Homework answers / question archive / Introductory sections:Briefly discuss the main issue(s) or topic(s)of the article/case

Introductory sections:Briefly discuss the main issue(s) or topic(s)of the article/case

Business

Share With

Introductory sections:Briefly discuss the main issue(s) or topic(s)of the article/case.Do yout hink the issues(s) are interesting? Discuss why or why not. You will find that author(s) of each paper discusses about prior studies or research on similar or related topic/issue(s). Discuss related issued and findings documented in the extant literature. Also discuss how this paper’s findings are linked with those of prior studies.Discuss how the paper contribute to the body of knowledge.Accounting Horizons Vol. 27, No. 4 2013 pp. 667–691 American Accounting Association DOI: 10.2308/acch-50513 Auditor Industry Expertise and Cost of Equity Jagan Krishnan, Chan Li, and Qian Wang SYNOPSIS: We examine the association between auditor industry expertise and clients’ cost of equity. Prior research suggests that industry experts are associated with higher earnings quality than non-experts. If such improved earnings quality were recognized by investors, we would expect it to be reflected in a lower cost of equity. Following recent research in this area, we distinguish between national-only, city-only, and joint citynational industry-expert auditors. Our results suggest that clients audited by city-only or joint city-national industry experts have a lower cost of equity. We also examine whether changing from non-expert (expert) to expert (non-expert) auditors result in a decrease (increase) in cost of equity. We find that when firms change from non-experts to city-only or joint city-national experts, their cost of equity is significantly decreased. Keywords: audit quality; cost of equity; auditor expertise; auditor change. INTRODUCTION T he purpose of this paper is to examine (1) the association between auditors’ industry expertise and their clients’ cost of equity capital, and (2) the change in cost of equity capital when companies switch from industry non-expert (expert) auditors to industry-expert (non-expert) auditors. Previous research has established a positive association between auditor industry expertise and earnings quality (Balsam et al. 2003; Krishnan 2003; Reichelt and Wang 2010). If such improved earnings quality were recognized by investors, we would expect it to be reflected in a lower cost of equity. In addition, prior work (Knechel et al. 2007) has documented a positive market reaction to auditor changes when the successor auditor is an industry expert and a negative market reaction when the successor auditor is not an industry expert. Thus, we would expect a decrease (an increase) in the cost of equity capital for companies that switch from an industry non-expert (industry expert) to an industry-expert (industry non-expert) auditor. We measure auditor industry expertise at both the city and national levels. Auditor industry expertise derives from the knowledge base and human capital invested in accounting professionals Jagan Krishnan is a Professor at Temple University, Chan Li is an Assistant Professor at the University of Pittsburgh, and Qian Wang is an Assistant Professor at Iowa State University. We thank two anonymous reviewers, Dana Hermanson (editor), Tom Adams, Thomas Omer, Mike Ettredge, Jere Francis, Jayanthi Krishnan, and workshop participants at the 2007 AAA Annual Meeting, The University of Kansas, and University of Pittsburgh for their helpful comments. Professor Krishnan acknowledges research support from Temple University Fox School of Business and Management’s Merves Research Fellowship. Submitted: October 2011 Accepted: April 2013 Published Online: May 2013 Corresponding author: Jagan Krishnan Email: krish@temple.edu 667 668 Krishnan, Li, and Wang (Francis et al. 2005). Therefore, such expertise can be obtained at both the local office level (‘‘city-level’’ industry expertise) and, by transfer of expertise to other offices, at the national level (‘‘national-level’’ industry expertise). Whereas earlier studies indicated that national-level expertise is important, recent work that examines both kinds of expertise suggests that city-level expertise, especially joint city-national expertise, may dominate. For example, Craswell et al. (1995) find that audit fees are positively associated with national-level industry expertise. However, based on a more disaggregated classification of expertise, Francis et al. (2005) find a positive association between audit fees and city-level expertise, either alone or jointly with national-level expertise. Likewise, while Balsam et al. (2003) and Krishnan (2003) show that national-level expertise is associated with lower discretionary accruals, Reichelt and Wang (2010) find that discretionary accruals are lower for clients of joint city-national experts and, much less consistently in their tests, for clients of city-only industry experts. We therefore examine whether a negative association between firms’ cost of equity capital and auditor industry expertise exists only when their auditors have city-level expertise, either alone or jointly with national expertise. We use a sample of publicly traded companies employing Big N auditors. Our sample period covers the years 2000 through 2008. For each of our cost of equity measures, we find that city-level auditor industry expertise, either alone or jointly with national expertise, is negatively and significantly associated with cost of equity. However, national-only expertise is negatively (and weakly) associated with cost of equity only for one of the three cost-of-equity measures. For the change analyses, we find that when firms change from non-experts to city-only ( joint) experts, their cost of equity capital is significantly decreased by 320 (200) basis points. We also find mild evidence that firms switching from joint city-national experts to non-experts experience an increase in the cost of equity. If industry-expert auditors contribute to lower cost of equity for their clients, we would expect firms to change auditors prior to issuing new equity. Therefore, we examine whether firms switch to industry-expert auditors before issuing new equity. We find firms that change their auditors from industry non-experts to industry experts are more likely to issue equity in the two years subsequent to the switch compared to firms that switch to other non-expert auditors. In additional analyses, we examine the associations for specific subsets of our sample. We find that the negative association between city-only and/or joint city-national expertise and cost of equity holds for clients that do not have international operations and for clients with low institutional ownership. Further, we find that small clients of national-only experts and joint city-national experts are associated with lower cost of equity capital than small clients of non-experts. In contrast, there is no difference in the cost of equity capital between large clients of industry experts and non-experts, suggesting that investors value auditors’ industry expertise more for small firms whose accounting information may be less transparent compared to large firms (Lang and Lundholm 1993). Taking all of our tests in combination, we conclude that clients of city-only and joint city-national industry-expert auditors enjoy lower cost of equity capital. Our study contributes to the literature on audit-quality effects of auditor industry expertise by examining investor perceptions as measured by the cost of equity. Although prior studies have extensively documented that industry-expert auditors have positive effects on earnings quality, whether shareholders recognize such positive effects has not been examined. Our results are consistent with the conjecture that city-level industry-expert auditors are associated with higher earnings quality, which lowers information risk, resulting in lower cost of equity. However, we acknowledge that we do not establish the lowering of information risk as the direct channel through which industry expertise affects the cost of equity capital. Our findings indicate that there are investor-perceived benefits to city-level auditor industry expertise in terms of lower cost of equity capital. Accounting Horizons December 2013 Auditor Industry Expertise and Cost of Equity 669 The results of this study should be of interest to policymakers. An August 2011 Public Company Accounting Oversight Board (PCAOB 2011) concept release asked for comments on the advisability of using mandatory auditor rotation to enhance auditor independence. It aroused concerns from commenters that such rotation could adversely affect clients benefiting from auditor industry specialization. Although these comments did not elaborate on the nature of the benefits from industry specialization, our evidence that equity investors value industry-expert auditors is relevant to the debate about the desirability of mandatory auditor rotation.1 While our focus is on the impact of auditor industry specialization on client’s cost of equity, Li et al. (2010) examine the association between auditor industry specialization and cost of long-term debt financing. As we discuss later, the two forms of financing differ and can have differential effects on cost of capital. In comparison to debt holders, equity holders are exposed to higher risk in case of firm bankruptcy. Li et al. (2010) argue that the information environment in the bond market is characterized by many information intermediaries who closely monitor the companies’ activities, making them potentially sensitive to the effects of auditor industry specialization. By contrast, the effect of industry specialization on cost of equity may depend on the characteristics of investors.2 Small and/or unsophisticated investors have fewer incentives to monitor companies’ activities and are less likely to recognize the impact of industry specialization on earnings quality. Therefore, ex ante, it is not clear whether the results for the bond market documented in Li et al. (2010) would hold in the equity market. Moreover, in addition to examining the association between industry expertise and cost of equity, we examine whether there is a decrease (increase) in cost of equity when companies switch from industry non-expert (expert) to industry-expert (non-expert) auditors. Thus, our study complements Li et al.’s (2010) study by focusing on the implications of audit quality for equity shareholders. The rest of the paper is organized as follows. In the next section, we review related research and develop our hypotheses. This is followed by a section that describes the sample and research methods. The final three sections contain the empirical findings, additional analyses, and our conclusions. PRIOR RESEARCH AND HYPOTHESES DEVELOPMENT Audit Expertise and Cost of Equity Independent audits provide a mechanism to reduce agency costs associated with information asymmetry between managers and investors (Jensen and Meckling 1976), thereby reducing investor uncertainty and lowering perceived risk (e.g., Watts and Zimmerman 1986). In other words, auditors provide assurance to financial reports, thus reducing information risk. Equity investors cannot diversify information risk; therefore, a higher-quality audit can compensate them by reducing information risk. Alternatively, with low assurance of information, investors would be skeptical of information provided by managers, demanding a higher rate of return to compensate for the risk of expropriation of their capital by managers. Therefore, the higher the audit quality (degree of assurance), the lower the rate of return that equity holders would require. 1 2 For example, the Royal Gold Corporation states that mandatory rotation would ‘‘make it more difficult for audit firms to build expertise in specialized areas of accounting, such as the mining industry’’ (PCAOB 2011). Likewise the American Institute of Certified Public Accountants (AICPA) is concerned that audit firm rotation may limit ‘‘institutional knowledge and industry specialization’’ that ‘‘is crucial to a high-quality audit’’ (Journal of Accountancy 2011). Easton (2007, 246) suggests that financial analysts use accounting estimates (among other sources) to base their recommendations. Moreover, Behn et al. (2008) provide evidence that the forecast accuracy of analysts is positively associated with non-Big N industry specialization. Thus, industry specialization is relevant to intermediaries and, likely, equity investors. Accounting Horizons December 2013 Krishnan, Li, and Wang 670 Previous work suggests that the benefit from auditing can be expected to vary with the quality of the auditor. Studies using auditor brand name (Big N versus non-Big N) as a proxy for audit quality have generally documented lower IPO underpricing, larger earnings response coefficients (ERCs), lower cost of debt, and lower cost of equity capital for clients of Big N auditors (e.g., Teoh and Wong 1993; Beatty 1989; Mansi et al. 2004; Khurana and Raman 2004). Another stream of research suggests that, in addition to Big N auditors, industry-expert auditors provide a higher level of assurance than non-experts. For example, Craswell et al. (1995) and Francis et al. (2005) document that industry-expert auditors charge audit fee premiums, indicating higher audit quality. Owhoso et al. (2002) show that auditors are better able to detect errors when they perform audit tasks in industries where they have expertise than in industries outside their expertise. Studies examining auditor industry expertise and earnings quality find that firms audited by industry experts have lower absolute discretionary accruals (Balsam et al. 2003; Krishnan 2003; Reichelt and Wang 2010) and larger ERCs (Balsam et al. 2003). Dunn and Mayhew (2004) find that analysts perceive disclosure quality as being higher for companies audited by industry experts. Behn et al. (2008) document that non-Big 5 industry experts are associated with ‘‘higher forecast accuracy and less forecast dispersion’’ than non-Big 5 non-experts. Finally, Knechel et al. (2007) find that the market reacts positively (negatively) when companies change from industry non-expert (industry expert) to industry-expert (non-expert) auditors. The positive (negative) market reaction could either be because of an increase (decrease) in expected future cash flows or because of a decrease (increase) in discount rate. We focus on the discount rate or cost of equity capital and argue that higher audit quality (through industry expertise) can reduce the cost of equity. We extend previous work to see if clients of industry-expert auditors are associated with lower cost of equity than non-expert auditors. As Boone et al. (2008) argue, higher audit quality leads to more transparent and reliable financial information and improves precision in firms’ earnings (Fortin and Pittman 2007), which reduces information risk, leading to lower cost of equity capital (Jensen and Meckling 1976; Watts and Zimmerman 1986; Lambert et al. 2007). That is, the discount factor that investors apply to future cash flows is reduced because of diminished information risk as a result of a high-quality audit. If investors view industry experts as providing higher-quality audits, we would expect firms audited by experts to have lower cost of equity. Hence, we hypothesize: H1: Cost of equity for clients of industry-expert auditors is lower than that for clients of nonindustry-expert auditors. We test H1 using the city-national industry-expert framework. As we discussed earlier, recent studies find that city-level experts, especially joint city-national experts have significant impact on audit fees and earnings quality. For example, Francis et al. (2005) document the existence of an audit price premium for both city-only and joint city-national industry experts but find no fee premium for national-only experts. Reichelt and Wang (2010) find that audit quality is higher when the auditor is a joint city-national industry expert. Specifically, they report that when the auditor is a joint citynational industry expert, their clients have lower abnormal accruals, their earnings are less likely to meet or beat analysts’ forecasts by one penny, and they are more likely to receive going-concern opinions. Reichelt and Wang (2010) also find, less consistently than their results for joint expertise, that city-only industry expertise is associated with lower abnormal accruals. We examine whether the impact of auditor industry expertise on clients’ cost of equity differs between national-, city-, and joint city-national-level industry experts. Auditor Industry Expertise, Auditor Changes, and Cost of Equity Previous studies have tested the brand-name hypothesis by examining the market reaction when a client changes from a non-brand-name auditor (non-Big N) to a brand-name auditor (Big N) and vice Accounting Horizons December 2013 Auditor Industry Expertise and Cost of Equity 671 versa. If brand-name auditors were perceived to be providers of higher-quality audits, then auditor changes from non-Big N to Big N auditors would trigger positive abnormal returns. Conversely, auditor changes from Big N to non-Big N auditors would result in negative abnormal returns. Early studies did not find evidence consistent with the brand-name hypothesis. Fried and Schiff (1981), using a sample of 48 companies that switched auditors during the 1972 through 1975 period, document a negative but insignificant market reaction to the switch from non-Big 8 to Big 8 audit firms. Similarly, Nichols and Smith (1983) and Klock (1994) report no association between abnormal returns and a switch from a non-Big 8 auditor to a Big 8 auditor. Knechel et al. (2007) extend the above studies by examining the market reaction to auditor changes between industry-expert and non-expert audit firms. They find that firms switching from non-expert auditors to expert auditors experience positive abnormal stock returns, while those switching from expert auditors to non-expert auditors experience negative abnormal stock returns around the date of the change in auditors.3,4 Their results are consistent with industry-expert auditors providing higher audit quality. Thus, we pose the following hypotheses: H2a: Cost of equity capital for clients that change from non-industry-expert auditors to industry-expert auditors would decrease following the switch. H2b: Cost of equity capital for clients that change from industry-expert auditors to nonindustry-expert auditors would increase following the switch. Related Work: Auditor Industry Expertise and Cost of Debt Li et al. (2010) examine the association between auditor industry expertise and cost of debt, another component (in addition to the cost of equity) of a firm’s overall cost of capital. We complement Li et al. (2010) by examining the association between industry expertise and cost of equity.5 While companies rely on both debt and equity capital, there are well-known differences between the two forms of capital (Mansi et al. 2004). Debt is a cheaper source of capital than equity, and the interest on debt is tax deductible. Interest on debt must be paid to debt holders, regardless of whether the company is profitable. By contrast, equity holders need not be paid dividends if the company’s financial condition is not strong. Moreover, debt holders have priority over equity holders in case of bankruptcy or liquidation, and the latter only have a residual claim on the company’s assets. Consequently, the risk borne by equity holders is much higher than that borne by debt holders. Thus, audit quality is important to both equity holders and debt holders, but for different reasons. Debt holders are concerned about the downside of not being paid, and do not profit when firms’ earnings exceed debt obligations considerably. High-quality auditors, such as industry specialists, enable equity investors to more accurately price securities for the portion of expected payoffs that exceeds debt obligations. In addition to the inherent differences in the nature of debt and equity capital, the information environment encountered by providers of each type of capital may also be different. Debt providers typically engage in significant monitoring of their debtors and may be sensitive to the quality of their debtors’ audits. By contrast, equity investors are sometimes a more diverse group, and smaller, 3 4 5 Unlike Knechel et al. (2007), who document negative market reaction to firms that downgrade auditors, Behn et al. (2008, 346, fn. 26) do not find an association between downgrades and forecast dispersion. Mansi et al. (2004) argue that ‘‘auditor changes are, however, often associated with other confounding events that can influence stock prices, which can make clean inferences difficult.’’ Several sources provide excellent discussion of these differences and of factors that influence the choice of capital structure (e.g., Brealey and Myers 1991; Hackel 2011). Accounting Horizons December 2013 Krishnan, Li, and Wang 672 unsophisticated investors are likely to be less sensitive to audit quality improvements stemming from auditors’ industry expertise. One would expect that higher-quality audits would have a negative impact on both cost of equity and cost of debt. However, given the differences between equity and debt (mentioned above), it is not clear, a priori, if the magnitude of each effect is the same.6 If auditor industry expertise affects the cost of equity differently than cost of debt, the overall benefit of industry-expert auditors can be different for firms with different capital structures. Our study provides evidence in this regard. RESEARCH DESIGN AND DATA The Ex Ante Cost of Equity Capital Following prior research (e.g., Khurana and Raman 2004; Hail and Leuz 2006), we use three models to compute the ex ante cost of equity capital: (1) the Easton (2004) model (RPEG); (2) the modified Easton (2004) model (RMPEG); and (3) the Ohlson and Juettner-Nauroth (2005) model (ROJN ):7 r?????????????????????????????? epstþ2 epstþ1 ; ð1Þ RPEG ¼ pt pt ¼ ROJN epstþ2 þ RMPEG dps0 epstþ1 ; R2MPEG ð2Þ s?????????????????????????????????????????????????????????????????? ? epstþ1 2 ¼aþ a þ g2 ðrf 0:03Þ ; pt ð3Þ where: epstþ1 ¼ the one-year-ahead-mean analysts’ earnings forecast per share; epstþ2 ¼ the two-year-ahead mean analysts’ earnings forecast per share;8 pt ¼ price per share of common stock in June of year t; rf ¼ risk-free rate equal to the yield on a ten-year Treasury note in June of year t; and dps0 ¼ dividends per share paid during year t1; dps0 a ¼ 0:5 ðrf 0:03Þ þ ; and pt g2 ¼ 6 7 8 ðepstþ2 epstþ1 Þ : epstþ1 In their discussion of value relevance, Holthausen and Watts (2001, 26) point out that lenders and investors may be interested in different numbers: ‘‘It is not apparent that the relevance of a given number would be the same for equity investors and lenders.’’ Botosan and Plumlee (2005) compare the consistency and predictability of different measures of ex ante cost of equity, and conclude that the Easton (2004) model outperforms other models. We also use long-run earnings forecasts (epstþ5 and epstþ4 ) in the Easton (2004) model (our Model 1) to construct RPEG. The results show that the coefficients on NATIONALONLY and CITYONLY are negative and significant, while the coefficient on JOINTEXPERT is negative, but not significant. Accounting Horizons December 2013 Auditor Industry Expertise and Cost of Equity 673 Auditor Industry Expertise The measurement of auditor industry expertise is based on audit firms’ market share in audit fees within industry groups classified by two-digit SIC codes (Ferguson et al. 2003; Francis et al. 2005). Auditors’ national industry expertise is measured using national market share rankings in industry groups. An audit firm is defined as a national industry expert if its market share is ranked number one at the national level. Auditors’ city-level industry expertise is measured using cityspecific market share rankings within industry groups. The client’s city is defined as the city where the auditor’s report is issued. An auditor is defined as a city industry expert if its market share in an industry is ranked number one at the city level. An auditor is defined as a joint city-national expert if its industry share is ranked number one at both national and city levels. We determine auditors’ national and city expertise in each industry for each of our sample years. Since the majority of national- or city-level industry-expert auditors (i.e., 99 percent at the national level and 96 percent at the city level) are Big N auditors, we focus our analyses on Big N auditors.9 Model We estimate the following regression model: R ¼ b0 þ b1 NATIONALONLY þ b2 CITYONLY þ b3 JOINTEXPERT þ b4 LNMV þ b5 LEVERAGE þ b6 BVMV þ b7 BETA þ b8 EPSSTDDEV þ b9 IDIORISK þ b10 RETURN þ b11 DEPENDENCE þ b12 TENURE þ b13 INSTOWN þ b14 EXGRW þ Year Dummies þ Industry Dummies; ð4Þ where: R ¼ client-specific ex ante cost of equity capital (RPEG/RMPEG/ROJN ) defined in Model (1)– Model (3); NATIONALONLY ¼ 1 if an auditor’s industry share is ranked one at the national level but not at the city level, 0 otherwise; CITYONLY ¼ 1 if an auditor’s industry share is ranked one at the city level but not at the national level, 0 otherwise; JOINTEXPERT ¼ 1 if an auditor’s industry share is ranked one at both national and city levels, 0 otherwise; LNMV ¼ natural logarithm of total market value; LEVERAGE ¼ total liabilities divided by total assets; BVMV ¼ book value of equity divided by market value of equity; BETA ¼ stock beta calculated over 36 months preceding the measurement of R; EPSSTDDEV ¼ analysts’ forecast dispersion measured by the standard deviation of analysts’ earnings forecasts; IDIORISK ¼ standard deviation of the residuals of the market model regression using monthly returns over the 36 months preceding the measurement of R; RETURN ¼ one-year stock return calculated over the 12-month period preceding the measurement of R; DEPENDENCE ¼ ratio of client’s total fees to the sum of total fees received by the practice firm’s office that conducted the audit; 9 The results are similar when we include non-Big N auditors’ clients. Accounting Horizons December 2013 Krishnan, Li, and Wang 674 TENURE ¼ number of years the incumbent auditor has audited the client; INSTOWN ¼ 1 if the market value of institutional shares is above median, 0 otherwise; and EXGRW ¼ difference between the mean analysts’ five- and four-year-ahead earnings forecasts scaled by the four-year-ahead earnings forecast. To test the effects of the three auditor industry-expertise variables, we employ a joint city-national framework using the three variables (NATIONALONLY, CITYONLY, and JOINTEXPERT) defined above. The base comparison group comprises clients of all non-expert auditors. Since the ex ante cost of equity capital is a measure of expected risk premium, we control for the following risk factors identified in prior studies (e.g., Gebhardt et al. 2001; Botosan and Plumlee 2005; Hail and Leuz 2006; Khurana and Raman 2004, 2006): client size, leverage, bookto-market ratio, beta, analysts’ forecast dispersion, idiosyncratic risk, recent one-year return, and earnings growth. We expect negative associations between cost of equity and size, and between cost of equity and stock returns. We expect positive associations between cost of equity and the other risk factors. We also control for two auditor-related factors. We control for auditor independence, because Khurana and Raman (2006) find that clients paying higher fees (a proxy for auditor independence) have higher cost of equity. We control for auditor tenure because Boone et al. (2008) find that auditor tenure is significantly associated with cost of equity.10 In addition, we include institutional ownership to control for the impact of active monitoring on the cost of equity (McConnell and Servaes 1990; Bushee 1998). Because industries differ on risk factors, we control for industry impacts by including Fama-French industry groups in our model. Finally, because each firm may exist in multiple years in our sample, we control for year effects and report t-statistics based on robust standard errors to control for firm clustering effects (Petersen 2009). Sample Selection Table 1 details the sample selection process. We start with companies covered by the Audit Analytics database for the years 2000 through 2008, for which information on auditor identity and the city where the audit opinion rendered is available. There are 98,028 company-year observations in our initial sample. We delete foreign companies, companies with foreign auditors, and companies in the financial services industry (SIC code 6000-6999). This process results in 57,324 company-year observations, which are used to calculate auditor industry market shares. Next, we compute measures of cost of equity. This computation requires availability of (1) one- and two-year-ahead analyst forecasts with positive growth in earnings forecasts on the I/B/E/ S database, and (2) closing prices at the end of fiscal years on the CRSP database. We also require the sample to have necessary financial data. These requirements reduce our sample to 15,362 company-year observations. Finally, we exclude 3,357 clients of non-Big N companies, resulting in a final sample of 12,005 observations. 10 Francis and Yu (2009) find auditor office size to be positively associated with audit quality. Thus, we control for the office size, which is measured as the natural log of aggregate audit fees for an audit office. Because the officesize variable is significantly correlated (correlation ¼ 0.49) with our auditor independence variable (DEPENDENCE), we exclude DEPENDENCE when we include office size. The results show that the coefficient on office size is positive but not significant when the dependent variable is RPEG and RMPEG, and is marginally significant when the dependent variable is ROJN. Joint city-national expert continues to be significantly associated with all three measures of cost of equity. The coefficient on NATIONALONLY is significant only for RPEG and the coefficient on CITYONLY is insignificant for all measures of cost of equity. Accounting Horizons December 2013 Auditor Industry Expertise and Cost of Equity 675 TABLE 1 Sample Selection Description Number of Obs. Companies in Audit Analytics (2000–2008) Less: Foreign companies, clients of foreign auditors, and companies with SIC codes 6000-6999 Sample used to compute auditor industry expertise Less: Companies not covered by I/B/E/S, or with negative earnings forecast growth or missing other financial data Subsample Less: Companies with non-Big N auditors Final sample (Big N auditors)a a 98,028 (40,704) 57,324 (41,962) 15,362 (3,357) 12,005 Big N auditors include Arthur Andersen LLP for the years 2000 through 2002. EMPIRICAL RESULTS Industry Expertise Measures As discussed, we construct our industry expertise measures for the population of firms with available data. However, we test our hypotheses using a smaller sample for which all variables are available. Recall that we define experts as auditors that are ranked number one at the nationalindustry or city-industry levels. In Table 2, we present the distribution of auditors ranked both number one and number two at the national (Panel A) and city (Panel B) levels during 2000 through 2008. In Panel A, PricewaterhouseCoopers has the highest number of industries in which they are experts across the nine years, followed by Ernst & Young, Deloitte & Touche, and KPMG. In Panel B, at the city level, Ernst & Young has the highest number of industries in which they are experts across the nine years, followed by PricewaterhouseCoopers, Deloitte & Touche, and KPMG. Both panels also indicate that the industry share of auditors ranked two is significantly smaller than that of auditors ranked one in each year. Over our sample period, the top ranked auditor has an average market share of 42.6 percent at the national level and 84.9 percent at the city level. The second ranked auditor has an average market share of 23.2 percent at the national level and 8.7 percent at the city level. We compared the distributions of national-level and city-level industry experts in Table 2 (based on our final sample) with those for the entire population of companies. The results (untabulated) are very similar, suggesting that the reduced sample does not change the industry distribution of auditor expertise. Table 3 reports the means for clients of non-experts, national-only experts, city-only experts, and joint city-national experts, and examines differences between them. Note that the three groups of experts are mutually exclusive: national experts are not city experts, city experts are not national experts, and joint city-national experts are not included in the first two groups. To mitigate the effect of outliers, all continuous variables are winsorized at 1 percent and 99 percent. Compared to non-experts, clients of national-only experts, city-only experts, and joint city-national experts all experience lower cost of equity for all three cost of equity measures (RPEG, RMPEG, and ROJN ), which provides initial support for our first hypothesis (H1). When we compare cost of equity among different types of industry experts, we find there is no difference between national-only and city-only experts (see Column 8). However, clients audited by joint city-national experts have Accounting Horizons December 2013 Krishnan, Li, and Wang 676 TABLE 2 Auditor Industry Expertise (n ¼ 12,005) Panel A: National Industry Expertise Rank 1 Rank 2 Number of Industry-Years for Which the Auditor Is Ranked 1 or 2 in Market Share Arthur Andersen LLP 23 Deloitte & Touche LLP 93 Ernst & Young LLP 125 KPMG LLP 53 PricewaterhouseCoopers LLP 149 Average Expert Market Share 42.6% Total No. of Industry-Years 15 72 125 86 105 23.2% 443 Panel B: City Industry Expertise Rank 1 Rank 2 Number of City-Industry-Years for Which the Auditor Is Ranked 1 or 2 in Market Share Arthur Andersen LLP 239 Deloitte & Touche LLP 1,164 Ernst & Young LLP 1,760 KPMG LLP 1,137 PricewaterhouseCoopers LLP 1,591 Average Expert Market Share 84.9% Total No. of City-Years Total No. of City-Industry-Years 102 377 537 436 443 8.7% 1,052 5,891 significantly lower cost of equity compared to those audited by either national-only (see Column 9) or city-only experts (see Column 10). This provides univariate evidence that investors value joint city-national industry expertise more than either national-only or city-only expertise. Compared to clients of non-experts, clients of national-only experts have higher leverage, larger beta, lower analysts’ earnings forecast dispersions, and lower forecasted earnings growth (see Column 5). Compared to clients of non-experts, clients of city-only experts are larger, have higher leverage, higher book-to-market ratios, lower beta, larger analysts’ earnings forecast dispersions, lower idiosyncratic risk, and lower forecasted earnings growth (see Column 6). They are also likely to contribute more revenues to auditors’ local offices, have longer relationships with their auditors, and have higher institutional ownership. The univariate comparisons between clients of joint citynational experts and non-experts (see Column 7) are similar to those between clients of city-only experts and non-experts, except that compared with clients of non-experts, clients of joint citynational experts have lower returns, and have similar book-to-market ratios. Clients of national-only experts differ from those of city-only experts (see Column 8) in all respects except book-to-market ratio, returns, institutional ownership, and forecasted earnings growth. Clients of national-only experts also differ from those of joint city-national experts (see Column 9) in all respects other than book-to-market ratio and institutional ownership. Similarly, clients of city-only experts differ from those of joint city-national experts (see Column 10) in all respects except analysts’ forecast dispersions, auditor independence, and institutional ownership. Accounting Horizons December 2013 Accounting Horizons December 2013 0.118 0.122 0.116 3434.030 6.603 0.452 0.483 0.982 0.052 0.180 0.149 0.066 11.009 0.493 0.214 0.114 0.119 0.112 3746.960 6.605 0.478 0.507 1.051 0.047 0.174 0.167 0.061 10.881 0.496 0.183 NATIONAL ONLY (n ¼ 904) (2) Mean 0.113 0.119 0.111 5448.300 7.049 0.507 0.505 0.935 0.058 0.159 0.150 0.151 11.648 0.519 0.182 CITY ONLY (n ¼ 5,186) (3) Mean 0.109 0.116 0.107 8835.090 7.439 0.542 0.487 0.806 0.058 0.143 0.120 0.158 12.715 0.519 0.167 4.214*** 2.741*** 3.991*** 6.130*** 11.622*** 10.597*** 2.292** 2.032** 3.017*** 5.637*** 0.082 18.105*** 3.102*** 2.262** 6.722*** Diff. between (1) and (3) (6) t-stat. 7.221*** 5.084*** 6.241*** 10.237*** 18.924*** 15.533*** 0.374 7.133*** 2.787*** 9.587*** 2.078** 18.049*** 7.103*** 2.011** 9.000*** Diff. between (1) and (4) (7) t-stat. Diff. between (2) and (4) (9) t-stat. 2.919*** 1.825* 2.516** 5.704*** 12.495*** 7.524*** 1.345 7.025*** 3.690*** 5.858*** 2.443** 11.492*** 5.042*** 1.249 2.391** Diff. between (2) and (3) (8) t-stat. 0.422 0.060 0.619 3.183*** 7.235*** 3.511*** 0.135 3.140*** 3.721*** 2.629*** 0.915 11.209*** 2.332** 1.305 0.257 Variable Definitions: RPEG ¼ client-specific ex ante cost of equity capital based on the Easton (2004) model; RMPEG ¼ client-specific ex ante cost of equity capital based on the modified Easton (2004) model; ROJN ¼ client-specific ex ante cost of equity capital based on the Ohlson and Juettner-Nauroth (2005) model; NATIONALONLY ¼ 1 if an auditor’s industry share is ranked 1 at the national level but not 1 at the city level, 0 otherwise; CITYONLY ¼ 1 if an auditor’s industry share is ranked 1 at the city level but not 1 at the national level, 0 otherwise; JOINTEXPERT ¼ 1 if an auditor’s industry share is ranked 1 at both the national level and the city level, 0 otherwise; MV ¼ total market value; LNMV ¼ natural logarithm of total market value; LEVERAGE ¼ total liabilities divided by total assets; BVMV ¼ book value of equity divided by market value of equity; BETA ¼ stock beta calculated over 36 months ending in the month of the fiscal year-end; EPSSTDDEV ¼ analysts’ forecast dispersion measured by the standard deviation of analysts’ forecasts; IDIORISK ¼ standard deviation of the residuals of the market model regression using monthly returns over the 36 months preceding the measurement of R; RETURN ¼ most recent one-year stock return calculated over the 12-month period preceding the measurement of R; DEPENDENCE ¼ ratio of client’s total fees to the sum of total fees received by the practice office that conducted the audit; TENURE ¼ number of years the incumbent auditor has audited the client; INSTOWN ¼ 1 if the percentage of institutional holding is above the median, 0 otherwise; and EXGRW ¼ difference between the mean analysts’ five- and four-year-ahead earnings forecasts scaled by the four-year-ahead earnings forecast. 2.103** 1.810* 1.952* 0.641 0.027 3.044*** 1.562 1.788* 1.720* 0.875 0.830 1.124 0.388 0.124 4.251*** Diff. between (1) and (2) (5) t-stat. Descriptive Statistics JOINT EXPERT (n ¼ 2,892) (4) Mean *, **, *** Significant at the 0.10, 0.05, and 0.01 level, respectively, (two-tailed). RPEG RMPEG ROJN MV LNMV LEVERAGE BVMV BETA EPSSTDDEV IDIORISK RETURN DEPENDENCE TENURE INSTOWN EXGRW NONEXPERT (n ¼ 3,023) (1) Mean TABLE 3 4.046*** 3.019*** 3.087*** 7.424*** 9.661*** 6.825*** 1.890* 5.673*** 0.129 4.915*** 2.571** 1.321 4.889*** 0.024 3.490*** Diff. between (3) and (4) (10) t-stat. Auditor Industry Expertise and Cost of Equity 677 Krishnan, Li, and Wang 678 TABLE 4 Pearson (above the Diagonal)/Spearman (below the Diagonal) Correlations Panel A: Correlations Variables RPEG to BVMV 1. 2. 3. RPEG RMPEG ROJN 0.881 (0.000) 0.945 (0.000) 0.846 (0.000) 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 0.906 (0.000) 0.911 (0.000) 0.007 (0.418) 0.003 (0.765) 0.059 (0.000) 0.386 (0.000) 0.010 (0.253) 0.239 (0.000) 0.254 (0.000) 0.076 (0.000) 0.350 (0.000) 0.090 (0.000) 0.076 (0.000) 0.085 (0.000) 0.044 (0.000) 0.272 (0.000) 0.855 (0.000) 0.001 (0.883) 0.006 (0.515) 0.045 (0.000) 0.371 (0.000) 0.110 (0.000) 0.295 (0.000) 0.157 (0.000) 0.094 (0.000) 0.257 (0.000) 0.014 (0.137) 0.039 (0.000) 0.070 (0.000) 0.082 (0.000) 0.164 (0.000) 0.011 (0.262) 0.004 (0.637) 0.053 (0.000) 0.358 (0.000) 0.027 (0.005) 0.222 (0.000) 0.228 (0.000) 0.102 (0.000) 0.316 (0.000) 0.103 (0.000) 0.076 (0.000) 0.092 (0.000) 0.127 (0.000) 0.240 (0.000) 4. NATIONALONLY 5. CITYONLY 6. JOINT EXPERT 0.003 (0.723) 0.004 (0.688) 0.004 (0.696) 0.003 (0.726) 0.005 (0.585) 0.008 (0.381) 0.249 (0.000) 0.054 (0.000) 0.042 (0.000) 0.046 (0.000) 0.161 (0.000) 0.491 (0.000) 0.249 (0.000) 0.161 (0.000) 0.065 (0.000) 0.029 (0.002) 0.006 (0.545) 0.040 (0.000) 0.029 (0.002) 0.021 (0.023) 0.010 (0.258) 0.108 (0.000) 0.021 (0.019) 0.009 (0.339) 0.015 (0.092) 0.491 (0.000) 0.025 (0.006) 0.029 (0.001) 0.010 (0.266) 0.010 (0.266) 0.033 (0.000) 0.020 (0.027) 0.025 (0.006) 0.153 (0.000) 0.006 (0.534) 0.014 (0.116) 0.023 (0.011) 0.142 (0.000) 0.115 (0.000) 0.004 (0.692) 0.079 (0.000) 0.039 (0.000) 0.095 (0.000) 0.019 (0.035) 0.083 (0.000) 0.055 (0.000) 0.010 (0.294) 0.097 (0.000) 7. LNMV 0.379 (0.000) 0.364 (0.000) 0.351 (0.000) 0.065 (0.000) 0.026 (0.004) 0.144 (0.000) 0.188 (0.000) 0.387 (0.000) 0.187 (0.000) 0.301 (0.000) 0.480 0.000 0.044 (0.000) 0.405 (0.000) 0.207 (0.000) 0.208 (0.000) 0.291 (0.000) 8. LEVERAGE 0.059 (0.000) 0.133 (0.000) 0.074 (0.000) 0.027 (0.003) 0.028 (0.002) 0.106 (0.000) 0.160 (0.000) 0.012 (0.175) 0.259 (0.000) 0.170 (0.000) 0.240 (0.000) 0.040 (0.000) 0.258 (0.000) 0.063 (0.000) 0.006 (0.509) 0.360 (0.000) 9. BVMV 0.261 (0.000) 0.302 (0.000) 0.247 (0.000) 0.008 (0.369) 0.021 (0.024) 0.012 (0.207) 0.398 (0.000) 0.038 (0.000) 0.029 (0.001) 0.033 (0.000) 0.005 (0.565) 0.106 (0.000) 0.006 (0.514) 0.018 (0.048) 0.084 (0.000) 0.165 (0.000) (continued on next page) To rule out the possibility that the differences do not drive the results, we include control variables in the multivariate models. Table 4 presents the Pearson and Spearman correlations among the dependent and independent variables. Among the test variables, only joint city-national expertise is significantly, negatively associated with all three measures of cost of equity. Firms with higher market value, lower leverage, lower book-to-market value, lower beta, lower analysts’ forecast dispersion, lower idiosyncratic risk, higher stock return, more fee dependence, longer auditor tenure, larger institutional ownership, Accounting Horizons December 2013 Auditor Industry Expertise and Cost of Equity 679 TABLE 4 (continued) Panel B: Correlations Variables BETA to EXGRW 10. BETA 1. 2. 3. 4. 5. 6. 7. 8. 9. 0.184 (0.000) 0.127 (0.000) 0.171 (0.000) 0.036 (0.000) 0.009 (0.320) 0.067 (0.000) 0.154 (0.000) 0.183 (0.000) 0.037 (0.000) 10. 11. 12. 13. 14. 15. 16. 17. 0.015 (0.103) 0.343 (0.000) 0.024 (0.010) 0.116 (0.000) 0.104 (0.000) 0.089 (0.000) 0.317 (0.000) 11. EPSSTDDEV 12. IDIORISK 0.116 (0.000) 0.105 (0.000) 0.151 (0.000) 0.030 (0.001) 0.023 (0.010) 0.017 (0.065) 0.176 (0.000) 0.142 (0.000) 0.028 (0.003) 0.006 (0.534) 0.231 (0.000) 0.198 (0.000) 0.206 (0.000) 0.024 (0.010) 0.013 (0.164) 0.069 (0.000) 0.426 (0.000) 0.094 (0.000) 0.133 (0.000) 0.161 (0.000) 0.111 (0.000) 0.214 (0.000) 0.027 (0.003) 0.172 (0.000) 0.072 (0.000) 0.192 (0.000) 0.210 (0.000) 0.067 (0.000) 0.215 (0.000) 0.214 (0.000) 0.179 (0.000) 0.449 (0.000) 13. RETURN 0.052 (0.000) 0.032 (0.000) 0.054 (0.000) 0.013 (0.167) 0.010 (0.263) 0.025 (0.005) 0.042 (0.000) 0.023 (0.012) 0.091 (0.000) 0.075 (0.000) 0.020 (0.027) 0.046 (0.000) 0.015 (0.096) 0.008 (0.392) 0.042 (0.000) 0.014 (0.119) 14. DEPENDENCE 15. TENURE 16. INSTOWN 0.032 (0.000) 0.021 (0.020) 0.018 (0.051) 0.085 (0.000) 0.107 (0.000) 0.088 (0.000) 0.293 (0.000) 0.170 (0.000) 0.004 (0.627) 0.086 (0.000) 0.121 (0.000) 0.096 (0.000) 0.016 (0.079) 0.089 (0.000) 0.074 (0.000) 0.085 (0.000) 0.025 (0.006) 0.004 (0.691) 0.063 (0.000) 0.258 (0.000) 0.094 (0.000) 0.046 (0.000) 0.111 (0.000) 0.044 (0.000) 0.185 (0.000) 0.032 (0.000) 0.104 (0.000) 0.091 (0.000) 0.107 (0.000) 0.147 (0.000) 0.009 (0.339) 0.014 (0.116) 0.010 (0.294) 0.191 (0.000) 0.001 (0.880) 0.109 (0.000) 0.047 (0.000) 0.095 (0.000) 0.221 (0.000) 0.011 (0.220) 0.066 (0.000) 0.083 (0.000) 0.129 (0.000) 0.022 (0.018) 0.206 (0.000) 0.081 (0.000) 0.151 (0.000) 17. EXGRW 0.323 (0.000) 0.278 (0.000) 0.319 (0.000) 0.004 (0.621) 0.021 (0.018) 0.056 (0.000) 0.257 (0.000) 0.170 (0.000) 0.011 (0.225) 0.178 (0.000) 0.112 (0.000) 0.199 (0.000) 0.041 (0.000) 0.071 (0.000) 0.113 (0.000) 0.097 (0.000) 0.035 (0.000) and lower forecasted earnings growth are more likely to have lower cost of equity. Although the correlations between a few variables are above 0.35, the highest variance inflation factor (VIF) observed in regressions is 2.525, which is well below the suggested multicollinearity problem threshold of ten (Marquardt 1980; Gujarati 1995). Regression Results for Auditor Industry Expertise and Cost of Equity Table 5 reports the regression results. The model includes the three mutually exclusive expert variables, NATIONALONLY, CITYONLY, and JOINTEXPERT. The coefficients on CITYONLY and JOINTEXPERT are significantly negative for all three cost-of-equity measures. The coefficients on NATIONALONLY are negative but marginally significant for only the RPEG Accounting Horizons December 2013 Krishnan, Li, and Wang 680 TABLE 5 Regression Results of the Relation between Auditor Industry Expertise and Client Cost of Equity Capital Dependent Variable RPEG Expected Sign Intercept NATIONALONLY CITYONLY JOINTEXPERT LNMV LEVERAGE BVMV BETA EPSSTDDEV IDIORISK RETURN DEPENDENCE TENURE INSTOWN EXGRW Year Dummies Industry Dummies Adj. R2 n ¼ 12,005 þ þ þ þ þ þ þ RMPEG ROJN Coeff. (1) t-stat. (2) Coeff. (3) t-stat. (4) Coeff. (5) t-stat. (6) 0.119 0.002 0.002 0.003 0.007 0.032 0.016 0.007 0.105 0.015 0.006 0.007 0.000 0.001 0.064 included included 30.125*** 1.336* 1.452* 1.902** 18.484*** 11.883*** 9.000*** 11.650*** 11.141*** 3.521*** 6.830*** 2.956*** 1.637* 1.180 18.984*** 0.113 0.002 0.002 0.003 0.007 0.047 0.026 0.005 0.093 0.009 0.003 0.007 0.000 0.001 0.059 included included 26.100*** 1.112 1.505* 1.818** 16.219*** 15.740*** 11.823*** 8.562*** 10.218*** 1.902** 2.401*** 2.645*** 2.289** 1.157 11.787*** 0.114 0.002 0.002 0.003 0.008 0.035 0.016 0.008 0.153 0.008 0.007 0.007 0.000 0.005 0.071 included included 25.318*** 1.151 1.724** 1.737** 16.876*** 11.283*** 7.656*** 11.970*** 13.265*** 1.577* 6.819*** 2.244** 1.486* 4.346*** 19.445*** 0.303 0.294 0.304 *, **, *** Significant at the 0.10, 0.05, and 0.01 level, respectively. Significance levels are based on one-tailed tests where signs are predicted, two-tailed otherwise. All t-statistics are based on the robust standard errors controlling for firm-clustering effects. See Table 3 for the description of other variables. Variable Definitions: Year Dummies ¼ 8 year dummies, coded 1 for the years 2001, 2002, 2003, 2004, 2005, 2006, 2007, and 2008, respectively, 0 otherwise; and Industry Dummies ¼ 47 industry dummies, based on the Fama-French (1997) 48-industry classification. measure (p-value ¼ 0.09). These results suggest that clients audited by city-only industry experts and by joint city-national industry experts have lower cost of equity than clients of non-expert auditors.11,12 The coefficients indicate that the cost of equity for clients of city-only experts, and joint city-national experts is lower by 20 and 30 basis points, respectively, than for clients of non11 12 Our results are similar when we estimate the models separately for each of our sample years. For example, using the RPEG measure, the Fama-MacBeth (Fama and MacBeth 1973) one-tailed p-values for NATIONALONLY, CITYONLY, and JOINTEXPERT are 0.054, 0.033, and 0.015, respectively. Our results are also similar when we use the Fama-French (Fama and French 1997) industry groups and the 27 industry groups used in Francis et al. (1999). Two alternative measures of industry expertise yield similar results to those reported here: (1) an auditor defined as an industry expert if its market share is more than 20 percent (25 percent in the post-SOX period) in that industry (Mayhew and Wilkins 2003; Dunn and Mayhew 2004), and (2) an auditor defined as an expert if it is top ranked and its market share is at least 10 percent higher than the second-ranked auditor in that industry (Mayhew and Wilkins 2003). Accounting Horizons December 2013 Auditor Industry Expertise and Cost of Equity 681 expert auditors. However, the coefficients on the three test variables do not significantly differ from each other. Among the control variables, companies’ cost of equity is negatively associated with size13 and recent returns, and is positively associated with leverage, book-to-market ratio, beta, analysts’ forecast dispersion, idiosyncratic risk, auditor tenure, auditor fee dependence, and forecasted earnings growth. Finally, one of the three COE measures (ROJN ) is negatively associated with institutional ownership.14 In sum, Table 5 provides new evidence on the associations between auditors’ city-only and joint city-national expertise and clients’ cost of equity. Complementing prior studies that find clients of auditors with joint city-national and city-only industry expertise pay a fee premium and have higher earnings quality (Francis et al. 2005; Reichelt and Wang 2010), our results indicate that those clients also have lower cost of equity. Change in Auditors and Change in Cost of Equity Next, we examine the change in cost of equity for clients with non-expert auditors that switched to expert auditors, as well as for clients with expert auditors that switched to nonexperts.15,16 If auditor industry expertise affects cost of equity (as we have hypothesized), we would expect the cost of equity to decrease (increase) when firms with non-experts (experts) switch to industry experts (non-experts). The change in cost of equity, ChgR, is defined as the change in the cost of equity from year t to year tþ1, where year t denotes the year of auditor change. To examine the change in cost of equity for clients switching from non-experts to experts, we focus on 152 firms that have non-expert auditors in year t and change auditors in year tþ1. The change in auditor expertise is measured by ChgEXP, coded 1 if the firm moved from a non-expert to an expert. We also decompose ChgEXP into three variables—ChgNATIONALONLY, ChgCITYONLY, and ChgJOINTEXPERT—to represent each individual type of expert to which the firms switched. Consistent with H2a, we expect negative coefficients on ChgEXP, 13 14 15 16 To rule out the possibility that the industry specialization variables are capturing nonlinearity in size, we include dummy variables for each LNMV decile, in place of LNMV. Our results continue to hold. The coefficient on tenure is unexpectedly positive. When we add a tenure-squared variable (following Boone et al. [2008]), the tenure variable becomes negative while the tenure-squared term is positive, which is consistent with Boone et al.’s findings. The coefficients and significance levels on the industry expertise variables are similar to our main results. Firms’ auditor choice decisions are likely to be based on cost-benefit analyses of auditors’ industry expertise. If factors that go into such analyses are also related to the cost of equity, and are not controlled for in our regressions, our results could be driven by endogeneity arising from omitted variables or self-selection bias. To the extent that such factors do not vary significantly across years, an analysis of change in cost of equity can help address concerns about endogeneity. In addition, to check for possible self-selection bias, we estimated the Heckman two-stage model. The first stage model for auditor choice is: NATIONALONLY/CITYONLY/JOINTEXP ¼ b0 þ b1LNAT þ b2ROA þ b3LIQUIDITY þ b4LEVERAGE þ b5BVMV þ b6BETA þ b7EPSSTDDEV þ b8IDIORISK þ b9RETURN þ b10DEPENDENCE þ b11TENURE þ b12INSTOWN þ b13EPSGRW þ Year Dummies. (ROA ¼ income before extraordinary item/total assets. LIQUIDITY ¼ current assets/current liabilities). We estimate three logit models using NATIONALONLY, CITYONLY, or JOINTEXP as the dependent variable to obtain three inverse Mills ratios. Next we include the three inverse Mills ratios in the second-stage model following Equation (4). Our results are similar. For example, for the RPEG measure, the p-values for NATIONALONLY, CITYONLY, and JOINTEXP are 0.102, 0.044, and 0.002, respectively. Previous studies that examine market reaction to auditor changes (between brand name and non-brand-name auditors or between specialist and non-specialist auditors) restrict their sample to companies that changed auditors. Auditor changes occur for many reasons (e.g., financial distress [Schwartz and Menon 1985]; client growth [Landsman et al. 2009]; litigation risk [Krishnan and Krishnan 1997]; audit fee [Ettredge et al. 2007; Rama and Read 2006]). Therefore, companies that switch auditors are inherently different from those that do not. We follow this literature in our research design choice of restricting the sample to those that changed their auditors. Accounting Horizons December 2013 Krishnan, Li, and Wang 682 TABLE 6 Descriptive Statistics of Change in Expert Status Variable Mean Median Minimum Maximum ChgEXP ChgNATIONALONLY ChgCITYONLY ChgJOINTEXPERT ChgNONEXP ReduNATIONALONLY ReduCITYONLY ReduJOINTEXPERT ChgRPEG ChgLNMV ChgLEVERAGE ChgBVMV ChgBETA ChgEPSSTDDEV ChgIDIORISK ChgRETURN ChgDEPENDENCE ChgTENURE ChgINSTOWN ICREPORT ChgEXGRW 0.136 0.034 0.058 0.045 0.103 0.014 0.065 0.023 0.001 0.129 0.006 0.017 0.013 0.016 0.019 0.134 0.017 6.061 0.014 0.083 0.027 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.002 0.098 0.007 0.004 0.017 0.000 0.010 0.054 0.001 4.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.210 2.640 0.397 1.436 3.105 0.600 0.523 3.838 0.992 32.000 1.000 0.000 1.638 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000 0.268 2.658 0.569 1.089 7.726 3.750 0.556 3.423 0.972 1.000 1.000 1.000 0.861 The sample (n ¼ 556) includes firm-year observations that had changed auditors between years t and tþ1. Variable Definitions: ChgEXP ¼ 1 if the change is from a non-expert to any expert, 0 if the change is from a non-expert to another non-expert; ChgNATIONALONLY ¼ 1 if the change is from a non-expert to a national-level expert, 0 otherwise; ChgCITYONLY ¼ 1 if the change is from a non-expert to a city-level expert, 0 otherwise; ChgJOINTEXPERT ¼ 1 if the change is from a non-expert to a joint expert, 0 otherwise; ChgNONEXP ¼ 1 if the change from any expert to a non-expert, 0 if the change is from an expert to another expert; ReduNATIONALONLY ¼ 1 if the change is from a national-level expert to a non-expert, 0 otherwise; ReduCITYONLY ¼ 1 if the change is from a city-level expert to a non-expert, 0 otherwise; ReduJOINTEXPERT ¼ 1 if the change is from a joint expert to a non-expert, 0 otherwise; and ICREPORT ¼ 1 if there is a reportable event related to internal control in the auditor change 8-K filings, 0 otherwise. For other variables, the prefix ‘‘Chg’’ indicates the change in the variable from the previous year. For example, ChgLNMV is the change in LNMV. ChgNATIONALONLY, ChgCITYONLY, and ChgJOINTEXPERT.17 Changes in other control variables are defined as the changes of variables from year t to year tþ1. We also include a dummy variable to measure the reportable internal control events disclosed in 8-K filings for auditor changes, to control for the effect of internal control quality on change in cost of equity capital. Descriptive statistics for the variables used in the change model are presented in Table 6. Among all auditor switches, about 13.6 percent of the companies switched from a non-expert to an industry expert and 10.3 percent switched from an industry expert to a non-expert. The change in cost of equity on average is 0.1 percent. 17 The number of changes from non-expert to national-only, city-only, and joint city-national experts are 19, 42, and 25, respectively. Accounting Horizons December 2013 Auditor Industry Expertise and Cost of Equity 683 TABLE 7 Association between Change in Expert Status and Change in Cost of Equity Capital Panel A: Change from Non-Experts to Experts Expected Sign Intercept ChgEXP ChgNATIONALONLY ChgCITYONLY ChgJOINTEXPERT ChgLNMV ChgLEVERAGE ChgBVMV ChgBETA ChgEPSSTDDEV ChgIDIORISK ChgRETURN ChgDEPENDENCE ChgTENURE ChgINSTOWN ICREPORT ChgEXGRW Year Dummies Industry Dummies Adj. R2 n þ þ þ þ þ þ ? þ Change to Any Expert Model (1) Coeff. t-stat. 0.008 0.023 0.320 2.630*** 0.018 0.015 0.004 0.002 0.030 0.144 0.005 0.008 0.000 0.011 0.002 0.047 included included 1.443* 0.368 0.122 0.157 0.733 2.899*** 1.222 0.188 0.506 0.786 0.109 3.394*** 0.408 152 Change to an Expert (by Type) Model (2) Coeff. t-stat. 0.006 0.001 0.032 0.020 0.025 0.035 0.021 0.001 0.037 0.119 0.005 0.006 0.000 0.031 0.004 0.044 included included 0.197 0.094 2.790*** 1.998** 2.024** 0.854 0.643 0.100 0.943 2.764*** 1.354* 0.124 0.619 1.818** 0.246 2.961*** 0.444 152 (continued on next page) The results for the multivariate model are presented in Table 7, Panel A. For brevity, we present results for the change analysis using the RPEG measure. Consistent with our expectations, Model (1) in Panel A shows ChgEXP is associated with a negative change in cost of equity one year after the auditor switch. Model (2) shows that changes from non-experts to city-level experts (ChgCITYONLY) or to joint experts (ChgJOINTEXPERT) are significantly associated with a decrease in cost of equity (by 320 and 200 basis points, respectively), while changes to nationalonly experts (ChgNATIONALONLY) are not associated with a change in cost of equity. To examine the change in cost of equity for clients switching from experts to non-experts, we focus on 404 firms that have expert auditors in year t and change auditors in year tþ1. The change in auditor expertise is measured by ChgNONEXP, coded 1 if the firm moved from an expert to a nonexpert. Again, we decompose ChgNONEXP into three variables—ReduNATIONALONLY, ReduCITYONLY, and ReduJOINTEXPERT—to represent each individual type of expert from which the firm switched. Consistent with H2b, we expect positive coefficients on ChgNONEXP, ReduNATIONALONLY, ReduCITYONLY, and ReduJOINTEXPERT.18 18 The number of changes from national-only, city-only, and joint city-national experts to non-experts are 19, 58, and 22, respectively. Accounting Horizons December 2013 Krishnan, Li, and Wang 684 TABLE 7 (continued) Panel B: Change from Experts to Non-Experts Expected Sign Intercept ChgNONEXP ReduNATIONALONLY ReduCITYONLY ReduJOINTEXPERT ChgLNMV ChgLEVERAGE ChgBVMV ChgBETA ChgEPSSTDDEV ChgIDIORISK ChgRETURN ChgDEPENDENCE ChgTENURE ChgINSTOWN ICREPORT ChgEXGRW Year Dummies Industry Dummies Adj. R2 n þ þ þ þ þ þ þ þ þ þ ? Change from Any Expert to a Non-Expert Model (3) Coeff. 0.003 0.005 0.017 0.042 0.028 0.002 0.010 0.019 0.004 0.014 0.000 0.015 0.010 0.029 included included t-stat. 0.364 0.677 1.413* 1.160 1.865** 0.557 2.849*** 0.464 0.884 0.924 1.280 1.010 1.081 1.507* 0.170 404 Change from an Expert (by Type) to a Non-Expert Model (4) Coeff. t-stat. 0.004 0.428 0.019 0.003 0.017 0.016 0.039 0.027 0.002 0.013 0.027 0.004 0.015 0.000 0.014 0.028 0.010 included included 0.649 0.353 1.432* 1.449* 1.096 1.870** 0.547 2.873*** 0.659 0.930 0.967 1.244 0.945 1.489 1.092 0.177 404 *, **, *** Significant at the 0.10, 0.05, and 0.01 level, respectively. Significance levels are based on one-tailed tests where signs are predicted, two-tailed otherwise. All t-statistics are based on the robust standard errors controlling for firm-clustering effects. Dependent variable is the change of cost of equity capital (ChgRPEG). See Tables 3 and 6 for the definition of the variables. Table 7, Panel B, presents the results. Model (3) in Panel B shows ChgNONEXP is not statistically significant. However, Model (4) shows that although changes from national-only and city-only experts to non-experts are not associated with increases in cost of equity, changes from joint city-national experts to non-experts (ReduJOINTEXPERT) are significantly (although with mild significance) associated with an increase in cost of equity, which provides further support that joint experts are associated with lower cost of equity.19 Change in Auditors and Subsequent Issuance of Equity To further investigate the direct relationship between cost of equity and industry expertise, we examine the association between auditor changes to industry experts and the subsequent issuance of 19 Results using the other two measures of cost of equity capital are similar. The coefficients on ChgEXP, ChgCITYONLY, and ChgJOINTEXPERT are negative and significant for both RMPEG ( p-values ¼ 0.010, 0.006, and 0.062, respectively) and ROJN ( p-values ¼ 0.017, 0.015, and 0.078, respectively). The coefficient on ReduJOINTEXPERT is positive and significant for both RMPEG and ROJN measures ( p-values ¼ 0.098 and 0.089, respectively). Accounting Horizons December 2013 Auditor Industry Expertise and Cost of Equity 685 equity. A large body of literature has identified multiple motivations for a change in auditors. If an association with industry-expert auditors could reduce cost of equity, we would expect firms that intend to issue equity to be more likely to change to expert auditors before the issuance, in order to enjoy the benefit of reduced cost of equity. We take firms that have non-expert auditors in year t and change auditors in year tþ1 (n ¼ 175) and examine their equity issuance during the first two years subsequent to the auditor change. We use the ratio of new issuance of common and preferred stock in the two years subsequent to an auditor change to total market value as our measure of equity issuance. The mean issuance of new equity is 2.3 percent for firms that change from non-experts to other non-experts and 5.3 percent for firms that change from non-experts to experts. The difference is significant with a pvalue of 0.047. The mean issuance of equity is 5.5 percent, 4.2 percent, and 6.5 percent for firms that change from non-expert auditors to national-only, city-only, and joint experts, respectively. Compared to switching to non-experts, the differences are significant with p-values of 0.068, 0.017, and 0.006, respectively. Thus, there is evidence that the prospect of a reduction in cost of equity may be one motivation for changing to expert auditors. However, as we noted above, there may be multiple reasons for auditor change, and change in cost of equity might be only one of them. ADDITIONAL ANALYSES Consistent with previous work (e.g., Reichelt and Wang 2010) that reports city-level expertise, especially joint city-national expertise, is associated with higher earnings quality, our findings above indicate that cost of equity is also lower for city-only and joint city-national experts. However, unlike prior work, we also find (albeit with relatively low statistical significance) that national-only experts also have lower cost of equity for one of our cost-of-equity measures. In this section, we conduct additional analyses to examine the effect of industry expertise on cost of equity in different settings. The Impact of Companies’ Foreign Operations Differences in the effectiveness of auditor industry expertise may depend on the extent of foreign operations. For clients with foreign operations, industry expertise includes an understanding of foreign operations. Therefore, city-level expertise may be less important for these clients, and national-level specialization—reflected in national knowledge-sharing practices and standardized industry-tailored audit programs—may be more relevant. In this section, we examine whether the association between auditor expertise and cost of equity extends to both firms operating internationally and firms operating only domestically. We split the total sample into two groups: U.S. companies with international operations and U.S. companies with purely domestic operations. Companies with at least one reportable segment in a foreign country and companies with no reportable foreign segments are designated as international companies and domestic companies, respectively. We estimate our model for each group.20 For brevity, we only report the results for the RPEG measure. The results for other cost of equity measures are similar. The results for the test variables (control variable coefficients are not tabulated) are reported in Table 8, Panel A. For companies with international operations, none of the coefficients on the industry expert variables is significant, indicating that for companies operating internationally, auditor industry expertise is not associated with the cost of equity, possibly because the 20 We also estimate the regressions using the full sample including interactions of the industry expertise variables with dummy variables for the sample partitions (e.g., foreign versus domestic firms; low versus high institutional holdings; and large versus small clients), and interactions of control variables with partition variables. Results of these regressions generally remain the same. Accounting Horizons December 2013 Krishnan, Li, and Wang 686 TABLE 8 Additional Analyses International Operations, Institutional Ownership, and Client Size Panel A: International Companies versus Domestic Companies Companies with International Operations NATIONALONLY CITYONLY JOINTEXPERT Controls Adj. R2 n Companies without International Operations Coeff. t-stat. Coeff. t-stat. 0.002 0.001 0.001 Included 1.222 0.820 1.210 0.001 0.002 0.004 Included 0.303 1.354* 2.183** 0.313 7,501 0.382 4,504 Panel B: High-Institutional-Ownership Companies versus Low-Institutional-Ownership Companies Companies with High-Institutional Ownership NATIONALONLY CITYONLY JOINTEXPERT Controls Adj. R2 n Companies with Low-Institutional Ownership Coeff. t-stat. Coeff. t-stat. 0.002 0.000 0.002 Included 0.690 0.051 1.109 0.003 0.003 0.003 Included 1.081 1.693** 1.414* 0.233 6,133 0.346 5,872 Panel C: Large Clients versus Small Clients Large Client NATIONALONLY CITYONLY JOINTEXPERT Controls Adj. R2 n Small Client Coeff. t-stat. Coeff. t-stat. 0.000 0.002 0.001 Included 0.161 1.193 0.644 0.004 0.002 0.002 Included 1.442* 1.029 1.842** 0.192 6,001 0.279 5,994 *, ** Significant at the 0.10 and 0.05 levels, respectively. Significance levels are based on one-tailed tests where signs are predicted, two-tailed otherwise. All t-statistics are based on the robust standard errors controlling for firm-clustering effects. The dependent variable is the cost of equity capital (RPEG). See Table 3 for the definition of other variables. complexities of international operations require auditor expertise unrelated to industry expertise. However, for domestic companies, the coefficients on CITYONLY and JOINTEXPERT are negative and significant, while the coefficient on NATIONALONLY is not significant. Thus, for firms operating only in the U.S., city-related industry expertise is associated with lower cost of equity, Accounting Horizons December 2013 Auditor Industry Expertise and Cost of Equity 687 possibly because the main audit work is done by the local offices where the companies’ headquarters are located. The Impact of Institutional Shareholders We also examine whether the association between cost of equity and auditor industry expertise differs across investor groups. Prior literature suggests that incentives to monitor corporate performance differ across investors. Institutional investors have greater incentive to monitor management because they potentially benefit the most from such monitoring (Brickley et al. 1988; Shleifer and Vishny 1986). Bushee (1998) finds that institutional investors constrain management discretion by reducing myopic investment behavior. To the extent that institutional investors attempt to constrain agency conflicts, they are either more likely than other investors to recognize the impact of industry expertise on earnings quality, or to substitute for auditor monitoring. To separate the effect of different types of investors, we divide the sample into two groups. Companies with high institutional ownership are those with a percentage of institutional shareholding above the sample median, and companies with low institutional ownership are those with a percentage of institutional shareholdings below the sample median. The results reported in Table 8, Panel B, indicate that, for firms with ‘‘high’’ institutional ownership, none of the industry expert variables is significant, suggesting auditor industry expertise is not associated with cost of equity. This is consistent with the conjecture that institutional ownership serves as an alternative monitoring mechanism for auditor industry expertise. For firms with ‘‘low’’ institutional ownership, CITYONLY and JOINTEXPERT are significantly negative, while NATIONALONLY is not significant. Thus, for firms with low institutional ownership, for which the monitoring effect from large investors is likely to be weak, city-level industry experts are associated with lower cost of equity. The Impact of Client Size Our industry expertise measures are based on audit fees and, therefore, auditors of large clients are more likely to be classified as experts. To differentiate the effect of industry expertise from client-size effects on cost of equity, we divide the sample into large and small clients using market value as the basis. Table 8, Panel C, reports the results. Two of the three industry specialization variables (NATIONALONLY and JOINTEXPERT) are significant for the small client subsample. None of the expert variables is significant for the large client subsample. Thus, the results imply that investors value auditors’ industry expertise more for small clients whose accounting information is less transparent than large clients’ (Lang and Lundholm 1993). The Impact of Accruals Quality Prior literature suggests a positive association between auditor expertise and accruals quality (Balsam et al. 2003; Krishnan 2003; Reichelt and Wang 2010) and a negative association between accruals quality and cost of equity capital (Francis et al. 2008). Therefore, we test whether accruals quality can explain the negative association between cost of equity capital and auditor industry expertise. We compute the Dechow and Dichev (2002) measure of accruals quality and add it to our model as an additional explanatory variable (Francis et al. 2008). Consistent with Francis et al. (2008), accruals quality (not tabulated) is negatively associated with cost of equity capital (p-value ¼ 0.001). The results for our industry expertise variables remain similar after controlling for accruals quality. For example, for the RPEG measure, the coefficient on JOINTEXPERT is significantly negative (p-value ¼ 0.025), and the coefficients on CITYONLY and NATIONALONLY Accounting Horizons December 2013 Krishnan, Li, and Wang 688 are marginally significant (p-values ¼ 0.095 and 0.111). Thus, industry expertise contributes to lowering the cost of equity capital, even after controlling for the effect of accruals quality. Results are similar for the other two cost-of-equity measures. City Defined at the MSA Level Following prior work (e.g., Reynolds and Francis 2001; Ferguson et al. 2003), we have defined City based on geographic location of the office. An alternative approach used in some studies (e.g., Li et al. 2010; Reichelt and Wang 2010) is to define City using the Metropolitan Statistical Area (MSA). Our results are similar when we define City based on the MSA approach. For example, for RPEG, the coefficient on JOINTEXPERT is negative and significant for the MSA level analysis (p-value ¼ 0.052). For the change analysis, the coefficients on ChgEXP, ChgCITYONLY, and ChgJOINTEXPERT are negative and significant (p-values ¼ 0.060, 0.004, and 0.070 respectively). CONCLUSIONS In this study, we examine whether companies that hire industry-expert auditors have lower cost of equity capital. We use the city-national framework of auditor expertise developed by Ferguson et al. (2003) and Francis et al. (2005) and examine the effects of national-only, city-only, and joint city-national industry expertise on cost of equity. Our main analysis provides evidence that, after controlling for other factors, the cost of equity capital is lower for companies that are audited by city-only and joint city-national industry experts. This is consistent with the assertion that investors’ positive perceptions about auditor reputation exist at the city level, either alone or jointly with national level. We also find that firms that move from non-experts to city-only and joint city-national industry experts experience a decline in cost of equity, and firms that switch from joint city-national experts to non-experts experience an increase in cost of equity. In addition, firms that switch from non-experts to experts issue more equity in the two years following the switch. Further analyses suggest that, while the effects of national-only and city-only experts vary across some subsets of our sample, joint city-national experts enjoy lower cost of equity in the majority of subsamples that we examined. However, for large companies and for companies that operate internationally or have high institutional ownership, the industry expertise variables are not significant, indicating that the effect of auditor industry expertise on cost of equity may not extend to international and large firms, and that sophisticated institutional investors may serve as substitutes for industry-expert auditors. Our study contributes to the literature by examining whether the enhanced quality of financial statements audited by industry-expert auditors impacts investors’ perceptions of financial reporting credibility as measured by the ex ante cost of equity. We extend the growing literature that focuses on the national-level and city-level dimensions of auditors’ industry expertise. Consistent with recent research, we find that the joint effect of the two forms of expertise (city-national) has the most consistent impact on the cost of equity capital. Regulators may believe that the interest of the public is better protected if industry-expert auditors are engaged. If the cost of equity is associated with industry expertise, then engaging an industry expert mitigates the mispricing of equity securities, ensures greater efficiency in capital markets, and reduces the social and political costs of audit failures and market failures. Then, an important question for regulators is: when do industry specialists matter? Our cross-sectional analyses provide some answers: industry specialists matter for smaller clients, companies with only domestic operations, and with low institutional ownership. Accounting Horizons December 2013 Auditor Industry Expertise and Cost of Equity 689 REFERENCES Balsam, S., J. Krishnan, and J. S. Yang. 2003. Auditor industry specialization and earnings quality. Auditing: A Journal of Practice & Theory 22 (2): 71–97. Beatty, R. P. 1989. Auditor reputation and the pricing of initial public offerings. The Accounting Review 64 (4): 693–709. Behn, B. K., J.-H. Choi, and T. Kang. 2008. Audit quality and properties of analyst earnings forecast. The Accounting Review 83 (2): 327–349. Boone, J. P., I. K. Khurana, and K. K. Raman. 2008. Audit firm tenure and the equity risk premium. Journal of Accounting, Auditing and Finance 23 (1): 115–140. Botosan, C. A., and M. A. Plumlee. 2005. Assessing alternative proxies for the expected risk premium. The Accounting Review 80 (1): 21–53. Brealey, R. A., and S. C. Myers. 1991. Principles of Corporate Finance. 4th edition. New York, NY: McGraw-Hill. Brickley, J. A., R. C. Lease, and C. W. Smith Jr. 1988. Ownership structure and voting on antitakeover amendments. Journal of Financial Economics 20 (January-March): 267–291. Bushee, B. J. 1998. The influence of institutional investors on myopic R&D investment behavior. The Accounting Review 73 (2): 305–333. Craswell, A. T., J. R. Francis, and S. L. Taylor. 1995. Auditor brand name reputations and industry specializations. Journal of Accounting & Economics 20 (3): 297–322. Dechow, P. M., and I. D. Dichev. 2002. The quality of accruals and earnings: The role of accrual estimation errors. The Accounting Review 77 (Supplement): 35–58. Dunn, K. A., and B. W. Mayhew. 2004. Audit firm industry specialization and client disclosure quality. Review of Accounting Studies 9 (1): 35–58. Easton, P. 2007. Estimating the cost of capital implied by market prices and accounting data. Foundations and Trends in Accounting 2 (4): 241–364. Easton, P. D. 2004. PE ratios, PEG ratios, and estimating the implied expected rate of return on equity capital. The Accounting Review 79 (1): 73–95. Ettredge, M., C. Li, and S. Scholz. 2007. Audit fees and auditor dismissals in the Sarbanes-Oxley era. Accounting Horizons 21 (4): 371–386. Fama, E. F., and J. D. MacBeth. 1973. Risk, return, and equilibrium: Empirical tests. Journal of Political Economy 81 (3): 607–636. Fama, E. F., and K. R. French. 1997. Industry costs of equity. Journal of Financial Economics 43 (2): 153– 193. Ferguson, A., J. R. Francis, and D. J. Stokes. 2003. The effects of firm-wide and office-level industry expertise on audit pricing. The Accounting Review 78 (2): 429–448. Fortin, S., and J. A. Pittman. 2007. The role of auditor choice in debt pricing in private firms. Contemporary Accounting Research 24 (3): 859–896. Francis, J. R., and M. D. Yu. 2009. Big 4 office size and audit quality. The Accounting Review 84 (5): 1521–1552. Francis, J. R., D. J. Stokes, and D. Anderson. 1999. City markets as a unit of analysis in audit research and the re-examination of Big 6 market shares. Abacus 35 (2): 185–206. Francis, J. R., K. Reichelt, and D. Wang. 2005. The pricing of national and city-specific reputations for industry expertise in the U.S. audit market. The Accounting Review 80 (1): 113–136. Francis, J., D. Nanda, and P. Olsson. 2008. Voluntary disclosure, earnings quality, and cost of capital. Journal of Accounting Research 46 (1): 53–99. Fried, D., and A. Schiff. 1981. CPA switches and associated market reactions. The Accounting Review 56 (2): 326–341. Gebhardt, W. R., C. M. C. Lee, and B. Swaminathan. 2001. Toward an implied cost of capital. Journal of Accounting Research 39 (1): 135–176. Gujarati, D. N. 1995. Basic Econometrics. 3rd edition. New York, NY: McGraw-Hill. Hackel, K. S. 2011. Security Valuation and Risk Analysis: Assessing Value in Investment Decision Making. 1st edition. New York, NY: McGraw-Hill. Accounting Horizons December 2013 690 Krishnan, Li, and Wang Hail, L., and C. Leuz. 2006. International differences in the cost of equity capital: Do legal institutions and securities regulations matter? Journal of Accounting Research 44 (3): 485–531. Holthausen, R., and R. Watts. 2001. The relevance of the value-relevance literature for financial accounting standard setting. Journal of Accounting & Economics 31 (1): 3–75. Jensen, M. C., and W. H. Meckling. 1976. Theory of the firm: Managerial behavior, agency costs, and ownership structure. Journal of Financial Economics 3 (4): 305–360. Journal of Accountancy. 2011. AICPA Letter to PCAOB Raises Concerns about Mandatory Audit Firm Rotation. Available at: http://www.journalofaccountancy.com/Web/20114896.htm Khurana, I. K., and K. K. Raman. 2004. Litigation risk and the financial reporting credibility of Big 4 versus non-Big 4 audits: Evidence from Anglo-American countries. The Accounting Review 79 (2): 473– 495. Khurana, I. K., and K. K. Raman. 2006. Do investors care about the auditor’s economic independence on the client? Contemporary Accounting Research 23 (4): 977–1016. Klock, M. 1994. The stock market reaction to a change in certifying accountant. Journal of Accounting, Auditing & Finance 9 (2): 339–347. Knechel, W. R., V. Naiker, and G. Pacheco. 2007. Does auditor industry specialization matter? Evidence from market reaction to auditor switches. Auditing: A Journal of Practice & Theory 26 (1): 19–45. Krishnan, G. V. 2003. Does Big 6 auditor industry expertise constrain earnings management? Accounting Horizons 17 (Supplement): 1–16. Krishnan, J., and J. Krishnan. 1997. Litigation risk and auditor resignations. The Accounting Review 72 (4): 539–560. Lambert, R., C. Leuz, and R. E. Verrecchia. 2007. Accounting information, disclosure, and the cost of capital. Journal of Accounting Research 45 (2): 385–420. Landsman, W. R., K. K. Nelson, and B. W. Rountree. 2009. Auditor switches in the pre- and post-Enron eras: Risk or realignment? The Accounting Review 84 (2): 531–558. Lang, M., and R. Lundholm. 1993. Cross-sectional determinants of analyst ratings of corporate disclosures. Journal of Accounting Research 31 (2): 246–271. Li, C., Y. Xie, and J. Zhou. 2010. National-level, city-level auditor industry specialization and cost of debt. Accounting Horizons 24 (3): 395–417. Mansi, S. A., W. F. Maxwell, and D. P. Miller. 2004. Does auditor quality and tenure matter to investors? Evidence from the bond market. Journal of Accounting Research 42 (4): 755–793. Marquardt, D. W. 1980. A critique of some ridge regression methods: Comment. Journal of the American Statistical Association 75 (369): 87–91. Mayhew, B. W., and M. S. Wilkins. 2003. Audit firm industry specialization as a differentiation strategy: Evidence from fees charged to firms going public. Auditing: A Journal of Practice & Theory 22 (2): 33–52. McConnell, J. J., and H. Servaes. 1990. Additional evidence on equity ownership and corporate value. Journal of Financial Economics 27 (2): 595–612. Nichols, D. R., and D. B. Smith. 1983. Auditor credibility and auditor changes. Journal of Accounting Research 21 (2): 534–544. Ohlson, J. A., and B. E. Juettner-Nauroth. 2005. Expected EPS and EPS growth as determinants of value. Review of Accounting Studies 10 (2-3): 349–365. Owhoso, V. E., W. F. Messier, Jr., and J. G. Lynch Jr. 2002. Error detection by industry-specialized teams during sequential audit review. Journal of Accounting Research 40 (3): 883–900. Petersen, M. A. 2009. Estimating standard errors in finance panel data sets: Comparing approaches. Review of Financial Studies 22 (1): 435–480. Public Company Accounting Oversight Board (PCAOB). 2011. Concept Release on Auditor Independence and Audit Firm Rotation. PCAOB Rulemaking Docket Matter No. 37. Available at: http://pcaobus. org/Rules/Rulemaking/Docket037/088_Royal_Gold.pdf Rama, D. V., and W. J. Read. 2006. Resignations by the Big 4 and the market for audit services. Accounting Horizons 20 (2): 97–109. Accounting Horizons December 2013 Auditor Industry Expertise and Cost of Equity 691 Reichelt, K. J., and D. Wang. 2010. National and office-specific measures of auditor industry expertise and effects on audit quality. Journal of Accounting Research 48 (3): 647–686. Reynolds, J. K., and J. R. Francis. 2001. Does size matter? The influence of large clients on office-level auditor reporting decisions. Journal of Accounting & Economics 30 (3): 375–400. Schwartz, K. B., and K. Menon. 1985. Auditor switches by failing firms. The Accounting Review 60 (2): 248–261. Shleifer, A., and R. Vishny. 1986. Large shareholders and corporate control. Journal of Political Economy 94 (3): 461–488. Teoh, S. H., and T. J. Wong. 1993. Perceived auditor quality and the earnings response coefficient. The Accounting Review 68 (2): 346–366. Watts, R. L., and J. L. Zimmerman. 1986. Positive Accounting Theory. New York, NY: Prentice Hall. Accounting Horizons December 2013