assets to current liabilities. Lev is the ratio of long-term debt to total assets. ROI is the ratio of earnings before interest and tax to total assets. SOE is an indicator variable that indicates the controlling shareholder is a government agency. OP is an indicator variable that indicates the audit report is modified. Loss is an indicator variable that indicates a loss in a given year. Cha is an indicator variable that indicates auditor change in a given year. ProvS is natural log of aggregate audit fees for all firms audited by the company’s auditor for each province. HHI is the Herfindahl concentration index per audit market. IPS is the audit fees an audit firm generates in an industry as a percentage of the total fees generated by an audit firm in a province. Local is an indicator vari- able that indicates audit offices and client firms are located in the same region. DS is the natural logarithm of number of business segments. China Journal of Accounting Studies 111 Table 6. Instrumental variables estimation. Dependent variable: Laf Coefficient z-value *** Spec 0.195 (3.73) *** CP 0.832 (2.72) *** Spec × CP −1.910 (−4.94) *** Scale −0.303 (−4.03) Spec × Scale −0.024 (−0.24) *** Scale × CP −3.617 (−3.97) *** Spec × Scale × CP 4.083 (4.27) *** Size 0.273 (29.77) Cata 0.050 (1.15) *** Qui −0.020 (−5.35) ** Lev −0.190 (−2.46) ROI 0.164 (1.72) *** SOE 0.146 (5.97) *** OP 0.063 (3.52) Loss −0.012 (−1.18) *** Cha 0.122 (11.23) *** ProvS −0.053 (−3.35) *** HHI −0.255 (−3.33) *** IPS 0.126 (5.06) Local 0.024 (0.88) *** DS 0.050 (5.39) Intercept 0.133 (0.57) Year Yes Industry Yes N 8,413 Adj. R 0.550 F 72.345 1st-Stage test statistic: F-statistic F = 609.39 J-statistic and p-Value J = 3.942，p = 0.414 Standard errors are adjusted for heteroscedasticity and clustering at firm-level and the adjusted t-statistics are *** ** * given in parentheses. , , and denote statistical significance at the level of 1%, 5%, and 10%, respectively. The dependent variable (Laf) is the natural log of audit fees (in thousands of dollars). Spec is the indicator vari- able coded 1 for auditors that are province-industry leaders and 0 otherwise. CP is competitive pressure, which is smallest absolute fee market share difference between the incumbent auditor and his closest competitor in an audit market, which is replaced by –CP in regression. Scale is percentile rank of the province-industry number of audit clients for each audit firm (variable values range from 0.01 to 1). Size is natural logarithm of total assets. Cata is the ratio of current assets to total assets. Quick is the ratio of the current assets to current liabilities. Lev is the ratio of long-term debt to total assets. ROI is the ratio of earnings before interest and tax to total assets. SOE is an indicator variable that indicates the controlling shareholder is a government agency. OP is an indicator vari- able that indicates the audit report is modified. Loss is an indicator variable that indicates a loss in a given year. Cha is an indicator variable that indicates auditor change in a given year. ProvS is natural log of aggregate audit fees for all firms audited by the company’s auditor for each province. HHI is the Herfindahl concentration index per audit market. IPS is the audit fees an audit firm generates in an industry as a percentage of the total fees gen- erated by an audit firm in a province. Local is an indicator variable that indicates audit offices and client firms are located in the same region. DS is the natural logarithm of number of business segments. scale than non-specialists. Panels B, C, and D of Table 3 also show how spec firms and non-spec firms differ in other dimensions. In comparison with non-spec firms, spec firms tend to have lower current assets to total assets (Cata), lower current assets to current liabilities (Qui), less frequently modified audit opinion (OP), and to be larger (Size), have higher debt (Lev), and higher performance (ROI). Table 4 presents correlations between the variables. The audit fee (Laf) is signifi- cantly and positively correlated with industry specialisation (Spec) and economies of scale (Scale), and the Pearson correlation is 0.259 and 0.085. The Pearson correlation 112 Chen and Ma Table 7. Coefficients on Spec for yearly regression. Dependent variable: Laf Coefficient t-value N Adj. R *** 2001 0.266 (8.13) 736 0.361 *** 2002 0.249 (8.12) 768 0.355 *** 2003 0.229 (8.10) 823 0.421 *** 2004 0.194 (6.94) 849 0.441 *** 2005 0.186 (7.09) 806 0.539 *** 2006 0.185 (6.49) 767 0.504 *** 2007 0.221 (7.60) 793 0.488 *** 2008 0.220 (8.39) 943 0.518 *** 2009 0.231 (9.06) 1,068 0.523 *** 2010 0.238 (9.92) 1,261 0.529 *** 2011 0.202 (9.89) 1,657 0.562 Standard errors are adjusted for heteroscedasticity and clustering at firm-level and the adjusted t-statistics are *** ** * given in parentheses. , , and denote statistical significance at the level of 1%, 5%, and 10%, respectively. The dependent variable (Laf) is the natural log of audit fees (in thousands of dollars). Spec is an indicator vari- able coded 1 for auditors that are province-industry leaders and 0 otherwise. between audit fee (Laf) and competitive pressure (CP)is –0.163, which is significantly negative. It is shown that the audit fee can be increased by developing industry special- isation, while it can be mitigated by competitive pressure at the same time. Table 5 reports the results from estimating equation (1). In Column 1, the adjusted R is 0.511, and the control variables are statistically significant in the expected direc- tion. Consistent with Fung et al. (2012), the coefficient on Spec measured by province clienteles is positive and significant (t = 16.12). The coefficient value indicates that all else being equal, an industry specialist charges a 22.1% higher audit fee than a non- specialist. The result shows that, on average, the industry specialist auditor at the prov- ince level can charge a higher audit fee. However, Wu and Zhang (2012) find that a specialist charges a 17.7% higher audit fee than a non-specialist using national clien- teles to measure industry specialisation. This suggests that industry specialisation as a differentiation strategy can satisfy clients’ needs and earn more audit fees, while the industry specialisation premiums must be underestimated if using national clienteles to measure auditor industry specialisation. It is therefore reasonable to examine the auditor industry specialisation premium at the province level. The above regression results pro- vide supportive evidence for our Hypothesis 1. An important feature of China’s audit market is its fierce competition. It is mean- ingful to examine the effect of the competitive pressure on the industry premium. The coefficient on CP is negative and significant at the 1% level in Column 2, which indi- cates competitive pressure from the closest competitor can force audit firms to reduce the industry specialisation premium. The coefficient for the interaction term Spec × CP is negative and significant in Column 3. Thus, some economic benefits produced from industry specialisation are passed on to clients when the specialist auditors are bearing greater competitive pressure. In summary, although audit firms can increase audit fees by developing industry specialisation, they will adjust their audit pricing decisions based on competitive pressure from their closest competitors. As a result, as the com- petitive pressure increases, audit firms will charge lower audit specialisation premiums to maintain their market share. The empirical results, which indicate the industry specialist auditors earn lower premiums under greater competitive pressure, provide supporting evidence for our Hypothesis 2. China Journal of Accounting Studies 113 Table 8. Robustness tests. Non-IPO firms subsample Full sample Dependent variable: Laf coefficient t-value coefficient t-value *** *** Spec 0.172 (4.75) 0.215 (5.52) *** *** CP 0.467 (3.74) 0.513 (3.83) *** *** Spec × CP −1.621 (−8.46) −1.945 (−10.69) *** *** Scale −0.149 (−3.47) −0.209 (−4.72) Spec × Scale −0.064 (−1.03) −0.100 (−1.56) *** *** Scale × CP −1.469 (−4.83) −1.775 (−5.44) *** *** Spec × Scale × CP 2.233 (6.26) 2.799 (7.70) *** *** Size 0.272 (30.92) 0.289 (32.45) Cata 0.045 (1.08) 0.023 (0.57) *** *** Qui −0.020 (−6.15) −0.017 (−5.82) ** *** Lev −0.183 (−2.44) −0.199 (−2.64) ** *** ROI 0.235 (2.52) 0.301 (3.19) *** *** SOE −0.054 (−3.53) −0.046 (−2.97) *** *** OP 0.136 (5.98) 0.147 (6.30) *** *** Loss 0.062 (3.63) 0.064 (3.71) *** Cha −0.014 (−1.43) −0.030 (−3.15) *** *** ProvS 0.117 (11.53) 0.138 (13.55) ** *** HHI −0.129 (−2.12) −0.173 (−2.77) *** *** IPS 0.130 (5.58) 0.153 (6.55) Local 0.022 (0.84) −0.007 (−0.30) *** *** DS 0.051 (5.79) 0.050 (5.37) ** Intercept 0.063 (0.30) −0.494 (−2.31) Year Yes Yes Industry Yes Yes N 9,749 11,113 Adj. R 0.546 0.593 F 76.191 103.172 Standard errors are adjusted for heteroscedasticity and clustering at firm-level and the adjusted t-statistics are *** ** * given in parentheses. , , and denote statistical significance at the level of 1%, 5%, and 10%, respectively. The dependent variable (Laf) is the natural log of audit fees (in thousands of dollars). Spec is the indicator var- iable coded 1 for auditors that are province-industry leaders and 0 otherwise. CP is competitive pressure, which is the smallest absolute fee market share difference between the incumbent auditor and his closest competitor in an audit market, and which is replaced by –CP in regression. Scale is percentile rank of the province-indus- try number of audit clients for each audit firm (variable values range from 0.01 to 1). Size is natural logarithm of total assets. Cata is the ratio of current assets to total assets. Quick is the ratio of the current assets to current liabilities. Lev is the ratio of long-term debt to total assets. ROI is the ratio of earnings before interest and tax to total assets. SOE is an indicator variable that indicates the controlling shareholder is a government agency. OP is an indicator variable that indicates the audit report is modified. Loss is an indicator variable that indicates a loss in a given year. Cha is an indicator variable that indicates auditor change in a given year. ProvS is natu- ral log of aggregate audit fees for all firms audited by the company’s auditor for each province. HHI is the Her- findahl concentration index per audit market. IPS is the audit fees an audit firm generates in an industry as a percentage of the total fees generated by an audit firm in a province. Local is an indicator variable that indi- cates audit offices and client firms are located in the same region. DS is the natural logarithm of number of business segments. Finally, we examine whether there is an interaction effect of CP and Scale on the industry specialisation premium. The question is whether the scale discount is more likely to be passed on when there is a less significant competitive pressure. The coeffi- cient on the interaction term Spec × Scale is –0.387 (Column 4), and it is negative and significant at 1% level, suggesting that a one-decile increase in scale is associated with a 7.82% (0.202 × 0.387) decline in audit fees. The coefficient on the interaction term Spec × Scale is –0.179, and it is negative and significant at 1% level (Column 5). The 114 Chen and Ma value of the coefficient indicates that a one-decile increase in scale is associated with a 3.62% decline in audit fees. Interestingly, the specialisation premium will be passed on to clients by scale discounts in both subsamples, although this is significantly stronger in the LP subsample, as indicated by the F-test (F = 5.036, p = 0.025). In addition, the coefficient on the interaction term Spec × Scale × CP is 2.618, and it is positive and significant at the 1% level (Column 6). The results indicate that the fierce competition in the audit market forces audit firms to give up more industry spe- cialisation premiums, the lower cost savings from the scale economy can be passed on to the clients when audit firms charge an equilibrium audit fee. This is because each audit firm can calculate the audit fee offered by the other audit firms, and respond with an audit fee that is attractive to the client but still leaves it with a monopolistic rent (Chan, 1995). Once the industry specialisation premiums reduced by competitive pres- sure are determined, audit firms can determine the extent to which scale economies may be passed on to clients. In sum, the greater competitive pressure comes from competitors, the smaller cost savings from economies of scale that industry specialist auditors pass on to their clients. The results provide strong evidence in favour of our Hypothesis 3. Another potential problem for interpreting our results is the issue of endogeneity. To address the concern of endogeneity, we employ the instrumental variable approach. In the spirit of Hoechle, Schmid, Walter, and Yermack (2012), Wintoki, Linck, and Netter (2012) and Jayaraman and Milbourn (2012), we use the industry average and lagged competitive pressure as instruments for competitive pressure. These instruments affect competitive pressure, but do not directly affect the audit fees, thus meeting the requirements for a valid instrument. Following Larcker and Rusticus (2010), we test whether the instruments can satisfy the relevance and exclusion conditions. Table 6 reports the regression results using instrumental variables. Specifically, the F-statistic in the first stage for weak identification is 609.39, indicating that the instru- ments adequately identify the model, and J-statistics (J = 3.942, p = 0.414) are not sig- nificant at conventional levels, meaning we cannot reject the null that the instruments are uncorrelated with the error terms and are correctly excluded from the second-stage regressions. This suggests instruments satisfy the relevance and exclusion conditions. The coefficient on the interaction term Spec × Scale × CP is 4.083 and significant at the 1% level (Column 1). These results are consistent with our earlier analyses. Thus, the substitutive effect of competitive pressure and economies of scale on the industry specialisation premium remains after addressing the potential endogeneity problem. We conduct three robustness tests to further test our hypotheses in this section. First, to avoid the undue influence of exogenous events in the given year over that period, following Fung et al. (2012), we perform the regression test yearly. Table 7 pre- sents the yearly coefficients for Spec. The results show that the specialisation premium is significant throughout the period and, as in Column 1 of Table 4. This suggests that the differentiation services are accepted by clients and will be paid more audit fees in the audit market, which is less sensitive to the exogenous events in the audit market. Next, we reconstruct the explanatory variables including Spec and CP, and re-examine our four hypotheses using the subsample excluding the IPO auditors. As the IPO firms consider their alternatives, even incumbent audit firms face pressure to provide competi- tive fees and services. The coefficient on the interaction term Spec × Scale × CP is positive and significant (Column 1 of Table 8). The empirical evidence shows industry specialists at the province level earn more audit fees than non-specialists, but competitive pressure has a significantly negative effect on the industry specialisation premium, and competitive China Journal of Accounting Studies 115 pressure is a substitute for economies of scale when moderating the industry specialisation premiums. Finally, we focus on a full sample including non-Big 4 and Big 4 auditors and we examine the effect of competitive pressure and economies of scale on the auditor specialisation premium using the full sample. The results are presented in the Column 2 of Table 8. The coefficient on the interaction term Spec × Scale × CP is significant at the 1% level, suggesting that the results are in accordance with earlier findings. 7. Conclusion We first construct an industry specialisation measure, economies of scale measure and competitive pressure measure at the province-industry level using a sample of China’s listed firms over the period 2001–2011. Our empirical results show that industry spe- cialist auditors (i.e., province-industry leaders) earn significant premiums (22.1%, on average) relative to non-specialist auditors. To be specific, we find that competitive pressure has a negative effect on the industry specialisation premium. We also document that specialist auditors pass on lower cost savings from economies of scale to clients when facing greater competitive pressure. Some relevant researches examine the audit pricing by Big N auditors in the American oligopolistic audit market, such as Numan and Willekens (2012a) and Fung et al. (2012), while we investigate audit pricing by non-Big 4 auditors in China’s audit market. The potential problem for interpreting our results is the issue of endogeneity. To address endogeneity, we estimate a two-stage-least-squares (2SLS) regression. Our results remain economically and statistically strong under the instrumental variable approach. To assess the sensitivity of our results, we conduct several robustness tests. First, considering the specificity of IPO audit, we reconstruct explanatory variables and re-examine our four hypotheses using the subsample excluding the IPO auditors and re-examine our research questions. Next, using the non-Big 4 and Big 4 full sample we re-examine our research questions. The empirical results are in accordance with our above-mentioned findings. These findings have an important implication for research on the auditor industry specialisation premium. Acknowledgements We appreciate helpful comments and suggestions from Liansheng Wu (Joint Editor), Xi Wu (Associate Editor), Jason Zezhong Xiao (Joint Editor) and two anonymous reviewers. This paper is the result of research supported by the National Natural Science Foundation of China (71263034), the Humanities and Social Science Project of the Ministry of Education of China (10XJC630003) and the Program of Higher-level Talents at Inner Mongolia University, China (Z20100103). References Bae, G. S., Choi, S. U., & Rho, J. H. (2012). Do industry specialist auditors improve investment efficiency? Working paper. Korea University, Korea, Indian School of Business, India, and Chungnam National University, Korea. Balsam, S., Krishnan, J., & Yang, J. S. (2003). Auditor industry specialization and earnings quality. AUDITING: A Journal of Practice & Theory, 22,71–97. Cachon, G. P., & Harker, P. T. (2002). Competition and outsourcing with scale economies. Management Science, 48, 1314–1333. 116 Chen and Ma Carson, E., Simnett, R., Soo, B. S., & Wright, A. M. (2012). Changes in audit market competi- tion and the Big N premium. AUDITING: A Journal of Practice & Theory, 31,47–73. Chan, D. K. (1995). Competition in auditing: A spatial approach (Unpublished doctoral disserta- tion). The University of British Columbia, Canada. Chan, D. K. (1999). “Low-balling” and efficiency in a two-period specialization model of audit- ing competition. Contemporary Accounting Research, 16, 609–642. Chen, S., & Ma, H. (2013). Auditor industry specialization, reputation, and pass-through effects of economies of scale. Auditing Research, (6),84–92. (in Chinese). Chen, S., Sun, S. Y. J., & Wu, D. (2010). Client importance, institutional improvements, and audit quality in China: An office and individual auditor level analysis. The Accounting Review, 85, 127–158. Choi, J. H., Kim, C. F., Kim, J. B., & Zang, Y. (2010). Audit office size, audit quality and audit pricing. AUDITING: A Journal of Practice & Theory, 29,73–97. Choi, J. H., Kim, J. B., Liu, X. H., & Simunic, D. A. (2009). Cross-listing audit fee premiums: Theory and evidence. The Accounting Review, 84, 1429–1463. Craswell, A. T., Francis, J. R., & Taylor, S. L. (1995). Auditor brand name reputations and indus- try specializations. Journal of Accounting and Economics, 20, 297–322. DeAngelo, L. (1981). Auditor size and audit quality. Journal of Accounting and Economics, 3, 183–199. Dunn, K., & Mayhew, B. (2004). Audit firm industry specialization and client disclosure quality. Review of Accounting Studies, 9,35–58. Ferguson, A., Francis, J. R., & Stokes, D. J. (2003). The effects of firm-wide and office-wide industry expertise on audit pricing. The Accounting Review, 78, 429–448. Fields, L. P., Fraser, D. R., & Wilkins, M. S. (2004). An investigation of the pricing of audit ser- vices for financial institutions. Journal of Accounting and Public Policy, 23,53–77. Francis, J. R., Reichelt, K., & Wang, D. (2005). The pricing of national and city-specific reputa- tions for industry expertise in the US audit market. The Accounting Review, 80,113–136. Francis, J. R., Stokes, D. J., & Anderson, D. (1999). City markets as a unit of analysis in audit research and the re-examination of Big 6 market shares. Abacus, 35, 185–206. Francis, J. R., & Yu, M. D. (2009). Big 4 office size and audit quality. The Accounting Review, 84, 1521–1552. Fresard, L., & Valta, P. (2013). Competitive Pressure and Corporate Investment: Evidence from Trade Liberalization. Working paper of University of Maryland, USA and HEC Paris, France. Fung, S. Y. K., Gul, F. A., & Krishnan, J. (2012). City-level auditor industry specialization, econ- omies of scale, and audit pricing. The Accounting Review, 87, 1281–1307. Gist, W. E. (1994). A research note on the relationship between regulation and audit firm size on audit fees. Journal of Accounting, Auditing and Finance, 9, 381–396. Hall, R. E. (1988). The relation between price and marginal cost in U.S. Industry. Journal of Political Economy, 96, 921–947. Han, H. L., & Chen, H. W. (2008). Is industry specialized a useful competitive strategy for accounting Firms? Evidence from Chinese audit market. Auditing Research, (1),53–60. (In Chinese). Harrison, A. E. (1994). Productivity, imperfect competition and trade reform. Journal of Interna- tional Economics, 36,53–73. Hoechle, D., Schmid, M., Walter, I., & Yermack, D. (2012). How much of the diversification dis- count can be explained by poor corporate governance? Journal of Financial Economics, 103, 41–60. Hotelling, H. (1929). Stability in competition. The Economic Journal, 39,41–57. Jaggi, B., Gul, F. A., & Lau, T. S. C. (2012). Auditor industry specialization, political economy and earnings quality: Some cross-country evidence. Journal of International Financial Man- agement & Accounting, 23,23–61. Jayaraman, S., & Milbourn, T. T. (2012). The role of stock liquidity in executive compensation. The Accounting Review, 87, 537–563. Johnstone, K. M., Bedard, J. C., & Ettredge, M. L. (2004). The effect of competitive bidding on engagement planning and pricing. Contemporary Accounting Research, 21,25–53. Junius, K. (1997). Economies of scale: A survey of the empirical literature. Working paper DGZ-DekaBank, Germany. China Journal of Accounting Studies 117 Klette, T. (1999). Market power, scale economies and productivity: estimates from a panel of establishment data. The Journal of Industrial Economics, 47, 451–476. Krishnan, G. V. (2003). Does big 6 auditor industry expertise constrain earnings management? Accounting Horizons, 17,1–16. Larcker, D., & Rusticus, T. (2010). On the use of instrumental variables in accounting research. Journal of Accounting and Economics, 49, 186–205. Li, C., Xie, Y., & Zhou, J. (2010). National level, city level auditor industry specialization and cost of debt. Accounting Horizons, 24, 395–417. Lu, Z. F., Wang, C. F., & Wu, L. N. (2012). Institutional improvement, client importance of busi- ness groups and modified audit opinion. Accounting Research, (10),71–96. (In Chinese). Minutti-Meza, M. (2013). Does auditor specialization improve audit quality? Journal of Account- ing Research, 51, 779–817. Numan, K., & Willekens, M. (2012a). Competitive pressure, audit quality and industry special- ization. Working paper of Katholieke Universiteit Leuven. Numan, W., & Willekens, M. (2012b). An empirical test of spatial competition in the audit mar- ket. Journal of Accounting and Economics, 53, 450–465. Petersen, M. A. (2009). Estimating standard errors in finance panel data sets: Comparing approaches. Review of Financial Studies, 22, 435–480. Simunic, D. A. (1980). The pricing of audit services: Theory and evidence. Journal of Account- ing Research, 18, 161–190. Simunic, D. A., & Wu, X. (2012). China-related research in auditing: A review and directions for future research. China Journal of Accounting Research, 2,1–25. Vives, X. (2008). Innovation and competitive pressure. The Journal of Industrial Economics, 56, 419–469. Wintoki, M. B., Linck, J. S., & Netter, J. M. (2012). Endogeneity and the dynamics of internal corporate governance. Journal of Financial Economics, 105, 581–606. Wu, X., & Zhang, J. S. (2012). Market position of Chinese local audit firms and associated economic return. Accounting Research, (7),80–89. (In Chinese).
http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png
China Journal of Accounting Studies
Taylor & Francis
http://www.deepdyve.com/lp/taylor-francis/competitive-pressure-economies-of-scale-and-auditor-industry-K5cBhcQuJN
