Director Capital and Corporate Disclosure Quality - CiteSeerX

6 downloads 238188 Views 368KB Size Report
Mar 26, 2009 - accounting reports, we argue that audit committee director capital should be ..... with a bachelor's degree, an MBA, a master's degree, a PhD or ...
Director Capital and Corporate Disclosure Quality David M. Reeba and Wanli Zhaob a

Fox School of Business, Temple University, Philadelphia, PA 19122 b Worcester Polytechnic Institute, Worcester, MA 01609 March 26, 2009

Abstract We investigate the role of director human and social capital on board oversight, focusing on one specific component of board efficacy, namely corporate disclosure quality. One role of the board involves setting corporate disclosure policy, thereby deciding upon the depth and degree of the firm’s disclosures. Since the audit committee is directly charged with overseeing the integrity of the accounting reports, we argue that audit committee director capital should be an especially important component of corporate disclosure quality. To test these ideas, we manually collect data on several different director characteristics and develop indices to measure the networking, educational, and experience capital of the board and audit committee. The results indicate that board capital is positively related to corporate disclosure quality, with differing key attributes for inside and outside directors. We also find that audit committee capital is associated with greater disclosure quality and lower external audit fees. In subsequent tests, we develop an instrument of director capital based on corporate headquarters accessibility and continue to find that board and audit committee capital are positively related to corporate disclosure quality. Further tests indicate that board and audit committee capital are related to shareholder value, which is, to some extent, attributable to greater disclosure quality. Interestingly, we also observe a substantial gap between accounting disclosures on director capital and the information that we collected; with this gap being positively related to shareholder value. Taken as whole, the evidence implies that board efficacy is a function of director capital because it affects the corporate disclosure quality. JEL: G34, M41 Corresponding author. Tel: 508-831-5185, Email: [email protected]

b

We would like to thank Stephen Asare, Zhaohui Chen, Victoria Dickinson, Ken Kopecky, Connie Mao, Ram Mudambi, Surjit Tinaikar, and seminar participants at the University of Florida for their helpful comments and suggestions.

1. Introduction Does director ability affect governance effectiveness? Surprisingly, little is known about the answer to this question as there is scarce empirical research on the relation between director ability and board effectiveness. Recent regulatory activity implies that one particular aspect of director is a critical aspect of governance efficacy; specifically that financial expertise is an important consideration for audit committee competency (e.g. Sarbanes-Oxley Act of 2002).

Empirical

research supports this notion, finding that shareholders benefit from financial literacy and expertise on the audit committee (Karamanou and Vafeas, 2005; Defond, Hann, and Hu, 2005). More generally, Becker (1964) suggests the productivity of agents (directors) depends on a broad range of attributes that comprise their human and social capital. Yet, the academic and popular press have primarily centered on director bias or incentives as the key issue in evaluating director efficacy. Although, potential conflicts of interest are important, we suggest that in the day-to-day operations of the board, director capital is of principal importance. One reason we believe that board capital is critical to shareholders is because they have little recourse for poor decision-making due to limited ability or skill. In contrast, in a robust legal system with relatively strong safeguards, shareholders and government regulators can sue or prosecute directors for board errors due to conflicts of interest. In addition, economic theory suggests that short- and long-term incentives can mitigate moral hazard problems and that agent ability is equally important to outside shareholders (Milgrom and Roberts, 1982; Fudenberg and Levine, 1992; Ely and Välimäki, 2003). Moreover, Hermalin and Weisbach (2006) suggest that while penalties or regulations can affect potential conflicts of interest in many circumstances, they have little impact on board ability. Casual inspection reveals that boards vary across firms with substantial differences among directors in terms of their human and social capital. Two firms with the same percentage of 1

independent directors, the same number of directors, and similar incentive mechanisms may not be equally effective because their directors have differing expertise, skills, and social contacts (Carpenter and Westphal, 2001). Holmström (1999) observes that the agent ability is a key consideration in purchasing labor services (i.e. hiring directors) and that reputation concerns are increasing in agent ability. In this context, we argue that the human and social capital of both the board and audit committee are important attributes for effective governance. More specifically, we posit that the education, work experience, and business networks of the board and audit committee are positively related to governance efficacy. Our primary analysis centers on one specific component of governance or board efficacy, namely corporate disclosure quality. We focus on the role of the board in setting corporate disclosure policy, thereby defining the depth and degree of disclosures for the firm. Starting from the perspective that high ability directors are concerned about safeguarding their own reputations by reducing asymmetric information between managers and outside investors, we test the notion that board of director capital is related to corporate disclosure quality. Since the audit committee is directly charged with overseeing the integrity of the accounting reports, we argue that audit committee director capital should be an especially important driver of corporate disclosure quality. We also examine the relation between director capital and shareholder value to examine the idea that director human and social capital influence governance efficacy. To empirically test our hypotheses that board and audit committee capital are related to corporate disclosure quality, we manually collect data on three categories of director characteristics and develop these measures into indices to capture director networking, educational, and experience capital. We compute total board and audit committee capital as the sum of our networking capital, educational capital, and experience capital measures. We also use an alternative approach to develop composite measures of board capital based on factor analysis. Our primary set of tests focus on the 2

relation between board (audit committee) capital and corporate disclosure quality, with further tests investigating the impact of each category of board (audit committee) capital. We measure corporate disclosure quality using a variation of the approach suggested by Kim and Verrecchia (2001) who observe that traded share volume has less association with share price in firms with poor disclosure quality. More specifically, we use an industry adjusted version of their measure to further isolate the impact of firm characteristics on disclosure quality from industry factors. We also use an alternative approach developed in Anderson, Duru, and Reeb (2008) who suggest that corporate opacity incorporates both a market scrutiny component and a disclosure quality component. They adopt a process to isolate the corporate disclosure quality from the overall corporate opacity using governance and accounting quality data. Both approaches provide similar inferences. We also examine the relation between audit committee capital and corporate disclosure quality. The audit committee is arguably one of the most important sub-committees of the board because it is tasked with overseeing the integrity of the financial reporting system. In addition, auditors explicitly consider audit committee composition in evaluating audit risk (Krishnan, 2005), suggesting that audit committee capital is an important component of disclosure quality. As such, we argue that any disclosure quality aspects of director capital should be evident in this sub-committee of the board. These arguments also suggest that audit fees should be decreasing in audit committee capital if auditors determine that high ability directors improve the integrity of the financial accounting process.

In another set of tests we differentiate between outside and inside directors

(e.g. inside networking capital, outside educational capital) in our examination of director capital and corporate disclosure quality. Using a sample comprised of the industrial firms (615 firms; 6,160 directors) in the Russell 1000 Index (as of 2003 and 2005), we find a significant and positive relation between board capital and corporate disclosure quality. After controlling for firm and industry characteristics, our tests 3

indicate that a one standard deviation increase in board capital is associated with about 21% greater corporate disclosure quality. We also find that while the educational and experience capital of the board as a whole has a significant relation with corporate disclosure quality, networking capital appears to have little relation to disclosure quality. Focusing on audit committee capital we find it is positively related to corporate disclosure quality. Specifically, we find that a one standard deviation greater level of audit committee capital is associated with about 27% greater disclosure quality. Considering how external auditors price their services, we also find that auditor fees are decreasing in audit committee capital, which is consistent with the notion that director capital improves the reliability of the financial accounting process, thereby reducing auditor risk. Further tests indicate that the educational and experience capital of the audit committee as a whole is also significantly related to corporate disclosure quality, while the networking capital is insignificant. Although the results are consistent with the notion that director capital facilitates board oversight and improves corporate disclosure quality, an alternative perspective is quite plausible. Namely, board and audit committee capital could instead be a function of corporate disclosure quality, indicating that directors with greater human capital seek and accept positions in more transparent firms. A noteworthy aspect of this reverse causality concern is that it is the converse of the conventional endogeneity argument (e.g., Hermalin and Weisbach, 1998). To provide further insights into this endogeneity concern, we develop a distinctive instrumental variable for director capital based on the distance between the corporate headquarters and the nearest metropolitan airport. We posit that traveling costs to board activities may be an important issue to directors, as more remote locations require greater time and energy to participate in corporate meetings and

4

functions1. The underlying notion is that company location affects the likelihood of attracting high ability directors since better directors are likely to have busy agendas. We argue this variable is a potentially viable instrument because it is related to board capital but is not subject to managerial influence (at least in the short term). We find consistent empirical results using ordinary least squares, fixed effects regressions, and 2-stage-least-squares instrumental variable approach. Still, we cannot rule out the possibility that more transparent firms are able to hire directors with greater human and social capital. We also decompose the directors into outsiders and insiders and examine the effects of each type of director capital. We find dissimilar results for insiders and outsiders regarding the impact of their human capital on disclosure quality. For outside directors, the aforementioned experience capital is the salient issue in evaluating their impact on corporate disclosure quality. However, education and networking capital have little explanatory power for outside directors. In contrast, the only important attribute of inside directors on corporate disclosure quality is their educational capital. On average, a one standard deviation higher insiders’ educational capital is associated with about 8.36% better disclosure quality.

Thus, the results suggest that the attributes important for

inside and outside directors are quite different, with their education and experience capital being respectively the most salient issues in understanding corporate disclosure quality. Next, we examine the impact of director capital on firm performance and find a significant relation between board capital and shareholder value. Our results indicate that a one standard deviation greater board capital is associated with about 16.16% greater Tobin’s Q. We interpret this

1Several issues potentially mitigate the effectiveness of this instrument. For instance, directors may have access to private aircraft, meetings may be regularly scheduled in other undisclosed locations, and firms can use video conferencing technology, all which may serve to reduce the potential usefulness of this prospective instrument. We conduct a series of specification tests on our instrumental variable and the structural model (Larcker and Rusticus, 2008) that suggest it is an appropriate approach. Nevertheless it is difficult to infer the nature of causality.

5

evidence to indicate that investors recognize the greater governance efficacy associated with directors of greater capital. We also find evidence to suggest that the primary channel for better governance is greater disclosure quality. Focusing on audit committee capital we also find that it is positively related to firm performance. Specifically, we find that a one standard deviation greater level of audit committee capital is associated with about 9.73% better firm performance. Next, we repeat the analysis with our instrumental variable for director capital and find similar results, that is, both board and audit committee capital are related to greater firm performance. Interestingly and maybe more importantly, we also observe a substantial gap between accounting disclosures on director capital in the corporate SEC filings and the information that we collected from various other sources for the Russell 1000 industrial firms. Of the 19 variables that we collected for our analysis, the typical firm reports only 3 or 4 of these variables for their directors. Consider Tyson Foods for instance, they have 10 directors but the only director information in their financial statements is their name, their last 5-year work experience, and their current corporate board memberships in other firms. The other remaining 16 variables were collected from LexisNexis, google.com, The Dun and Bradstreet Reference Book of Corporate Management, Who’s Who in Finance and Industry and MergentOnline. Although, we were able to collect the data with substantial effort, time and confusion, our analysis suggests this important information to outside shareholders should be disclosed in the financial statements. We further test the performance implication of not just having greater board capital but on voluntary disclosing this information to outside investors in the accounting reports. We find that reporting this greater director capital is also important and that such information is associated with better firm performance. We conclude that shareholders may be better served with additional accounting disclosures concerning the human and social capital of corporate directors. Yet, it is puzzling why firms with greater board and audit committee capital

6

don’t routinely disclose greater information concerning director human capital. One potential explanation is that firms are concerned about other firms hiring their directors away. The findings of this paper have some important policy implications. The Sarbanes-Oxley Act seeks to improve corporate governance practices through greater board independence and legal liability under the assumption that governance crises stem from conflicts of interest. Our results suggest that while mandates about greater board independence may be important, director ability is at least as important to shareholders who are concerned about disclosure quality and firm performance. Thus, this study provides empirical evidence to document that better boards are related to greater governance efficacy and shareholder value. A second important aspect of the results is that they provide indirect evidence on the endogenous board argument articulated by Hermalin and Weisbach (1998).2 Hermalin and Weisbach (1998) observe that weak firms may attempt to hire the best directors in order to improve firm governance, implying a negative relation between director ability and firm performance. However, our empirical results indicate that director capital is positively related to both disclosure quality and shareholder value. Consequently, if causality runs counter to our hypothesis that director capital improves governance efficacy, and instead the best directors are simply attracted to the best firms, then the results are inconsistent with the conventional board endogeneity argument. A third implication is that while mandates about financial expertise appear to be effective as we find they are associated with greater disclosure quality, our evidence suggest that outside shareholders’ interest be promoted even more with other types of human capital that directors bring to the firm. Oddly enough, in most states there are greater restrictions and regulations concerning 2

Several prior studies indicate that director independence is an important element in protecting minority shareholders in specific instances (Brickley and James, 1987; Beasley, 1996; Carcello and Neal, 2000; and Klein, (2002). However, in general, there is little cross-sectional evidence of a board independence having a positive effect on firm performance (Hermalin and Weisbach, 1998). Recent empirical research finds the relation between firm performance and board independence varies across firms depending on corporate complexity and management influence (Coles, Daniel and Naveen, 2008).

7

the training and examination of barbers and cosmetologist than there are relating to board of director knowledge, skills, or qualifications. Finally, this paper raises the question of whether it would be appropriate to call for greater disclosures concerning director qualifications. We find a substantial gap between the human capital information disclosed in the proxy statements and the information that we were able to uncover about directors. As such, investors may be better served with additional disclosures about the human capital of both inside and outside directors of the firm. Overall, our analysis suggests that director capital is a crucial factor to governance efficacy and thus important to outside shareholders’ interest.

2. Motivation and Hypotheses Development The board of directors is considered one of the most important governance devices in monitoring the behavior of the CEO. Academic work on board effectiveness emphasizes the role of board composition in alleviating conflicts of interest, focusing on how board independence affects firm performance. For instance, there is substantial evidence on the role of independent directors in protecting outside shareholders in extraordinary situations such as with the integrity of accounting process, adoption of poison pills, and mergers and acquisitions (Beasley, 1996; Brickley, Coles and Terry, 1994; Cotter, Shivdasani and Zenner, 1997). While well-publicized cases of corporate fraud and malfeasance indicate the need for guarding shareholders’ interest and reducing bias,3 we argue that director capital should be of at least equal importance in the day-to-day operations of the firm. We posit that a crucial difference between bias and ability is that for board/director bias (or conflicts of interest problems), shareholders and government agencies can seek redress through the legal system, while there is limited recourse for 3

For instance, Ribstein (2002) suggests that the common belief is that in the Enron case, the directors failed to detect and prevent the fraud due to their close ties to the management. Vice Chancellor of the Court of Chancery Leo E. Strine (2002) suggests that “independent directors…have a trivial risk of legal liability”.

8

poor ability. Among class action litigation against boards of directors, almost all of them have accused the board of “breach of fiduciary duty” due to conflicts of interest (Helland, 2006). In addition, the monetary compensation for shareholders from settlements in cases of conflicts of interest is significant. For instance, according to Stanford Securities Class Action Clearinghouse, the WorldCom and Enron litigations have reached settlement agreements of $6.128 and $4.760 billion, respectively. Other high-profile cases see similar numbers. In contrast to conflicts of interests, there is no easy recourse for director errors due to insufficient ability or skill. Veasey and Di Guglielmo (2005) suggest that the legal system encourages “sound structure and process, good disclosure and fair elections” in the boardroom so as to establish director discretion and minimize liability. As such, the courts focus on the directors’ intention and loyalty to serve the shareholders, not the outcome of the decisions. Even “stupid” decisions should not be part of the legal liability of the directors (Veasey and Di Guglielmo, 2005) even though they do harm the shareholders’ value.4 In addition, recent SEC reforms increasingly seek harsher penalties for director misconduct that stems from director bias/conflicts of interest problems, but have little to regulate concerning director ability. In the absence of any regulation concerning director qualifications in public firms, there is little, if any remedy, for shareholders if directors are unskilled. The Sarbanes-Oxley Act (2002) requires that firms have “financial expert(s)” on their audit committee because those directors presumably are better able to prevent managerial misconduct on accounting issues. As an illustration, firms reacting to the Sarbanes-Oxley Act and SEC regulations 4

For instance, another high-profile case is the Disney’s board’s decision to hire the former president, Michael Ovitz, who was paid a no-fault termination package of $130 million after 1 year on the job. In subsequent attempts at litigation, none of the directors were found to breach the fiduciary duty (no conflicts of interest). However, the court did note that although the board’s decision clearly fell short of best practices it did not represent grounds for legal action because there was no bias in the decision-making process. The courts have consistently have ruled that the key issue in shareholders seeking legal redress is the intention and/or bias of the board’s decision-making, not their ability to make “good” decisions. The implication is that the judicial system does not offer recourse to shareholders concerning director ability.

9

explicitly state their criteria for director selection. Observations of these criteria show that firms not only require outside directors to be independent but to also possess a high level of working experience, education, professional training, and reputation. Such trends also have caught a growing attention from academics. Defond et al (2005) find evidence that financial expertise of the audit committee matters in respect to alleviating agency issues. Brickley et al (1999) suggest that firms do consider their abilities when selecting retired CEOs to serve on their boards. The implication is that director ability is an important component of board efficacy. Perhaps one of the most important tasks of the board is developing the firm’s corporate disclosure policy. Healy and Palepu (2001) observe that disclosure quality is a central issue for effective governance to outside shareholders. Dunn and Mayhew (2004) report that disclosure quality adds value to shareholders, which Klein (2002) suggests is a function of director characteristics. Furthermore, directors with greater human and social capital may have both ability and the incentive to reduce asymmetric information issues to order to protect their own reputations. Stein (1989) observes that fear of replacement can influence agent behavior; suggesting that high ability directors will be more willing to have more open disclosure strategies in comparison to lower ability directors. From an outside shareholder perspective, being able to accurately assess firm activity through meaningful financial reports is an important product of the governance process as it facilitates external monitoring and the market for corporate control (Sloan, 2001). If director human and social capital improves governance efficacy, then we expect to observe a positive relation between board capital and corporate disclosure quality, which leads to our first set of hypotheses: Hypothesis 1a: Board capital is positively related to corporate disclosure quality. Hypothesis 1b: Each of the individual board capital components (education, experience, and networking capital) is positively related to corporate disclosure quality.

10

We also examine audit committee capital as it may be the most important monitoring device of managerial opportunism within the board (Klein, 1998). One charge of the audit committee is overseeing the reliability of the financial accounting process (Anderson et al. 2004). The audit committee serves to maintain the integrity of financial reporting process and the compliance with legal and regulatory requirements. Kothari et al. (2002) observe that accounting information can be considered reliable if it is free from error or bias (i.e. it is accurate). Prior research on audit committee focus on audit committee composition and generally concludes that a more independent audit committee may be associated with less corporate fraud and earnings manipulations, as well as higher financial reporting integrity (Beasley, 1996; Klein, 2002; Anderson et al, 2004;). Recent research and regulatory activity also emphasize the importance of audit committee competence in protecting shareholder interests. Anderson et al (2004) suggest that audit committee constitution is an important element in financial reporting integrity, leading to a lower cost of debt. Defond et al (2005) report that audit committee members’ financial expertise is an important aspect of monitoring effectiveness. Karamanou & Vafeas (2005) also find that a financially literate audit committee improves the information disclosure quality. Further, to be better able to search for, recognize, and identify managerial misconduct from complex financial reporting and managerial presentations and proposals, the audit committee members who have abundant working experience in a variety of industrial and corporate environments should have better knowledge of the corporate operations and thus be better able to fulfill the task. In other words, directors with experience as corporate insiders may be better able to understand their peers. Consequently, we argue that the human and social capital of the audit committee is positively related to corporate disclosure quality. Auditors may also explicitly consider audit committee human and social capital when evaluating audit risk. Krishnan (2005), for instance, indicates that one aspect of client risk from an auditor perspective is the composition of the audit committee. Houston (1999) 11

finds that senior auditors explicitly consider client risk is assessing audit fees, while Seetharaman et al. (2002) report that auditors facing greater litigation risk charge higher audit fees. Hay et al. (2006) note that recent research has demonstrated that auditors charge higher fees to firms where they perceive higher client risks. Thus, our arguments imply that that greater human and social capital on the board will lead to greater disclosure quality and lower audit fees. This leads to our second set of hypotheses: Hypothesis 2a: Audit committee capital is positively related to corporate disclosure quality. Hypothesis 2b: Audit committee capital is negatively related to audit fees.

Next, we decompose board capital into insiders’ capital and outsiders’ capital, since extant research usually assumes that they have differing characteristics and roles in board monitoring. We argue that the impact of the human and social capital of these two director types may differ due to differing director roles. Corporate insiders are thought to have detailed knowledge and understanding about the firm’s operations, while independent directors act as monitors (Fama and Jensen, 1983). Therefore, we predict that the human capital proxies (education and work experience) are more important for inside directors, while both human and social capital are important for outside directors who are concerned with their reputation. More specifically, we suggest that for insiders, their educational capital and experience capital may be important contributors to an effective board. As the corporate insiders are the information initiators and decision-makers of the firm, the specific types of human capital which improves their information processing and decisionmaking ability should be important. Presumably, educational capital is related to intellectual rigor to learn and process information. In addition, a greater scope of working experience may enable the managers to be better able to gather, filter and process various information embedded in complex environment. Better knowledge of firm’s operations through working experience may also improve

12

the communications and discussions with the outside directors (e.g., when they ask question about the operations and projects). In addition, managers with greater working experience may be better able to judge and incorporate outsiders’ view and opinions into further actions. This leads to our third hypothesis:

Hypothesis 3: The human capital of inside directors (education and experience capital) is positively related to corporate disclosure quality;

Outsiders, on the other hand, are distinct from insiders in that they are thought to monitor corporate insiders, providing an objective perspective on managerial decision-making. Fama (1980) argues that outside directors are especially concerned with their external reputations which increase their incentive to monitor managers. In this context, director networking ties may serve as a proxy for managerial concerns with reputation. Furthermore, outside directors who have experience in a variety of industrial and corporate environments may be better able to gather, judge and process the reports, plans and proposals from the corporate insiders and be more alert to recognize any potential problems. Therefore, greater human capital of the outside directors should be positively related to corporate disclosure quality. This leads to our fourth hypothesis:

Hypothesis 4: The human and social capital of the outsider directors on the board (education, experience, and networking capital) is positively related to corporate disclosure quality;

Our final arguments focus on the value implications of board and audit committee capital, which we argue should facilitate director monitoring and advising. Implicit in our prior hypotheses is that director capital improves governance efficacy by improving disclosure quality and improving internal monitoring. Although, prior empirical work focuses on director bias, the ability and skills of the agents are also likely to have a substantial impact on their governance efficacy and the day-to-day

13

operations of the firm. Directors bring a variety of resources to the firm that may facilitate the evaluation of senior managers and governance of the firm. Coase (1976) for instance, suggests that agent ability is a first order effect, while conflicts of interest are likely to be a second order effect in understanding agent behavior and firm performance.5 In his model, reputation has a greater impact on effort for high ability agents than in does for low ability agents. Fama (1980) argues that directors are indeed worried about their reputations, while Fich (2005) and Perry and Peyer (2005) find evidence consistent with a reputation concern. One implication of these arguments is that director ability is an important issue in mitigating moral hazard problems on the board, as high ability directors perform better for a given effort level and also have reputational incentives to provide greater effort. Another channel through which the directors can have a positive impact on shareholder value is through advising (Westphal, 1999).

Such functions are accomplished through the

assortment of the directors’ experience, reputation and networking connections, as well as their affiliations with government and other stakeholders (Baysinger and Butler, 1985; Carpenter and Westphal, 2001). Firms need a variety of resources to deal with contingencies involving various stakeholders and evaluating long-term strategic plans. For example, lawyers on a board may help with regulatory and legal issues directors from strategically-related organizations may provide better long-term advice, bankers enhance access to external financing (Pfeffer, 1972; Booth and Deli, 1999; Carpenter and Westphal, 2001; Güner, Malmendier, and Tate, 2005). Consequently, we argue that ability is a paramount concern in the advising managers, in developing and evaluating strategic plans, and in choosing among competing investment opportunities. This leads to our final hypotheses.

5

Incentives and effort are only important if one has the requisite ability. Consider, for instance, the notion of one of the authors joining a major league baseball team (e.g. the Yankees). The compensation contract could indicate that he is paid $1 million per homerun and $100,000 per base hit. Since the actions are observable, monitoring is relatively simple. Despite the favorable incentive package and a fervent desire to perform well it is hard to imagine that either of us would contribute much to team performance.

14

Hypothesis 5a: Board capital is positively related to shareholder value. Hypothesis 5b: Audit committee capital is positively related to shareholder value.

3. Data, Board Capital Indices and Descriptive Statistics Due to the need to manually collect data on each director, we limit our analysis to the industrial firms in the Russell 1000 Index (2003 and 2005). In addition, we require a firm to list their directors in the proxy statement and we focus on non-controlled companies. The final sample contains 615 industrial firms. We use the firms’ proxy statements from 2004 as the beginning point to identify the directors during fiscal 2003 (and 2006 for 2005). To include those directors who actually served in 2003 and/or 2005 but were absent from the 2004 or 2006 proxy statements due to death, resignation or other reasons, we also conduct a thorough search on the proxy statements of the preceding year. The proxy statements have information of each director’s name, age, gender, and director tenure. Regarding the human and social capital information of the directors, there is a large variation across firms in terms of their disclosure of this information. In their proxy statements, firms often disclose directors’ most recent work experience and current external corporate board membership. We manually searched for information on each director’s work experience, lifetime board experience, educational background, professional training, industrial association affiliations and honors and awards from various sources, including LexisNexis, google.com, The Dun and Bradstreet Reference Book of Corporate Management and Who’s Who in Finance and Industry and MergentOnline. The data on the firm-level variables are from CompuStat, ExecuComp and Dun & Bradstreet’s American Corporate Families and International Affiliations.

15

We require the directors to serve at least 3 months in a fiscal year to be counted in our sample. In Appendix A, we report descriptive statistics for the 6,160 directors in our sample for the board as a whole, the audit committee, and based on whether they are insiders or outsiders. Panel A presents the number and percentage of directors with the specific educational characteristics. Among the directors in the sample firms, 84% of them have at least a bachelor’s degree, while 60% report some form of graduate degree (or simply indicate a higher degree).

Panel B presents

descriptive statistics on various aspects of director experience capital. One interesting pattern is that it appears that outside directors have considerable more experience capital than inside directors. However, Panel C indicates that outside directors have worked for only slightly more firms and have held fewer high level corporate posts. As mentioned above, human and social capital information about directors disclosed in proxy statements is rather limited. For instance, most of the sample firms do not disclose directors’ educational background and/or specific degrees obtained. The quality/consistency of the information disclosed about directors also varies within a given firm with some directors having greater information provided about their human and social capital. In addition, even for the same director (804 of them are overlap across firms in our sample), different firms may disclose different levels of information about director capital.

3.1. Director Capital Information In our analysis, board capital is measured by the human and social capital of all the directors in the firm. Human capital consists of education and work experience that individual directors obtain cumulatively. Social capital is a measure of the directors’ external network with outside institutions, corporations, or governments. The specific variables used to measure the board capital and other variables of interest are discussed below. We develop indices for the networking, 16

educational, and experience capital of the board using these variables. We then aggregate these three indices to form a composite measure of total board capital. We use several director characteristics to develop our measure of networking or social capital. The first characteristic is the current number of boards a director sits on during a given year. The second is the current number of nonprofit boards a director sits on and the third is the number of corporate board memberships. Non-profit boards often have a more diverse composition of directors and thus greater networking ties (O’Regan and Oster, 2005). We measure the number of non-profit boards that a director has served on in the past but is no longer a current member. Next, we include any current or prior government position (e.g. department secretary, member of the Joint Chiefs of Staff or other elected/nominated official). Finally, we capture director information on government board service.

Government board membership is designated or nominated by

government agents (such as the president, governor or mayor) or agencies (such as Congress or state and municipal governments). Such networking ties with the government could be very important because such directors can provide counseling on government-related issues such as environmental regulation changes or potential commercial contracts.

In section 3.2 we use this data to form a

measure of networking capital. We also collect data on the educational capital of directors of the board. We use dummy variables to indicate whether a director has obtained a bachelor’s degree, master’s degree, law degree or medical degree, as well as a PhD degree. The individual directors’ educational status are then aggregated to represent the overall board’s educational capital, by the number of firm’s directors with a bachelor’s degree, an MBA, a master’s degree, a PhD or another higher degree (master’s or above). This data is used to form a measure of educational capital. The third type of board capital is the directors’ working experience, which we define as the diversity of the directors’ expertise established during their careers. Diversity of directors’ 17

professional background enriches the resources the board can bring to the boardroom while counseling the CEO on various corporate matters. We measure directors’ expertise by searching their working history and classifying it into five areas such as the number of directors who have been a partner in a law firm, have investment bank/venture capital firm expertise, management consulting experience, accounting firm expertise, or academic experience. We also capture the director information on professional certification, such as CPA, CFA or certified fraud examiner. Professional certification may proxy for the qualification of the directors in the areas of auditing/financing. Next, we collect the number of positions higher than vice president (Chemmanur & Paeglis, 2005) that directors have held during their lifetime. Additionally, we count the number of firms with which the directors have worked during their lifetime. Another potential director characteristic is national-level honors and awards (such as Hall of Fame, Lord, etc.) or the membership in professional or industrial association affiliations, such as IEEE or the American Bar Association, the directors have.

3.2. Forming Indices of Board Capital We develop 3 board capital indices from the above proxies using a ranking methodology (Patil and Taillie, 1982). We rank all firms by the aggregate measure of their directors for each of the proxies and then divide the firms into 20 percentiles, i.e., firms are assigned a score from 1 to 20 according to their ranking percentile for each individual proxy variable. We repeat this process for all the variables and finally we add up the ranking scores for each firm and rank the total scores again from 1 to 20 to get the aggregate director capital index. As such, each firm has a board capital index ranging from 1 to 20. For example, we have 6 variables to proxy the board networking capital. We rank all the firms by each one of 6 variables from high to low and divide the firms into 20 percentile groups. Therefore, each firm is assigned a ranking number, from 1 to 20, according to their ranks in 18

each of the proxies. Finally, we add up the rankings of the 6 proxies to construct a single networking capital ranking index for all the sample firms. Likewise, we construct the education capital index and experience capital index in a similar fashion. The correlations between the three components (networking, education and experience capital indices) of board capital are 0.66, 0.45 and 0.49. The advantage of the indexing methodology is that it may be able to limit bias due to extreme values in the proxies. On the other hand, indexing may also “shrink” the variation of the proxy variables. As such, we also adopt factor analysis in the robustness check section. We find qualitatively similar results using either approach. Finally, we add up the ranking scores of the three components of board capital (networking, education and experience capital) for each firm and rank the total scores again to get the aggregate board capital index. As such, each firm has a board capital index ranging from 1 to 20. Developing a single board capital index allows us to test the effects of directors’ human capital and social capital on corporate disclosure quality and firm performance in a parsimonious way. We will also test our hypotheses regarding the impact of each board capital component on corporate disclosure quality. Indeed, we suggest that the board capital components may have different effects on firm performance, as they conceptually measure different aspects of directors’ human and social resources.

3.3 Dependent Variable and Control Variables 3.3.1 Primary Measure of Disclosure Quality We use an industry-adjusted variation of the approach suggested by Kim and Verrecchia (2001) as our primary measure of the corporate disclosure quality. They argue that since security prices are affected by both disclosed (public) information and private-held information, we can infer the magnitude of private information by observing the relationship between share price changes and 19

share-trading volume. More specifically, Kim and Verrecchia (2001) posit that due to adverse selection, the cost for the informed traders will be larger when the corporate disclosure quality is poor. This happens because the information disclosed to the public is incorporated into the trading volume and the price. When the firm discloses less, the traders who rely on the trading activities will be hurt because the trading activities reflect less information of the firm. Consequently, trading volume will have less association with security prices in firms with poor disclosure quality. This notion leads to the following proxy of corporate disclosure quality, which is the coefficient of the trading volume in the following regression,

Ln

Pt − Pt −1 = β0 + β1 (VOLt − AVGVOL) + ε t Pt −1

(1)

where, Pt: daily stock closing price; VOLt: daily trading volume of the stock in thousand shares; AVGVOL: the average of daily stock trading volume within the last 6-month (we use 182 days) in thousand shares. We run the regression for every stock in CRSP database for 2003 and 2005 respectively. The coefficient of the trading volume, β1, is the proxy for corporate disclosure quality. To control for the impact of industry disclosure on firm disclosure, we use the raw β1 minus the industry-median value of β1. When the firm discloses less, the industry-adjusted β1 will be lower because the traders will rely less on the trading activities as a source for corporate information. That is, the public information incorporate in the trading activities will only be a smaller portion of the overall information set for the company, which consists both public and private information. Note that the coefficient is an

20

inverse measure of disclosure quality6. Therefore, we use the inverse of this coefficient for each firm as our primary measure of disclosure quality. This provides a measure where disclosure quality is increasing as the number gets larger.

3.3.2 Alternative Measure of Disclosure Quality We also use an alternative measure of corporate disclosure quality. Specifically, we use the method suggested in Anderson et al. (2008) whereby they decompose corporate opacity into disclosure quality and market scrutiny components. They use a two step process. In the first step, they create a measure of corporate opacity based on the number of analysts, analyst forecast errors, bid-ask spreads, and trading volume. In the second step, they use this corporate opacity measure as the dependent variable in a regression with the corporate governance quotient from ISS, earnings quality, and the change in earnings per share as the independent variables. They observe that the residual from this regression captures the portion of the corporate opacity attributable to market scrutiny, while the rest represents disclosure quality (Anderson et al. 2008). We use the same procedure except we estimate disclosure quality as the inverse of their measure, so that a higher value represents greater disclosure quality.

3.3.3 Other Variables We also include a set of control variables. First, we include a Herfindahl index based on sales to measure the competitiveness of industry, as higher Herfindahl index denotes less competition within the same industry. It is important to control for industry competition because the collateral cost associated with disclosure may be greater in a more competitive industry. We include board

6

Also for ease of presentation we multiply the coefficient by -1,000 so that bigger value means higher disclosure quality.

21

size and board independence to control for the effects from other board characteristics on corporate disclosure quality. Yermack (1996) posits that bigger boards are more subject to managerial manipulation and control and thus are less effective in protecting shareholders’ interest, suggesting a negative association between board size and disclosure quality. Prior literature suggests that more independent boards may be better able to monitor the managers and implement corporate policies and practice align with the shareholders’ interest. We expect a positive association between board independence and corporate disclosure quality. Firm size is another potential control variable as larger firms may more information available because of greater information intermediary coverage. Leverage is measured by total liabilities divided by total assets. Debtholders may especially pay attention to the disclosure quality of the firm so as to ensure the operational conditions and the ability of the firm to pay off the debt. We also include R&D, which is measured by Research & Development expense divided by total sales. R&D involves a longer-term investment of resources and the outcome may be hard to predict, leading to a higher possibility of extreme scenarios and uncertainty. Furthermore, intangible investment such as R&D may be more subject to managerial manipulations of accounting earnings. As such, we expect a negative association between R&D and disclosure quality. We include the lag ROA, which is the income before extraordinary items divided by total sales, to control for the effects of firm performance on disclosure quality. Institutional investors may have better access to information as well as better information processing ability. In addition, their stakes in the firm make them more vulnerable to information asymmetry or adverse selection problem, leading to greater diligence on corporate disclosure practice from the institutional investors.

We use total common equity

ownership as the measure for institutional investors’ influence. Finally, we include the stock return volatility to control for firm risk.

22

Panel A in Table I gives summary statistics for firm characteristics, including the board capital indices and some of the control variables used in our tests. The average disclosure quality is 0.40 and the median 0.52. The average board in our sample has about 10 directors, with 75% of them being independent. On average over 70% of the common equity is owned by institutional investors. Finally, as the board capital index is structured as ranking score for the firms, from 1 to 20, the mean board capital index is 10.29. Similarly, the mean of networking, education and experience capital index is around 10.4 and the standard deviation of all the board capital indices is in the range of 4.6 – 4.8. In Panel B of Table I, we present the univariate tests on the means of disclosure quality between firms with high rankings in each of the board capital indices and those with low rankings. As the rankings range from 1 to 20, “High” refers to rankings equal to or higher than 16 and “Low” equal to or lower than 5. Comparing firms with high and low capital index scores, we find that high board capital firms have statistically better disclosure quality than firms with low board capital (0.59 of disclosure quality for high firms and 0.29 for low firms). We apply the same univariate tests on the three individual components of board capital and find that firms with higher networking, education and experience capital have better disclosure quality than low capital firms. In addition, firms with higher Audit Committee Capital also have better disclosure quality than firms with low Audit Committee Capital. Further tests reveal that firms with higher audit networking, education, or experience capital show significant greater disclosure quality. Finally, when we examine the insiders’ and outsiders’ capital, we find that firm disclosure quality differs most dramatically when outsiders’ capital are different, with a gap of 0.34 between high outsiders’ capital firms and low ones. These univariate statistics are consistent with our central hypothesis that director capital is an important factor in evaluating board and audit committee efficacy on corporate disclosure practices. 23

4. Board Capital and Disclosure Quality 4.1 Multivariate Regression Results Our first test examines the relation between board capital and disclosure quality. We use the following specification to test hypothesis 1a. Disclosure quality = α + β1 × BoardCapital + β 2 × OutsiderRatio + β 3 × FirmSize + β 4 × Leverage + β 5 × BoardSize + β 6 × R & D + β 7 × ROA + β 8 × INSTOWN + β 9 × RISK + β10 × SalesHHI + ξ We use two measures for corporate disclosure quality, one being the inverse of the Kim and Verrecchia (2001) measure and other in Anderson, et al (2008). Both measures are industry-adjusted. Our primary variable of interest is board capital, which captures the human and social capital of all the directors on the board. Column 1 in Table II presents the results of the primary regression using the Hubert-White Sandwich estimator clustered on the firm. The results indicate that board capital is positively related to firm disclosure quality. Specifically, we find that the coefficient estimate of board capital is 0.020 and statically significant (t-values of 2.68) at 1% level. A positive coefficient estimate for board capital is consistent with Hypothesis 1a, which we interpret to suggest that board capital improves governance efficacy through promoting a greater disclosure quality.

These results are also

economically significant. Specifically, the results indicate that after controlling for other factors, a 10% increase of board capital is associated with a 5.1% higher disclosure quality7. In column 2 of Table II we include firm level fixed-effects (the Hausman test suggests firmlevel fixed effects are more appropriate than random effects although both approaches yield similar conclusions) and find qualitatively similar results to those in column 1. More specifically, these 7

We calculate this as follows. Start from the mean point of board capital which is 10.29 and 10% increase from the mean is 11.32. The associated decrease of the dependent variable is 1.029*(0.020) = 0.02058, which is 5.145% increase from the mean of the dependent variable (0.02058/0.40).

24

results suggest that after controlling for firm-specific time-invariant characteristics, the coefficient estimate of board capital is also positive (0.018) and significantly different from zero (t-value of 2.94). Economically, the results are also similar to those from column 1. Again, we interpret these results to suggest that board capital improves governance efficacy. One potential concern is that the results are driven by endogeneity. In other words, our findings can be simply interpreted as a “matching” story, i.e., high capital directors deliberately choose those firms with better disclosure quality to join. Although, this explanation is the opposite of the standard board endogeneity argument, it is certainly a plausible alternative. We attempt to address this endogeneity concern by an instrumental variable approach. We posit that the directors, especially those with greater capital, are typically busy individuals; leading them to consider the traveling difficulties associated with their board activities. More specifically, we use the distance between the firm’s corporate headquarters and the nearest metropolitan airport as a proxy for the traveling difficulties. The metropolitan areas include the following 10 cities: New York City, Los Angeles, Chicago, Washington-Baltimore, San Francisco, Philadelphia, Boston, Detroit, Dallas, and Houston. Other than traveling consideration, metropolitan areas also provide greater networking opportunities and diversity of other activities and may be more desirable for directors, than rural areas. Because the distance between the headquarters location and major airports is typically not a managerial choice in the short/medium term, we argue it makes a potentially important and robust instrument in our tests. We measure Distance as the number of miles between the firms’ headquarters and the nearest major airport (using Google Maps). A list of the firms and the distance to the closest major airport is available upon request. In the first stage of the process, we use the instrument variable, Distance, to explain the crosssection board capital index. We find that Distance is negative and significantly associated with board

25

capital. 8 Specifically, we find that every 100% increase of the distance is associated with a 4.4% decrease of board capital, confirming our suggestion that directors do consider traveling costs when choose the firms they want to serve. Given that the sample average distance to the nearest metropolitan airport is 195 miles and the range stretches from 2.7 miles to 1,093 miles, we believe that transportation difficulties may provide a meaningful and exogenous instrument for board capital. Column 3 of Table II presents the regressions’ results for the second stage of the tests. The results are qualitatively similar to both the basic OLS regression results and the firm-level fixedeffects regression results. In the 2SLS regressions, we find that the coefficient estimate of board capital is 0.028 and significant at 1% level (t-value of 3.36). In column 4 & 5, we present the fixedeffects and 2SLS results using the Anderson et al (2008) measure of disclosure quality. We again find that board capital shows positive relations with disclosure quality. Overall, the results in Table II are consistent with Hypothesis 1a.

4.2. Components of Board Capital and Disclosure Quality In order to examine the relation between the individual board capital components and disclosure quality (hypothesis 1b) we use the following specification. In the first stage test we use the following specification, Board Capital = α + β1 Distance + β2 Outsider Ratio + β3 Firm Size + β4 Risk + β5 Leverage + β6 Board Size + β7 R&D + β8 ROA + β9 INSTOWN + ζ. The coefficient is 0.0023. From the mean distance value of 195 miles, a 100% increase of the distance to 390 miles is associated with a decrease of board capital of 0.4485 from the mean value (10.29), which is a 4.4% decrease. We conduct a series of specification tests on our instrumental variable and the structural model. Our first test indicates that the error term of the second-stage regression is unrelated to the instrument, suggesting that the instrument is valid. We then apply the Hausman endogeneity test which examines whether the IV estimator significantly differs from the OLS estimator, indicating an endogeneity concern (Hayashi, 2000; Beaver, McAnally and Stinson, 1997). The results of the test provide a χ2-statistic of 2.904 with an associated p-value of 9%; indicative of endogeneity concerns and thus the use of a 2SLS approach to estimate the relation between firm performance and board heterogeneity. Third, we conduct the under-identification and weak-identification tests to test the strength of the instrument variable and we find that both tests generate highly significant statistics. For instance, the weak-identification test has the Kleibergen-Paap F-statistic of 28.45, which is well above the threshold suggested in Stock and Yogo (2005), suggesting sufficient explanatory power for the instrument. 8

26

Disclosure quality = α + β1 × Networking + β 2 × Educational + β3 × Experience + β 4 × OutsiderRatio + β5 × FirmSize + β 6 × Leverage + β 7 × BoardSize + β8 × R & D + β9 × ROA + β 10 ×INSTOWN + β11 × RISK + β12 × SalesHHI + ξ

(2)

Our arguments suggest that the components of board capital (networking, education and experience) may have different impacts on disclosure quality. Table III shows that both the time clustered and firm fixed-effects models generate qualitatively and quantitatively similar results. We find that networking capital is negative but insignificant. In contrast, the educational and experience capital components are both found to be significantly related to disclosure quality.9 Taken together these results imply that directors’ educational as well as experience capital assist them in shaping the corporate disclosure practice.

However, these results could stem from our approach in

decomposing board capital into educational, experience, networking capital. There is no uniform agreement on how to categorize some specific measures of director and human capital. As such, any arbitrary grouping of these characteristics, such as ours, should be interpreted with caution. As the three components exhibit relatively high correlations, we also orthogonized each of the measures and found similar results.

4.3. Audit Committee Capital and Disclosure Quality In this section, we examine the relation between audit committee capital and firm disclosure quality (hypothesis 2a). An audit committee with high-capital members should be better able to request the relevant information and identify any potential loopholes in information disclosure practice and improve the reliability of financial reports. To test the relation between audit committee capital and disclosure quality, we use the following specification.

9

We also perform the tests with each component of board capital separately. We find that networking capital is positive and insignificant (coefficient of 0.003 and t = 1.22) but education capital and experience capital are both positive and significant (t-values of 2.37 and 3.70 respectively).

27

Disclosure quality = α + β1 × AuditComCapital + β 2 × OutsiderRatio + β3 × FirmSize + β 4 × Leverage + β5 × BoardSize + β 6 × R & D + β7 × ROA + β 8 ×INSTOWN + β9 × RISK + β10 × SalesHHI + ξ

(3)

We show the results in the first two columns in Table IV. In column (1), using the HubertWhite Sandwich estimator clustered on time, we find the coefficient estimate of audit committee capital is 0.028 and the t-value is 1.88. The firm level fixed-effects regression in column (2) generates similar estimate (coefficient estimate 0.025 and t-value 2.04). These findings suggest that hypothesis 2a is generally supported and more specifically, we find that a 10% increase of audit committee capital is associated with 6.5% greater firm disclosure quality. In order to examine the relation between the individual audit capital components and disclosure quality, we replace the audit committee capital component in equation 3 with each of the three subcomponents. Our arguments suggest that the components of audit committee capital (networking, education and experience) may have different impacts on disclosure quality. We present the results in column (3) and (4) in Table IV. We find that audit committee networking capital is insignificant while the other two components are significant. More specifically, the OLS regression in column (3) show the coefficient estimate of networking capital is 0.001 and insignificant at 10% level (t-value of 0.34). The educational capital of the audit committee has a coefficient estimate of 0.005 and is significant at 10% level (t value = 1.85). In addition, the coefficient estimate on the experience capital of the audit committee is 0.015 and significant at 5% level (t value = 2.33). Fixed-effects regression in column (4) present very similar results. These numbers suggest that a one standard deviation difference in audit committee educational capital is associated with 5.8% difference in disclosure quality, while audit committee experience capital provides impact of more than double magnitude. Overall, the findings are consistent with the notion

28

that different components of audit committee capital have different effects on corporate disclosure practice. Next, we apply our test using a different dependent variable, audit fee charged by auditors of the firm (hypotheses 2B). We suggest that if audit committee capital implies lower risk for external auditors, then we should observe a negative relation between audit committee capital and the audit fee. Columns (5)-(6) show the results. More specifically, both OLS and fixed-effects regressions show that firms with higher audit committee capital will have lower audit fees. We interpret it as additional evidence suggesting that audit committee capital is an important factor for its efficacy.

4.4. Outsiders vs. Insiders Next, we explore the disclosure quality implications of inside and outside director capital (hypotheses 3 and 4). Our arguments suggest that outside director networking ties or social capital may be an especially important aspect of directors’ reputational concerns. While, our previous results are insignificant regarding networking capital, this could occur because we have not separated inside and outside directors into separate groups. In addition, the human capital of inside directors may be especially important as they control the day-to-day operations of the firm and they can play a critical role in determining corporate disclosure quality. We adopt the same process of first aggregating the individual director measures into board level measures. We then apply the ranking methodology and develop the board capital indices for outsiders and insiders, respectively. Table V shows the regression results. In column (1), we show the results when we put in the aggregate board capital indexes for outsiders and insiders. We find that our main findings are primarily driven by the outsiders’ capital, not the insiders’, consistent with the notion that corporate disclosure quality is a crucial aspect of board governance and monitoring. In column (2) and (3), we examine the specific components of board capital of outsiders and 29

insiders. These findings imply that among outsiders, their experience seems to be the primary factor related to greater disclosure quality. For insiders, we find that it is both of their education and experience capital which are related to corporate disclosure quality. Consistent with the results for the board as whole, the networking capital of both insiders and outsiders are not related to firm disclosure quality. Overall, these findings are consistent with the notion that different types of directors bring in different types of board capital to the board room. Shareholders seem to benefit most by bringing in independent monitors with sufficient experience and expertise so that they are able to dissimilate information to the outside. Under the premise that educational capital reflects the innate ability of directors, the results suggest that firms with high ability managers are less inclined to keep shareholders in the dark about corporate activities. In general, these results suggest that Hypotheses 3 and 4 are not supported. However, these results also indicate that shareholders value different characteristics for outsiders and insiders. Overall, these results suggest that shareholders match up the specific attributes of the directors’ human/social capital with their job descriptions for the firm.

4.5. Director Capital and Shareholder Value We propose that the fullness of director human capital brings a variety of resources to the board deliberations and therefore improves board governance and ultimately firm performance. In Table VI, we present the results of our tests that examine the firm performance implication of board (audit committee) capital. In columns (1) – (3), we show that board capital index is positively associated with Tobin’s Q, regardless of the different methodologies we apply. These results are also economically significant. Specifically, the results indicate that after controlling for other factors, a 20% increase of board capital is associated with a 7.99% higher firm performance. In columns (4) – (6), we examine audit committee capital and we find similar results with those of board capital. 30

Specifically, we find that the results hold when we use OLS, firm fixed-effects or 2SLS regressions. Economically, we find that a 20% increase of audit committee capital is associated with a 4.99% greater firm performance. Overall, Table VI results are consistent with Hypothesis 5. A central theme of our analysis is that board capital improves corporate governance by increasing corporate disclosure quality. The implication is that the channel for the improved firm performance from greater director capital in Table VI is through corporate disclosure quality. To gain some insights on this potential interpretation, we decompose corporate disclosure quality into the part attributable to board capital and the part from other attributes of the firm. Specifically, we regress corporate disclosure quality on board capital (without control variables). The residual from this specification can be interpreted as the portion of disclosure quality not attributable to board capital (we label this Supplementary Disclosure Quality). In contrast, the predicted value from this specification gives a proxy of disclosure quality that is attributable to board capital (we label this Board Disclosure Quality). In appendix B, we report the results of three additional tests using these new variables in the performance regressions shown in Table VI. The first column in Appendix B adds supplementary disclosure as an additional potential control variable. Column 2 removes board capital and replaces it with board disclosure quality. In Column 3, we include board capital, board disclosure quality, and supplementary disclosure quality. Taken as whole, the results in this appendix are consistent with the notion that board capital improves firm performance through greater corporate disclosure quality.

4.6 Voluntary Disclosure of Directors’ Capital Information and Firm Performance We observe a huge variation across firms in terms of their disclosure practice of directors’ human and social capital information. In light of recent regulatory reform by the SEC and stock exchanges, calling for greater disclosure of information on board composition and “financial 31

expertise”, it may be beneficial to extend such requirement to other aspects of director human and social capital. Yet, some argue that regulatory actions on corporate boards may not be in the best interest of the shareholders (Hermalin & Weisbach, 2006) because the market mechanism may be better able to price different practices of the firm. Given the possible negative impact on firm performance by imposing mandatory disclosure on the firm, shareholders may be better off if the government leaves such decision to the firm and the market. As such, we test the relation between voluntary disclosures of director human and social capital and firm performance. We focus on the disclosure of directorships, working experience and education information in the proxy statement and measure the corresponding gap with the information we collect from other sources. For instance, we use a dummy variable to indicate whether a firm discloses board seats (corporate and non-profit) information of the directors.10 For education, we use the percentage of directors whose educational background is disclosed. For working experience, we measure the disclosure gap as the coefficient of variation of the number of years of each director’s working background. Table VII presents the results. In columns (1) – (3), we put each category of director capital information gap in the regression separately, while in column (4) we put all the variables in together. The regression results suggest that disclosures on working experience (t value = 1.85), educational information (t value = 3.07) and non-profit boards (t value = 1.99) are all positively associated with firm performance.11 More specifically, we find that in firms that disclose directors’ educational background they have a higher Tobin’s Q of 0.507 than firms that do not. Firms disclosing nonprofit board seats information enjoy a higher Tobin’s Q of 0.147. However, we do not find that the

10

We use some discretion in the process as we observe that firms may disclose board seats for certain directors but not for some others. We thus designate the dummy to be one if the majority directors’ board seats are disclosed. We use the following regression: Adj. Tobin’s Q = α + β1 WorkingGap + β2 EducationGap + β3 BoardSeatsGap + β4 NonprofitGap + control variables +ζ.

11

32

disclosure gap of corporate board seats is associated with firm performance. Overall, we interpret our findings as evidence supporting the notion that the market values the directors’ capital information and is consistent with our findings that such information is important to board efficacy. 4.7. The Impact of Financial Expertise As we measure a comprehensive set of human and social capital of the directors, our findings may be also consistent with the notion that it is actually just the accounting/financial expertise of the directors which promotes disclosure quality. To explore this possibility, we include accounting/financial expertise in the regression along with board (audit committee) capital. We measure accounting/financial expertise by the number of directors with CFO experience, with CPA/CFA certificate, with accounting/auditing firm partner experience, or designated by the board to be a “financial expert”. We find that our measure of financial/accounting expertise is positive and significantly associated with disclosure quality. However, the board (audit committee) capital remains qualitatively and quantitatively similar to our prior results (see appendix C). We interpret these results to suggest that the “fullness” of director human and social capital plays an important role in corporate disclosure policy; beyond just the financial expertise mandated by the SEC and stock exchanges.

4.8 Alternative Measure of Board Capital - Factor Scores Our analysis is based on developing indices of board capital. However, another approach that may be appropriate is to use factor analysis to develop each capital measure. For each type of board capital, we apply factor analysis to the multiple proxies within each group of variables and calculate the factor score to represent each type of board capital. The goal of using factor analysis is to “extract” the underlying single measure from a bunch of noisy proxies.

33

We have 6 variables to potentially use in developing the board networking capital measure. Factor analysis shows that a single factor explains 83.4% of the total variation among the proxies. The Eigen value of the first factor is 3.9355 and the second factor’s Eigen value is 0.5200, which is consistent with the idea that they capture the same underlying issue. Similar to the networking capital, factor analysis of board education capital proxies shows that a single factor explains 83.9% of the total variation. The Eigen value of the first factor is 3.1789 and the second is 0.6663. Finally, factor analysis of the experience capital proxies shows that a single factor explains 87.2% of the total variation among the proxies. The first factor has an Eigen value of 1.6208 and the second 0.5253. After the factor analysis, we calculate the factor scores for each underlying factor. The respective factor score is the numerical representation of the underlying board capital. The correlation between the three board capital factor scores is 0.48, 0.40 and 0.39. Comparing the factor scores to the composite indices that we computed earlier, it seems that indices “shrink” the outliers due to extreme values among the proxy variables. However, using these factor scores in each of tests leads to qualitatively and quantitatively similar results.

5. Conclusion and Discussion In this study, we suggest that on a day-to-day basis, directors’ capital should be of great importance to shareholders because they have little recourse for poor decision-making due to limited ability or skills. In contrast, in a robust legal system with relatively strong safeguards, shareholders and the SEC can sue or prosecute directors for board errors due to conflicts of interest. In addition, economic theory implies that agent ability is quite important to outside shareholders (Milgrom and Roberts, 1982) which is consistent with conventional thinking that ability is an important factor in explaining performance. Building on these ideas, we posit that shareholders focus on the ability of directors in evaluating board effectiveness. We primarily focus on one key aspect of board and audit 34

committee efficacy, corporate disclosure quality. We use an industry adjusted version of the Kim and Verrecchia (2001) as our primary proxy for disclosure quality and we find similar results using alternative Anderson et al (2008) measure for disclosure quality. We define board capital as the aggregate of the individual directors’ human capital and social capital. To test the notion that board capital is positive related to corporate disclosure quality, we collect detailed data on director’s educational background, working experience and networking ties of 615 industrial firms of the Russell 1000 Index as of 2003 and 2005. We then develop these various proxies into indices for board networking capital, educational capital and experience capital, as well as a single aggregate of board capital index. We use a similar approach to measure audit committee capital. Our univariate and multivariate tests indicate that board capital is positively associated with corporate disclosure quality, even after controlling for other factors that influence disclosure policy and using an econometrically attractive instrument of director capital. Splitting board capital into educational, experience, and networking capital, we find director educational and experience capital appear to have the greatest impact on corporate disclosure policy. In contrast, director networking capital is unrelated to corporate disclosure quality.

Decomposing board capital into insiders’ and

outsiders’, we find that the relation board capital and disclosure quality is a function of outside director experience and inside director educational capital. These results suggest that shareholders should weigh the attributes of insiders and outsiders differently, although both groups can contribute to greater corporate disclosure quality. In addition, we find that audit committee capital is positively related to firm performance. Again, the most important factor for these directors is human capital as opposed to their social capital. We also find significant evidence that both board and audit committee capital are related to greater firm performance. We interpret this to suggest that shareholders appreciate a variety of skills 35

by directors, with greater corporate disclosure quality acting as a channel for improved corporate governance and firm performance. Our final set of tests indicate that the market rewards those firms that do a better job of disclosing information about the human and social capital of their directors. This raises the question of why firms, with high capital boards, do not make greater efforts to disclose this information.

REFERENCES Anderson, Ronald C., Augustine Duru, and David M. Reeb, 2008, Founders, heirs, and corporate opacity in the U.S. Journal of Financial Economics, forthcoming. Anderson, Ronald C., Sattar A. Mansi, and David M. Reeb, 2004, Board characteristics, accounting report integrity, and the cost of debt. Journal of Accounting & Economics 37, 315-342. Baysinger, Barry D., and Henry N. Butler, 1985, Corporate governance and the board of directors: Performance effects of changes in board composition. Journal of Law, Economics & Organization 1, 101-124. Beasley, Mark S., 1996, An empirical analysis of the relation between the board of director composition and financial statement fraud. Accounting Review 71, 443-465. Beaver, William H., Mary Lea McAnally, and Christopher H. Stinson, 1997, The information content of earnings and prices: A simultaneous equations approach, Journal of Accounting & Economics 23, 53-81. Becker, Gary, 1964, Human Capital. New York: Columbia University Press. Booth, James R., and Daniel N. Deli, 1996, Factors affecting the number of outside directorships held by CEOs. Journal of Financial Economics 40, 81-104. Brickley, James A., and Christopher M. James, 1987, The takeover market, corporate board composition, and ownership structure: The case of banking. Journal of Law and Economics 30, 161180. Brickley, James A., Jeffrey L. Coles, and Rory L. Terry, 1994, Outside directors and the adoption of poison pills. Journal of Financial Economics 35, 371-390. Carcello, Joseph V., and Terry L. Neal, 2000, Audit committee composition and auditor reporting. Accounting Review 75, 453-467. Carpenter, Mason A., and James D. Westphal, 2001, The strategic context of external network ties: Examining the impact of director appointments on board involvement in strategic decision making. Academy of Management Journal 44, 639-660. Chemmanur, Thomas J., and Imants Paeglis, 2005, Management quality, certification, and initial public offering. Journal of Financial Economics 76, 331-368. Coase, Ronald H., 1976, Adam Smith’s view of man. Journal of Law and Economics 19, 529-546. 36

Coles, Jeffrey L., Naveen D. Daniel, and Lalitha Naveen, 2008. Boards: Does one size fit all? Journal of Financial Economics 87, 329-356. Cotter, James F., Anil Shivdasani, and Marc Zenner, 1997, Do independent directors enhance target shareholder wealth during tender offers? Journal of Financial Economics 43, 195-218. Defond, Mark L., Rebecca N. Hann, and Xuesong Hu, 2005, Does the market value financial expertise on audit committees of boards of directors? Journal of Accounting Research 43, 153-193. Dunn, Kimberly A., and Brian W. Mayhew, 2004, Disclosure quality and auditor choice. Review of Accounting Studies 9, 35-58. Ely, Jeffrey C., and Juuso Välimäki, 2003, Bad reputation. Quarterly Journal of Economics 118, 785-814. Fama, Eugene F., 1980, Agency problems and the theory of the firm. Journal of Political Economy 88, 288-307. Fama, Eugene F., and Michael C. Jensen, 1983, Separation of ownership and control. Journal of Law and Economics 26, 301-325. Fich, Eliezer, 2005, Are some outside directors better than others? Evidence from director appointments by Fortune 1000 firms. Journal of Business 78, 1943-1971. Fudenberg, Drew, and David K. Levine, 1992, Maintaining a reputation when strategies are imperfectly observed. Review of Economic Studies 59, 561-579. Güner, A. Burak, Ulrike Malmendier, and Geoffrey Tate, 2008, Financial expertise of directors. Journal of Financial Economics 88, 323-354. Hay, David C., W. Robert Knechel, and Norman Wong, 2006, Audit fees: A meta-analysis of the effect of supply and demand attributes. Contemporary Accounting Research 23, 141-191. Hayashi, Fumio, 2000, Econometrics. Princeton University Press: Princeton, NJ. Healy, Paul M., and Krishna G. Palepu, 2001, Information asymmetry, corporate disclosure, and the capital markets: A review of the empirical disclosure literature. Journal of Accounting & Economics 31, 405-440. Helland, Eric, 2006, Reputational penalties and the merits of class-action securities litigation. Journal of Law and Economics 49, 365-395. Hermalin, Benjamin E., and Michael S. Weisbach, 1998, Endogenously chosen boards of directors and their monitoring of the CEO. American Economic Review 88, 96-118. Hermalin, Benjamin E., and Michael S. Weisbach, 2006, A framework for assessing corporate governance reform. Available at SSRN: http://ssrn.com/abstract=881581. Holmström, Bengt, 1999, Managerial incentive problems: A dynamic perspective. Review of Economic Studies 66, 169-182. Houston, Richard W., 1999, The effects of fee pressure and client risk on audit seniors’ time budget decisions. Auditing 18, 70-86. Karamanou, Irene, and Nikos Vafeas, 2005, The association between corporate boards, audit committees, and management earnings forecasts: An empirical analysis. Journal of Accounting Research 43, 453-486.

37

Kim, Oliver, and Robert E. Verrecchia, 2001, The relation among disclosure, returns, and trading volume information. Accounting Review 76, 633-654. Klein, April, 2002, Audit committee, board of director characteristics, and earnings management. Journal of Accounting & Economics 33, 375-400. Klein, April, 1998, Firm performance and board committee structure. Journal of Law and Economics 41, 275-303. Krishnan, Jayanthi, 2005, Audit committee quality and internal control: An empirical analysis. Accounting Review 80, 649-675. Larcker, David F., and Tjomme O. Rusticus, 2008, On the use of instrumental variables in accounting research. Available at SSRN: http://ssrn.com/abstract=694824. Milgrom, Paul, and John Roberts, 1982, Predation, reputation and entry deterrence. Journal of Economic Theory 27, 280-312. O’Regan, Katherine, and Sharon M. Oster, 2005, Does the structure and composition of the board matter? The case of nonprofit organizations. Journal of Law, Economics, & Organization 21, 205227. Patil, G.P., and C. Taillie, 1982, Diversity as a concept and its measurement. Journal of the American Statistical Association 77, 548-561. Perry, Tod, and Urs Peyer, 2005, Board seat accumulation by executives: A shareholder’s perspective. Journal of Finance 60, 2083-2123. Pfeffer, Jeffrey, 1972, Size and composition of corporate boards of directors: The organization and its environment. Administrative Science Quarterly 17, 218-228. Ribstein, Larry E., 2002, Market vs. regulatory responses to corporate fraud: A critique of the Sarbanes-Oxley Act of 2002. Journal of Corporation Law 28, 1-67. Seetharaman, Ananth, Ferdinand A. Gul, and Stephen G. Lynn, 2002, Litigation risk and audit fees: Evidence from UK firms cross-listed on US markets. Journal of Accounting & Economics 33, 91115. Sloan, Richard G., 2001, Financial accounting and corporate governance: A discussion. Journal of Accounting & Economics 32, 335-347. Stanford Securities Class Action Clearinghouse, 2008, http://securities.stanford.edu. Stock, J.H., and M. Yogo, 2005, Testing for weak instruments in linear IV regression. In D.W.K. Andrews and J.H. Stock, eds. Identification and inference for econometric models: Essays in honor of Thomas Rothenberg. Cambridge: Cambridge University Press, 2005, pp. 80–108. Working paper version: NBER Technical Working Paper 284. http://www.nber.org/papers/T0284. Strine Jr., Leo E., 2002, Derivative impact? Some early reflections on the corporation law implications of the Enron debacle. Business Lawyer 57, 1371-1402. Veasey, E. Norman, and Christine T. Di Guglielmo, 2005, What happened in Delaware corporate law and governance from 1992-2004? A retrospective on some key developments. University of Pennsylvania Law Review 153, 1399-1512.

38

Westphal, James D., 1999, Collaboration in the boardroom: Behavioral and performance consequences of CEO-board social ties. Academy of Management Journal 42, 7-24. Yermack, David, 1996, Higher market valuation of companies with a small board of directors. Journal of Financial Economics 40, 185-211.

39

Table I – Board Capital Information Panel A: Firm Characteristics Networking Capital is developed as an index on the following proxies, Total boards, Nonprofit board, Nonprofit ever, Company board, Politician, and Government board. Education capital is an index developed based on proxies of bachelor’s, Law, PhD, MBA, Master’s and Higher degree. Experience Capital is an index developed on the proxies of Firms worked, Higher posts, Certification, Law, Consulting, Accounting, IB/venture, Academia, honors & awards, and Associations. Board Capital is an index developed from Networking, education and Experience capital indices. The detailed description of each proxy and the methodology of developing indices are provided in the text. Mean Median Max Min Standard Deviation Disclosure Quality 0.40 0.52 2.87 -10.15 0.74 Networking Capital 10.45 10.00 20.00 1.00 4.77 Experience Capital 10.42 10.00 20.00 1.00 4.75 Education Capital 10.46 10.00 20.00 1.00 4.75 Board Capital 10.29 10.00 20.00 1.00 4.60 Number Directors 9.71 9.00 20.00 5.00 2.30 Percentage Independent 75.06 77.12 100.00 30.77 12.34 Leverage 0.52 0.53 1.00 0.00 0.23 Institutional ownership 0.72 0.74 1.00 0.25 0.16 Risk (stock return volatility) 0.42 0.37 3.50 0.12 0.22 SalesHHI 0.40 0.34 1.00 0.10 0.24

Panel B: Univariate Statistics High refers to the firms with a capital index equal or bigger than 16 and Low includes those with a capital index equal or lower than 5. The p-value is from t-test for equal means of disclosure quality between those two groups of firms. Disclosure Quality High Capital Low Capital t-test p-value Board Capital 0.59 0.29 0.000 Networking Capital 0.59 0.28 0.000 Education Capital 0.58 0.32 0.000 Experience Capital 0.49 0.32 0.002 Audit Committee Capital 0.53 0.33 0.000 Audit Networking 0.52 0.34 0.001 Audit Education 0.54 0.35 0.000 Audit Experience 0.49 0.33 0.012 Insiders’ Capital 0.48 0.32 0.011 Outsiders’ Capital 0.62 0.28 0.000

40

Table II Board Capital and Disclosure Quality This table presents the results of regressing board capital on disclosure quality with various control variables. The dependent variable is industry-adjusted disclosure quality. For column (1) – (3), we use the Kim & Verrecchia (2001) measure for disclosure quality. In column (4) – (5), we use the Anderson, et al (2008) measure. Board capital is an index developed from the underlying proxies. Firm size is the logarithm of annual sales. Board size is the number of directors. Outsider ratio is the number of outside directors divided by the board size. Leverage is total liabilities divided by total assets. ROA is the net income before extraordinary items divided by total assets. R&D is the research & development expenses divided by total sales. INSTOWN is the total percentage equity ownership of all the institutional investors. Risk is the volatility of the stock return. SalesHHI is a Herfindahl index based on sales. Column (1) shows the OLS results. Column (2) shows firm fixedeffects results. Column (3) shows the results of 2SLS regression results. Statistical significance at the 10, 5 and 1 percent levels is indicated by *, **, ***, respectively. t-statistics (by Huber-WhiteSandwich estimates, allowing for correlation of all observations of a given firm) are in parentheses. Dependent Variable: Variable Intercept Board Capital Outsider ratio Firm size Leverage Board size R&D ROA INSTOWN Risk SalesHHI Firm Fixed Effects Observations Adjusted R2

Disclosure Quality (KV, 2001) (1) (2) OLS Fixed Effects -1.237*** -1.184*** (-5.05) (-4.89) 0.020*** 0.018*** (2.68) (2.94) 0.336 0.178 (1.01) (1.03) 0.225*** 0.212*** (4.97) (4.19) -0.449*** -0.455*** (-2.61) (-3.28) -0.033** -0.032** (-2.32) (-2.39) 0.516** 0.409** (2.23) (2.06) -0.631 -0.493 (-0.87) (-0.77) -0.075 -0.055 (-0.50) (-0.37) 0.710*** 0.715*** (2.99) (3.14) 0.064 0.065 (1.35) (1.29) No Yes 1,196 1,196 0.33 0.29

(3) 2SLS -2.068*** (-5.23) 0.028*** (3.36) 0.093 (1.31) 0.224*** (3.60) -0.392*** (-3.05) -0.019** (-2.19) 0.480** (2.39) -0.591 (-0.96) -0.025 (-0.66) 0.721*** (4.08) 0.035 (1.23) No 1,196 0.34

Disclosure Quality (ADR, 2008) (4) (5) Fixed Effects 2SLS -2.690*** -2.451*** (-4.55) (-4.67) 0.262** 0.250** (2.32) (2.25) 0.603** 0.757*** (2.15) (2.65) 0.038* 0.036* (1.75) (1.80) -0.410** -0.392** (-2.13) (-1.99) -0.015* -0.014* (-1.76) (-1.85) 0.345 0.101 (1.54) (1.44) 1.703 2.018* (1.47) (1.94) -0.079 -0.162 (-0.31) (-0.34) 0.190 0.162 (0.88) (0.72) 0.188* 0.113 (1.78) (1.23) Yes No 1,056 1,056 0.27 0.30

41

Table III Board Capital Components and Disclosure Quality This table presents the results of regressing board capital components on firm disclosure quality. The dependent variable is industry-adjusted disclosure quality. The components of board capital are networking capital, educational capital and experience capital, which are indices based on 19 different director characteristics. Firm size is the logarithm of annual sales. Board size is the number of directors. Outsider ratio is the number of outside directors divided by the board size. Leverage is total liabilities divided by total assets. ROA is the net income before extraordinary items divided by total assets. R&D is the research & development expenses divided by total sales. INSTOWN is the total percentage equity ownership of all the institutional investors. Risk is volatility of stock return. SalesHHI is a Herfindahl index based on sales. Statistical significance at the 10, 5 and 1 percent levels is indicated by *, **, ***, respectively. t-statistics (by Huber-White-Sandwich estimates, allowing for correlation of all observations of a given firm) are in parentheses.

Dependent variable: Disclosure Quality Variable Intercept Networking capital Educational capital Experience capital Outsider ratio Firm size Leverage Board size R&D ROA INSTOWN Risk SalesHHI Observations Adjusted R2

Time Clustered -1.312*** (-4.32) 0.004 (0.99) 0.012** (2.15) 0.011** (2.32) 0.103 (1.01) 0.232*** (5.14) -0.416*** (-2.70) -0.040** (-2.10) 0.517** (2.26) -0.634 (-0.98) -0.054 (-0.61) 0.685*** (3.01) 0.052 (0.88) 1,196 0.30

Firm Fixed-Effects -1.522*** (-5.38) 0.002 (0.98) 0.013** (2.54) 0.013*** (3.32) 0.078 (0.83) 0.235*** (5.22) -0.357*** (-2.79) -0.043** (-2.52) 0.486** (2.41) -0.615 (-1.02) -0.065 (-0.44) 0.676** (3.07) 0.058 (0.95) 1,196 0.25

42

Table IV Audit Committee Capital and Disclosure Quality This table presents the results of regressing audit committee capital on firm disclosure quality. The dependent variable is industry-adjusted disclosure quality in column (1)-(4) and log of audit fee in (5)-(6). The components of audit committee capital are networking capital, educational capital and experience capital, which are indices based on 19 different director characteristics. Firm size is the logarithm of annual sales. Board size is the number of directors. Outsider ratio is the number of outside directors divided by the board size. Leverage is total liabilities divided by total assets. ROA is the net income before extraordinary items divided by total assets. R&D is the research and development expenses divided by total sales. INSTOWN is the total percentage equity ownership of all the institutional investors. Risk is the volatility of stock return. SalesHHI is a Herfindahl index based on sales. Statistical significance at the 10, 5 and 1 percent levels is indicated by *, **, ***, respectively. t-statistics (by Huber-White-Sandwich estimates, allowing for correlation of all observations of a given firm) are in parentheses. Dependent variable: Disclosure Quality Audit Fee (ln) Variable Intercept AuditComCapital

(1) Time Clustered -1.461*** (-4.27) 0.028* (1.88)

AuditComNetworking AuditComEducational AuditComExperience Outsider ratio Leverage R&D Firm size Board size ROA INSTOWN Risk SalesHHI Observations Adjusted R2

0.102 (0.52) -0.376** (-2.47) 0.545** (2.57) 0.233*** (3.50) -0.013 (-1.03) -0.443 (-0.74) -0.175 (-1.02) 0.689*** (3.54) 0.021 (0.19) 1,196 0.28

(2) (3) Fixed-effects Time Clustered -1.520*** -1.478*** (-4.50) (-4.25) 0.025** (2.04) 0.001 (0.34) 0.005* (1.85) 0.015** (2.33) 0.098 0.099 (0.79) (0.52) -0.364** -0.373** (-2.44) (-2.42) 0.562*** 0.551*** (2.63) (2.61) 0.236*** 0.234*** (2.92) (3.83) -0.012 -0.014 (-0.82) (-1.03) -0.416 -0.422 (-0.72) (-0.73) -0.171 -0.173 (-1.00) (-1.10) 0.685*** 0.688*** (3.47) (3.53) 0.017 0.020 (0.15) (0.19) 1,196 1,196 0.26 0.30

(4) Fixed-effects -1.487*** (-4.42)

-0.003 (-0.76) 0.006** (1.99) 0.017** (2.40) 0.084 (0.43) -0.357** (-2.38) 0.544*** (2.60) 0.239*** (3.49) -0.013 (-1.20) -0.458 (-0.77) -0.192 (-1.11) 0.686*** (3.56) 0.018 (0.17) 1,196 0.28

(5) (6) Time Clustered Fixed-effects 9.762*** 9.174*** (8.64) (8.50) -0.007** -0.013*** (-2.44) (-2.62) . . .

.

.

.

0.522** (2.07) 0.265 (1.38) 0.983** (2.27) 0.479*** (4.57) 0.024* (1.88) -0.673** (-2.07) -0.712*** (-3.29) -3.256*** (-4.52) -0.363** (-2.45) 1,190 0.63

0.214* (1.77) 0.224 (1.57) 0.902** (2.05) 0.585*** (3.87) 0.036* (1.70) -0.502* (-1.75) -0.594*** (-2.93) -1.674*** (-3.45) -0.151** (-2.32) 1,190 0.55

43

Table V Outsiders’ and Insiders’ Capital on Disclosure Quality The dependent variable is industry-adjusted disclosure quality. Independent variables include outsiders’ and insiders’ networking capital, educational capital and experience capital. Firm size is the logarithm of annual sales. Board size is the number of directors. Outsider ratio is the number of outside directors divided by the board size. Leverage is total liabilities divided by total assets. ROA is the net income before extraordinary items divided by total assets. R&D is the research & development expenses divided by total sales. INSTOWN is the total percentage equity ownership of all the institutional investors. Risk is volatility of stock returns. SalesHHI is a Herfindahl index based on sales. Statistical significance at the 10, 5 and 1 percent levels is indicated by *, **, ***, respectively. t-statistics are Huber-White-Sandwich using firm clustering.

Dependent Variable: Disclosure Quality

Variable Intercept Outsidercapital Insidercapital Networking Capital (outsiders)

Time Clustered (1) -1.292*** (-3.09) 0.014*** (2.99) 0.006 (1.00) .

Educational Capital (outsiders)

.

Experience Capital (outsiders)

.

Networking Capital (insiders)

.

Educational Capital (insiders)

.

Experience Capital (insiders)

.

Outsider ratio Firm size Leverage Board size R&D ROA INSTOWN Risk SalesHHI Observations Adjusted R2

0.058 (0.77) 0.225*** (4.67) -0.385*** (-3.01) -0.035* (-1.82) 0.486** (2.27) -0.463 (-0.79) -0.186 (-1.09) 0.681*** (3.53) 0.031 (0.29) 1,196 0.35

Time Clustered (2) -1.487*** (-4.18) . . -0.001 (-0.33) 0.009 (1.36) 0.008** (2.09) -0.004 (-1.21) 0.008** (2.19) 0.008** (2.18) 0.230 (0.99) 0.232*** (3.99) -0.349*** (-2.97) -0.040** (-2.42) 0.468** (2.13) -0.482 (-0.83) -0.211 (-1.21) 0.668*** (3.44) 0.014 (0.14) 1,196 0.33

Fixed-effects (3) -1.490*** (-5.03) . . -0.001 (-0.23) 0.008 (1.42) 0.010** (2.22) -0.003 (-0.97) 0.007** (2.05) 0.008** (2.07) 0.109 (0.69) 0.223*** (3.62) -0.333*** (-3.00) -0.037** (-2.07) 0.400** (2.02) -0.459 (-0.68) -0.194 (-1.07) 0.642*** (3.20) 0.015 (0.22) 1,196 0.31

44

Table VI Board Capital and Firm Performance This table presents the results of regressing board capital on firm performance and various control variables. The dependent variable is industry-adjusted Tobin’s Q. BoardCapital (AuditCapital) is an index developed from the underlying proxies. Firm size is the logarithm of annual sales. Board size is the number of directors. Outsider ratio is the number of outside directors divided by the board size. Leverage is total liabilities divided by total assets. ROA is the net income before extraordinary items divided by total assets. R&D is the research & development expenses divided by total sales. INSTOWN is the total percentage of institutional ownership. Statistical significance at the 10, 5 and 1 percent levels is indicated by *, **, ***, respectively. t-statistics are Huber-White-Sandwich estimates with firm clustering.

Dependent variable: Industry-adjusted Tobin’s Q Variable Intercept BoardCapital AuditCapital Outsider ratio Firm size Leverage Board size R&D ROA INSTOWN Observations Adjusted R2

(1) OLS 3.530*** (5.02) 0.030*** (2.82) .

(2) Fixed Effects 2.457*** (4.70) 0.026** (2.33) .

(3) 2SLS 4.682*** (5.16) 0.033** (2.34) .

0.122 (1.35) -0.211*** (-4.57) -0.254 (-0.72) -0.066*** (-3.76) 1.677*** (3.33) 6.334*** (4.33) -1.375 (-1.12) 1,220 0.41

0.229 (1.44) -0.200*** (-4.16) -0.232 (-0.75) -0.052*** (-2.87) 1.357*** (3.50) 6.215*** (4.18) -1.223 (-1.32) 1,220 0.34

0.560 (1.05) -0.117** (-2.11) -0.038 (-0.90) -0.120 (-1.04) 1.479** (2.33) 9.086*** (7.21) -0.716 (-1.55) 1,220 0.39

(4) OLS 3.678*** (4.56) . 0.018*** (2.84) 0.066 (1.11) -0.167*** (-3.20) -0.154 (-0.45) -0.070*** (-3.22) 1.505** (2.22) 6.945*** (4.67) -1.121 (-1.43) 1,220 0.34

(5) Fixed Effects 3.420** (2.22) . 0.016*** (3.02) 0.100 (1.09) -0.206*** (-2.99) -0.132 (-1.04) -0.064*** (-3.33) 1.003*** (3.01) 5.002*** (3.78) -0.978 (-1.44) 1,220 0.30

(6) 2SLS 4.995*** (6.03) . 0.020** (2.22) 0.105 (1.22) -0.223*** (-3.35) -0.165 (-1.22) -0.085*** (-3.78) 1.395*** (3.53) 6.783*** (5.21) -0.886 (-1.52) 1,220 0.35

Table VII Voluntary Disclosure of Board Capital and Firm Performance This table presents the results of regressing voluntary disclosure of board capital on firm performance with various control variables. The dependent variable is industry-adjusted Tobin’s Q. There are three categories of information gap. BoardSeatsGap is a dummy variable, which equals 1 if the firm discloses corporate board seats information for the majority of the directors. NonProfitGap is 1 if the firm discloses non-profit board seats for the majority of the directors. EducationGap is the percentage of the directors whose education information is disclosed. WorkingGap is coefficient of variation of the number of years of each director’s working background. Firm size is the logarithm of annual sales. Board size is the number of directors. Outsider ratio is the number of outside directors divided by the board size. Leverage is total liabilities divided by total assets. ROA is the net income before extraordinary items divided by total assets. R&D is the research & development expenses divided by total sales. INSTOWN is the total percentage equity ownership of all the institutional investors. Statistical significance at the 10, 5 and 1 percent levels is indicated by *, **, ***, respectively. t-statistics (by Huber-White-Sandwich estimates, allowing for correlation of all observations of a given firm) are in parentheses. Dependent variable: Industry-adjusted Tobin’s Q Variable (1) (2) (3) (4) Intercept 4.727*** 4.394*** 4.233*** 4.508*** (6.71) (6.51) (5.57) (5.69) WorkingGap 0.198* 0.185* (1.91) (1.85) EducationGap 0.499*** 0.507*** (2.68) (3.07) BoardSeatsGap 0.175 0.100 (0.49) (0.26) NonprofitGap 0.167* 0.147** (1.76) (1.99) Outsider ratio 0.421 0.419 0.361 0.496 (1.05) (1.08) (0.93) (1.25) -0.089*** -0.075*** -0.081*** -0.083*** Board size (-2.93) (-2.65) (-2.80) (-2.78) Firm size -0.289*** -0.263*** -0.275*** -0.284*** (-4.94) (-4.72) (-4.88) (-4.87) Leverage 0.371 0.335 0.316 0.385 (0.63) (0.58) (0.54) (0.66) R&D 1.017* 1.012* 1.050* 0.932* (1.93) (1.81) (1.86) (1.80) ROA 5.629*** 5.482*** 5.489*** 5.564*** (3.91) (3.94) (3.87) (3.87) INSTOWN -1.583*** -1.507*** -1.345*** -1.741*** (-2.96) (-3.01) (-2.65) (-3.30) Observations 1,220 1,220 1,220 1,220 2 Adjusted R 0.35 0.36 0.35 0.36

46

Appendix A - Director Statistics Panel A: Director Educational Backgrounds This panel shows the percentage of individual directors with different level of education within the board, the audit committee, the insiders and outsiders, respectively. “Higher degree” refers to a master’s degree or above. Board Audit Committee Insiders Outsiders Bachelor’s 84.37 85.50 79.16 85.99 MBA 28.39 32.74 24.76 29.54 Master’s 55.67 58.12 48.65 57.90 Law 11.56 10.54 9.81 12.11 PhD 10.23 10.37 6.63 11.34 Higher degree 60.03 62.20 51.49 62.73 Panel B: Director Experience Capital Certified means directors with professional certifications (e.g. CPA). Professional Experience includes directors with working experience as a partner/executive in law firms, accounting firms, consulting firms, and/or investment banking firms. Professor means directors with academic experience. Honors & awards refer to national-level honors and awards (e.g. Hall of Fame). Associations refers to the membership in a professional or industrial association. Politician means directors with prior government position experience. Government board refers to the membership of directors on any government board, which is designated or nominated by government agents or agencies (such as Congress or state and municipal governments). Board Audit Committee Insiders Outsiders Certified 5.84 9.82 5.01 6.11 Professional Experience 40.93 50.20 23.54 46.78 Honors & awards 4.24 3.06 2.91 4.66 Associations 10.94 11.72 9.07 11.53 Professor 11.15 11.51 3.99 13.43 Politician 8.85 8.33 3.11 10.68 Government board 13.65 13.57 6.63 15.87 Panel C: Director Networking Capital Firms worked refers to the number of firms the directors have worked with. Higher posts means the number of higher posts (higher than vice president) a director has obtained in lifetime. Total boards means the total number of boards (corporate and nonprofit) a director currently sits on. NONPROFIT measures the number of nonprofit boards a director currently serves. Nonprofit ever means the number of all the nonprofit boards a director has ever served on. Company boards measures the number of corporate boards a director currently serves. Board Audit Committee Insiders Outsiders Firms worked 2.77 2.91 2.49 2.87 Higher posts 5.37 5.32 5.43 5.36 Total boards 3.13 3.22 2.33 3.39 Nonprofit boards 1.25 1.27 1.09 1.31 Nonprofit ever 2.70 2.84 1.84 2.97 Company boards 1.87 1.95 1.23 2.08

47

Appendix B - Board Capital, Board Disclosure and Firm Performance This table presents the results of regressing board capital and board disclosure on firm performance and various control variables. The dependent variable is industry-adjusted Tobin’s Q. BoardCapital is an index developed from the underlying proxies. BoardDisclosure is the predicted value and SupplementDisclosure the residual from the following regression, DisclosureQuality = β1 × networking + β 2 × education + β 3 × exp erience + ξ Firm size is the logarithm of annual sales. Board size is the number of directors. Outsider ratio is the number of outside directors divided by the board size. Leverage is total liabilities divided by total assets. ROA is the net income before extraordinary items divided by total assets. R&D is the research & development expenses divided by total sales. INSTOWN is the total percentage of institutional ownership. Risk is stock volatility. Statistical significance at the 10, 5 and 1 percent levels is indicated by *, **, ***, respectively. t-statistics are Huber-White-Sandwich estimates with firm clustering.

Dependent variable: Industry-adjusted Tobin’s Q Variable Intercept BoardCapital BoardDisclosure SupplementDisclosure Outsider ratio Firm size Leverage Board size R&D ROA INSTOWN Observations Adjusted R2

(1) 2.043*** (3.27) 0.013** (2.32) . 0.165*** (4.10) -0.195 (-0.53) -0.130*** (-3.25) 0.099 (0.30) -0.062* (-1.96) 1.331* (1.85) 9.113*** (6.03) -0.597* (-1.68) 1,196 0.34

(2) 2.065*** (3.52) . 0.118** (2.30) 0.165*** (4.10) -0.197 (-0.53) -0.117*** (-3.24) 0.107 (0.22) -0.063* (-1.89) 1.331* (1.84) 9.293*** (6.04) -0.567* (-1.67) 1,196 0.35

(3) 2.026*** (3.77) 0.002 (1.05) 0.173*** (3.00) 0.120*** (4.09) -0.204 (-0.50) -0.119*** (-3.24) 0.131 (0.24) -0.062* (-1.90) 1.330* (1.84) 9.111*** (6.03) -0.561* (-1.69) 1,196 0.35

48

Appendix C – Impact of Financial Expertise In this table we regress board capital on disclosure quality and separately include financial expertise of the audit committee. The dependent variable is industry-adjusted disclosure quality. AccntExp is the number of directors with CFO experience, with CPA/CFA certificate, with accounting/auditing firm partner experience, or designated as “financial expert”. Firm size is the logarithm of annual sales. Board size is the number of directors. Outsider ratio is the number of outside directors divided by the board size. Leverage is total liabilities divided by total assets. ROA is the net income before extraordinary items divided by total assets. R&D is the research and development expenses divided by total sales. INSTOWN is the total percentage equity ownership of all the institutional investors. Risk is the volatility of stock return. SalesHHI is a Herfindahl index based on sales. Statistical significance at the 10, 5 and 1 percent levels is indicated by *, **, ***, respectively. t-statistics (by Huber-White-Sandwich estimates, allowing for correlation of all observations of a given firm) are in parentheses.

Dependent variable: Disclosure Quality Variable Intercept BoardCapital AccntExp Outsider ratio Leverage R&D Firm size Board size ROA INSTOWN Risk SalesHHI Observations Adjusted R2

(1) OLS -1.299*** (-4.05) 0.016*** (3.00) 0.031** (2.37) 0.035 (0.78) -0.437*** (-3.05) 0.516** (2.02) 0.225*** (3.60) -0.038** (-2.09) -0.939 (-1.00) -0.077 (-0.45) 0.729*** (3.14) 0.064 (0.69) 1,196 0.34

(2) 2SLS -1.410*** (-3.85) 0.018*** (3.24) 0.035** (2.27) 0.033 (1.00) -0.458*** (-3.27) 0.480** (2.39) 0.212*** (3.81) -0.039** (-2.34) -0.827 (-0.93) -0.054 (-0.46) 0.721*** (4.09) 0.057 (0.77) 1,196 0.36

49