BOARD CHARACTERISTICS AND AUDIT FEES Joseph V. Carcello University of Tennessee 601 Stokely Management Center Knoxville, TN 37996-0560 (865) 974-1757 (865) 974-4631 (Fax) jcarcell@ Dana R. Hermanson Kennesaw State University Terry L. Neal University of Kentucky Richard R. Riley, Jr. West Virginia University April 2000 Acknowledgements: We thank Bruce Behn and Roger Hermanson, our co-authors on a related prior paper, as well as Mark Beasley, Mike Dugan, Rob Ingram, Zoe-Vonna Palmrose, Sean Peffer, Bob Ramsay, Larry Rittenberg, Gary Taylor, and workshop participants at the University of Alabama and University of Kentucky. We also gratefully acknowledge Lei Chen, Roy Clemons, Brandon Leffew, and Sharon Sexton for their research assistance. This paper can be downloaded from the Social Science Research Network Electronic Paper Collection:
BOARD CHARACTERISTICS AND AUDIT FEES Abstract This paper examines the relationship between board characteristics and external audit fees for Fortune 1000 companies. Competing arguments exist regarding the possible relationship between board characteristics and fees. One view is that a more independent, diligent, and expert board would be more concerned with effectively discharging its monitoring role and would be more supportive of the external audit function. Such a board would be likely to insist on enhanced audit scope, thus increasing the audit fee. An alternative view is that a more independent, diligent, and expert board would reduce the auditor’s assessment of control risk and would substitute some of its own monitoring for the monitoring of the auditor. This would reduce audit effort, thus decreasing the audit fee. Consistent with the first view, we find significant positive relationships between board independence, diligence, and expertise and audit fees. A more independent, diligent, and expert board does not appear to substitute for audit effort; rather, such a board may complement auditor oversight. These results add to the growing body of literature documenting a relationship between corporate governance mechanisms and various facets of the financial reporting and audit processes. In addition, the results add to our understanding of the determinants of audit fees. 2
BOARD CHARACTERISTICS AND AUDIT FEES The quality of board oversight has received increasing attention in recent years. Directors are being held to higher performance expectations, and many groups have called for greater emphasis on the independence, diligence and expertise of corporate board members (., NACD 1996). Prior research documents several outcomes associated with “higher quality” boards, including a reduced incidence of financial statement fraud (Beasley 1996) and fewer SEC enforcement actions for earnings manipulation (Dechow, Sloan, and Sweeney 1996). To provide greater insight into board monitoring, this paper examines the relationship between board characteristics and external audit fees for Fortune 1000 companies. More specifically, we examine the relationship between board independence, diligence, and expertise and audit fees. Competing arguments exist regarding the possible relationship between board characteristics and fees. One view is that a more independent, diligent, and expert board will be more concerned with effectively discharging its monitoring role. A key component of a board’s monitoring function involves overseeing the entity’s financial reporting process. The board typically relies on the external auditor to attest to the fairness of the entity’s reported financial results. As such, a board that is more concerned with fulfilling its monitoring role may be more supportive of the external audit function. A board that supports the external audit 3
function is likely to insist on enhanced audit scope (quantity of audit services provided), thus increasing the audit fee. Under this view, after controlling for other variables typically associated with audit fees, we would expect a positive relationship between the board characteristics examined and An alternative view is that a more independent, diligent, and expert board reduces the auditor’s risk and substitutes some of its own monitoring for the monitoring of the auditor. In terms of risk reduction, a more independent, diligent, and expert board may reduce the auditor’s assessment of control risk, thus allowing the auditor to reduce the scope of work and the fee. In addition, to the extent that a more independent, diligent, and expert board performs extensive monitoring activities of its own, these activities may substitute for some of the auditor’s effort, thus reducing the fee. Under this alternative view, after controlling for other variables typically associated with audit fees, we would expect a negative relationship between the board characteristics examined and fees. We measure board independence as the percentage of outside (., non-management) directors on the board. Board diligence is measured by referring to the number of board meetings that occurred during the year. We measure board expertise based on the average number of other director positions held by non-management directors. 4
We find significant positive relationships between board independence, diligence, and expertise and audit fees. The results are robust across numerous sensitivity tests. These results are consistent with board independence, diligence, and expertise being complements with audit effort, rather than substitutes. In addition, the results add to the growing body of literature documenting a relationship between corporate governance mechanisms and various facets of the financial reporting and audit processes, as well as to our understanding of the determinants of audit fees. The next section presents further background and develops the hypotheses. The following sections present the data and model, results, and conclusion. BACKGROUND AND HYPOTHESES The National Association of Corporate Directors (NACD) recently stated that a “professional boardroom culture requires that the governance process be collectively determined by individual board members” who possess the following characteristics -- independence, diligence, and expertise (NACD (1996, vii)).2 A number of other studies (., Conger, Finegold, and Lawler 1998; John and Senbet 1998; Lorsch 1995) also mention director independence, diligence, and expertise as the key ingredients necessary for the board to effectively discharge its monitoring The board of directors typically collaborates with management in selecting the external auditor (often subject to shareholder ratification). Since the auditor is 5
to look to the board as its client, it is reasonable to expect the board to review the overall planned audit scope and proposed audit fee (Blue Ribbon Committee 1999; Public Oversight Board 1994). Given the board’s oversight of the financial reporting and audit processes, as well as prior literature linking certain board characteristics to adverse financial reporting outcomes (Beasley 1996; Dechow, Sloan, and Sweeney 1996), we believe that it is appropriate to explore the link between board characteristics and audit fees. The following sections develop the hypotheses related to board independence, diligence, and expertise. Independence Board independence reflects the extent to which the board is independent of company management. A number of studies document a positive relationship between board independence and actions that are in the best interests of shareholders. See, for example, Brickley, Coles, and Terry (1994) (adoption of poison pills); Byrd and Hickman (1992) (abnormal returns associated with tender offers); Cotter, Shivdasani, and Zenner (1997) (tender offer premiums, revisions, and shareholder gains); Kosnik (1987) (greenmail payments); Rosenstein and Wyatt (1990) (share-price reactions to appointment of outside directors); and Tufano and Sevick (1997) (management fees to shareholders in mutual funds). Consistent with Beasley (1996), we view the board as more independent when the percentage of outside (., non-management) directors on the board is 6
greater. Our definition of an outside director is consistent with how national stock exchanges in the . define an outside director (Beasley 1996, 447). Since the owners of most large . corporations are separated from firm management, there is a possible incentive for managers to misreport financial results for opportunistic purposes (Chow 1982; Jensen and Meckling 1976; Watts and Zimmerman 1983). Outside directors, as representatives of shareholders, have a particularly strong incentive to prevent and detect such opportunistic reporting behavior by management (Fama and Jensen 1983). This incentive is driven, at least in part, by a stringent liability regime in the . Those outside directors who fail to exercise reasonable care in discharging their monitoring responsibilities are subject to severe sanctions (see Cordtz and Reingold 1993; Gilson 1990; Parker 1998; Sahlman 1993). An extreme case of opportunistic reporting behavior by management is fraudulent financial reporting. Consistent with the view that outside directors are more effective monitors of the financial reporting process, Beasley (1996) finds an inverse relationship between the percentage of outside directors on the board and the incidence of fraudulent financial reporting. Dechow, Sloan, and Sweeney (1996) find a similar result in their study of SEC enforcement actions related to earnings overstatements. A control mechanism for reducing the likelihood of fraudulent reporting -- and opportunistic reporting behavior in general -- is to purchase a greater quantity 7
of external audit services. Therefore, one may view outside directors as more concerned with enhancing the scope of the external audit than are management directors, who face greater conflicts of interest. Given the incentives that outside directors have to ensure reliable financial reporting, one view is to expect more independent boards to support the purchase of a greater amount of external auditing services, thus leading to increased audit fees. Alternatively, there are arguments suggesting a negative relationship between board independence and audit fees. First, if outside directors are more effective monitors, their enhanced monitoring efforts might substitute for some of the auditor’s effort, thus decreasing the fee. As an analogy, Feltham, Hughes, and Simunic (1991) suggest that audit quality and the ownership level retained by an entrepreneur are possible substitutes -- the oversight provided by the owner may reduce the need for some of the auditor’s Second, a more independent board may reduce the auditor’s assessment of control risk. Assessed control risk may decline given that the quality of the board is a component of a strong control environment (“tone at the top”) (COSO 1992). Under the audit risk model, lower control risk would be expected to result in fewer audit hours, reducing the audit fee. The discussion above leads to the first null hypothesis. H1: There is no relationship between the percentage of outside directors on the board and the external audit fee. 8
Diligence The diligence of the board includes factors such as the number of board meetings and the behavior of individual board members surrounding such meetings (., preparation prior to meetings, attentiveness and participation during meetings, and post meeting follow-up). The only one of these factors that is publicly observable is the number of board meetings. Lipton and Lorsch (1992) suggest that a major impediment to board effectiveness is a lack of time to complete board duties. In addition, prior studies (Conger, Finegold, and Lawler 1998; Pound 1995; Vafeas 1999) suggest that an increase in the number of board meetings can increase board effectiveness. One view is that a board that demonstrates greater diligence in discharging its responsibilities -- as measured by the number of board meetings -- will seek an enhanced level of oversight of the financial reporting process. As such, we would expect more diligent boards to support the purchase of a greater amount of external auditing services, resulting in higher audit fees. Alternatively, one might argue that a greater number of board meetings signals an enhanced degree of oversight by the board, and this enhanced oversight may substitute for some of the auditor’s effort, thus decreasing the fee. In addition, a more diligent board may reduce the auditor’s assessment of control risk, also reducing the audit fee. The discussion above leads to the second null hypothesis. 9
H2: There is no relationship between the number of board meetings and the external audit fee. Expertise The expertise of board members is a critical component in assuring that the monitoring role of the board is effectively discharged. Although there is no universal definition of board expertise, a number of studies argue that those directors who sit on multiple boards have made a significant investment in developing reputation capital as decision experts (., Fama 1980; Fama and Jensen 1983). In a manner generally consistent with the extant literature (., Beasley 1996; Cotter, Shivdasani, and Zenner 1997), we measure other directorships as the average number of outside directorships held in other corporations by non-management directors. One view is to expect directors who hold multiple directorships, and who presumably possess greater expertise, to have a better understanding of the value of the audit and, therefore, to be more supportive of the audit process. In addition, directors who hold multiple directorships have more to lose from opportunistic financial reporting behavior by management. For example, Gilson (1990) finds that directors who resign from bankrupt firms or from firms that privately restructure debt hold approximately one-third fewer directorships three years after their departure. Therefore, we would expect boards where multiple directorships 10
are common to be more supportive of the purchase of a greater amount of external auditing services, resulting in higher audit fees. Alternatively, one might argue that a greater number of other directorships signals enhanced quality of oversight by the board (., greater expertise results in higher quality oversight), and this enhanced oversight may substitute for some of the auditor’s effort, thus decreasing the fee. In addition, a more expert board may reduce the auditor’s assessment of control risk, also reducing the audit fee. The discussion above leads to the third null hypothesis. H3: There is no relationship between the average number of outside directorships held in other corporations by non-management directors and the external audit fee. DATA AND MODEL DEVELOPMENT A questionnaire was sent to the controllers of all Fortune 1000 companies (financial institutions and private companies were excluded) asking them to provide the amount of their external audit Beatty (1993) found a high degree of correspondence between mandated public disclosures of audit fees (in the IPO prospectus) and audit fees disclosed via a questionnaire. This high degree of correspondence lends some support to the self-reported audit fee To test the three hypotheses, we estimate the following model using OLS regression: LNFEE = ß0 + ß1 PCTOUTSIDE + ß2 NUMBODMTG + 11
ß3 DIRECTORSHIPS + ß4 SEGMT + ß5 SQSUBS + ß6 FOREIGN + ß7 LOSS + ß8 UTIL + ß9 RECINT + ß10 INVINT + ß11 LNASSETS + ε The dependent and predictor variables are defined as follows: LNFEE. Consistent with most recent studies on audit fees (., see Craswell, Francis, and Taylor 1995; Francis and Simon 1987; Palmrose 1986; Simon and Francis 1988; Turpen 1990), the dependent variable is LNFEE, the natural log of audit fees (first expressed in thousands of dollars).7 PCTOUTSIDE. We measure board independence by computing the percentage of outside (., non-management) directors on the board. NUMBODMTG. We measure board diligence as the number of meetings of the full board as disclosed in the proxy. DIRECTORSHIPS. We measure board expertise as the average number of outside directorships held in other corporations by non-management directors. We gathered data for the three test variables from the proxy statement filed immediately prior (typically seven to eight months) to the financial statement date for which we have audit fee data. By examining the proxy statement filed prior to the financial statements for which we have fee data, we have the best measure of the independence, diligence, and expertise of the board that was in place at the time the audit scope and audit fee were set. 12
In addition to the three test variables of interest, we control for the effects of other variables that have been found in prior literature to affect audit fees. These variables are: SEGMT. The number of business segments previously has been used (., Simon 1985) to provide a measure of the complexity of the entity’s operations. We expect a positive relationship between the number of business segments and audit fees. SQSUBS. In addition to the number of business segments, prior studies (Simon 1985; Simunic 1980) often have included the number of consolidated subsidiaries as a measure of business complexity. Consistent with more recent studies (., Craswell, Francis, and Taylor 1995; Francis and Simon 1987; Simon and Francis 1988; Turpen 1990), this variable is defined as the square root of the number of consolidated subsidiaries. We expect a positive relationship between the square root of the number of consolidated subsidiaries and audit fees. FOREIGN. The extent of the entity’s foreign operations also is related to the complexity of its operations. This variable represents the proportion of the entity’s total assets derived from foreign operations (Simunic 1980; Turpen 1990). We expect a positive relationship between the extent of foreign operations and audit fees. LOSS. If the entity has suffered a loss from continuing operations during any of the three preceding years, the auditor’s risk increases. This variable is 13
coded one if the entity has experienced a loss from continuing operations during any one of the three preceding years (0 otherwise). Consistent with prior studies (Simunic 1980; Turpen 1990), we expect a positive relationship between LOSS and audit fees. UTIL. Prior studies (Simunic 1980; Turpen 1990) have found that firms in the utility industry are particularly sensitive to audit fees. We expect a negative relationship between UTIL (coded 1 = utility, 0 = other) and audit fees. RECINT. Both this variable and the next have been included in a number of previous studies (Maher et al. 1992; Simon 1985; Simunic 1980) as predictors of audit fees. Receivables intensiveness is computed by dividing accounts receivable by total assets. We expect a positive relationship between audit fees and the percentage of total assets represented by receivables. INVINT. Inventory intensiveness is measured by dividing inventory by total assets. We expect a positive relationship between audit fees and the percentage of total assets represented by inventory. LNASSETS. Prior audit fee studies have found the natural log of total assets to be a highly important predictor of audit fees (., see Craswell, Francis, and Taylor 1995; Francis and Simon 1987; Simon and Francis 1988). We define LNASSETS as the natural log of total assets (first expressed in millions of dollars). We expect a positive relationship between client size and audit fees. 14
RESULTS Response Rate The adjusted sample size was 760 firms after deleting 40 non-public companies and 200 financial institutions. We received information on audit fees from 331 companies. We excluded 31 firms for which complete financial data were not available. We also excluded four firms that did not use a Big 6 The usable response rate was 39 We attempted to gather the relevant proxy statement for the 296 firms for which we had usable audit fee data. We were unable to obtain the relevant proxy statement for 35 firms,10 and three firms did not disclose the number of board meetings, leaving a final sample of 258 firms. Panel A of table 1 provides further details on the sample. A distribution of respondents, by industry, is presented in panel B of table ___________________ Insert table 1 about here ___________________ Descriptive Statistics Descriptive statistics on the variables used in the regression model are presented in table 2. The average audit fee for the sample companies was $ million (client assets averaged $ billion (not listed in table 2)). The mean percentage of outside board members was 75 percent. At a minimum, 17 percent of board members were outsiders, and there were some boards with all outside directors. The average number of board meetings was (minimum of three 15
meetings, maximum of 17). Finally, the average number of outside directorships in other firms held by outside directors was slightly more than two (minimum of zero, maximum of five).12 ___________________ Insert table 2 about here ___________________ A correlation matrix of the dependent and independent variables is presented in table 3. LNFEE is positively correlated with the percentage of outsiders on the board, the number of board meetings, and the average number of outside directorships held in other firms by outside directors (p < .01). Also, larger firms tend to have more independent boards, boards that meet more often, and boards where outside directors hold more outside directorship positions in other corporations (p < .01). ___________________ Insert table 3 about here ___________________ Audit Fee Regressions Results from the audit fee regressions appear in table To most clearly illustrate the effect of board independence, number of board meetings, and average number of outside directorships held in other firms by the firm’s outside directors, we ran two Model 1 regresses LNFEE only on the eight control variables included in the regression model. Model 2 regresses LNFEE on the control variables and on the three test variables of interest (PCTOUTSIDE, 16
NUMBODMTG, and DIRECTORSHIPS). We compare the R-square from the two models to determine if the R-square from Model 2 is significantly higher than Model 1’s R-square. Model 1, which regresses LNFEE on control variables derived from the extant literature, is significant (p < .001), and the adjusted R-square is 69 percent. As expected, several of the control variables are positively associated with audit fees (SEGMT, SQSUBS, FOREIGN, LOSS, RECINT, and LNASSETS). Audit fees are negatively associated with utilities (UTIL). The coefficient on INVINT is insignificant. ___________________ Insert table 4 about here ___________________ Model 2, which regresses LNFEE on control variables derived from existing literature and the three test variables of interest, is significant (p < .001), and the adjusted R-square is 71 percent. Consistent with the view that board independence, diligence, and expertise are complementary with audit effort, there are significant positive relationships between the percentage of outsiders on the board, the number of board meetings, and the average number of outside directorships held in other firms by outside directors and audit fees (p < .01 for PCTOUTSIDE and NUMBODMTG; p < .05 for DIRECTORSHIPS). In terms of the magnitude of the three test variables’ effects, starting at the average values of the independent variables, a 10% (20%) increase in each of the three test variables 17
would be expected to yield an $85,000 ($179,000) increase in audit fees. Given the overall average fee of $ million in this sample, such increases appear meaningful. The results for the control variables generally are consistent with those reported in Model Finally, the incremental F-statistic associated with testing Model 2 against Model 1 is significant at p < .01. Additional Analysis -- Independent and Gray Directors Outside directors can be categorized as either gray directors or as independent, outside directors (Baysinger and Butler 1985; Carcello and Neal 2000; Vicknair, Hickman, and Carnes 1993). A gray director is a non-management director who has either economic or personal ties to the firm or the firm’s management. Gray directors include former officers or employees, relatives of management, professional advisors to the firm (., consultants, bank officers, legal counsel), officers of significant suppliers or customers of the firm, and interlocking directors. An independent, outside director has no ties to the firm other than the directorship position held. The treatment of gray directors in the literature is not uniform. A number of studies -- including the majority of published studies in accounting -- treat gray directors as outside directors rather than as management directors (., Beasley 1996; Beasley and Salterio 1999; Dechow, Sloan, and Sweeney 1996; Rosenstein and Wyatt 1990). However, a large number of other studies conclude exactly the 18
opposite, that gray directors act more like management directors than outsiders (., Brickley, Coles, and Terry 1994; Byrd and Hickman 1992; Carcello and Neal 2000; Cotter, Shivdasani, and Zenner 1997). To be consistent with the existing accounting literature, we initially examined the relationship between the percentage of non-management directors (., gray directors and independent, outside directors are grouped together) and audit fees (table 4). As a supplemental analysis, we now consider the relationship between audit fees and gray directors and independent, outside directors separately. Consistent with Beasley (1996), we partition the PCTOUTSIDE variable into two new variables. The first variable, PCTINDEPENDENT, measures the percentage of directors classified as independent, outside directors. These directors have no ties with the entity or the entity’s management other than their service as directors. The second variable, PCTGRAY, measures the percentage of gray directors (as defined above). Gray directors have some economic and/or personal tie to the firm or its Results from this analysis appear in table 5. ___________________ Insert table 5 about here ___________________ There is a positive relationship between the percentage of independent directors on the board and audit fees (p < .01). There is a marginally significant 19
positive relationship between the percentage of gray directors on the board and audit fees (p < .10). This result suggests that gray directors act more like independent directors than like management directors with respect to audit fees. Sensitivity Tests Other firm governance mechanisms, audit firm characteristics, and client characteristics may be correlated with the test variables of interest and with audit fees. To address this possibility, we individually added several other variables to the model. The variables added were board size, the cumulative stock ownership percentage of outside directors, whether the entity’s CEO also serves as chairman of the board, director tenure, individual audit firm effects (using dummy variables), audit firm industry specialization,17 and client leverage. Only one of the additional variables tested, outsider director stock ownership percentage, was even weakly associated with audit fees (marginal positive relationship, p < .10). All others were not associated with audit fees. In each sensitivity test, the coefficient on PCTOUTSIDE was significantly positive at the .03 level or less, the coefficient on NUMBODMTG was significantly positive at the .01 level, and the coefficient on DIRECTORSHIPS was significantly positive at the .03 level or less. The results from these numerous sensitivity tests suggest that the positive relationships between the three test variables and audit fees are robust. 20
CONCLUSION This study represents an initial attempt to study the relationships between various board characteristics and audit fees. Although the audit fee literature is quite developed, no prior study examines the links between various corporate governance mechanisms and fees. Also, to the extent that the extant accounting literature has examined corporate governance mechanisms at the board level, most of the focus has been on board composition. This study extends this literature by testing proxies for board diligence and expertise. The results are consistent with the view that board independence, diligence, and expertise are complementary with audit effort, rather than substitutes for audit work. First, a greater percentage of outside directors on the board is associated with higher audit fees. Second, there is a positive relationship between the number of board meetings and audit fees. Third, there is a positive relationship between the average number of outside directorships in other firms held by the entity’s outside directors and audit fees. As an exploratory study, this paper is subject to a number of limitations. The sample is limited to very large public, non-financial companies. Therefore, the extent to which the results apply in other settings is uncertain. More importantly, we have only documented an association between certain board characteristics and audit fees. There may be exogenous factors at the entity level that are correlated both with the board characteristics examined in this study and 21
with audit fees. We have attempted to control for determinants of audit fees documented in the literature, as well as for other possible governance mechanisms that may be correlated with the test variables and audit fees. There can be no guarantee that we have been successful in ruling out all plausible correlated omitted variables. The results add to a growing body of literature that finds a link between corporate governance mechanisms and various facets of the financial reporting and audit processes (., Beasley 1996; Dechow, Sloan, and Sweeney 1996). Given the heightened interest of the accounting profession, the business community, and regulators in the relationship between corporate governance and financial reporting quality, we believe that the relationship between corporate governance mechanisms and other facets of the audit process continues to be a fruitful area of inquiry. 22
REFERENCES Baysinger, B. D., and H. N. Butler. 1985. Corporate governance and the board of directors: Performance effects of changes in board composition. Journal of Law, Economics and Organization 1 (Fall): 101-124. Beasley, M. S. 1996. An empirical analysis of the relation between the board of director composition and financial statement fraud. The Accounting Review 71 (October): 443-465. Beasley, M. S., and S. Salterio. 1999. Cross-sectional variation in Canadian audit committee composition and member experience. Working paper, North Carolina State University and University of Waterloo. Beatty, R. P. 1993. The economic determinants of auditor compensation in the initial public offerings market. Journal of Accounting Research 31 (Autumn): 294-302. Blue Ribbon Committee. 1999. Report and Recommendations of the Blue Ribbon Committee on Improving the Effectiveness of Corporate Audit Committees. New York: NYSE and NASD. Breusch, T., and A. . The Lagrange multiplier test and its application to model specification in econometrics. Review of Economic Studies 47 (January): 239-253. Brickley, J. A., J. L. Coles, and R. L. Terry. 1994. Outside directors and the adoption of poison pills. Journal of Financial Economics 35 (June): 371-390. Byrd, J. W., and K. A. Hickman. 1992. Do outside directors monitor managers? Evidence from tender offer bids. Journal of Financial Economics 32 (October): 195-221. Carcello, J. V., and T. L. Neal. 2000. Audit committee composition and auditor reporting. Working paper, University of Tennessee and University of Kentucky. Chow, C. W. 1982. The demand for external auditing: Size, debt, and ownership influences. The Accounting Review 57 (April): 272-291. Committee of Sponsoring Organizations of the Treadway Commission (COSO). 1992. Internal Control – Integrated Framework. New York: COSO. 23
Conger, J. A., D. Finegold, and E. E. Lawler, III. 1998. Appraising boardroom performance. Harvard Business Review 76 (January-February): 136-148. Cordtz, D., and J. Reingold. 1993. The vanishing director. Financial World 162 (October 12): 22-26. Cotter, J. F., A. Shivdasani, and M. Zenner. 1997. Do independent directors enhance target shareholder wealth during tender offers? Journal of Financial Economics 43 (February): 195-218. Craswell, A. T., J. R. Francis, and S. L. Taylor. 1995. Auditor brand name reputations and industry specialization. Journal of Accounting and Economics 20 (December): 297-322. Dechow, P. M., R. G. Sloan, and A. P. Sweeney. 1996. Causes and consequences of earnings manipulation: An analysis of firms subject to enforcement actions by the SEC. Contemporary Accounting Research 13 (Spring): 1-36. Deis, D. R., and G. Giroux. 1996. The effect of auditor changes on audit fees, audit hours, and audit quality. Journal of Accounting & Public Policy 15 (Spring): 55-76. Fama, E. F. 1980. Agency problems and the theory of the firm. Journal of Political Economy 88 (April): 288-307. Fama, E. F., and M. C. Jensen. 1983. Separation of ownership and control. Journal of Law & Economics 26 (June): 301-325. Feltham, G. A., J. S. Hughes, and D. A. Simunic. 1991. Empirical assessment of the impact of auditor quality on the valuation of new issues. Journal of Accounting and Economics 14 (December): 375-399. Francis, J. R., and D. T. Simon. 1987. A test of audit pricing in the small-client segment of the . audit market. The Accounting Review 62 (January): 145-157. Gilson, S. C. 1990. Bankruptcy, boards, banks, and blockholders. Journal of Financial Economics 27 (October): 355-387. 24
Jensen, M. C., and W. H. Meckling. 1976. Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of Financial Economics (October): 305-360. John, K., and L. W. Senbet. 1998. Corporate governance and board effectiveness. Journal of Banking & Finance 22 (May): 371-403. Kosnik, R. D. 1987. Greenmail: A study of board performance in corporate governance. Administrative Science Quarterly 32 (June): 163-185. Lipton, M., and J. Lorsch. 1992. A modest proposal for improved corporate governance. Business Lawyer 48 (November): 59-77. Lorsch, J. W. 1995. Empowering the board. Harvard Business Review 73 (January-February): 107-117. Maher, M. W., P. Tiessen, R. Colson, and A. J. Broman. 1992. Competition and audit fees. The Accounting Review 67 (January): 199-211. Monks, R., and N. Minow. 1995. Corporate Governance. Cambridge, MA: Blackwell. National Association of Corporate Directors. 1996. Report of the NACD Blue Ribbon Commission on Director Professionalism. Washington, .: NACD. O’Keefe, T. B., R. D. King, and K. M. Gaver. 1994. Audit fees, industry specialization, and compliance with GAAS reporting standards. Auditing: A Journal of Practice & Theory 13 (Fall): 41-55. Palmrose, Z-V. 1986. Audit fees and auditor size: Further evidence. Journal of Accounting Research 24 (Spring): 97-110. Parker, S. 1998. The effect of audit committee characteristics on financial reporting. Working paper, University of Oregon. Pound, J. 1995. The promise of the governed corporation. Harvard Business Review 73 (March-April): 89-98. Public Oversight Board. 1994. Strengthening the Professionalism of the Independent Auditor. Stamford, CT: POB. 25
Rosenstein, S., and J. G. Wyatt. 1990. Outside directors, board independence, and shareholder wealth. Journal of Financial Economics 26 (August): 175-91. Sahlman, W. A. 1990. Why sane people shouldn’t serve on public boards. Harvard Business Review 68 (May-June): 28-35. Simon, D. T. 1985. The audit services market: Additional empirical evidence. Auditing: A Journal of Practice & Theory 5 (Fall): 71-78. Simon, D. T., and J. R. Francis. 1988. The effects of auditor change on audit fees: Tests of price cutting and price recovery.” The Accounting Review 63 (April): 255-269. Simunic, D. A. 1980. The pricing of audit services: Theory and evidence. Journal of Accounting Research 18 (Spring): 161-190. Tufano, P., and M. Sevick. 1997. Board structure and fee-setting in the . mutual fund industry.” Journal of Financial Economics 46 (December): 321-355. Turpen, R. A. 1990. Differential pricing on auditors’ initial engagements: Further evidence. Auditing: A Journal of Practice & Theory 9 (Spring): 60-76. Vafeas, N. 1999. Board meeting frequency and firm performance. Journal of Financial Economics 53 (July): 113-142. Vicknair, D., K. Hickman, and K. Carnes. 1993. A note on audit committee independence: Evidence from the NYSE on “grey” area directors. Accounting Horizons 7 (March): 53-57. Watts, R. L., and J. L. Zimmerman. 1983. Agency problems, auditing, and the theory of the firm: Some evidence. Journal of Law & Economics 26 (1983): 613-633. White, H. 1980. A heteroskedasticity-consistent covariance matrix and a direct test for heteroskedasticity. Econometrica 48 (May): 817-838. 26
ENDNOTES 27
TABLE 1 Sample Description PANEL A: SAMPLE DETERMINATION Fortune 1000 1000 Non-public companies (40) Financial institutions (200) Adjusted sample 760 Responding companies 331 Missing financial data (31) Non-Big Six auditors (4) Usable responses 296 Less: Proxy information not available (35) Board meeting data not available (3) Final sample 258 PANEL B: SAMPLE INDUSTRY DISTRIBUTION (n = 258) Number of SIC Codes Industry Observations 1000 – 1999 Mining, Construction 14 2000 – 2999 Manufacturing – Food, textiles, lumber, chemicals 80 3000 – 3999 Manufacturing – Rubber, metal, machinery, equipment 90 4000 – 4999 Transportation, Communication, Utilities 42 5000 – 5999 Wholesale, Retail 21 7000 – 9999 Services 11 Total 258 28
TABLE 2 Descriptive Statistics (n = 258) Standard Variable Mean Deviation Minimum Median Maximum AUDITFEE 1,502 2,409 45 800 26,000 LNFEE PCTOUTSIDE NUMBODMTG DIRECTORSHIPS SEGMT SQSUBS FOREIGN LOSS UTIL RECINT INVINT LNASSETS AUDITFEE = audit fee in thousands of dollars LNFEE = natural log of audit fee in thousands of dollars PCTOUTSIDE = percentage of non-management board members NUMBODMTG = number of board of director meetings DIRECTORSHIPS = average number of outside directorships in other firms held by outside directors SEGMT = number of business segments SQSUBS = square root of number of consolidated subsidiaries FOREIGN = foreign assets / total assets LOSS = 1 if firm had a loss from continuing operations during the past three years, else 0 UTIL = 1 if firm operates in the utility industry, else 0 RECINT = accounts receivable / total assets INVINT = inventory / total assets LNASSETS = natural log of total assets in millions of dollars 29
TABLE 3 Correlation Matrix (n = 258) PCTOUTSIDE NUMBODMTG DIRECTORSHIPS SEGMT SQSUBS FOREIGN LOSS UTIL RECINT INVINT LNASSETS LNFEE PCTOUTSIDE NUMBODMTG DIRECTORSHIPS SEGMT SQSUBS FOREIGN LOSS UTIL RECINT INVINT LNFEE = natural log of audit fee in thousands of dollars PCTOUTSIDE = percentage of non-management board members NUMBODMTG = number of board of director meetings DIRECTORSHIPS = average number of outside directorships in other firms held by outside directors SEGMT = number of business segments SQSUBS = square root of number of consolidated subsidiaries FOREIGN = foreign assets / total assets LOSS = 1 if firm had a loss from continuing operations during the past three years, else 0 UTIL = 1 if firm operates in the utility industry, else 0 RECINT = accounts receivable / total assets INVINT = inventory / total assets LNASSETS = natural log of total assets in millions of dollars 30
TABLE 4 Audit Fee Regression Results (n = 258) (Dependent Variable = LNFEE) Model #1 Model #2 Pred. Variable Sign Coefficient t-statistic Coefficient t-statistic Constant ? -2 .73 *** *** PCTOUTSIDE ? *** NUMBODMTG ? *** DIRECTORSHIPS ? ** SEGMT + *** *** SQSUBS + ** * FOREIGN + *** *** LOSS + ** * UTIL - *** *** RECINT + ** *** INVINT + LNASSETS + *** *** Model F *** *** Adjusted R-square .69 .71 Model F Tests: Model 2 vs. 1, F = *** LNFEE = natural log of audit fee in thousands of dollars PCTOUTSIDE = percentage of non-management board members NUMBODMTG = number of board of director meetings DIRECTORSHIPS = average number of outside directorships in other firms held by outside directors SEGMT = number of business segments SQSUBS = square root of number of consolidated subsidiaries FOREIGN = foreign assets / total assets LOSS = 1 if firm had a loss from continuing operations during the past three years, else 0 UTIL = 1 if firm operates in the utility industry, else 0 RECINT = accounts receivable / total assets INVINT = inventory / total assets LNASSETS = natural log of total assets in millions of dollars ******, , Statistically significant at less than the .10, .05, .01 level, based on one-tailed (two-tailed) tests for variables whose relation to the dependent variable is (is not) predicted. 31
TABLE 5 Audit Fee Regression Results (n = 258) Substitution of PCTINDEPENDENT and PCTGRAY for PCTOUTSIDE (Dependent Variable = LNFEE) Pred. Variable Sign Coefficient t-statistic Constant ? *** PCTINDEPENDENT ? *** PCTGRAY ? * NUMBODMTG + *** DIRECTORSHIPS + *** SEGMT + *** SQSUBS + * FOREIGN + *** LOSS + * UTIL - *** RECINT + *** INVINT + LNASSETS + *** Model F *** Adjusted R-square .71 LNFEE = natural log of audit fee in thousands of dollars PCTINDEPENDENT = percentage of board members who are independent, outside directors PCTGRAY = percentage of board members who are gray directors NUMBODMTG = number of board of director meetings DIRECTORSHIPS = average number of outside directorships in other firms held by outside directors SEGMT = number of business segments SQSUBS = square root of number of consolidated subsidiaries FOREIGN = foreign assets / total assets LOSS = 1 if firm had a loss from continuing operations during the past three years, else 0 UTIL = 1 if firm operates in the utility industry, else 0 RECINT = accounts receivable / total assets INVINT = inventory / total assets LNASSETS = natural log of total assets in millions of dollars ****, Statistically significant at less than the .10, .01 level, based on one-tailed (two-tailed) tests for variables whose relation to the dependent variable is (is not) predicted.
1 We cannot rule out the possibility that a more independent, diligent, and expert board simply would exhibit less price resistance (., audit fees are higher without any increase in audit scope). We would need data on audit hours to conclusively rule out this possibility, and we do not have access to audit hours incurred. However, prior studies suggest that audit effort and audit fees are highly correlated. Deis and Giroux (1996) find that audit hours and audit fees move together, and O’Keefe, King and Gaver (1994) use audit fees as a proxy for labor (audit effort). 2 The NACD report also lists director integrity and a realization and acceptance by the board that they have a function separate from management (NACD 1996, vii). However, since these board attributes are not as directly observable as are independence, diligence, and expertise, we do not consider them in this study. 3 We do not consider any possible link between audit fees and either the existence or the composition of the entity’s audit committee. All sample firms maintain an audit committee, and 95 percent of the firms in the sample had audit committees composed entirely of outsiders. Given the almost complete lack of variation on the percentage of outside directors on the audit committee, testing the relationship between audit committee composition and audit fees is not practical. 2
4 We recognize that outside directors and entrepreneurs have very different levels of company-specific knowledge; therefore, it probably is more likely that an entrepreneur’s monitoring could substitute for auditor effort. 5 Financial institutions were excluded consistent with prior research (Francis and Simon 1987; Simon 1985; Simon and Francis 1988). Private firms were excluded due to the lack of publicly available information needed to compute various control variables. 6 With the exception of audit fees, all of the other variables in the model were gathered from public sources (., Compact Disclosure, 10-Ks, annual reports, and proxy statements). 7 Audit fee models typically take one of two forms -- either the dependent variable is audit fees deflated by assets (., Simunic 1980) or the natural log of audit fees (., Francis and Simon 1987). As a sensitivity test, we define the dependent variable as audit fees deflated by assets. Similar to Simunic, we first regressed LNFEE on LNASSETS. The resulting coefficient associated with LNASSETS, .57, was used to scale total assets in computing a new dependent variable measure, SIMFEE. SIMFEE is thus computed as: Audit Fees / Total . When we replace LNFEE with SIMFEE in the table 4 model (removing LNASSETS from the right-hand side of the equation), there were significant positive relationships between PCTOUTSIDE (p < .01) and NUMBODMTG (p < .01) and SIMFEE. 3
However, there was not a significant relationship between DIRECTORSHIPS and SIMFEE (p = .15). As a further sensitivity test, we use a two-stage regression approach. We first regressed LNFEE on LNASSETS and kept the residual. The residual was then used as the dependent variable in the table 4 model (with LNASSETS excluded from the model). Consistent with the results in table 4, there were significant positive relationships (all p < .05) between all three test variables (PCTOUTSIDE, NUMBODMTG, DIRECTORSHIPS ) and audit fees. 8 There may be systematic differences between the types of companies that retain Big 6 / non-Big 6 auditors. These differences could be correlated with both board characteristics and audit fees. Given the small number of responding Fortune 1000 companies retaining non-Big 6 firms (n = 4), we chose to exclude these observations to ensure that the audit service providers were relatively homogeneous (., all Big 6 firms). 9 The usable response rate, 39 percent (296 / 760), compares favorably with other studies where audit fees are self-reported. Simunic (1980) reported a usable response rate of 33 percent, and Palmrose (1986) reported a response rate of approximately 30 percent. Also, we utilized the early-late approach to test for non-response bias. There were no significant differences (p < .05) between the early and late groups for any of the variables. 4
10 We contacted Primark / Disclosure to obtain proxy statements. Primark / Disclosure is the designated commercial repository for all SEC filings. 11 To address potential industry effects on audit fees, the regression model was run with industry dummy variables added (one for each one-digit SIC, with the 7000-9999 SIC companies in the intercept). The results for the three test variables (PCTOUTSIDE, NUMBODMTG, and DIRECTORSHIPS) were qualitatively unchanged. 12 Some have suggested (Monks and Minow 1995; Pound 1995) that holding an excessive number of directorship positions can prevent a director from effectively discharging his or her duties. For example, Monks and Minow (1995) suggest that a prudent director will sit on no more than three boards. As a sensitivity test, we exclude all observations where the average number of outside directorships in other firms held by outside directors exceeds three (n = 46). There continues to be a significant positive relationship between all three test variables and audit fees (p = .03 or less). 13 To evaluate the possible influence of multicollinearity on the regression results, we examined the VIFs (variance inflation factors) and condition indices. All VIFs were less than two, and the condition index was less than , suggesting that the influence of multicollinearity was minimal. 5
14 We tested for the presence of heteroskedasticity and significant outliers. The results of the Breusch-Pagan (1980) test indicate mild heteroskedasticity. As a result, all standard errors are developed using White’s (1980) heteroskedasticity-consistent covariance matrix. To evaluate the influence of particular observations on the regression results, DFBetas were calculated and examined. No influential observations were identified. This result also was checked visually by examining plots of standardized residuals. 15 Some audit fee studies have controlled for whether the auditor is “new” (., has accepted the engagement within the past three years) (Simon and Francis 1988) and whether the auditor issued a modified report (., Francis and Simon 1987; Simon and Francis 1988; Simunic 1980). We add these variables to the models presented in table 4. Neither variable is significant, and the results for the test variables of interest (PCTOUTSIDE, NUMBODMTG, and DIRECTORSHIPS) are qualitatively unchanged. 16 The mean percentage of independent directors is 59 percent (minimum of zero, maximum of 92 percent). The mean percentage of gray directors is 16 percent (minimum of zero, maximum of 71 percent). 17 We used both continuous and dichotomous measures. The continuous measure of industry specialization was developed by summing total client revenues audited by each Big 6 firm within each 3-digit SIC code and then dividing by total 6
revenues for all companies within that SIC code. We also characterized auditors as a specialist / non-specialist using the dichotomous approach utilized by Palmrose (1986). 7
