Big 4 Audit Fee Premiums for National and Office-Level Industry Leadership in the United Kingdom*

Transcription

1 Big 4 Audit Fee Premiums for National and Office-Level Industry Leadership in the United Kingdom* by Ilias G. Basioudis Aston Business School Aston University Birmingham B4 7ET United Kingdom and Jere R. Francis University of Missouri-Columbia and University of Melbourne College of Business 432 Cornell Hall Columbia, MO USA Draft Date: May 12, 2005 Version 7.2 Please Do Not Quote Without Permission *We appreciate comments on earlier versions of the paper presented at the 2004 annual meetings of the European Accounting Association, British Accounting Association, and the National Audit Conference of the British Accounting Association. Contact Author: Ilias G. Basioudis i.g.basioudis@aston.ac.uk Tel:

2 Big 4 Audit Fee Premiums for National and Office-level Industry Leadership in the United Kingdom SUMMARY. The pricing of Big 4 industry leadership is examined for a sample of U.K. publicly-listed companies. We find no evidence that Big 4 national industry leadership commands a fee premium despite the fact that industry leaders have an average market share of 45 percent of industry audit fees. As in other countries, the lead engagement office on a U.K. audit is normally located in the same city as the client s corporate headquarters, and we do find a significant premium of 15 percent for city-specific industry leaders relative to other Big 4 auditors. Thus the U.K. evidence is that Big 4 auditor reputations for industry expertise are based solely on office-level industry leadership. This result differs from Australian and U.S. audit markets in which the pricing of auditor reputations is jointly affected by national industry leadership and city-specific industry leadership. In addition, we find that all Big 4 auditors earn a premium relative to smaller (third-tier) accounting firms, although Big 4 city-specific industry leaders have a much larger premium (38 percent versus 19 percent). In addition, Big 4 cityspecific industry leaders have a premium of 18 percent relative to second-tier national firms (Grant Thornton and BDO Seidman), but other Big 4 firms (non-city leaders) do not. In sum, the U.K. data suggest a three-level hierarchy of audit quality based on differential audit pricing: (1) Big 4 city-specific industry leaders, (2) second-tier national firms and other Big 4 auditors (noncity leaders), and (3) smaller third-tier accounting firms.

3 Big 4 Audit Fee Premiums for National and Office-Level Industry Leadership in the United Kingdom Introduction This study analyzes the effect of Big 4 industry leadership on audit pricing for U.K. listed companies using the joint national-city framework developed by Ferguson et al. (2003). 1 There is evidence from Australia, Hong Kong and the U.S. that a Big 4 accounting firm s national industry leadership results in higher fees relative to other Big 4 auditors (Craswell et al. 1995; DeFond et al. 2000; Francis et al. 2005a), and we investigate if there is evidence of a similar premium for industry leadership in the U.K. However, a recent line of research has begun investigating if reputations of Big 4 accounting firms and the pricing of audits are driven by office-level industry leadership in specific cities rather than a firm s overall national industry leadership. The lead engagement office on an audit is typically located in the same city as the client s corporate headquarters, and there is evidence from both the Australian and U.S. audit markets that Big 4 accounting firms do not earn a premium for national industry leadership alone unless they are also the city-specific industry leader where the client is headquartered (Ferguson et al. 2003; Francis et al. 2005a). 2 We investigate if there is evidence of a similar joint national-city pricing effect in the UK audit market. It is not obvious that audit pricing in the U.K. will parallel the U.S. and Australia. The U.K. is geographically a much smaller country with one dominant city (London). As a result there may be less pronounced differences between offices in different cities and audits may 1 The term Big 4 is used because data in the study are from which is after the collapse of Arthur Andersen. The Big 4 accounting firms are Deloitte Touche, Ernst & Young, KPMG, and Pricewaterhousecoopers. 2 The broader issue is whether industry expertise of Big 4 auditors is a firm-wide or office-specific phenomenon. Ferguson et al. (2003) argue that it is firm-wide if the expertise of office-based professionals is captured and distributed throughout other offices of Big 4 firms by knowledge sharing practices, and alternatively that it is officespecific if industry expertise is indelibly local in character and tied to local engagement offices servicing clients generally headquartered in the same locale. A fee premium for national leadership would support the firm-wide view, while a premium for city-specific leadership would support the office-specific argument.

4 2 in fact be centrally conducted out of London. In other words, the national-city framework may not necessarily explain auditor reputations and audit pricing in geographically smaller and more economically centralized countries such as the U.K. What do we find? Surprisingly, the U.K. audit market is not dominated by the city of London, or by London offices of accounting firms, to the degree that was expected. Only 39 percent of the 907 sample firms (n=351) are headquartered in London, so the majority of U.K. companies are located elsewhere and are audited by non-london offices of accounting firms. There are 39 cities with two or more company headquarters in the sample, and the five largest audit markets for listed companies after London are Birmingham (71), Leeds (70), Manchester (59), Bristol (33) and Nottingham (29). 3 On average, the top-ranked auditor per industry has a national market share of 45 percent of total industry audit fees, and the second-ranked firm has a national market share of 22 percent, averaged across all 28 industries in the study. While the top two auditors nationally are clearly dominant, there is no evidence that either of the top two firms earn a premium relative to other Big 4 auditors. However, it is a different story when we examine the pricing of city-specific industry leadership. On average, Big 4 city-specific industry leaders earn a premium of 15 percent relative to other Big 4 auditors. Further analysis indicates that auditors that are joint city-national industry leaders do not earn a higher premium than city leaders alone, a result which differs from Ferguson et al. (2003) and Francis et al. (2005a). We conclude that U.K. audits are priced as if Big 4 reputations for industry expertise are driven solely by office-level industry leadership in city-specific audit markets. For completeness we also report a series of 3 Audit fees are somewhat more concentrated in London, but still total only 55 percent of audit fees in the sample of 907 listed companies. However, because London is so much larger than any other city in the sample we do include a London indicator variable to control for systematic London effects on audit pricing. Audit fees are significantly higher in London, but otherwise have no effect on the pricing of industry expertise.

5 3 tests comparing Big 4 fees with those of non-big 4 auditors. The main result from these tests is that Big 4 city-level industry leaders have a premium relative to both second-tier national auditors and smaller third-tier auditors, but Big 4 non-city leaders do not have a premium over second-tier firms. These results may help to explain the mixed results in prior U.K. research with respect to Big 4 audit fee premia. 4 Before proceeding it is important to understand what can be learned from audit pricing research. Systematically higher audit fees by specific classes of accounting firms provide evidence (albeit indirectly) that higher quality audits are provided by these auditors, ceteris paribus. Companies that voluntary purchase higher-priced audits are presumably paying for a higher-quality service since any licensed auditor can legally conduct an audit. There is a substantial body of empirical evidence (mainly from the United States) supporting that the large international accounting firms do provide higher quality audits, a viewpoint that has long been advanced in the research literature (e.g., DeAngelo 1981). 5 In addition, there is evidence that audits by Big 4 industry experts provide higher quality audits than those of other Big 4 firms. For example, Balsam et al. (2003) and Krishnan (2003) document that abnormal accruals are 4 Prior U.K. studies have investigated large firm pricing relative to other auditors, but no studies have examined differential pricing of industry expertise among the dominant large firms. U.K. studies have reported mixed results on whether or not the large international accounting firms (now Big 4) earn a premium relative to other accounting firms. For example, Chan et al. (1993), Pong and Whittington (1994) and Ezzamel et al. (1996) report fee premiums, but other studies by Brinn et al. (1994), Che-Ahmad and Houghton (1996) and Ezzamel et al. (2002) fail to find evidence of a large-firm premium. 5 For example, the large international accounting firms are sued less frequently and have fewer sanctions by the Securities and Exchange Commission, both of which suggest a lower rate of audit failure (Palmrose 1988; Feroz et al. 1991); auditors of the largest accounting firms are more likely to issue non-clean audit reports which suggests a more cautious and conservative reporting model (Francis and Krishnan 1999), and there is evidence that modified audit reports issued by the large accounting firms are more informative to investors (Weber and Willenborg 2003). In addition, the earnings of companies audited by the largest auditors show less evidence of discretionary earnings management (Becker et al. 1998; Francis et al. 1999). Companies with higher agency costs have a greater need for credible monitoring and are more likely to be audited by the largest accounting firms (Francis and Wilson 1988; DeFond 1992); and companies going public have severe information asymmetry problems and there is evidence the largest accounting firms reduce information asymmetry and IPO underpricing (Beatty 1989). Finally, the stock market response to earnings surprises is significantly higher for companies audited by the largest accounting firms (Teoh and Wong 1993).

6 4 smaller for companies audited by national industry leaders, which implies less aggressive accounting policies and higher quality earnings. More recently, Francis et al. (2005b) document that clients of Big 4 city-specific industry leaders have smaller abnormal accruals and higher earnings quality than the clients of other Big 4 auditors, but there is no evidence of a nationallevel industry effect on earnings quality after controlling for Big 4 city-level industry leadership. We discuss the implications of differential audit quality at the end of the paper. 6 The remainder of the paper is organized as follows. The sample, data and audit fee model are described in the next section, followed by results of the Big 4 industry leadership tests. Big 4 versus non-big 4 tests are then reported, and the paper concludes with a discussion of the study s implications. Sample, Data and Audit Fee Model Our sample comprises companies listed on the London Stock Exchange. Global Access and FAME databases provide financial and other data for the financial year. 7 This is the first U.K. study of audit pricing using fee data following the collapse of Arthur Andersen and the absorption of Andersen clients by other accounting firms. The initial sample includes around 2,200 companies. After excluding listed companies that provide financial and other services, and companies with missing data, the final sample consists of 907 publicly-listed companies and is 6 These arguments are not pejorative to non-big 4 accounting firms or to Big 4 auditors that not industry leaders. Presumably all auditors of publicly-listed companies meet minimum professional and legal standards. However, mandated minimum standards do not preclude individual accounting firms from developing reputations for expertise that exceed minimum standards. In addition, there is no reason to presume all companies have a uniform demand (and willingness to pay) for higher audit quality which explains observed cross-sectional differences in the demand for and supply of audits by different accounting firms such as the Big 4 or industry experts. In addition, the argument does not mean that audits of all Big 4 firms or Big 4 industry experts are of higher quality; rather, the argument simply means that on average these audits are likely to be of higher quality. 7 Global Access is a product of Thompson Financial Inc., and FAME is an acronym for Financial Analysis Made Easy, a comprehensive data base for U.K. private and publicly-listed companies maintained by Bureau Van Dijk.

7 5 summarized in Table 1. 8 Table 1 reports that Big 4 auditors performed 70 percent of audits in the sample of 907 companies and received 86 percent of audit fees. KPMG and Pricewatercoopers (PWC) are the two leading firms in the UK market, auditing 41 percent of clients nationally and earning 57 percent of total audit fees in the sample. The two second-tier auditors, BDO Seidman and Grant Thornton, conducted 13 percent of audits and received 7 percent of audit fees, and 58 third-tier firms audited 17 percent of the sample and received 7 percent of audit fees. [Insert Tables 1 and 2 Here] Table 2 reports the sample distribution and national industry leaders for the 28 London Stock Exchange industry codes applicable to the sample. On average, the national industry leader has 45 percent of industry fees while the second-ranked auditor has 22 percent of industry fees. A Big 4 accounting firm is the top-ranked auditor in all cases except three small industries: SIC codes 1 (n=11), 40 (n=3) and 80 (n=3). In these three industries, there is no designated Big 4 national industry leader in the analysis. As note above, KPMG and PWC are the two leading firms in the UK market with a combined market share of 57.4 percent of sample audit fees. Therefore it is not surprising that KPMG or PWC are the top-ranked firm in 19 of 28 industries, and the second-ranked firm in 18 of 28 industries. Despite their overall market and industry dominance, there is no evidence of a firm-specific audit fee premium for KPMG and PWC relative to other Big 4 auditors. 9 8 Companies that provide financial and other services are typically excluded from audit fee studies because financial statement data and related ratios that are used in audit fee models are qualitatively different for this sector than for other companies. 9 This conclusion is based on an audit fee regression using all 907 observations in which we create separate auditor indicator variables for each Big 4 firm and a variable for second-tier auditors, along with the set of control variables in equation (1) later in the paper. We find that Deloitte Touche has a significantly smaller premium than the other Big 4 firms, but coefficients for the remaining three Big 4 firms are not significantly different from one another. We also test if each Big 4 auditor has a premium relative to the second-tier firms, and none of the individual Big 4 coefficients is significant (p>.10) which again indicates no firm-specific premium for KPMG or PWC.

8 6 There are 39 cities in the sample having 2 or more audits (n=884), with the remaining 23 observations located in smaller cities having only one observation. City-specific industry leadership is based on an accounting firm s share of industry fees per city. As indicated before, the lead engagement office is normally in the locale as the client s corporate headquarters. Following Reynolds and Francis (2000), the accounting firm s lead engagement office is identified from the office-specific letterhead used for the audit report. This data was hand collected in order to accurately measure each office s share of city-level industry fees and to determine the city-level industry leader for each unique city-industry combination in the sample. Descriptive statistics are reported in Table 3 for the full sample (n=907), clients of Big 4 auditors (n=631), clients of second-tier national firms, and clients of third-tier accounting firms (n=162). Big 4 firms audit larger clients (LTA and SQRTSUBS), issue fewer modified audit reports (OPINION), have relatively more nonaudit fees (LNAF) and proportionately fewer clients in London (LONDON). The audit clients of non-big 4 firms have lower liquidity risk (QUICK) but are less profitable (ROI) and make more losses (LOSS) compared to Big 4 clients. The remaining variables (CATA, DE, FOREIGN, BUSY) are comparable across auditor groups. [Insert Table 3 Here] The primary analysis is based on companies with Big 4 auditors and tests if Big 4 industry leaders have a fee premium relative to other Big 4 firms. To accomplish this, the sample of 631 companies having Big 4 audits is further reduced to 506 observations by requiring a minimum of two city-specific observations per industry to be included in the sample. The reason for this additional screen is that cities with only one listed company in an industry may not have a competitive audit market, although the results are comparable if the full sample of 631 observations is used. The reduced sample of 506 observations has 21 unique cities and 125 unique city-industry

9 7 combinations with an average of four observations per city-industry combination. City-specific industry leadership in these 125 city-industry combinations is distributed among the Big 4 firms as follows: Deloitte Touche (25), Ernst & Young (17), KPMG (38), and Pricewaterhousecoopers (45). Audit Fee Model Following prior research, a cross-sectional audit fee regression model is used to estimate Big 4 audit fee premia for industry leadership (Craswell et al. 1995; Ferguson and Stokes 2002). Audit fee regression models use a set of variables to control for cross-sectional differences in factors that affect fees such as client size, audit complexity and auditor client risk sharing (Simunic 1980). These models have good explanatory power (adjusted r-squares of 0.70 and higher) and have been robust across different samples, time periods, countries, and sensitivity analyses for model misspecification (Francis and Simon 1987; Chan et al. 1993). The experimental indicator variable in equation (1) below is denoted AUDITOR and represents the various codings of Big 4 industry leaders used in the model estimations. The research design in equation (1) tests for differential Big 4 audit fees after controlling for other factors affecting fees. Formally, the test determines if there is a significant positive intercept shift (higher fees) in the fitted regression model for observations audited by Big 4 industry leaders. The OLS regression model is formally specified as follows in equation (1): LAF = b 0 + b 1 LTA + b 2 SQRTSUBS + b 3 CATA + b 4 QUICK + b 5 DE + b 6 ROI + b 7 FOREIGN + b 8 OPINION + b 9 YE + b 10 LOSS + b 11 LNAF + b 12 LONDON + b 13 AUDITOR + e (1) where: LAF = natural log of audit fees in thousands of GB Pounds, LTA = natural log of total assets in thousands of GB Pounds, SQRTSUBS = square root of total subsidiaries, CATA = ratio of current assets to total assets, QUICK = ratio of current assets (less inventories) to current liabilities, DE = ratio of long-term debt to total assets, ROI = ratio of earnings before interest and tax to total assets,

10 8 FOREIGN OPINION BUSY LOSS LNAF LONDON AUDITOR e = proportion of total sales from foreign operations, = indicator variable, 1 = qualified audit report, = indicator variable, 1 = December 31 st or March 31 st year end, = indicator variable, 1=loss in any of the past three years, = natural log of nonaudit fees (in thousands of GP Pounds) paid to the auditor, = indicator variable if observation is a London-based company, = experimental auditor indicator variable (specification varies across models), = error term. Equation (1) is estimated as an industry fixed effects model to control for mean differences in fees across the 28 industries in the sample. For brevity, the industry indicator variables are not reported. With respect to the 12 control variables in the model, higher fees are expected (positive signs) for larger clients (LTA), for greater audit complexity (SQRTSUBS and FOREIGN), for greater audit risk (CATA, DE and LOSS), and for London-based companies (LONDON) due to higher salaries and other costs. A positive sign is expected for OPINION as prior studies document higher fees associated with modified opinions, possibly due to more investigative efforts in such circumstances. Given prior research, a positive association is also expected between audit fees and nonaudit fees (Whisenant et al. 2003). Lower fees (negative signs) are expected for higher values of QUICK and ROI. QUICK is an audit risk variable and clients with a smaller quick ratio are riskier (less liquid) and therefore are expected to have higher audit fees. Prior studies find that clients with higher ROI have lower fees, which is consistent with auditor client risk sharing, i.e., more profitable clients pose less risk to the auditor, resulting in lower fees. Finally, clients with December 31 st or March 31 st fiscal year ends (BUSY) are expected to have higher fees as these are the predominant busy seasons in the U.K. National and City-Specific Industry Leadership Tests Three models are estimated following the specifications in Ferguson et al. (2003). The sample comprises 506 companies with Big 4 auditors and the models test if Big 4 industry

11 9 leaders (defined in various ways) have higher fees than other Big 4 firms. Model 1 tests the effect of national-level industry leadership (alone) on differential Big 4 audit pricing. In model 1 the auditor test variable is coded one for 170 observations in which the Big 4 auditor is the national industry leader based on Table 2, and the default comparison group is all of the remaining 336 observations not having Big 4 industry leaders. Model 2 tests the effect of cityspecific leadership (alone) on differential Big 4 pricing. The auditor test variable is coded one for the 237 observations audited by Big 4 city-specific industry leaders, and the default comparison group is the remaining 269 observations not audited by city-specific industry leaders. Model 3 is the primary focus and uses three auditor indicator variables based on the joint national-city framework in Ferguson et al. (2003): Big 4 auditors that are both national industry leaders and city-specific industry leaders (n=118); Big 4 auditors that are national leaders but not city-specific industry leaders (n=52); and Big 4 auditors that are city-specific leaders but not national industry leaders (n=119). The default comparison group is Big 4 auditors that are neither national nor city-specific industry leaders (n=217). Results of the three model estimations are reported in Table 4. All three models are significant at p<.001 with adjusted R-squares of The control variables are significant at p<.05 (two-tail) in the expected direction, with the exception of CATA, QUICK, ROI, OPINION, and LOSS which are insignificant at p>.10 (two-tail). [Insert Table 4 Here] Model 1 is a test of national leadership and the auditor indicator variable is insignificant (p=.626). A second specification is estimated with an additional auditor indicator variable for the second-ranked auditor in the industry. In this model neither the top-ranked nor second-

12 10 ranked auditor indicator variable is significant at conventional levels. We conclude that national industry leadership alone does not result in an audit fee premium in the U.K. audit market. Model 2 is a test of city-specific industry leadership and the auditor indicator variable is positive and significant (p=.006). The coefficient value is.141 which equates to an average audit fee premium of 15 percent. 10 We also test if the second-ranked auditor in the city-specific industry has a fee premium, in addition to the city leader, but the second-ranked auditor variable is insignificant. Model 2 provides evidence that Big 4 reputations based on office-level industry leadership in specific cities drive the pricing of Big 4 industry expertise in the U.K. Ferguson et al. (2003) and Francis et al. (2005a), document a more complex relation in which there is a joint reputation effect from both national and city-specific leadership, and model 3 determines if this is also the case in the U.K. Model 3 shows that city-specific industry leadership, both alone and joint city-national industry leadership, leads to significantly higher audit fees. However, an F-test indicates there is so significant difference between these two auditor coefficients (F-ratio=0.18, p=.6687), which means that joint national-city leadership adds nothing over and above city leadership alone in the pricing of industry expertise. Finally, the coefficient on the auditor variable for national leaders that are not city-specific leaders is insignificant (p=.980). We conclude from model 3 that there is no evidence that Big 4 reputations for national industry leadership are priced in the U.K. audit market. However, the evidence does indicate that local office reputations based on city-specific industry leadership result in a significant audit fee premium Following Craswell et al. (1995, 307), the percentage magnitude of the positive intercept shift on the dependent variable (natural log of audit fees) is defined as e z - 1, where z is the auditor coefficient value in the regression model. 11 To assure the significance of city-specific industry leadership is not driven by KPMG and PWC, the two largest firms in the U.K., indicator variables are added to model 2 for KPMG and PWC. Results of this estimation are qualitatively the same as those of model 2 reported in Table 4: the city-specific variable is significant at p=.02 and the coefficient is 0.124, while coefficients on the indicator variables for KPMG and PWC are insignificant (p>.10).

13 11 Big 4 Versus Non-Big 4 Tests Table 5 reports a set of Big 4 versus non-big 4 audit fee comparisons based on the evidence in Table 4 that the Big 4 city-specific industry leaders have systematically higher fees than other Big 4 auditors. Three models are reported in Table 5. Model 1 uses all observations in the original sample of 907 companies in which there are a minimum of two observations per industry at the city level (n=711). We delete a further 18 observations from 14 unique cityindustry combinations in which a non-big 4 auditor is the city industry leader since the purpose of our study is to test Big 4 industry leadership, although the results are comparable if we retain these observations and classify them as non-industry leaders. Model 1 compares the fees of Big 4 auditors with fees of all non-big 4 auditors. Model 2 drops observations with third-tier auditors, and makes a direct comparison of the fees of Big 4 auditors and the second-tier firms Grant Thornton and BDO Seidman. Finally, model 3 uses the same observations as in model 1, but adds an additional auditor indicator variable for second-tier accounting firms. [Insert Table 5 Here] The results for model 1 indicate that Big 4 city-specific industry leaders have a premium of 27 percent over all other auditors, but other Big 4 auditors (non-city leaders) do not have a significant premium. Model 2 indicates there is a positive and significant (p=.06, two-tail) premium for Big 4 city-specific industry leaders relative to second-tier auditors, but there is no premium for other Big 4 auditors (non-city leaders). The coefficient is which represents a premium of 18 percent. Finally, model 3 indicates that Big 4 city-specific industry leaders have the largest premium relative to third-tier auditors. The coefficient is which represent a premium of 38 percent over the fees of third-tier auditors. In addition, the premiums for other Big 4 auditors (non-city leaders) and for second-tier auditors are both significant but much

14 12 smaller in magnitude (19 and 25 percent). An F-test (F-ratio=0.14, p=.5058) indicates there is no significant difference in the premiums of Big 4 non-city leaders and second-tier auditors, relative to third-tier auditors. We conclude from the models reported in Table 5 that all Big 4 accounting firms have a premium relative to third-tier auditors but Big 4 city-specific industry leaders have a much larger premium, and that Big 4 auditors have a premium over second-tier national firms only if they are a city-specific industry leader. Discussion and Conclusion This study adds to the global evidence that office-specific industry leadership affects a Big 4 accounting firm s reputation for industry expertise in the sense that it is priced in the audit market and results in higher observable fees relative to other Big 4 auditors. What is unique in the U.K. audit market is that national industry leadership appears to have no affect whatsoever on audit pricing. In contrast, the evidence from Australia and the U.S. indicates that the pricing of Big 4 industry expertise is based on joint reputation effects of national and city-specific industry leadership (Ferguson et al. 2003; Francis et al. 2005a). The U.K evidence indicates that office-specific industry leadership alone results in audit fee premia for industry expertise. What are the implications of these findings? First, an industry expertise premium for Big 4 city-specific industry leaders implies differential audit quality relative to other Big 4 auditors. There is evidence from U.S. companies that earnings are of higher quality when audited by Big 4 city-specific industry leaders (Francis et al. 2005ab), and this begs the question of whether there is a similar affect on the earnings quality of U.K. companies where the lead engagement office of the Big 4 auditor is the city-specific industry leader.

15 13 There are also implications for accounting firms, regulators and investors. If office-level industry leadership drives the reputation for industry expertise, and if this reputation is priced in the audit market, then Big 4 accounting firms have economic incentives to become office-level industry leaders in city-specific audit markets and to actively market their local reputations for industry expertise. This is in contrast to current marketing strategies of Big 4 firms which tend to emphasize their national and global expertise, despite the empirical evidence now from three different countries that local office reputations for industry expertise also matter and may be just as important as national reputations, perhaps even more important, at least in the U.K. The implication for regulators and investors is that audits may be of higher quality when conducted by industry experts, which implies that earnings may also be of higher quality. If this is the case, it may be more effective if regulators place relatively more emphasis (monitoring and investigations) on companies that are not audited by industry experts because the earnings of such companies are more likely to have potential misstatements (Francis et al. 2005ab). Finally, the message for investors is that earnings reports may be more credible and less likely to be materially misstated if audited by industry experts, and this would affect the usefulness of earnings reports in assessing firm performance and the pricing of equity securities.

16 14 References Balsam, S., J. Krishnan, and J. Yang Auditor Industry Specialization and Earnings Quality. Auditing: A Journal of Practice and Theory 23: Beatty, R Auditor reputation and the pricing of initial public offerings. The Accounting Review 64: Becker, C., M. DeFond, J. Jiambalvo, and K.R. Subramanyam The effect of audit quality of earnings management. Contemporary Accounting Research 15: Brinn, T., M. J. Peel, and R. Roberts Audit fee determinants of independent and subsidiary unquoted companies in the UK: an explanatory study. British Accounting Review 26: Chan, P., M. Ezzamel, and D. Gwilliam Determinants of audit fees for quoted UK companies. Journal of Business Finance and Accounting 20: Che-Ahmad, A. and K. A. Houghton Audit fee premiums of Big Eight firms: evidence from the market for medium-size UK auditees. International Accounting, Auditing & Taxation 5 (1, June): Craswell, A., J. Francis, and S. Taylor Auditor brand name reputations and industry specializations. Journal of Accounting and Economics 20: DeAngelo, L Auditor size and audit quality. Journal of Accounting and Economics 1: DeFond, M The association between changes in client firm agency costs and auditor switching. Auditing: A Journal of Practice and Theory 11: DeFond, M., J. Francis, and T. Wong Auditor industry specialization and market segmentation: evidence from Hong Kong. Auditing: A Journal of Practice and Theory 19: Ezzamel, M., D. R. Gwilliam, and K. M. Holland On the relationship between the pricing of audit and non-audit services: some empirical evidence from publicly quoted UK companies. Accounting and Business Research 27 (1): Ezzamel, M., D. R. Gwilliam, and K. M. Holland The relationship between categories of non-audit services and audit fees: evidence from UK companies. International Journal of Auditing 6 (1, March): Ferguson, A., and D. Stokes Brand name audit pricing, industry specialization and leadership premiums post Big 8 and Big 6 mergers. Contemporary Accounting Research 19:

17 15 Ferguson, A., J. Francis, and D. Stokes The effects of firm-wide and office-level industry expertise on audit pricing. The Accounting Review 78: Feroz, E., K. Park, and V. Pastena The financial and market effects of the SEC s accounting and auditing enforcement releases. Journal of Accounting Research 29: Francis, J The effect of audit firm size on audit prices: a study of the Australian market. Journal of Accounting and Economics 6: Francis, J., and J. Krishnan Accounting accruals and auditor reporting conservatism. Contemporary Accounting Research 16 (Spring): Francis, J., E. Maydew and H.C. Sparks The role of big six auditors in the credible reporting of accruals. Auditing: A Journal of Practice and Theory 18: Francis, J., K. Reichelt, and D. Wang. 2005a. The pricing of national and city-specific reputations for industry expertise in the U.S. audit market. The Accounting Review 80: Francis, J., K. Reichelt, and D. Wang. 2005b. Is earnings quality higher when auditors are cityspecific industry leaders? Working Paper (University of Missouri). Francis, J., and D. Simon, A test of audit pricing in the small-client segment of the U.S. audit market. The Accounting Review 62: Francis, J., and E. Wilson Auditor changes: a joint test of theories relating to agency costs and auditor differentiation. The Accounting Review 63: Krishnan, G Does big 6 auditor industry expertise constrain earnings management? Accounting Horizons 17 (Supplement): Palmrose, Z An analysis of auditor litigation and audit service quality. The Accounting Review 63: Pong, C. M. and G. Whittington The determinants of audit fees: some empirical models. Journal of Business Finance and Accounting, 21 (December): Reynolds, J.K., and J. Francis Does size matter? the influence of large clients on officelevel auditor reporting decisions. Journal of Accounting and Economics 30 (December): Simunic, D. A The pricing of audit services: theory and evidence. Journal of Accounting Research 18 (1):

18 16 Teoh, S., and T.J. Wong Perceived audit quality and the earnings response coefficient. The Accounting Review 68: Weber, J., and M. Willenborg Do expert informational intermediaries add value? evidence from auditors in microcap IPOs. Journal of Accounting Research 41 (September): Whisenant, S., S. Sankaraguruswamy, and K. Raghunandan Evidence on the joint determination of audit and non-audit fees. Journal of Accounting Research 41 (September): White, H., A heteroskedasticity-consistent covariance matrix estimator and a direct test for heteroskedasticity. Econometrica 48:

19 17 Table 1 Audit Fees and Sample Distribution Based on UK data n % Audit Fees % (GBP 000s) Big % 108, % Second Tier * % 8, % Third Tier ** % 8, % KPMG % 37, % PWC % 34, % DT % 22, % EY % 14, % GT % 4, % BDO % 4, % % GBP125, % * Second-tier firms are GT and BDO ** There are 58 small third-tier firms in the sample Definitions of Big 4 and Second-Tier Accounting Firms: DT = Deloitte & Touche EY = Ernst & Young KPMG = KPMG PWC = Pricewaterhousecoopers GT = Grant Thornton BDO = BDO Seidman

20 18 Table 2 National Industry Leadership Based on UK Audit Fees SIC London Stock Exchange Industry Name Observations per SIC Audit Fees (000 GBP) per SIC Industry Leader (#1) and market share of audit fees Industry Leader (#2) and market share of audit fees 1 Farming 11 1, MS (45.58%) EY (20.01%) 11 Mining 39 3, PWC (43.24%) EY (20.89%) 15 Manufacturing - Food 29 9, KPMG (79.02%) EY (8.27%) 17 Manufacturing - Clothing 29 3, DT (47.26%) PWC (21.40%) 20 Manufacturing - Wood 9 2, KPMG (37.45%) PWC (28.59%) 22 Printing & Publishing 37 6, DT (47.11%) PWC (21.71%) 24 Chemicals 30 5, KPMG (38.08%) PWC (31.87%) 25 Manufacture - Plastics 14 2, KPMG (39.77%) PWC (25.96%) 26 Manufacture - Ceramics 11 1, PWC (67.89%) KPMG (15.67%) 27 Manufacture - Metals 74 9, KPMG (30.39%) PWC (29.82%) 30 Manufacture - PC, TV, 21 2, KPMG (29.64%) DT (29.44%) 31 Electrical 34 5, DT (25.56%) PWC (19.58%) 33 Manufacture - Medical instruments 31 3, PWC (33.33%) DT (27.62%) 34 Manufacture - Transport 14 4, PWC (57.20%) DT (13.97%) 36 Manufacture - Other 28 4, KPMG (47.81%) DT (18.75%) 40 Utility Electricity AK (89.29%) PWC (8.48%) 41 Utility Water DT (31.93%) PWC (28.11%) 45 Construction 47 10, KPMG (45.30%) PWC (16.06%) 50 Wholesale 98 13, PWC (28.25%) KPMG (27.92%) 52 Retail 61 5, PWC (35.74%) DT (27.90%) 55 Leisure - Hotels, Restaurants 35 3, KPMG (31.84%) PWC (28.74%) 60 Transport 26 5, PWC (40.29%) KPMG (22.27%) 64 Utility - Telecommunications DT (44.16%) PWC (21.43%) 71 Support activities - Advert, Rent 56 8, PWC (45.19%) KPMG (33.96%) 72 Software publishing 84 7, EY (26.85%) PWC (19.79%) 80 Education P K F (59.43%) KPMG (33.39%) 83 Health activities PWC (79.67%) EY (4.80%) 92 Media - TV, Radio, Sport Stadiums 53 3, KPMG (36.78%) DT (18.61%) Total (Mean) 907 GBP125,908 (45.14%) (22.32%) Accounting Firm Definitions: DT = Deloitte & Touche EY = Ernst & Young KPMG=KPMG PWC = Pricewaterhousecoopers, P K F = Pannell Kerr Forster MS = Moore Stephens AK = A K & Co

21 23 Table 3 Descriptive Statistics for Sample Data Total Sample (n = 907) Big 4 Sample (n = 631) Mean Median Std. Dev. Q1 Q3 Mean Median Std. Dev. Q1 Q3 LAF LTA SQRTSUBS CATA QUICK DE ROI FOREIGN OPINION BUSY LOSS LNAF LONDON Second Tier Sample (n = 114) Third Tier Sample (n = 162) Mean Median Std. Dev. Q1 Q3 Mean Median Std. Dev. Q1 Q3 LAF LTA SQRTSUBS CATA QUICK DE ROI FOREIGN OPINION BUSY LOSS LNAF LONDON Variable Definitions: LAF = natural log of audit fees in thousands of GB Pounds LTA = natural log of total assets in thousands of GB Pounds SQRTSUBS = square root of total subsidiaries CATA = ratio of current assets to total assets QUICK = ratio of current assets (less inventories) to current liabilities DE = ratio of long-term debt to total assets ROI = ratio of earnings before interest and tax to total assets FOREIGN = proportion of total sales from foreign operations OPINION = indicator variable, 1 = qualified audit report BUSY = indicator variable, 1 = December 31 st or March 31 st year end LOSS = indicator variable, 1=loss in any of the past three years LNAF = natural log of nonaudit fees (in GP Pounds) paid to the auditor LONDON = indicator variable if observation is a London-based company.

22 Table 4 Effects of Big 4 National and City-Specific Industry Leadership on Audit Fee Premia Relative to Big 4 Nonleaders 24 Model 1 Model 2 Model 3 Exp. Sign Estimate t Sig. Estimate t Sig. Estimate t Sig. Control Variables: (Constant) +/ LTA SQRTSUBS CATA QUICK DE ROI FOREIGN OPINION BUSY LOSS LNAF LONDON Experimental variables: National Leader (n=170) City-Specific Industry Leader (n=237) Joint National and City-Specific Industry Leader (n=118) City-Specific Industry Leader but not National Leader (n=119) National Industry Leader but not City-Specific Industry Leader (n=52) F statistic (p-value) (<0.0001) (<0.0001) (<0.0001) Adjusted R Sample size Big 4 sample only. All p-values are two-tail. Industry fixed effects variables not reported for brevity, and t-statistics and significance levels are calculated using White (1980) corrected standard errors. Control variables : LAF = natural log of audit fees in thousands of GB Pounds, LTA = natural log of total assets in thousands of GB Pounds, SQRTSUBS = square root of total subsidiaries, CATA = ratio of current assets to total assets, QUICK = ratio of current assets (less inventories) to current liabilities, DE = ratio of long-term debt to total assets, ROI = ratio of earnings before interest and tax to total assets, FOREIGN = proportion of total sales from foreign operations, OPINION = indicator variable, 1 = qualified audit report, BUSY = indicator variable, 1 = December 31 st or March 31 st year end, LOSS = indicator variable, 1=loss in any of the past three years, LNAF = natural log of nonaudit fees (in thousands of GP Pounds) paid to the auditor, LONDON = indicator variable if observation is a London-based com pany.

23 25 Table 5 The Effect of Big 4 Industry Leadership on Audit Fee Premia Relative to Non-Big 4 Auditors Model 1 Model 2 Model 3 Estimate t Sig. Estimate t Sig. Estimate t Sig. Control Variables: (Constant) LTA SQRTSUBS CATA QUICK DE ROI FOREIGN OPINION BUSY LOSS LNAF LONDON Experimental variables: Big 4 City-Specific Industry Leaders (n=237) All Other Big 4 Auditors (n=269) Second-Tier Auditors (n=78) F statistic (p-value) (.<001) (.<001) (.<001) Adjusted R Sample size All p-values are two-tail. Industry fixed effects variables not reported for brevity, and t-statistics and significance levels are calculated using White (1980) corrected standard errors. Control variables : LAF = natural log of audit fees in thousands of GB Pounds, LTA = natural log of total assets in thousands of GB Pounds, SQRTSUBS = square root of total subsidiaries, CATA = ratio of current assets to total assets, QUICK = ratio of current assets (less inventories) to current liabilities, DE = ratio of long-term debt to total assets, ROI = ratio of earnings before interest and tax to total assets, FOREIGN = proportion of total sales from foreign operations, OPINION = indicator variable, 1 = qualified audit report, BUSY = indicator variable, 1 = December 31 st or March 31 st year end, LOSS = indicator variable, 1=loss in any of the past three years, LNAF = natural log of nonaudit fees (in thousands of GP Pounds) paid to the auditor, LONDON = indicator variable if observation is a London-based company.