Artikel Audit UAS - Working Paper Series WP 2003-04 Boards...

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: Working Paper Series WP 2003-04 Boards of Directors, Audit Committees, and the Information Content of Earnings Kirsten L. Anderson, Daniel N. Deli and Stuart L. Gillan http://www.lerner.udel.edu/ccg/ Boards of Directors, Audit Committees, and the Information Content of Earnings Kirsten L. Anderson, Daniel N. Deli and Stuart L. Gillan* Georgetown University, Washington DC 20057 Arizona State University, Tempe, AZ 85287 University of Delaware, Newark, DE 19716 First Draft: June 2002 This Draft: September 2003 ABSTRACT We examine the relation between the information content of earnings (earnings response coefficients) and board and audit committee structure for a broad sample of more than 1,200 firms during 2001. We find that earnings are more informative the greater is the independence and the activity of the full board. Moreover, firms that have separated the CEO and board Chair positions appear to have more informative earnings. We also find that audit committee characteristics influence the information content of earnings. In particular, firms that have smaller audit committees have more informative earnings. However, we find that audit committee independence incremental to the independence of the full board is unrelated to the information content of earnings. Keyword: Earnings, Corporate Boards, Audit Committees, Outside Directors JEL Classification: G32 Correspondence to: Daniel N. Deli, Arizona State University, Department of Finance, Tempe, AZ 85287 Tel: (480) 965-4603 FAX: (480) 965-8539. * 1 This paper was started while Gillan was with the TIAA-Cref Institute. Gillan acknowledges partial support from the Weinberg Center for Corporate Governance in the Lerner College of Business and Economics at the University of Delaware and the KPMG Audit Committee Institute. The views expressed in this paper are those of the authors and do not necessarily reflect those of TIAA-CREF or the KPMG Audit Committee Institute. Boards of Directors, Audit Committees, and the Information Content of Earnings ABSTRACT We examine the relation between the information content of earnings (earnings response coefficients) and board and audit committee structure for a broad sample of more than 1,200 firms during 2001. We find that earnings are more informative the greater is the independence and the activity of the full board. Moreover, firms that have separated the CEO and board Chair positions appear to have more informative earnings. We also find that audit committee characteristics influence the information content of earnings. In particular, firms that have smaller audit committees have more informative earnings. However, we find that audit committee independence incremental to the independence of the full board is unrelated to the information content of earnings. 1. Introduction We examine the informativeness of annual earnings announcements for a large sample of publicly traded corporations as a function of corporate governance. Specifically, we focus on earnings response coefficients as a function of board and audit committee characteristics, including independence, size, activity, and the separation of the CEO and Board Chair positions. Our research is conducted against the backdrop of increased regulation of US corporate governance. In 1999 the Securities and Exchange Commission adopted rules requiring that audit committees consist entirely of independent directors, have at least three financially literate directors, and at least one director with an accounting or related financial management background [Chung and Kallapur (2002)]. Moreover, recent events and associated regulatory changes have provided the impetus for an even greater focus on the monitoring role of boards and audit committees. Indeed, the Sarbanes Oxley Act of 2002 represents a sea-change in the Federal oversight of the corporate governance process, oversight which has traditionally been the domain of the states.1 SEC rulemaking seeking to implement aspects of Sarbanes-Oxley highlights two important issues. First, the accuracy and reliability of financial information used in investment decision-making is critical to the integrity of US financial markets. Second, the roles of boards and audit committees are considered pivotal in ensuring the integrity of corporate financial statements. As stated in the SEC’s 2003 proposing release pertaining to audit committees: The board of directors, elected by and accountable to shareholders, is the focal point of the corporate governance system. The audit committee, composed of members of the board of 1 The NYSE and NASDAQ have also proposed changes to listing standards including changes to board membership and board committee structures. directors, plays a critical role in serving as a check and balance on a company's financial reporting system.2 Thus, the goal of increased regulation of corporate governance is to increase the quality of financial statement information. Underlying these changes is the presumption that corporate governance affects the informativeness of reported earnings. Furthermore, regulatory reforms appear to be based on an objective function of maximizing earnings informativeness. In some sense, the issue at hand is the normative question of “What should corporate governance look like in order to maximize the information content of earnings?”3 That question, however, is being asked and answered in a near vacuum of information on how corporate governance actually affects the information content of earnings. Although the link between reported earnings and market valuation has long been recognized in the literature [e.g., Ball and Brown (1968), Beaver (1968)], we currently lack a rich set of positive theories addressing how corporate governance actually affects the informativeness of reported earnings [Jensen (1983)].4 We examine a sample of more than 1,200 publicly traded firms with annual shareholder meetings during the calendar year 2001 and find the following. We find that earnings response coefficients increase as corporate boards become more independent. We also find that more active boards are associated with higher earnings response coefficients. Moreover, the effects of board independence and activity are contingent upon one another. For example, our results suggest that increasing board independence has the greatest positive effect on the information 2 Securities and Exchange Commission Proposed Rule: Standards Relating To Listed Company Audit Committees, 17 CFR PARTS 228, 229, 240, 249 and 274 [RELEASE NOS. 33-8173; 34-47137; IC-25885; File No. S7-02-03] RIN 3235-AI75. 3 Jensen (1983) notes that the choice of objective function is, itself, a normative issue. 2 content of earnings when board activity is low. In addition, it appears that firms in which the Chief Executive Officer (CEO) and board chair (COB) positions are separated have more informative earnings. These findings suggest that the full board plays an important role in safeguarding the informativeness of the reported earnings numbers We also find evidence suggesting that audit committees play a role in determining the information content of the earnings. When considering audit committee characteristics in isolation, we find that more independent and more active audit committees are associated with more informative earnings. Also, smaller audit committees are associated with more informative earnings. In a combined focus incorporating board and audit committee characteristics, however, we do not find a relation between audit committee independence and the information content of earnings. Although we have no direct evidence, we suggest that the observed lack of a relation between audit committee independence and the information content of earnings may result from the incremental benefits of greater independence being offset by a loss of expertise pertaining to firm financial reporting systems and internal controls. We make two important contributions to the literature. Our first contribution arises from a focus on earnings informativeness, rather than on “earnings quality” when examining the effectiveness of boards and their audit committees. This provides a different perspective on the issues relative to many earlier studies. Notably, much of the prior literature addressing board and audit committee effectiveness has been hampered by the need to develop an appropriate characterization of earnings quality (typically a measure of accruals). Irrespective of the quality of earnings, the central issue surrounding reported accounting numbers is how well they inform 4 This begs the question of just what the relevant objective function is. For example, it is typical to treat the objective function of the firm as the maximization of shareholder value. Maximizing shareholder value and maximizing the informativeness of reported earnings are not necessarily consistent with one another. 3 the financial markets with respect to the true economic condition of the firm.5 In other words, earnings may be of low quality in the sense that they deviate substantially from cash flows, but they may still inform the market as to underlying firm value. It is this basic intuition that we rely on to motivate our examination of the information content of earnings as a function of board and audit committee characteristics. Our second contribution is that we provide further evidence on the cross-sectional determinants of earnings response coefficients. Previous research suggests that earnings response coefficients are a function of several factors including growth opportunities, risk, size, and auditor quality amongst others.6 We extend this work by providing evidence on how certain corporate governance characteristics, specifically those pertaining to the independence and activity of corporate boards and audit committees, affect cross-sectional variation in earnings response coefficients. The paper proceeds as follows. In section 2 we provide background and develop our hypotheses. In section 3 we describe the data and provide our univariate and multivariate analyses. Section 4 concludes. 2. Background Fama and Jensen (1983b) contend that boards play an important role in monitoring managerial actions. Empirical evidence consistent with that contention is extensive [e.g., Weisbach (1988), Byrd and Hickman (1992), Shivdasani (1993), Brickley, Coles, and Terry (1994)]. Research has increasingly focused on the importance of the board and board 5 6 We abstract from the role of earnings numbers in contracting. See, for example, Easton and Zmijewski (1989), Defond and Park (2001), and Teoh and Wong (1993) amongst others. 4 committees in corporate governance.7 For example, Jensen (1993) argues that more independent as well as smaller boards are more effective monitors. Empirical evidence [Yermack (1996) and Vafeas (2000)] supports this view. Vafeas (1999) concludes that board activity is an important dimension of board operations and reports that meeting frequency is related to ownership structure and firm valuation. Moreover, when examining the link between board committee structure and firm performance, Klein (1998) reports that the independence of key board committees is related to firm value. We extend this literature by focusing on the role of boards and audit committees in monitoring the accounting process and in assuring the veracity of accounting numbers. Separating the CEO and COB positions may also strengthen corporate governance [Jensen (1993)]. The evidence on this issue is, however, mixed. Goyal and Park (2002) report that the sensitivity of CEO turnover to performance is lower when the same individual is both CEO and COB. In contrast, Brickley, Coles and Jarrell (1997) suggest that, while there are potential benefits of separating the CEO and COB positions, there are also potential costs in that for some firms CEO/Chair duality may be an efficient outcome. Given the increased focus on separating the two positions and the mixed evidence in the literature, we also examine the association between earnings informativeness and CEO/Chair duality. 2.1. The Role of Audit Committees in the Valuation Process Although recent events have led to an increased focus on audit committees, interest in their efficacy is not new. There has been notable demand for regulatory oversight of audit committee composition as a means of advancing audit committee effectiveness. During the late 7 Bushman and Smith (2003) provide an extensive review and discussion of corporate governance and financial information. 5 eighties, the Treadway Commission (1987) recommended that, “The board of directors of all public companies should be required by SEC rule to establish audit committees composed solely of independent directors.” Similarly, the Commission on Public Trust and Private Enterprise (2003) recommends that “Members of the audit committee must be independent and have both knowledge and experience in auditing financial matters.” Moreover, “…the board of directors should assess the independence and the qualifications of the members of the audit committee to ensure that each qualifies for membership on the committee.” The Commission on Public Trust goes further to advocate the requirements of Sarbanes-Oxley that “…the audit committee be responsible for the appointment, compensation, and oversight of the work of the auditors, and that the outside auditors report directly to the audit committee.” The responsibility of the audit committees for the internal audit process is also highlighted in the report. Specifically that “The internal auditor should have a direct line of communication and reporting responsibility to the audit committee, and he or she should attend all regularly scheduled audit committee meetings, report on the status of audits conducted by the internal audit group, report to the committee on other matters that the internal auditor, in his or her judgment, believes should be brought to the audit committee’s attention, and meet with the audit committee in executive session.” Put another way, the existence of an independent audit committee serves to enhance the veracity of both the internal and external auditing processes. Enhancing the veracity of the external auditing process helps guarantee the accuracy of firm financial statements and allows the capital markets to place greater reliance on reported results. Early studies focused on the role of boards and audit committees in deterring managerial actions leading to fraud or SEC enforcement actions. For example, Beasley (1996) finds that independent boards are important in deterring accounting fraud, but that audit committees 6 provide no incremental monitoring. Dechow, Sloan, and Sweeney (1996) examine firms subject to SEC enforcement actions for GAAP violations leading to overstated earnings and find that such actions are more likely with insider dominated boards and CEO/Chair duality. Prior research has also focused extensively on accruals to measure earnings quality and earnings management. Firms with large accruals (often in absolute value terms) are assumed to have low-quality earnings numbers and firms with smaller accruals are thought to have higher quality earnings numbers.8 Others have addressed the link between the quality of accounting numbers and corporate governance. For example, in finding a negative relation between board and audit committee independence and abnormal accruals, Klein (2002) suggests that more independent boards are more effective monitors. Similarly, Peasnell, Pope and Young (2001) suggest that, in a sample of U.K. firms, the likelihood of income increasing accruals declines as board independence increases. More recently, Bowen, Rajgopal, and Venkatachalam (2002) report evidence of a weak relation between aggressive accounting choice and poorer corporate governance.9 Some researchers, however, have questioned the appropriateness of using accruals as a measure of earnings quality. Pike (2003), for example, notes the implicit assumption in this research that discretionary accruals are related to opportunistic earnings management may not be the case. Others suggest methodological problems in accruals research. Collins and Hribar (2002) contend that earnings management tests based on accruals calculated from balance sheet values are potentially contaminated by measurement error in accruals estimates. Kothari, Leone, 8 Francis, LaFond, Olsson, and Schipper (2002) note that earnings management papers typically predict the direction of earnings management and focus on signed abnormal accruals. In contrast, earnings quality research typically focuses on unsigned abnormal accruals, with many such measures being derived from Jones (1991). 7 and Wasley (2002) suggest that performance-matching improves inferences but may still be problematic in some samples. Beaver (2002) and Kothari (2001) discuss the implications of these and other empirical considerations in capital markets research.10 As an alternative to focusing on accruals, researchers have examined earnings response coefficients (ERCs). This approach focuses on how the market reacts to unexpected earnings conditioned on characteristics of the firm and the contracting environment. This strand of the literature has taken several forms. First, there is a focus on the relationship between ERCs and earnings quality (including the pricing of discretionary accruals). Easton and Zmijewski (1989) report that ERCs are associated with risk, growth, and earnings variability. More recently, Defond and Park (2001) argue that when earnings surprises contain abnormal accruals, anticipated accrual reversals affect ERCs. Similarly, Core, Guay, and Kothari (2001) and Huson, Scott, and Wier (2001) argue that the existence of potentially dilutive securities attenuates the earnings-return relationship. Alternative approaches to measuring quality include those of Imhoff (1992) who finds that firms with higher analyst perceptions of earnings quality produce larger earnings response coefficients than their low quality counterparts. Similarly, Imhoff and Lobo (1992) report that ERCs increase with increased dispersion of analyst forecasts. Second, papers have focused on earnings informativeness and corporate governance features of the contracting environment. For example, Teoh and Wong (1993) suggest that ERCs for Big Eight audit clients are larger than those for Non Big Eight clients. Choi and Jeter (1992) report that ERCs decline after the issuance of qualified audit reports. Other studies have examined aspects of ownership structure. Warfield, Wild and Wild (1995) find that managerial 9 The authors also find 1) weak associations between governance characteristics and the tendency to report small positive surprises and 2) stronger evidence of a relation between poorer governance quality and aggressive accrualbased smoothing. 8 ownership is positively associated with earnings’ explanatory power for returns. For East Asian corporations, Fan and Wong (2002) report that earnings are less informative in the presence of concentrated ownership, pyramidal ownership structures, and cross-holdings. Similarly, Francis, Schipper and Vincent (2002) find evidence suggesting that a separation of voting rights and cash flow rights in US firms with dual class stock is associated with lower earnings informativeness. Finally, several papers have examined the link between ERCs and corporate governance by focusing on the board or audit committee [Wild (1994, 1996) and Vafeas (2000)]. The papers by Wild indicate that earnings are more informative for firms that voluntarily establish audit committees during the 1966-1980 period. Vafeas studies 350 large firms during the 1990-1994 period and finds that earnings informativeness is unrelated to board independence but increases as board size decreases.11 We extend the literature examining the link between corporate governance and earnings quality by focusing on the market’s reaction to earnings numbers conditioned on corporate governance characteristics. We rely on the intuition that, despite their quality, earnings may still inform the market as to underlying firm value. Specifically, we examine earnings informativeness as a function of board and audit committee characteristics including board and audit committee size, independence, activity, and CEO/Chair duality. Moreover, we do so by studying a large sample of firms during 2001 - a period of heightened focus on corporate governance and, in particular, an increased focus on the role of corporate boards and audit committees. Also see reviews by McNichols (2000) and Healy and Whalen (1999). Our paper is also related to recent research documenting relations between boards or audit committees and external auditors and audit fees [e.g., Antle, Gordon, Narayanmoorthy, and Zhui (2002), Asbaugh, Lafond and Mayhew (2003), Carcello and Neal (2000, 2003), Frankel, Johnson and Nelson (2002), and Larcker and Richardson (2003)]. 11 10 9 2.1 Hypothesis Development We argue that the markets’ perception of accounting numbers is a function of firms’ corporate governance. In very general terms, then, our hypothesis is that the stronger the firm’s corporate governance, the less noise there is in accounting numbers. This suggests that earnings response coefficients (the market’s reaction to unexpected earnings) should be greater for firms with better governance. In particular, we focus on corporate governance in terms of the independence, size, and activity of both the full board and the audit committee. To that end, we have three central hypotheses. Stated in the null form: H1: Earnings response coefficients are unrelated to board and audit committee independence. H2: Earnings response coefficients are unrelated to board and audit committee size. H3: Earnings response coefficients are unrelated related to board and audit committee activity (the number of reported meetings). Our priors are that ERCs are positively related to independence, negatively related to size, and positively related to activity. There are several other considerations in our analysis. First, we recognize that board and audit committee attributes may be interdependent. For example, the effect of increasing audit committee independence may depend on audit committee activity. Consequently, we examine interaction terms among our measures of independence and activity. Second, although our focus is on the effects of board and audit committee characteristics on the information content of earnings, other aspects of the governance environment may also be relevant. In particular, proponents of governance reform have advocated separating the CEO and Chairman of the Board positions to strengthen corporate governance. As discussed previously, 10 the evidence on this issue is mixed. Thus, we also investigate whether or not CEO duality is associated with earnings informativeness. Our fourth hypothesis, also stated in null form, is: H4: Earnings response coefficients are unrelated to CEO duality. Finally, previous research indicates that other factors may influence the earnings-return relationship, thus we incorporate additional control variables into our analysis as discussed in Section 3 below. 3. Data and Empirical Analysis In this section we describe our data and present empirical tests of our hypotheses. 3.1 Sample The data we use to describe boards and audit committees comes from the Investor Responsibility Research Center (IRRC). Our sample includes firms with annual shareholder meetings during the calendar year 2001. Therefore, these firms have fiscal year-ends in either 2000 or 2001. The IRRC data provides us with board characteristics including board member affiliation, ownership, and the frequency of board and audit committee meetings. The IRRC sample is drawn from those firms that constitute the S&P Super 1,500 and other large widely followed firms. We eliminate those observations that do not have all of the data items available for the multivariate analysis (see below). Further, we eliminate those observations where boards met more than twelve times in a year (6% of firms) or where audit committees met more than seven times in a year (6.5% of firms). We do so to eliminate the potential influence of unusual corporate events (e.g., financial distress) on our analysis.12 The resulting sample covers a broad spectrum of 1,241 firms with market values of equity ranging from a $11 million to $476 billion. Other financial characteristics of the sample are reported in panel B of table 1. 11 We examine several characteristics of boards and audit committees and the relations of those characteristics to the information content of earnings. One feature that we consider is the relative independence of the board and audit committee members. Characterizing the independence of the board and audit committee members requires some taxonomy for independence. We follow the taxonomy typically used in the literature and focus on the nature of the board member’s affiliation with the firms on whose boards they sit. Independent outside directors are those who have no relation with the firm other than their appointment as a director. That is, they are not current or former employees, they are not related to managers, their firms do not do business with the firm on whose board they sit, etc. Directors who are current employees of the firm are classified as insiders. Affiliated, or gray directors, are those directors who, while not current employees of the firm, have a business relation with the firm or were previously employed by the firm. Because of the potential for affiliated outside directors to act in the interests of management much of our focus is on independent outside directors (as opposed to affiliated and independent outside directors combined). Panel A of table 1 provides descriptive statistics for the sample boards and audit committees. Panel A reports that the boards in our sample have, on average, 63.3% independent directors. The median is 66.7%. Audit committees are 88.2% independent on average (100% median). Contrasting the results for the full board with those of the audit committee suggests that audit committees are characterized by a concentration of independent board members. We also conjecture that board and audit committee activity, as captured by the number of meetings that each holds, are potentially relevant to the information content of earnings. Panel A of table 1 reports that the full board meets an average (median) of 6.6 (6) times per year. Audit committees 12 Our results are, however, unchanged when we include observations with extreme numbers of board or audit 12 meet less frequently with the average (median) number of audit committee meetings being 4 (4). Finally, board and audit committee average sizes are 9.4 members and 3.6 members, respectively. Nearly 27% of firms have different individuals occupying the CEO and Board Chair positions. As discussed above, previous research suggests that separation of the CEO and Chairman of the Board positions may affect firm value [Jensen (1993) and Goyal and Park (2002)]. If co-location of the CEO and Chairman positions leads to a poorer contracting environment, then the market may find any given level of unexpected earnings less compelling than if the positions were separated. We expressly examine this hypothesis in our empirical tests. Ultimately, we want to examine the information content of earnings announcements as a function of board and audit committee characteristics. To do so, we require some model of investor expectations to capture the relation between new information coming to the market and the market’s response to that new information. We rely on two different measures of expected earnings throughout our analysis—a seasonal random walk [consistent with recent studies such as Francis, Schipper and Vincent (2003)] and the most recent analyst forecasts [see for example, Teoh and Wong (1993)]. Using these proxies for expected earnings we calculate our measure of unexpected earnings as the difference between reported earnings and expected earnings scaled by the closing stock price on the last day of the fiscal year. In the analysis that follows, we report only those results for our analysis using the seasonal random walk model of expectations, but we note in the text when the results differ using mean analyst forecasts. Table 2 reports descriptive statistics for our measure of unexpected earnings across deciles formed on the level of unexpected earnings. Overall, the mean (median) unexpected committee meetings. 13 earnings (unscaled by price) for our sample is -0.09 (0.07). Looking at the distribution of mean unexpected earnings across unexpected earnings deciles suggests that firms experienced greater negative earnings surprises than positive earnings surprises. One potential problem in the analysis that follows is that earning surprises may be a function of board and audit committee characteristics. Following the research design of Teoh and Wong (1993), we assume that the precision of the signal contained in announced earnings is affected by board and audit committee characteristics. It is possible, however, that the reported earnings may, in some sense, be biased and that the degree of bias could be related to board and audit committee characteristics. We examine that possibility in table 3. In table 3 there is no discernable link between board and audit committee characteristics and unexpected earnings deciles. 3.2 The information content of earnings We hypothesize that the information content of earnings, specifically the amount of value-relevant information that the market derives from a given amount of unexpected earnings, is a function of board and audit committee characteristics. Following the methodology of Teoh and Wong (1993), we examine the extent to which the information content of accounting earnings is contingent upon board and audit committee characteristics. To do so we use regressions of the following general format: CAR it = α + βUE it + ∑ δ j UE it Governance jit + ∑η k UE it Control kit + ε it j k (The Governance terms in the above regression specification include each of the relevant board and audit committee characteristics, and CEO/Chair duality.) The marginal effect of unexpected earnings on abnormal returns is captured by the following relation: 14 ∂ (CAR it ) ˆ ˆ = β + ∑ δˆjGovernance jit + ∑ηk Controlkit ∂ (UE it ) j k ˆ Earnings response coefficient research typically focuses on the regression coefficient, β , to capture the relation between unexpected earnings and returns. Our regression specification allows for that relation to be conditional on the board and audit committee characteristics (and a group of control variables). Thus, our regression specification allows us to examine the relation between unexpected earnings and the market’s response, conditional on the various characteristics of the board and audit committee. Additionally, recent research suggests that there are potentially important interaction effects between board and audit committee independence and board and audit committee activity [Abbot, Parker, Peters, and Raghunandan (2001), Carcello, Hermanson, Neal, and Riley (2000), Klein (2000, 2002)]. Consequently, we include interactions terms for board independence and board activity and for audit committee independence and audit committee activity in our empirical analysis. While we are interested in the effect of board and audit committee characteristics on the information content of earnings, it is important that we control for other factors that may be related to the information content of earnings. For example, the market’s response to a given piece of earnings information is likely to be contingent on the market’s perception of the persistence of those new earnings [Kormendi and Lipe (1987), Collins and Kothari (1989), Lipe (1990)]. Collins and Kothari (1989) contend that the ratio of a firm’s market value to its book value is a positive function of earnings persistence. Smith and Watts (1992) argue that marketto-book ratios proxy for a firm’s growth opportunities and growth opportunities may also affect the information content of earnings. Thus, we use the market-to-book ratio to proxy for crosssectional variation in the persistence of earnings and firm growth opportunities. We calculate the 15 market-to-book ratio as the market value of equity divided by the book value of common shareholders’ equity. Firm size may also be an important determinant of the information content of earnings announcements. For example, Atiase (1985) suggests that just prior to earnings announcements the market has better information for larger firms than for smaller firms. Greater pre-earnings announcement disclosure would attenuate the market’s response to a given earnings announcement. Thus, we would expect that a given level of earnings surprise would be greeted with a greater market response for large firms than for small firms. We calculate firm size as the market value of equity plus the book value of long-term debt plus the book value of preferred stock.13 We also suspect that the firm’s information environment affects the market’s assessment of earnings surprises (Imhoff and Lobo (1992)). In particular, if the information environment is relatively poor, then the market will have difficulty in interpreting earnings announcement. That being the case, a given level of unexpected earnings should convey more information (i.e., be more of a surprise) for those firms with less volatile returns. We proxy for the information environment by calculating the standard deviation of daily returns over the firm’s prior fiscal year. Each of the above control variables captures some aspect of the firm that may be related to the market’s response to a given level of unexpected earnings. It may be the case that the market’s perception of the information content of earnings is also affected by what we might generally refer to as the contracting environment. Our interest is in the incremental effects associated with various characteristics of boards and audit committees. Other aspects of the 16 contracting environment may also be relevant. For example, when managerial and investor incentives are better aligned, the market may find a given level of unexpected earnings more compelling than it would if director and officer ownership was low. Thus, we control for ownership in the firm by officers and directors in the regression analysis. 3.3 Board characteristics and the information content of earnings We examine the effect of various board characteristics on the information content of earnings in table 4. In the first three columns of table 4 we introduce each board characteristic individually and in the final regression we include all of the board characteristics. The results reported in the first and fourth regressions of table 4 suggest that earnings announcement returns are positively related to the independence of full boards. This suggests that a given level of unexpected earnings generates a greater market reaction (in the same direction) the more independent is the full board. Put differently, earnings are more informative for firms whose boards are relatively independent. This result is consistent with our hypothesis that the relation between unexpected earnings and abnormal returns depends upon board independence. We also hypothesize that board activity (as measured by the number of board meetings) is related to the information content of earnings. Active boards are thought to be indicative of more credible earnings numbers, all else equal. The results in table 4 suggest that board activity is, in fact, relevant to the information content of earnings. In the final regression of table 4, higher levels of board activity lead to greater market reactions to a given level of unexpected returns (pvalue = .057). These results suggest that the information content of earnings increases with board activity. 13 Because firm size is highly skewed, we use the natural log of firm size in the multivariate analysis. 17 It is interesting to note, however, that the results on the interaction term between board independence and the number of board meetings suggest a more complex relationship between board independence and activity, and the information content of earnings. For example, in the last regression of table 4 the coefficient on the interaction term between board independence and the number of meetings is negative and statistically significant (p-value = .027). This result suggests that the effect of each of the board characteristics on the information content of earnings is dependent upon the other. For example, the effect of an increase in board independence is contingent upon board activity. Consider two boards with relatively few independent directors where one meets infrequently and the other meets with greater frequency. Increasing independence will tend to have a larger effect on the information content of earnings for the board that meets infrequently than it will for the board that meets with greater frequency. To see this, remember that the marginal effect of unexpected earnings on abnormal returns is given by the following: ∂ (CAR it ) ˆ ˆ ˆ = β + δ 1PctIndep + δˆ2 NumMtgs + δˆ3 (PctIndep * NumMtgs) + δˆ4 BrdSize + δˆ5SepCEO & Chair + ∑ηkControlkit . ∂ (UEit ) k The effect of board independence, for example, on the relation between abnormal returns and unexpected earnings is thus captured by the following: ∂ (CAR it ) ˆ ˆ ∂ = δ1 + δ 3 (NumMtgs) ∂PctIndep ∂(UEit ) In this case the estimated coefficient on the number of meetings, δˆ3 , is negative. The negative coefficient suggests that the relation between unexpected earnings and abnormal returns increases with the level of board independence, but that the relation is less positive when the number of board meetings is large and more pronounced when the number of board meetings is small. 18 The relation between unexpected earnings and abnormal returns does not appear to be contingent on board size. In none of the specifications of table 4 is the coefficient on board size statistically significant. Board size, therefore, does not appear to affect the information content of earnings. This finding contrasts with that of Vafeas (200) who reports that ERCs increase as board size decreases. We also note that the separation of the CEO and Chair positions appears to be relevant to the information content of earnings. For a given level of positive unexpected earnings, the market reacts more favorably when firms have separated the CEO and Chair positions than when they have not (p-values ranging from .027 to .112). This result is consistent with the argument that boards are more effective when the positions of CEO and Chair are split. The market finds earnings from a firm that splits the positions to be more compelling than earnings from a firm with CEO/Chair duality. 3.4 Audit committee characteristics and the information content of earnings We consider the effect of various audit committee characteristics on the information content of earnings in table 5. As was the case with the full board, we examine the relevance of independence, activity, and size in affecting the information content of earnings. The first audit committee characteristic that we examine is audit committee independence. When we include only those terms involving audit committee independence (the first regression in table 5), we find little relation between audit committee independence and the information content of earnings. When we include all of the audit committee characteristics in the final regression of table 5, the coefficient on audit committee independence is positive and statistically significant (p-value = .017). This result suggests that earnings surprises from firms with more independent audit 19 committees are more informative than earnings surprises from firms with less independent audit committees. Like full board activity, the information content of earnings appears to be positively related to audit committee activity (p-value = .053). The relation using analyst forecasts as our measure of expected earnings is positive as well, but at a lower level of statistical significance (pvalue = .159). Also similar to the interaction between board independence and board activity, the results in the last column of table 5 suggest that the effects of audit committee independence and activity are conditional upon one another. The relation persists when using analyst forecasts, albeit at a lower level of significance (p-value = .166). Increasing audit committee independence has the greatest effect on the information content of earnings when audit committee activity is low. When audit committees are active, however, the relative effect of increased audit committee independence on the information content of earnings is reduced. Our earlier results on board characteristics suggest that the information content of earnings is unrelated to the size of the full board. This does not appear to be the case for audit committee size. In the final regression of table 5 we find that the information content of earnings is negatively related to audit committee size (p-value = .046).14 This result suggests that the market perceives the earnings number to be more informative when audit committees are smaller. One question that arises is the potential for nonlinearities. For example, Klein (2002) finds that evidence of non-linearity in the relation between audit committee independence and earnings quality (as captured by discretionary accruals). She finds that discretionary accruals are greatest when audit committees are least independent and that discretionary accruals decline with 20 audit committee independence. She does not find, however, any difference in discretionary accruals for firms with completely independent audit committees and those that do not have completely independent audit committees. Those results suggest the possibility that the relation between the information content of earnings and audit committee independence may be nonlinear. To address the possibility that the relation is non-linear we add a term to our previous analysis equal to the squared independence of the audit committee. The squared term offers one mechanism by which we might identify a non-linear relation between the information content of earnings and audit committee composition. The coefficient on this new term is negative, but statistically insignificant, suggesting that non-linearity (at least as it would be captured by a squared term) is not an issue in the relation between audit committee independence and earnings informativeness. Similar to the analysis of full boards, our focus on audit committee suggests audit committee characteristics are relevant to the information content of earnings. Audit committee independence appears to be important in that the information content of earnings increases with the degree of audit committee independence. 3.5 The incremental effect of board and audit committee characteristics on the information content of earnings Our previous analysis focuses on board and audit characteristics individually. We now turn our attention to the incremental effect each has on the information content of earnings. In table 6 we regress earnings announcement returns on board and audit committee characteristics together. Our earlier inferences with respect to the relevance of board characteristics remain. 14 The relation is statistically insignificant using analyst forecasts as our model of earnings expectations. 21 Even while controlling for audit committee characteristics, board independence and board activity remain important determinants of the information content of earnings. The information content of earnings is positively related to board independence and (weakly) to board activity (pvalues = < .001 and .135, respectively). These results suggest that the board plays an important role in the market’s perception of the information content of earnings. Audit committee size continues to affect the information content of earnings when we control for the characteristics of the full board. Specifically, the negative and significant (p-value = .001) coefficient on audit committee size suggests that the market finds the earnings of firms with smaller audit committees to be more informative than the earnings of firms with larger audit committees. This result is consistent with smaller committees being more effective in their governance roles and thus, the market lends greater credibility to the earnings numbers of firms with small audit committees relative to the earnings of firms with large audit committees. It is also interesting to note in table 6 that audit committee independence is unrelated to the information content of earnings. This result runs counter to our earlier result in table 5. Table 7 suggests a possible explanation. In table 7, the Pearson (Spearman) correlation between board independence and audit committee independence is .566 (.497). The coefficient on audit committee independence represents the influence of audit committee independence that is not related to full board independence on the information content of earnings. A large portion of audit committee independence is explained by full board independence. The portion of audit committee independence that is unrelated to full board independence has no effect on the information content of earnings. Results in table 1 suggest that audit committees tend to be more independent than full boards. That is, there is some level of audit committee independence above and beyond that of the full board. That result combined with the results in table 6 suggests 22 little benefit associated with the incremental independence of audit committees—at least with respect to the information content of earnings. Fama and Jensen (1983b), Klein (1998), and Klein (2002) suggest a possible explanation. They contend that non-independent directors, such as former employees, offer expertise about the firm that is unavailable from independent outside directors. In the context of the information content of earnings, it may be the case that a former employee (e.g., a retired CFO) has detailed knowledge of a firm’s accounting systems and internal controls. That knowledge leaves the former employee with a greater ability to “ask the right questions.” That specialized knowledge may make up for any perceived loss in independence. In our sample the median number of audit committee members is three. Less-than-complete audit committee independence implies (at most) independence of 66.7%. As noted earlier, boards in our sample are 63.3% independent, on average. Our results appear to suggest that complete independence of the audit committee does not noticeably increase the information content of earnings. Again, one possible explanation is that complete independence could exclude potential committee members with valuable knowledge with respect to the accounting systems and internal controls of the firm. That is, a better informed, but slightly less independent audit committee may produce earnings numbers with comparable information content to those produced by a more independent, but less well informed counterpart. It should be noted, however, that our results are merely suggestive and caution should be taken in their interpretation. 4. Conclusions Prior work suggests that a number of corporate governance characteristics influence earnings management and potentially the informativeness of earnings. We extend the literature 23 by examining earnings informativeness as a function of corporate governance characteristics. Specifically, we focus on how the market reacts to earnings surprises conditioned on the independence, activity, and size of both boards and audit committees. Furthermore, we examine whether or not separating the role of CEO and Chairman of the Board is related to the informativeness of earnings. Our findings suggest that characteristics of both the full board and the audit committee are indeed related to the information content of earnings. Focusing on the full board we find that earnings informativeness increases with board independence and activity, but is unrelated to board size. Furthermore, the relationship between board independence and activity is complex. The information content of earnings increases with independence regardless of board activity, but it increases more if board activity is low. In contrast, the information content of earnings increases with activity regardless of independence, but increases more if board independences is low. We also find that separating the roles of CEO and Chair is associated with increased earnings informativeness. When focusing on the audit committee our results differ somewhat. Specifically, when examining only audit committee independence and activity, we find that the information content of earnings increases with audit committee independence and activity. We find strong evidence that smaller audit committees are associated with earnings more informative earnings. When we examine board characteristics and audit committee characteristics together, however, we find that audit committee independence unrelated to the independence of the full board is unrelated to the information content of earnings. Put another way, some element of audit committee independence is explained by board independence. Audit committee independence beyond this level appears to have no effect on earnings informativeness. 24 Our findings have implications for regulators. First, our results inform the debate over proposed changes pertaining to board and audit committee membership. Moreover, our analysis highlights that the relationships are complex in that elements of the full board and the audit committee appear to interact. One interpretation of our findings is that audit committee independence beyond that of the full board does not increase the information content of earnings. We conjecture that a single non-independent audit committee member (such as a former employee) may offer expertise with respect to the accounting systems and internal controls that would be unavailable if the audit committee was completely independent. If this is the case, mandating that audit committees be completely independent may offer little incremental benefit with respect to the information content of earnings. 25 REFERENCES Abbot, Lawrence J., Susan Parker, Gary Peters, and Kannan Raghunandan, 2001, The effect of Audit Committee Characteristics and Non-Audit Fees on Audit Fees, working paper, University of Memphis. Antle, Rick, Elizabeth Gordon, Ganapathi Narayanmoorthy, and Ling Zhui, 2002, The joint determination of audit fees, non-audit fees, and abnormal accruals, working paper, Yale School of Management. Ashbaugh, Hollis, Ryan LaFond, and Brian W. Mayhew, 2003, Do non-audit services compromise auditor independence? Further evidence, forthcoming Accounting Review. Ball, Ray and Phillip Brown, 1968, An empirical evaluation of accounting income numbers. Journal of Accounting Research, 6, 159–177. Beaver, William H., 1968, The information content of annual earnings announcements. Journal of Accounting Research Supplement, 6, 67–92. Beasley, Mark, 1996, An empirical analysis of the relation between the board of director composition and financial statement fraud, The Accounting Review 71, 443-465. Beaver, William H. 2002, Perspectives on recent capital market research, Accounting Review, 77, 2, 453-474. Bradbury, Michael E., 1990, The incentives for voluntary audit committee formation, Journal of Accounting and Public Policy 9, 19-36. Brickley, James, Jeffrey Coles and R. Terry, 1997, Leadership structure: Separating the CEO and chairman of the board, Journal of Corporate Finance, 3, 3, 189-220. Brickley, James, Jeffrey Coles, and R Terry, 1994, Outside directors and the adoption of poison pills, Journal of Financial Economics 35, 371-390. Bushman, Robert M., and Abbie J. Smith, 2003, Financial accounting information and corporate governance, Journal of Accounting and Economics, 32, 237-333. Byrd, John W., and Kent Hickman, 1992, Do outside directors monitor management? Evidence from tender offer bids, Journal of Financial Economics, 32, 2, 195-221. Carcello, Joseph V., Dana R. Hermanson, Terry L. Neal, and Richard R. Riley, Jr., 2000, Board characteristics and audit fees, working paper, University of Tennessee. Carcello, Joseph V., and Terry L. Neal, 2000, Audit committee composition and auditor reporting, Accounting Review, 75, 4, 453-467. 26 Carcello, Joseph V., and Terry L. Neal, 2003, Audit committee characteristics and auditor dismissals following “new” going-concern reports, Accounting Review, 78, 1, 95-117. Choi, Sung K., and Debra C. Jeter, 1992, The effects of qualified audit opinions on earnings response coefficients, Journal of Accounting and Economics, 15, 2, 229-248. Chung, Hyeesoo and Sanjay Kallapur, 2003, Client importance, non-audit fees, and abnormal accruals, forthcoming Accounting Review. Commission on Public Trust and Private Enterprise, 2003, Findings and Recommendation, Conference Board, New York, Dechow, Patricia, Richard Sloan, and Amy Sweeney, 1996, Causes and consequences of earnings manipulation: An analysis of firms subject to enforcement actions by the SEC, Contemporary Accounting Research 13, 1-36. DeFond Mark L., and Chul W. Park, 1997, Smoothing income in anticipation of future earnings, Journal of Accounting and Economics 23, 115-139. Easton, Peter D., and Mark E. Zmijewski, 1989, Cross-sectional variation in the stock market response to accounting earnings announcements, Journal of Accounting and Economics, 117141. Eichenseher, John W. and David Shields, 1985, Corporate director liability and monitoring preferences, Journal of Accounting and Public Policy 4, 13-31. Fan, Joseph, P.H., and T.J. Wong, 2002, Corporate ownership and the informativeness of accounting earnings in East Asia, Journal of Accounting and Economics, 33, 401-425. Fama, Eugene F. and Michael C. Jensen, 1983a, Agency problems and residual claims, Journal of Law and Economics 26, 327-349. Fama, Eugene F. and Michael C. Jensen, 1983b, Separation of ownership and control, Journal of Law and Economics 26, 301-325. Francis, Jennifer, Ryan LaFond, Per Olsson, and Katherine Schipper, 2002, The market pricing of earnings quality. Working paper, Duke University. Francis, Jennifer, Katherine Schipper, and Linda Vincent, 2003, Earnings and dividend informativeness when cash flow rights are separated from voting rights, working paper, Duke University. Frankel, Richard M., Marylin F. Johnson, and Karen K. Nelson, 2002, The relationship between auditors' fees for non-audit services and earnings quality, Accounting Review, 77, 71-105. 27 Goyal, Vidhan K., and Chul Park, 2002, Board leadership structure and CEO turnover, Journal of Corporate Finance, 8, 1, 49-66. Healy, Paul, and James Wahlen, 1999, A review of the earnings management literature and its implications for standard setting, Accounting Horizons, 13, 4, 365-383. Hribar, Paul, and Daniel W. Collins, 2002, Errors in estimating accruals: Implications for empirical research, Journal of Accounting Research, 40, 1, 105-134. Imhoff, Eugene A. Jr.,1992, The relation between perceived accounting quality and economic characteristics of the firm, Journal of Accounting and Public Policy, 11, 97-118. Imhoff, Eugene A. Jr., and Gerald J. Lobo, 1992, The effect of ex-ante earnings uncertainty on earnings response coefficients, Accounting Review, 67, 2, 427-439. Jensen, Michael C., 1983, Organization theory and methodology, Accounting Review, 58,2, 319339. Jensen, Michael C. and William H. Meckling, 1976, Theory of the firm: Managerial behavior, agency costs and ownership structure, Journal of Financial Economics 3, 305-360. Klein, April, 1998, Firm performance and board committee structure, Journal of Law and Economics 41, 275-303. Klein, April, 2002, Audit committee, board of director characteristics and earnings management, Journal of Accounting and Economics, 33, 375-400. Kothari, S.P., 2001. Capital markets research in accounting, Journal of Accounting and Economics 31, 105-231. Kothari, S.P., A.J. Leone, and C.E. Wasley, 2002, Performance matched discretionary accrual measures, Working Paper, MIT. Larcker, David F., and Richardson Scott A., 2003, Corporate governance, fees for non-audit services and accrual choices, working paper, the Wharton School. McNichols, Maureen F., 2000, Research design issue in earnings management studies, Journal of Accounting and Public Policy, 19, 313-345. Menon, Krishnagopal and Joanne Deahl Williams, 1994, The use of audit committees for monitoring, Journal of Accounting and Public Policy 13, 121-139. 28 Peasnell, Ken, Peter Pope, and Steven Young, 2001, Board monitoring and earnings management: Do outside directors influence abnormal accruals?, working paper, Lancaster University Management School. Pike, Joel, E., 2003, Audit quality and the provision of non-audit services: Evidence from the property-casualty insurance industry, Working paper, university of Wisconsin-Madison. Pincus, Karen, Mark Rusbarsky, and Jilnaught Wong, 1989, Voluntary formation of corporate audit committees among NASDAQ firms, Journal of Accounting and Public Policy 8, 239-265. Smith, Clifford W., Jr. and Ross L. Watts, 1992, The investment opportunity set and corporate financing, dividend, and compensation policies, Journal of Financial Economics 32, 263-292. Teoh, Siew Hong and T.J. Wong, 1003, Perceived auditor quality and the earnings response coefficient, Accounting Review, 68, 2, 346-366. Titman, Sheridan and Brett Trueman, 1986, Information quality and the valuation of new issues, Journal of Accounting and Economics 8, 159-172. Treadway Commission, 1987, Report of the National Commission on Fraudulent Financial Reporting. Vafeas, N., 1999, Board meeting frequency and firm performance, Journal of Financial Economics, 53, 1, 113-142. Vafeas, N., 2000, Board structure and the informativeness of earnings, Journal of Accounting and Public Policy, 19, 139-160. Warfield, Terry D., Wild, John J., and Kenneth L. Wild, 1995, Managerial ownership, accounting choices, and informativeness of earnings, Journal of Accounting and Economics, 20, 61-91. Weisbach, Michael, 1988, Outside directors and CEO turnover, Journal of Financial Economics, 20, 431-460. Wild, John J., 1994, Managerial accountability to shareholders: Audit committees and the explanatory power of earnings for returns, British Accounting Review, 26, 353-374. Wild, John J., 1996, The audit committee and earnings quality, Journal of Accounting, Auditing & Finance 11, 2, 247-276. Yermack, David, 1996, Higher market valuation for firms with a small board of directors, Journal of Financial Economics 40, 185-211. 29 Table 1 Descriptive Statistics Panel A: Board and Audit Committee Composition Statistics Mean Number of Board Members Employee Directors Affiliated Directors Independent Directors Percent of Independent Directors Percent of Firms with a Separate CEO and Chairman Number of Meetings per Year Percent of Equity Held by Directors and Officers Number of Audit Committee Members Percent of Independent Audit Committee Members Number of Audit Committee Meetings Per Year 9.39 1.97 1.43 6.00 63.31 26.91 6.64 12.75 3.64 88.15 3.97 Median 9.00 2.00 1.00 6.00 66.67 --6.00 6.20 3.00 100.00 4.00 Std. Dev 2.97 1.18 1.49 2.74 18.45 --2.23 17.01 1.17 19.59 1.40 Min. 4.00 0.00 0.00 0.00 0.00 --1.00 0.00 1.00 0.00 1.00 Max. 30.00 10.00 9.00 22.00 93.75 --12.00 97.60 11.00 100.00 7.00 Our board data is for 1,241 firms that filed proxy statements in 2001 (from the S&P Super 1,500 and other large firms covered by the Investor Responsibility Research Center (IRRC). Employee directors are those directors who are current employees of the firm. Affiliated directors are those directors who, while not current employees of the firm, have a business relationship with the firm or were previously employed by the firm. Independent directors are those directors who have no other relation with the firm except in their capacity as a director. 30 Table 1, cont. Descriptive Statistics Panel B: Firms’ Financial Statistics Mean Total Assets ($ in millions) Book Value of Stockholders’ Equity ($ in million) Basic Earnings Per Share Excluding Extraordinary Items Market Value of Equity ($ in millions) Market to Book Ratio Leverage Size ($ in millions) 9,566.40 1,871.05 1.34 7,873.88 3.52 0.20 11,322.22 Median 1,394.33 519.46 1.34 1,280.50 2.21 0.18 1,906.12 Std. Dev. 44,609.05 5,057.48 1.86 29,331.98 9.99 0.18 45,876.25 Min. 27.87 -1,160.70 -11.48 11.40 -81.81 0.00 39.45 Max. 902,210.00 70,757.00 14.34 476,140.56 231.44 1.42 731,623.25 Our financial data is for 1,241 firms that filed proxy statements in 2001 (from the S&P Super 1,500 and other large firms) and comes COMPUSTAT. Leverage is long-term debt divided by total assets. Size is the book value of both current and long-term debt plus the book value of preferred stock plus the market value of common equity. 31 Table 2 Descriptive Statistics of Unexpected Earnings by Decile Unexpected Earnings Decile 1 2 3 4 5 6 7 8 9 10 Mean -3.27 -1.02 -0.48 -0.21 -0.01 0.04 0.29 0.49 0.87 2.33 Median -2.45 -0.99 -0.471 -0.22 0.00 0.15 0.28 0.48 0.83 2.15 Std. Dev. 2.06 0.22 0.11 0.07 0.05 0.04 0.05 0.07 0.19 0.91 Our data is for 1,241 firms that filed proxy statements in 2001 (from the S&P Super 1,500 and other large firms) as covered by the Investor Responsibility Research Center (IRRC). Earnings expectations are calculated using a seasonal random walk model. Unexpected earnings are ranked into deciles (1=smallest and 10=largest). Descriptive statistics are calculated within unexpected earnings deciles. 32 Table 3 Descriptive Statistics by Unexpected Earnings Deciles Decile of Unexpected Earnings 1 2 3 4 5 6 7 8 9 10 Percent of Independent Directors 63.02 (66.67) 63.19 (66.67) 61.97 (62.50) 62.41 (66.67) 63.27 (66.67) 64.12 (70.00) 62.97 (63.64) 65.21 (66.67) 62.86 (66.67) 64.12 (66.67) Number of Board Meetings 7.17 (7.00) 6.49 (6.00) 6.71 (6.00) 6.77 (6.50) 6.60 (6.00) 6.50 (6.00) 6.37 (6.00) 6.66 (6.00) 6.45 (6.00) 6.70 (6.00) Percent of Independent Directors on Audit Committee 86.01 (100.00) 88.37 (100.00) 90.52 (100.00) 85.62 (100.00) 91.46 (100.00) 90.09 (100.00) 86.65 (100.00) 86.61 (100.00) 87.94 (100.00) 88.29 (100.00) Number of Audit Committee Meetings 4.40 (4.00) 4.06 (4.00) 3.95 (4.00) 3.79 (4.00) 3.67 (4.00) 3.80 (4.00) 4.01 (4.00) 4.06 (4.00) 3.91 (4.00) 4.02 (4.00) Audit Committee Size 3.59 (3.00) 3.531 (3.00) 3.70 (3.00) 3.60 (3.00) 3.68 (3.00) 3.86 (4.00) 3.71 (3.00) 3.64 (3.00) 3.60 (3.00) 3.50 (3.00) Board Size 9.10 (9.00) 9.24 (9.00) 8.98 (8.00) 9.40 (9.00) 9.40 (9.00) 9.89 (10.00) 9.90 (10.00) 9.88 (9.00) 9.15 (9.00) 9.00 (9.00) Our data is for 1,241 firms that filed proxy statements in 2001 (from the S&P Super 1,500 and other large firms) as covered by the Investor Responsibility Research Center (IRRC). Earnings expectations are calculated using a seasonal random walk model. Ten deciles are created based on unexpected earnings. Means (medians) for various board and audit committee characteristics are then calculated within unexpected earnings deciles. 33 Table 4 Full Board Regression Results Intercept UEP BrdIndep BrdMeetings BrdIndep*BrdMeetings BrdSize SepCEO&Chair D&OHold lnSize MktBk StdDev Adj. R2 P-value for F-test 0.022 (0.1120) 0.000 (0.6670) 0.020 (0.0222) 0.000 (0.8877) -0.354 (0.5421) 0.0256 0.0001 0.027 (0.0488) 0.000 (0.9079) 0.017 (0.0519) 0.001 (0.6686) -0.528 (0.3654) 0.0211 0.0001 -0.000 (0.2820) 1.000 (0.0001) -0.179 (0.0255) 0.002 (0.0019) -0.009 (0.0455) 0.000 (0.0170) 0.003 (0.3924) 0.032 (0.0266) -0.000 (0.5609) 0.013 (0.1275) 0.001 (0.4126) -0.708 (0.2318) 0.0172 0.0002 1.000 (0.0001) -0.063 (0.3732) 1.000 (0.0001) -0.059 (0.4224) 0.999 (0.0001) -0.395 (0.0046) 0.006 (0.0015) 0.027 (0.0570) -0.001 (0.0267) -0.002 (0.7105) 0.025 (0.0823) 0.001 (0.3478) 0.020 (0.0292) 0.000 (0.9618) -0.501 (0.4005) 0.0289 0.0001 Our sample consists of 1,241 firms that filed proxy statements in 2001 (from the S&P Super 1,500 and other large firms) as covered by the Investor Responsibility Research Center (IRRC). The dependent variable is the cumulative abnormal return from day -1 through day +1 where day 0 is the earnings announcement date. Unexpected earnings are calculated using a seasonal random walk model. All independent variables (except unexpected earnings) are interacted with unexpected earnings. UEP is the unexpected earnings scaled by the price at the fiscal year-end. BrdIndep is the number of independent board members scaled by the total number of board members. BrdSize is the total number of board members. BrdMeetings is the number of meetings held by the board during the year. SepCEO&Chair is a dummy variable equal to 1 if the CEO and Chairman are different people and zero otherwise. D&OHold is the percentage of stock held by directors and officers. lnSize is the natural log of (current debt + long-term debt + market value of equity + preferred stock). MktBk is the market to book ratio at fiscal year-end. StdDev is the standard deviation of the previous fiscal year’s returns. p-values are in parentheses. 34 Table 5 Audit Committee Regression Results Intercept UEP AudIndep AudMeetings AudIndep*AudMeetings AudSize SepCEO&Chair D&OHold lnSize MktBk StdDev Adj. R2 P-value for F-test 0.042 (0.0029) -0.000 (0.7514) 0.018 (0.0561) 0.001 (0.5219) -0.468 (0.4456) 0.0180 0.0002 0.041 (0.0069) -0.000 (0.6589) 0.016 (0.1024) 0.001 (0.5242) -0.644 (0.2830) 0.0169 0.0003 -0.000 (0.6068) 1.000 (0.0001) -0.130 (0.1513) 0.001 (0.1542) -0.009 (0.4033) 0.000 (0.2428) -0.008 (0.4384) 0.035 (0.0087) -0.000 (0.4393) 0.015 (0.0774) 0.001 (0.3284) -0.935 (0.0988) 0.0171 0.0002 1.000 (0.0001) -0.059 (0.4109) 1.000 (0.0001) -0.015 (0.8368) 1.000 (0.0001) -0.368 (0.0237) 0.004 (0.0165) 0.065 (0.0525) -0.001 (0.0560) -0.024 (0.0458) 0.017 (0.3466) -0.000 (0.7894) 0.020 (0.0435) 0.001 (0.3710) -0.295 (0.6339) 0.0218 0.0001 Our sample consists of 1,241 firms that filed proxy statements in 2001 (from the S&P Super 1,500 and other large firms) as covered by the Investor Responsibility Research Center (IRRC). The dependent variable is the cumulative abnormal return from day -1 through day +1 where day 0 is the earnings announcement date. Unexpected earnings are calculated using a seasonal random walk model. All independent variables (except unexpected earnings) are interacted with unexpected earnings. UEP is the unexpected earnings scaled by the price at the fiscal year-end. AudIndep is the percentage of the audit committee that consists of independent directors. AudSize is the number of board members serving on the audit committee. AudMeetings is the number of times per year the audit committee meets. lnSize is the natural log of (current debt + long-term debt + market value of equity + preferred stock). MktBk is the market to book ratio at fiscal year-end. StdDev is the standard deviation of the previous fiscal year’s returns. p-values are in parentheses. 35 Table 6 Full Board and Audit Committee Regression Results Intercept UEP BrdIndep BrdMeetings BrdIndep*BrdMeetings BrdSize AudIndep AudMeetings AudIndep*AudMeetings AudSize SepCEO&Chair D&OHold lnSize MktBk StdDev Adj. R2 p-value for F-test 1.000 (0.0001) -0.405 (0.0308) 0.008 (0.0001) 0.022 (0.1353) -0.001 (0.0532) 0.005 (0.4253) 0.000 (0.9100) 0.036 (0.3007) -0.000 (0.3665) -0.049 (0.0011) -0.037 (0.1201) 0.001 (0.0978) 0.022 (0.0294) 0.000 (0.7440) -0.683 (0.2771) 0.0498 0.0001 Our sample consists of 1,241 firms that filed proxy statements in 2001 (from the S&P Super 1,500 and other large firms) as covered by the Investor Responsibility Research Center (IRRC). The dependent variable is the cumulative abnormal return from day -1 through day +1 where day 0 is the annual earnings announcement date. Unexpected earnings are calculated using a seasonal random walk model. All independent variables (except unexpected earnings) are interacted with unexpected earnings. UEP is the unexpected earnings scaled by the price at the fiscal year-end. BrdIndep is the number of independent board members scaled by the total number of board members. BrdSize is the total number of board members. BrdMeetings is the number of meetings held by the board during the year. AudIndep is the percentage of the audit committee that consists of independent directors. AudSize is the number of board members serving on the audit committee. AudMeetings is the number of times per year the audit 36 committee meets. SepCEO&Chair is a dummy variable equal to 1 if the CEO and Chairman are different people and zero otherwise. D&OHold is the percentage of stock held by directors and officers. lnSize is the natural log of (current debt + long-term debt + market value of equity + preferred stock). MktBk is the market to book ratio at fiscal year-end. StdDev is the standard deviation of the previous fiscal year’s returns. p-values are in parentheses. 37 Table 7 Full Board and Audit Committee Correlations BrdIndep BrdIndep BrdMeetings BrdSize AudIndep AudMeetings AudSize SepCEO&Chair D&OHold 0.171 (0.0001) 0.112 (0.0001) 0.497 (0.0001) 0.051 (0.0739) 0.344 (0.0001) -0.164 (0.0001) -0.448 (0.0001) 0.129 (0.0001) 0.048 (0.0936) 0.223 (0.0001) 0.182 (0.0001) -0.021 (0.4685) -0.234 (0.0001) -0.036 (0.2081) 0.082 (0.0040) 0.502 (0.0001) -0.010 (0.7274) -0.178 (0.0001) -0.001 (0.9721) 0.005 (0.8661) -0.063 (0.0277) -0.152 (0.0001) 0.072 (0.0109) -0.002 (0.9445) -0.116 (0.0001) -0.095 (0.0008) -0.300 (0.0001) 0.085 (0.0027) BrdMeetings 0.159 (0.0001) BrdSize 0.092 (0.0012) 0.132 (0.0001) AudIndep 0.566 (0.0001) 0.049 (0.0852) -0.009 (0.7385) AudMeetings 0.068 (0.0163) 0.237 (0.0001) 0.102 (0.0003) 0.009 (0.7413) AudSize 0.309 (0.0001) 0.166 (0.0001) 0.487 (0.0001) 0.045 (0.1137) 0.060 (0.0332) SepCEO&Chair -0.150 (0.0001) -0.024 (0.3934) -0.015 (0.6006) -0.071 (0.0124) -0.006 (0.8365) -0.085 (0.0029) D&OHold -0.414 (0.0001) -0.165 (0.0001) -0.066 (0.0195) -0.147 (0.0001) -0.092 (0.0012) -0.188 (0.0001) 0.057 (0.0442) This table shows the Pearson correlation coefficients in the upper triangle and Spearman correlation coefficients in the lower triangle. P-values are in parentheses. Our sample consists of 1,241 firms that filed proxy statements in 2001 (from the S&P Super 1,500 and other large firms) as covered by the Investor Responsibility Research Center (IRRC). BrdIndep is the number of independent board members scaled by the total number of board members. BrdSize is the total number of board members. BrdMeetings is the number of meetings held by the board during the year. AudIndep is the percentage of the audit committee that consists of independent directors. AudSize is the number of board members serving on the audit committee. AudMeetings is the number of times per year the audit committee meets. SepCEO&Chair is a dummy variable equal to 1 if the CEO and Chairman are different people and zero otherwise. D&OHold is the percentage of stock held by directors and officers. p-values are in parentheses. 38 ...
View Full Document

This note was uploaded on 02/24/2011 for the course ECON 201 taught by Professor Beasly during the Spring '11 term at American Intl. University.

Ask a homework question - tutors are online