Director Reputation and Settling-Up: Evidence from Internal and External Labor Markets

Transcription

1 Director Reputation and Settling-Up: Evidence from Internal and External Labor Markets Mark A. Chen, Qinxi Wu, and Evgenia Zhivotova* January 2017 ABSTRACT We examine the effects of individual reputation on career outcomes in the labor market for corporate directors. Using unique data on prestigious, individual-level awards, we isolate the career effects of reputational shocks by employing a number of identification strategies. We find that an enhanced reputation increases a director s likelihood of sitting on the board of a firm with larger size, more public prestige, greater social responsibility, and less fraud. A positive shock to reputation allows a director to upgrade his or her portfolio by gaining new, more desirable board seats. Moreover, improved reputation raises the likelihood that a director holds key internal board leadership roles. Overall, our findings suggest the presence of strong reputational rewards and ex post settling-up in the internal and external markets for boardroom talent. JEL Classification Codes: G30, G31, G34 Keywords: Boards, directors, reputation, career concerns, internal labor markets, external labor markets, settling-up *Chen is from the J. Mack Robinson College of Business, Georgia State University, Atlanta, GA 30303, USA. Tel.: ; machen@gsu.edu. Wu is from the J. Mack Robinson College of Business, Georgia State University, Atlanta, GA 30303, USA. Tel.: ; qwu4@gsu.edu. Zhivotova is from the University of Amsterdam, The Netherlands; E.Zhivotova@uva.nl. We thank Conrad Ciccotello, Daniel Greene, Lixin Huang, Omesh Kini, Ron Masulis, Daniel Urban, and seminar participants at Georgia State University, Clemson University, and the 2016 Greater China Area Finance Conference for valuable comments and suggestions. For helpful conversations regarding director awards, we are grateful to Lori Smith, Executive Director of the Forum for Corporate Directors, Judy Warner, Editor-in-Chief of NACD Directorship magazine, and Heather Wolf, Program Director of the Financial Times Outstanding Directors Exchange Program. All errors and omissions are the sole responsibility of the authors.

2 1. Introduction What incentives do corporate directors have to monitor management on behalf of the firm s shareholders? At least since Fama (1980), it has been argued that an ex post settling-up mechanism exists in the director labor market under which individuals are motivated to build reputations as monitoring experts in order to gain future directorships. While a number of studies have examined the impact of firm-level events on the careers of directors, precise tests of the ex post settling-up hypothesis have been lacking due to the difficulty of observing individual directors actions or performance that might lead to greater or lesser reputational capital. As Jiang (2015) and others point out, most research examining boards has focused on observable firm-level or board-level characteristics, and thus it remains a challenge to clearly attribute outcomes in the director labor market to the performance or capabilities of individual directors rather than the endogenous, firmlevel choice of board structure. In this paper, we provide novel evidence on reputational concerns and settling-up for individual corporate directors. Specifically, we construct a unique dataset that details the occurrence of prestigious national awards given to individual directors for excellence in the boardroom. Using these awards to capture positive shocks to reputation, we study the causal effects of enhanced reputational capital on directors labor market opportunities. We find strong evidence that winning an award leads to rewards in the labor market in the form of greater numbers of board seats and seats at larger firms. In addition, we document that reputable directors enjoy greater opportunities in the form of an increased likelihood of winning board seats at firms that exhibit a high degree of public status and prestige. Awardees are also more likely than comparable nonawardees to hold key leadership roles, including chairman roles on key board committees and of the board as a whole.

3 A key advantage of using awards to capture directors reputational shocks is that the data can help overcome the identification problem stemming from heterogeneity at the director level (e.g., unobserved differences in ability, talent, or network size). The panel of individual awards data allows us to control for director fixed effects in regression analysis, thereby removing likely confounders that would arise in tests linking firm-level performance to director-level rewards or penalties in the labor market. Since the director awards are central to our study, it is critical to establish that our results still hold despite the fact that the awards themselves could be endogenously determined by observable or unobservable factors. Indeed, we observe in our sample that award winners differ substantially from the much larger population of non-awardee directors along various dimensions such as age, experience in the boardroom, gender, existing reputation, number of board seats, median size of affiliated companies, and the extent of their professional and educational networks. Unobservable factors could also drive the selection of awardees while at the same time affecting labor market outcomes. We therefore adopt two complementary approaches in order to strengthen identification. First, we use a propensity-score matching approach (Rosenbaum and Rubin, 1983). We select, for each awardee, a group of non-awardee directors who have an ex ante similar likelihood of winning an award based on a large set of observable director, board, and firm characteristics. 1 The set of characteristics used in the matching includes not only age, experience, and gender, but also an extensive set of portfolio measures that characterize a director s workload and their professional 1 The propensity-score matching approach controls for potentially confounding variables that are empirically observable. This approach is particularly well-suited for our current study given that, unlike most managerial or board decisions, award decisions are made by outside second parties (e.g., award committees or advisory board members). Thus, the decision-makers with respect to awards, like researchers, must rely mainly on public information to assess director quality. See Malmendier and Tate (2009) for a similar justification of the use of propensity-score matching estimators in their study of CEOs business awards. 2

4 connections to other directors and specific members of the award selection committees in a given year. We confirm that the matching procedure works well: within the matched sample of awardees and non-awardees, the various individual-level and firm-level covariates have essentially no ability to predict who will go on to win an award. Second, we use an instrumental variables approach to address potential endogeneity due to unobserved heterogeneity. Our choice of instrument is motivated by institutional details surrounding the director awards. Specifically, we make use of the fact that award selection committees designate winners based on nominations received from individuals executives and other board members who serve on a board alongside the nominee. We exploit the structure of indirect connections to quantify the extent to which a director has colleagues who were associated with other director awards in the past. The fact that a director X serves on a board with another director Y who helped nominate a third party for an award in the past serves as a valid instrument because it is unlikely that the occurrence of the third-party award directly affects the labor market outcomes for director X. In the first part of the main analysis, we use panel data regressions to examine the effects of winning an award on the number and prestige of seats in directors portfolios. We document that winning an award has a significant effect not only on the number of seats in a director s portfolio, but also on the likelihood of serving on more desirable seats as measured in several different ways. Indeed, awardee status increases the likelihood that a director holds a seat at a company that is large; one that has been named as one of the world s most admired companies; or one that has been recognized as a top global brand or an outstanding corporate citizen. To gain a better understanding of what drives directors gain of prestigious board seats, we conduct a relative comparison of the prestige of a newly gained seat to the prestige of a director s pre-existing seats. We construct a prestige index that incorporates all of our positive firm-level 3

5 desirability characteristics: firm size as well as being the recipient of a prestigious company award. We then examine whether winning an award affects the likelihood that a director s new seat upgrades his or her seat portfolio. We find a strong dichotomy: an award increases the probability of securing a new affiliation that is better (more prestigious than the existing median), while it decreases the probability that the director gains an inferior (i.e., below-median) board seat. Given that directors appear to rationally prioritize among multiple directorships (Masulis and Mobbs (2014)), our finding that awardee directors eschew less desirable board seats suggests that they actively manage their time demands and substitute towards either more leisure or more rewarding directorships. A busy, in-demand awardee director may also benefit from the positive reputational shock by being able to choose board opportunities more carefully. With a strong external signal of quality in the form of an award, a director no longer needs to hold as many directorships to signal quality; he or she can substitute towards leisure while still maintaining his or her reputation level. Thus, to complement our findings on prestigious directorships, we also investigate whether new awardees have a greater tendency to forego new board seats at undesirable firms. Consistent with the idea that awardees can afford to be choosier, we find that awards are associated with a reduced probability of gaining directorships at firms recently experiencing fraud-related financial restatements. Finally, we explore the question of whether winning an award confers benefits to a director in the form of key internal leadership roles on the board. Key leadership roles on the board can be seen as an extension of the external labor market for directorships and, like prestige at the firm level, may represent a reward for high reputation. We consider three types of internal roles: committee chair roles on major board committees (audit, nominating/governance, and compensation); lead or presiding directorships; and board chairmanships. Committees are where 4

6 boards do most of their work (Adams, Hermalin, and Weisbach, 2010). 2 Thus, serving as chairman for a major board committee can bring prestige and higher director pay. Likewise, lead directorships and board chairmanships can bring a director high visibility, responsibility, and status. 3 Appointing an individual to one of these roles is a strong endorsement of the individual s ability to manage the board and to interface with shareholders and executive management. Our instrumental variables regressions reveal that awardees are significantly more likely than non-awardees to hold the role of audit committee chair, compensation committee chair, and board chair. We interpret these findings as evidence that the ex post settling-up mechanism for directors likely extends into internal labor markets and that reputable directors can be rewarded in various ways apart from simply accumulating more board seats. Overall, this study contributes in several ways to the literature on director labor markets and reputational concerns. First, our approach provides a powerful new method of measuring changes in director reputation at the individual level. Much of the existing literature on director labor market consequences focuses on specific types of firm-level reputational shocks, such as dividend cuts (Kaplan and Reishus, 1990), fraud (Fich and Shivdasani, 2007), option backdating (Ertimur, Ferri, and Maber, 2012), accounting restatements (Srinivasan, 2005), and bankruptcy (Gilson, 1990). Collectively, these papers and others find mixed evidence on the impact of firm events on directors subsequent labor market opportunities. 4 In comparison, our analysis focuses on director- 2 Prior work has shown that the structure of key board committees can affect firm performance (Klein (1998)), earnings quality (Anderson, Deli, and Gillan (2003)), and the process of nominating new directors to the board (Shivdasani and Yermack (1999)). 3 In 2014, about 54% of S&P 500 firms had a lead director on the board (Ernst & Young (2014)). Lead directors have assumed an increasing number of roles and responsibilities in recent years, including conducting performance evaluations of the CEO and other directors, calling meetings of the entire board, retaining outside consultants for the board, liaising between the CEO and independent directors, handling communications between investors and the board, and giving guidance or assurances to independent auditors (Penbera (2009)). 4 Some studies document that directors tend to lose more outside directorships after negative firm outcomes, including poor firm-level stock performance (Yermack, 2004), financial fraud lawsuits (Fich and Shivdasani, 2007), dividend cuts (Kaplan and Reishus, 1990), and decisions to retain state antitakeover provisions (Coles and Hoi, 2003). Also, directors who are nominated for re-election during a proxy contest tend to lose more board seats than directors not up 5

7 level shocks, which allows us to more precisely attribute changes in reputation to the individuals who are likely to benefit from improved labor market opportunities. As a result, we are able to implement cleaner tests of ex post settling-up. The results of our work also help give a richer, more complete picture of the types of rewards that directors can experience in the labor market. We show that focusing on the quantity of board seats reveals only part of the picture; the quality of board seats is also important to consider when assessing that benefits that directors can reap by improving their reputations. Indeed, our evidence suggests that, in some cases, directors with good reputation can afford to forego board seats of companies that are tainted by accounting fraud. Our findings also clarify the interpretation of previous empirical findings in the literature on labor markets for directors. While the negative firm-level outcomes studied in existing papers may indeed lower directors reputation, an observed loss of board seats after a negative firm-wide shock does not necessarily imply a reputation effect. As a number of authors have recognized, a negative firm outcome could cause a director on the board to become busier, more distracted, less willing to face the risk of lawsuits, or less able to handle the workload of multiple board commitments. These effects could cause a director to voluntarily seek out fewer board seats, confounding any ex post settling-up penalty from reputational effects. Since our study links positive reputational shocks to gains in large-firm directorships and key board roles, our findings are largely immune to the concern that increased risk-aversion or busyness are driving the results. for re-election (Fos and Tsoutsura, 2014). On the other hand, there is no evidence to suggest that directors are more likely to lose outside board seats after being individually named in lawsuits (Brochet and Srinivasan, 2014), serving on the compensation committee of a firm involved in options backdating (Ertimur, Ferri, and Maber, 2012), or presiding over value-destroying acquisitions (Harford and Schonlau, 2013). 6

8 The rest of the paper is organized as follows. In Section 2, we describe the director awards data and the construction of the propensity-score-matched sample of non-awardees. Section 3 describes our instrumental variables approach and presents our main results. Section 4 concludes. 2. Data and Sample description 2.1. Sample Construction Our sample is constructed mainly relying on the BoardEx database. BoardEx provides data on individual director characteristics and board attributes, allowing us to construct a panel data from 1999 to Although we do have director award events prior to 1999, we exclude them from the sample as coverage in the BoardEx database effectively begins in Based on other studies relying on BoardEx (e.g., Fracassi and Tate, 2012; Engelberg, Gao and Parsons, 2013) and our own investigation, BoardEx s coverage of U.S. public firms is highly incomplete for However, in robustness tests, we have verified that our main results are unchanged if 1999 is excluded. We obtain data on firm characteristics and stock performance from the CRSP/Compustat merged database. This filters out boards of small firms and private firms. We believe it is reasonable, in the context of studying reputation effects of awards, to focus on boards of relatively larger firms. This is because larger firms, being more visible and prestigious, provide stronger reputation incentives and thus encourage directors to prioritize their efforts on these boards (Masulis and Mobbs, 2014, 2015). Finally, we hand-match information from the BoardEx database with information on director awards 5 by company affiliation and director name. 5 A detailed description of the director award data is presented in section

9 2.2. Director awards Our study relies on a manually collected list of awards given to outstanding individual directors during the time period We focus on awards that are national in scope, open to public-company directors in the U.S., and aimed at recognizing individuals who exemplify excellence in the boardroom. Based on discussions with industry professionals and perusals of news articles, trade publications, and corporate governance websites, we identify four distinct types of awards or honors that fulfil these criteria: 1) the Outstanding Director Award sponsored by Financial Times (FT-ODX); 2) the Director of the Year Award given at the annual dinner event hosted by the Forum for Corporate Directors (FCD); 3) director awards given by the National Association of Corporate Directors (NACD); and 4) appointment to an NACD Blue Ribbon Commission. Each of these four awards is national, very prestigious, and highly selective. To be considered for any of these awards, nominees need to exhibit extraordinary qualities of courage and leadership under duress, facilitate superior board performance by bringing out the best in directors. As evident in the nomination process for each award, the qualities that are recognized in the directors are general in nature rather than attributable merely to service on a particular board. For example, one of the nomination criteria for the Outstanding Director of FT-ODX program is that a nominee needs to have shown consistent and strong contributions that have moved the needle to boards of the companies served. The ODX recognizes outstanding directors for their ability to challenge management decisions, for being respected by the CEO and fellow directors for their courage, integrity and consensus building skills, and naturally for clearly aligned interests with shareholders 6. J. Michael Cook, President of the ODX Institute, states that Our honorees have 6 These are directors who are dedicated to good corporate governance and who, as Business Week has reported of our efforts, are the stars of the boards who stand up for shareholders and stand up to management. 8

10 been selected by their peers for demonstrating courage, diligence and leadership in the boardroom. They have made an indelible mark on the companies they serve." For NACD awards, the nomination criteria are based on the NACD s promoted standards for integrity, mature confidence, informed judgement, and high performance standards. 7 Although the FCD does not publish formal criteria for nominations, Peter Craig, FCD Chairman, emphasizes that Directors of the Year honorees are awarded for the highest standards of corporate ethics and best governance practices for board of directors of both public and privately-held corporations and institutions. The national scope of the awards in our dataset is apparent from the calls for nominations that invite supporting letters from all over the country. Nominations can originate from any source for each type of award, namely from CEOs, fellow directors. For the ODX award, a nomination needs to be seconded by at least one independent director who sits on the board with the candidate. For the NACD award, each nomination requires three recommendation letters that highlight the most recent achievements or the entire career history of achievements. Winners of each of the awards are typically selected from the nominees by an independent committee, and the final decision is made by the board of directors of each respective organization. For example, the editorial board (the ODX team and editorial of the Agenda Week magazine) together with its advisory board are ultimately responsible for naming each individual class of outstanding directors of the FT-ODX. Winners of NACD awards are selected by an independent committee that is annually formed and consist of four to five members. The committee members names are not publicly known, but all members are deemed to be independent and diverse, and 7 For instance, Kenneth Dale, the CEO of NACD, notes that these awards are important because they provide deserved recognition to four individuals who have made incredible contributions to board effectiveness, corporate performance and shareholder value. The winners are great representatives of so many directors across the country who are performing at very high levels in challenging corporate environments. 9

11 often are previous winners of an award. FCD forms a special Director of the Year committee that is responsible for selection of the awardees and consists of the members of the FCD board of directors and individuals who applied to participate in the selections process on a yearly basis. NACD Blue Ribbon Commissions (BRCs) are usually comprised of directors and professionals who are deemed to be experts in a given year s focus area for the Commission. Based on our readings, many of the Blue Ribbon Commissioners are named as recipients of awards by the NACD or other organizations within a few years after serving as a Commissioner. Therefore, we consider a BRC appointment to be effectively an award that functions as an endorsement of a director s expertise in a particular area. We use various sources to collect lists of recipients of these awards, including news articles, magazine stories, archived internet websites, BRC online reports, and hard copies of lists provided by representatives of award organizations. 8 (For details regarding the collection of awards data, including award announcement dates and the names of awardees, refer to Appendix B.) We exclude awardees who are not corporate directors of public companies in the U.S. or for whom we cannot obtain information on director and board characteristics. As a result of these criteria, we obtain a list of 731 awardees. Table 1 presents the award frequencies at individual director level. In all, 663 individuals won a certain type of awards within one sample year, 31 directors received 2 different types of awards in a year, and 2 individuals were granted 3 different awards within one year. With regard to award frequencies by type, 398 out of 731 (54%) individuals are NACD award winners, 135 individuals received NACD BRC, 124 directors were granted FT-ODX award, while FCD award were given to 74 individual directors. [Insert Table 1 about here] 8 We thank Judy Warner, Editor-in-Chief of NACD Directorship magazine, for graciously providing us with some of the data on NACD awardees and Blue Ribbon Commissioners. 10

12 Figure 1 provides a histogram of award frequencies, by sample year and award type. No NACD award was granted in No Blue Ribbon Commission was convened in In 2007 and 2008, the number of NACD awardees exhibits a sizeable increase because new sub-categories of awards were introduced during that timeframe. [Insert Figure 1 about here] 2.3. Measures of Board Desirability To measure the outcomes of a director s enhanced reputation, a straightforward method would be tracking the net change in the total number of board seats held. As suggested by recent work (Masulis and Mobbs, 2014, 2015), however, directors are more likely to devote their time and attention to relatively prestigious boards in their directorship portfolio. In spirit of this finding, we believe it is important to take board desirability into account since, regardless the number of directorships, gaining a board seat in a prestigious firm can also be considered a type of labor market rewards. We use different measures to proxy for board desirability. Following Masulis and Mobbs (2014), we construct our first board-prestige measure as whether a firm is categorized as large relative to sample directors board seats. Specifically, a large firm is defined as a firm with market capitalization in the top decile of the sample within a given year. Another type of desirability measures is whether a firm was the recipient of one or more prestigious company awards in the past 3 years. Since honoree firms are, in general, well-known and outstanding, serving on the board of such a firm could bring considerable visibility and reputation to a director. We set a 3-year window because intuitively the influence of awards could decreases over time. Similar to director awards, we focus on influential awards that are national in scope, potentially open to all the public firms in the U.S. We chose three company awards which aim at recognizing firms from different aspects: 1) the Fortune Most Admired Companies, 2) the 11

13 Best Global Brands, and, 3) the Best Corporate Citizens. Since 1983, the Fortune magazine provides a ranking list of America s Most Admired Companies based on a Corporate Reputations Survey conducted by Hay Group. Honorees are selected based on 8 attributes, including the quality of management, the quality of products (services), and so on. The Best Global Brand ranking is provided by Bloomberg Business Week. Criteria include market scale, transparency, profitability, and so on. The Best Corporate Citizen is granted by the CR magazine and aims to recognize firms that well fulfill corporate responsibility. For details regarding the company awards, refer to Appendix B2. Our third measure is the non-reliance financial restatements which signal potential financial frauds or accounting application failure, and thus can be used to proxy for the undesirability of a firm. Specifically, we generate an indicator equal to 1 if a firm made nonreliance financial restatements in the past 3 years. We obtain non-reliance financial restatements from Audit Analytics for sample firms. Audit Analytics also provides the disclosure date and the information on restatement categories (e.g., financial fraud, board s involvement, material accounting and clerical application errors) Descriptive statistics Director, Firm, and Board Characteristics Our overall sample consists of 581,617 director-firm-year observations (70,245 firm-year observations and 456,877 director-year observations) over the period from 1999 to Table 2 reports selected summary statistics for director, board, and firm characteristics for all sample directors and firms. [Insert Table 2 Here] 12

14 Panel A of Table 2 presents summary statistics for firm and board characteristics. Consistent with previous literature, our sample boards, on average, have 8.22 directors, 70% of who are identified as independent according to the applicable NYSE or Nasdaq regulatory definitions. The average market cap of sample firms is 4,068, while the average total asset is 8,142, both in million dollars. The average book-to-market ratio, industry adjusted ROA and industry adjusted monthly return is 0.53, and 0.104, respectively. With regard to the measures of desirability, 3.3%, 3.3% and 1.1% of our sample firms won the AMAC award, the BGB award, and the BCC award, respectively, in the 3-year window before a given year, while 1.7% of sample firms made at least 1 financial restatement in the past 3 years. As shown in Panel B of Table 2, our sample directors are around 60 and hold 1.4 directorships on average, 90% of whom are male. The percentage of sample directors hold a leadership role (chairman, CEO, lead director, committee chair) on at least one board is 14%, 14.4%, 3.2%, and 29.4%, respectively. We measure director tenure, board size, market capitalization, are measured as medians across all board affiliations of an individual director. The average of median director tenure is 7.99 in our sample. The median board size, book-to-market, ROA and stock return of board affiliations are 9.1, 0.58, and 0.1, respective Awardees vs. Matched Non-Awardees Intuitively, awardees are different from the other non-award winners in many dimensions. It is important to control for these differences in estimating the effects of winning an award as the biases likely reflect sample selection based on observable director, firm, and board characteristics. One possible approach to controlling for sample selection bias based on observables would be to compare an award winner with other linked non-awardee directors who shared a board seat with the awardee at the time of the award. Such an approach would partly mitigate the concern that winning an award is influenced by firm performance or other firm and board characteristics. 13

15 However, there are two reasons we do not employ this approach. First, although the non-awardee group share at least one board seat with an awardee, in general they may have other board affiliations that are not shared with awardees. Second, using the linked non-awardee group as a benchmark would not adequately control for the differences in individual director characteristics between the two groups. In order to isolate the treatment effect of winning an award from a potential selection effect, we need to construct a matched sample of directors who have similar personal characteristics and board portfolio characteristics so that winning an award is effectively random across directors. Our empirical approach is to construct a benchmark control sample using propensity score matching (PSM) as in Rosenbaum and Rubin (1983) and Abadie and Imbens (2007). This approach involves matching each treated director (award-winner) with one or more non-treated directors (non-winners) who have a similar propensity of winning, where the propensity score is estimated from covariates that are observed at the time of the award. While the PSM approach does not directly address sample selection based on unobservable factors, this type of sample selection is less of a concern for the current setting as the final decisions of award-granting organizations are arguably driven by what they can observe, such as directors personal characteristics or corporate outcomes. 9 The first step to implementing the PSM approach is to estimate a logit regression relating the probability that a director wins an award with director and firm characteristics. This regression includes all directors in all months in which an award was given, with the exception of non- 9 Malmendier and Tate (2009), in their study of CEO business awards, provide a similar justification for the use of propensity-score matching and other matching methods. 14

16 awardee linked directors who share a board with an awardee. 10 The dependent variable in the regression is a dummy variable equal to 1 if a director wins an award in that month. It is apparent that both director and firm characteristics are relevant covariates to include in the regression. We include as covariates director age, gender, median board size, longest tenure at any currently-held board seat, and number of board seats at public companies. Measures of directors network connections are also included. Following prior literature (e.g., Cohen et al.), we include the log-transformed number of other directors to whom an individual has professional ties or shares the same alma mater. We also include the median size (in market capitalization) of firms in a director s portfolio, as awardees are typically affiliated with larger firms than are the nonawardees. Additional binary variables in the regression are used to capture whether an individual director holds certain key roles on at least one portfolio firm: committee chairmanships, lead directorships, CEO roles, or board chair roles. We also include the median book-to-market ratio, median return on assets, and median adjusted stock return as measured by the industry-adjusted (at the 2-digit SIC level) stock return over 12 months. Year and award-type dummies are also included. [Insert Table A2 Here] Table A2 reports the estimation results of the logit regression. Directors of large and better performing firms have higher probability to win an award. Older directors who have an important board role at their firms are also more likely to win awards. Most of the estimates are significant at the 1% level. 10 The rationale for excluding non-awardees who are linked (i.e., who share a board with an awardee) is that an award could have spillover effects that cause all members of a board to work harder, gain prestige, etc. As detailed by Rubin (1980), spillover treatment effects can lead to violation of the Stable Unit Treatment Value Assumption, giving rise to biased PSM estimates of treatment effects. 15

17 To implement the propensity-score matching, we rely on a nearest-neighbor matching estimator with replacement. We choose a 20-to-1 matching ratio that results in an overall sample of 12,096 directors (576 awardees and 11,520 non-awardees). In the literature on propensity score matching, it is recognized that the choice of the number of matches to use involves a trade-off between bias and variance (Rubin and Thomas, 2000). While there is no clear prescription for the optimal number of matches to use for each treated observation, utilizing more matches can decrease variance in the estimates of treatment outcomes. In general, it is recommended to use as many matches as possible as long as covariates are balanced, variance does not increase, and the treatment and control observations have common support across covariates. 11 After estimation, we run several tests to check whether the covariates are balanced. T-tests for differences across each of the covariates indicate no significant differences between the treatment and control samples: the largest absolute t-statistic is 1.27 (see Table A3, Appendix A). Since t-statistics are sensitive to the sample size, we also check standardized differences as in Rosenbaum and Rubin (1985), conduct tests of variance ratios, re-estimate the logit regression on the matched sample, and perform an F-test. The results from all of these tests indicate that the covariates in the matched sample are well-balanced. As a next step, we draw yearly data on individual, firm, and board attributes for each director in our matched sample. Doing so generates a panel which consists of 153,873 director-firm-year observations (7,820 awardee observations and 146,053 non-awardees). Although awardees and their matched counterparts have no significant difference around the award year, they could have very different individual and firm attributes in the panel data. Table 3 reports selected summary 11 We have verified that the main results of our analysis are qualitatively similar if we use fewer matches, e.g., a nonawardee to awardee ratio of 9:1 or 5:1. 16

18 statistics for director, board, and firm characteristics for awardee directors and matched nonawardees. [Insert Table 3 Here] Compared to matched non-awardees, award-winning directors are older, have longer tenure, have less board seats at quoted companies, and are more frequently appointed as board chairs and the CEO. It is also apparent from the table that the median market capitalization of awardees affiliated boards is larger than that of their matched counterparts boards. The last column of the table displays p-values for t-tests of differences between the groups of directors. Overall, the awardees are still different from the matched non-awardees in many dimensions. Thus, it is important to further control for these potential sample selection biases in our empirical tests. 3. Results 3.1 Baseline Analysis In this section, we examine the implications of winning an award for directors overall reputation as reflected in the number and prestige of their directorships. Perhaps the most straightforward measure of an increase in director reputation is the net change in the total number of board seats held. Recent work, however, shows that the relative prestige of board seats in a director s portfolio, as measured by market capitalization, is a highly relevant determinant of where directors choose to devote their time and attention (Masulis and Mobbs, 2014, 2015). Accordingly, we also examine changes in the prestige of directors board appointments, a proxy for the overall reputational consequences of an award. In the spirit of Masulis and Mobbs (2014, 2015), we use the median market capitalization as one of the measures for firm prestige. We also include three national firm awards (i.e., the AMAC award, the BCC award, and the BGB award) 17

19 as measures for firm reputation since wining these awards brings considerable visibility and renown to honorees. On the other hand, we introduce financial restatements as a measure for a loss of reputation. We first provide a graphical analysis of overall outcomes for awardees and matched nonawardees on a time-series univariate basis. Our analysis in this section is based entirely on the PSM sample described in section 2. Figure 2 displays the change in the number of board seats at quoted companies held by awardees and matched non-awardee directors from year t-4 to t+4, a sufficient window of time to track the trend. The left-hand side vertical axis is for plots of awardees, while the right-hand side one is for matched non-awardees. [Insert Figure 2 Here] Overall, the figure shows no evidence that awardees experienced a larger increase in the number of board seats compared to matched non-awardees after an award. Instead, the two groups have a similar trend over time, despite of the fact that awardees holding more seats all through the 8-year window. Before an award, both awardees and their matched peers reduced the number of board seats held. On the other hand, the number of board seats of both groups slightly increased during the first two years after an award and then dropped. Non-awardees experienced a slightly smaller drop in the number of board seats compared to awardees during the last two years. Next, we consider changes in the number of board seats at prestigious (or tainted) firms. Figure 3 presents the change in the percentage of awardees and matched non-awardees holding the following types of board seats from year t-4 to year t+4: boards seats at large firms, board seats in a firm that won the AMAC award, the BBC award, or the BGB award in the past 3 years, and board seats in firms which had non-reliance financial restatements in the past 3 years. As in Figure 18

20 2, the left-hand side vertical axis is for plots of awardees, while the right-hand side one is for matched non-awardees. [Insert Figure 3 Here] While Figure 2 provides scant evidence that winning an award translates into a greater number of board seats, there is notable difference in the change in prestige of directorships for the two groups. Figure 3A suggests that a larger portion of awardees hold at least one board seat in large firms compared to their non-awardee counterparts all across the 8-year window. More importantly, the difference between awardees and the matched counterparts gradually enlarged after an award. In addition, the percentage increases in a larger portion during the post-award period than the pre-award period. Similarly, Figure 3B-3D show that awardees, compared to matched non-awardees, are more likely to hold a board seat in prestigious firms as measured by winning the respective type of company awards. Such portion increases in a larger scale for awardees than non-awardees after an award. On the other hand, Figure 3E displays a nonmonotonic change in the portion of directors serving in firms that made financial restatements during years after an award. The percentage of awardees not only is higher but also increases in a larger scale than non-awardees during 3 years after an award, and experiences a bigger drop in the last year. A possible reason could be that firms with financial restatements may have a stronger demand for outstanding directors to solve the problems. Overall, evidence provided by these graphs is consistent with the argument that awardee directors are likely to enjoy an enhanced reputation in the form of holding board seats at larger, more prominent firms 3.2. Multivariate Analysis of the Number and Desirability of Board Seats Next, we turn to regression analysis to more carefully examine the association between awards and the number and quality of a director s board seats. Recall from Section 2 that 19

21 propensity-score matching was used to construct the non-awardee sample. Hence, the observable characteristics of directors and their firms by design should have very little explanatory power for awards. Nevertheless, these characteristics, along with other characteristics observed after an award, could still have an influence on directors labor-market outcomes. [Insert Table 4 Here] Table 4 shows estimates from OLS fixed-effects regressions that examine the impact of winning an award on either the number of board seats (column (1)), the presence of at least one prestigious seat (columns (2) through (5)), and the presence of a tainted seat (column (6)). Each regression is estimated with ordinary least squares and includes director fixed effects and year fixed effects. We note that the numbers of observations is slightly smaller for some regressions due to missing values for a few variables. Column (1) includes the same set of regressors as used in the propensity-score estimation (Table A2), except that here we combine together the major committee chair roles (audit, compensation, or nominating/governance) and do not distinguish between award types. In line with the univariate comparisons above, the estimates show that awards have no significant effect on the change in the number of board seats. The coefficient estimates for many of the control variables are consistent with intuition. For example, the regression shows that directors who have long tenures at their boards or who hold a large number of board seats tend to be more likely to have at least one prestigious seat. This is not surprising since, all else equal, a larger portfolio of seats is more likely to exhibit at least one seat with a given characteristic. Directors who hold a board chairmanship see a more positive net change in board seats; this is consistent with the idea that prominent board roles give directors visibility and networking opportunities that can pay off in terms of future opportunities. 20

22 Among the various measures of board seat desirability, only the Most Admired Company measure has a significant coefficient estimate on Post (coefficient = 0.056, significant at the 1% level). Taken at face value, this finding is consistent with ex post settling-up in the labor market, and it suggests that reputational effects are economically meaningful Endogeneity of Director Awards The above regression results establish that awardees experience a net increase in the median size of their directorships, but the evidence does not control for unobserved, time-varying influences on the probability of winning an award. We therefore turn to an instrumental variables approach that can address the potential problems of omitted variables and selection based on unobservable characteristics of directors or firms. Our choice of instrument is motivated by institutional details surrounding the director awards. As a practical matter, the selection committees for the awards used in this study designate winners based on nominations received from individuals executives and other board members who are current or past colleagues of the nominee. We exploit this fact and quantify the extent to which a director has colleagues who were associated with other director awards in the past. For instance, a director, X, may serve on a board with another director, Y, who shared board service with a third director who won an award. To construct our instrument, we calculate, for a given director in year t, the total number of external award events with which the director s current colleagues were associated with in year t This constructed variable arguably fulfills the exclusion restriction. Indeed, it seems a priori 12 In constructing the instrument, we exclude any connecting intermediary peers who share more than one board seat with the current director. 21

23 unlikely that the occurrence of the third-party award directly affects the labor market outcomes for director X. [Insert Table 5 Here] Table 5 reports the results from Two-Stage Least Squares (2SLS) estimation of the analogue to Table 4. The results stand in sharp contrast to those in Table 4. The number of board seats is positively affected by winning an award. Also, among the types of prestigious seats (Columns (2) through (5)), all types are more likely to be present in the portfolio of an awardee. In contrast, Column (6) shows that seats related to recent accounting fraud are less likely to be held by an awardee. Overall, this provides consistent evidence that award-winners are able to reap benefits from their improved reputation, but they also intentionally forego less desirable directorships. [Insert Table 7 Here] 3.4 Newly-Gained Board Seats and Relative Desirability Tables 6 and 7 analyze the incidence and properties of newly-gained board seats. The results of the regression models make it apparent that awardees are not only maintaining a larger stock of prestigious seats, but they are able to gain new ones more easily than before winning an award. Apart from gaining large board seats, a director may also be able to upgrade his or her portfolio if a new seat is more prestigious than existing ones. Masulis and Mobbs (2014) provide evidence to suggest that directors are more willing to forego a board seat at a poorly-performing firm when the seat is ranked lower (in terms of firm market capitalization) in the director s portfolio. Interestingly, Column (5) of Table 7 reveals that an award makes a director less likely to gain a 22

24 new board seat that is associated with recent financial-statement fraud. This suggests that positive reputational shocks cause a director to not only have expanded opportunities, but also to become choosier and more willing to forego less desirable seats. While we believe that all directors to some extent have incentives to relinquish less prominent seats, there are three reasons to expect that award-winners will be particularly aggressive in giving up such seats. First, an award serves as some proof of a director s ability, so the reputational signaling benefit from continuing to hold multiple board seats (Fama and Jensen, 1983) is lessened. Second, winning an award can lead directly to a new board appointment at an exceptionally large and visible firm, prompting a time- and attention-constrained director to relinquish a smaller seat out of necessity. Third, even if an award does not translate into new, prestigious board appointments in the short run, it may increase the odds that a director gains such roles in the future. 3.5 Internal Leadership Roles In addition to gaining opportunities to join the boards of large companies, awardee directors could also experience ex post settling up within the directorships that they already hold. Thus, it is of interest to examine whether award-winners benefit by having the opportunity to ascend to key board leadership roles such as lead director, presiding director, or chairman of a major committee. 13 As with labor market outcomes in the broader market for directorships, individuals undoubtedly face tradeoffs in deciding whether to take on or relinquish key internal roles at a given directorship. We hypothesize that winning an award can have different effects on internal labor market prospects. On the one hand, an award can elevate a director s standing relative to his peers 13 In our analysis, we do not examine whether an awardee gains the role of chairman of the board as many firms do not have separate CEO/board chairman roles. 23

25 on the board, making it more likely that he is chosen for top roles on the board. Yet, at the same time, receiving an award brings the director new external opportunities at larger firms, which may increase the opportunity cost of serving in a leadership position. The net effect of these two opposing considerations increased stature on the board versus better outside opportunities on an awardee s internal roles is ultimately an empirical issue. [Insert Table 8 Here] In Table 8, we carry out a test of our hypothesis by examining the presence of key board leadership roles for the sample of board seats that directors held between t-4 and t+4. We estimate a series of 2SLS regressions in which the dependent variable is equal to one if a director is observed to hold a particular key role on the board in a given year. The first four columns examine committee chair roles. In columns (1) and (2), the coefficient estimates for Post are positive and significant at 1%. This constitutes our main evidence that ex post settling-up for awardees extends to internal director labor markets. The regressions show, as expected, that larger boards are associated with a lower probability that any one director holds a key leadership role. Columns (3) and (4) show that awardees are less likely to keep committee chair roles than are non-awardees. The estimated coefficients on Awardee are about -0.33, significant at 5%. This lends empirical support to our hypothesis that awards cause directors to re-prioritize their internal and external opportunities, which often leads them to be more inclined to relinquish their leadership roles. Finally, in Columns (5) and (6), we consider the internal labor market for lead/presiding directorships and board chairman roles. In Column (6), we also require that the CEO be separate from the chairman in order for directors to be possibly able to occupy the chair role. The regression results show that awardees are more likely to hold the chairman role, reinforcing the idea that ex post settling up extends to the internal director labor markets. 24

26 5. Conclusion As Fama (1980) argues, outside corporate directors have an incentive to develop reputations as experts in monitoring and decision control, and a good or bad reputation can be rewarded or punished accordingly in the market for future directorships. Directorships are valuable to outside directors because they can confer prestige, reputation, and learning or networking opportunities (Fama and Jensen (1983)). Motivated by these arguments, research has examined whether directors do indeed experience a greater loss of directorships after negative outcomes at the firm level. However, evidence of ex post settling up for directors has been mixed, due in part to the lack of data on reputational shocks to individual directors. This study examines the labor market consequences of a particular type of shock to individual reputation: winning a prestigious, national award for excellence in the boardroom. These awards allow us to investigate the consequences of reputation not only on the number, size, and prestige of an individual s new directorships, but also on whether the individual tends to hold leadership positions in the internal labor market. To identify the causal effects of individual reputation, we employ propensity-score matching, director fixed effects, and a new instrumental variable approach based on a director s indirect connections to third-party directors who won awards in the recent past. Controlling for various director, board, and firm characteristics, including stock performance, awardees are significantly more likely to obtain a new, prestigious board seat in any given year compared to their non-awardee counterparts. At the same time, awardees are significantly less likely to take on a new board seat at a firm that is tainted by recent fraud-related restatement. We also find that awardees are more likely to obtain new directorships at firms that score higher on a prestige index compared to the median index value among their pre-existing board seats. 25

27 Awardees are also more likely to hold key leadership roles at their board appointments, suggesting that ex post settling-up for directors manifests itself in both internal and external labor market opportunities. 26

28 References Adams, R., B. Hermalin, and M. Weisbach, 2010, The Role of Boards of Directors in Corporate Governance: A Conceptual Framework and Survey, Journal of Economic Literature, 48, Anderson, K., D. Deli and S. Gillan, 2003, Boards of Directors, Audit Committees, and the Information Content of Earnings, working paper Abadie, A., and G. Imbens, 2007, Bias Corrected Matching Estimators for Average Treatment Effects, NBER Working Paper. Brochet, F., and S. Srinivasan, 2014, Accountability of independent directors: Evidence from firms subject to securities litigation. Journal of Financial Economics, 111, Coles, J., and C. Hoi, 2003, New evidence on the market for directors: Board membership and Pennsylvania Senate Bill 1310, The Journal of Finance, 58, Engelberg, J., P. Gao, P., and C. Parsons, 2013, The Price of a CEO's Rolodex, Review of Financial Studies, 26(1), Ernst & Young, 2014, Let's talk: governance Navigating the company-investor engagement landscape, Ertimur, Y., F. Ferri, and D. Maber, 2012, Reputation penalties for poor monitoring of executive pay: Evidence from option backdating, Journal of Financial Economics, 104, Fahlenbrach, R., A. Low, A., and R. Stulz, 2010, The dark side of outside directors: Do they quit when they are most needed? NBER working paper Falato, A., D. Kadyrzhanova, and U. Lel, 2014, Distracted directors: Does board busyness hurt shareholder value? Journal of Financial Economics, 113, Fama, E., 1980, Agency Problems and the Theory of the Firm, Journal of Political Economy, 8, Fama, E. and M. Jensen, 1983, Separation of ownership and control, Journal of Law and Economics, 26, Fich, E. M., 2005, Are Some Outside Directors Better than Others? Evidence from Director Appointments by Fortune 1000 Firms. The Journal of Business, 78, Fich, E. M., and A. Shivdasani, 2007, Financial fraud, director reputation, and shareholder wealth, Journal of Financial Economics, 86, Fos, V., and M. Tsoutsoura, 2014, Shareholder democracy in play: Career consequences of proxy contests, Journal of Financial Economics, 114,

29 Fracassi, C., and G. Tate, 2012, External networking and internal firm governance, The Journal of Finance, 67, Gilson, S. C., 1990, Bankruptcy, boards, banks, and blockholders: Evidence on changes in corporate ownership and control when firms default, Journal of Financial Economics, 27, Gow, D., S. Shin, and S. Srinivasan, 2014, Consequences to directors of shareholder activism, HBS working paper. Harford, J., 2003, Takeover bids and target directors incentives: The impact of a bid on directors wealth and board seats, Journal of Financial Economics, 69, Harford, J., and R. Schonlau, 2013, Does the director labor market offer ex post settling-up for CEOs? The case of acquisitions. Journal of Financial Economics, 110, Jiang, W., H. Wan, and S. Zhao, 2015, Reputation concerns of independent directors: Evidence from individual director voting, Review of Financial Studies, 29 (3), Kaplan, S., and D. Reishus, 1990, Outside directorships and corporate performance, Journal of Financial Economics, 27, Klein, A., 1998, Firm Performance and Board Committee Structure, Journal of Law and Economics, 41, Lel, U., and D. Miller, 2014, The Market for Director Reputation Around the World: Evidence from Shocks to International Reputation and Experience, working paper. Levit, D., and N. Malenko, 2014, The labor market for directors and externalities in corporate governance, working paper. Malmendier, U., and G. Tate, 2009, Superstar CEOs. The Quarterly Journal of Economics, 124, Masulis, R., and S. Mobbs, 2011, Are all inside directors the same? Evidence from the external directorship market. The Journal of Finance, 66, Masulis, R., and S. Mobbs, 2014, Independent director incentives: Where do talented directors spend their limited time and energy? Journal of Financial Economics, 111, Masulis, R., and S. Mobbs, 2015, Independent director reputation incentives: Major board decisions and corporate outcomes, working paper. Milbourn, T. T., 2003, CEO reputation and stock-based compensation. Journal of Financial Economics, 68,

30 Nguyen, B. D., 2006, Is more news good news? Media coverage of CEOs, firm value, and rent extraction, working paper. Penbera, J., 2009, What lead directors do, MIT Sloan Management Review, Rosenbaum, R., and D. Rubin, 1983, The central role of the propensity score in observational studies for causal effects, Biometrika 70, Rosenbaum, R., and D. Rubin, 1985, Constructing a control group using multivariate matched sampling methods that incorporate the propensity score. The American Statistician 39, Rubin, D., 1980, Randomization Analysis of Experimental Data: The Fisher Randomization Test Comment, Journal of the American Statistical Association 57, Rubin, D., and N. Thomas, 2000, Combining propensity score matching with additional adjustments for prognostic covariates, Journal of the American Statistical Association 95, Shivdasani, A., and D. Yermack, 1999, CEO Involvement in the Selection of New Board Members: An Empirical Analysis, The Journal of Finance, 54, Srinivasan, S., 2005, Consequences of financial reporting failure for outside directors: Evidence from accounting restatements and audit committee members. Journal of Accounting Research, 43, Wade, J. B., J. Porac, T. Pollock, and S. Graffin, 2006, The burden of celebrity: The impact of CEO certification contests on CEO pay and performance, Academy of Management Journal, 49, Yermack, D., 2004, Remuneration, retention, and reputation incentives for outside directors, The Journal of Finance, 59,

31 Appendix A Table A1: Variable Definitions and Data Sources Variable Source Definition Director Characteristics Awardee Various sources (see text and Appendix B) Indicator variable: equals 1 if a director was an awardee in year t. Age BoardEx The age of the director in year t-1. Board chair role BoardEx Indicator variable: equals 1 if a director held the chairman role on the board. CEO role BoardEx Indicator variable: equals 1 if a director was CEO at the company associated with his or her board seat. Committee chair role BoardEx Indicator variable: equals 1 if a director was chairman of any of the key board committees: audit, nominating/governance, compensation. Lead director role BoardEx Indicator variable: equals 1 if a director was the board s lead director. Any board chair role BoardEx Indicator variable: equals 1 if a director held the chairman role at any one of his or her board seats in year t-1. Any CEO role BoardEx Indicator variable: equals 1 if a director held the CEO role in any one of his or her affiliated companies in year t-1. Any committee chair role BoardEx Indicator variable: equals 1 if a director held the chair role in any of the three key committees at any one of his or her board seats in year t-1. Any lead director role BoardEx Indicator variable: equals 1 if a director was the lead director at any one of his or her board seats in year t-1. Longest tenure BoardEx The director s longest tenure (log-transformed) at any board seat held as of year t-1. Professional connections BoardEx The number of other individuals with whom the director shares an employment or board link at the same quoted firm (overlap during 10 years prior to award month) Professional connections (w/ selection committee members) BoardEx The number of award selection committee members with whom the director shares an employment or board link at the same quoted firm (overlap during 10 years prior to award month) Alma mater connections BoardEx The number of other individuals with whom the director shares an alma mater Alma mater connections (w/ selection committee members) BoardEx The number of award selection committee members with whom the director shares an alma mater 30

32 Table A1, Continued Median board size BoardEx The median value of board size across the director s affiliated boards. Median book-to-market ratio Compustat The median value of book-to-market ratio across the director s affiliated firms. Median industry-adjusted stock return CRSP The median value of industry-adjusted stock return across the director s affiliated firms. Returns are adjusted using median returns at the 2-digit SIC industry level. Median log market capitalization CRSP The median value of log-transformed market capitalization across the director s affiliated firms. New seat gained BoardEx Indicator variable: equals 1 if a director gained at least one new board seat between year t-1 and year t+2 relative to an award event. Number of board seats BoardEx Log-transformed number of quoted board seats held in Year t-1. Firm and Board Characteristics Board size BoardEx The number of directors on the board at fiscal-year end. Book-to-market ratio Compustat (Book value per share)/ (stock price per share), measured at the beginning of each fiscal year. Independent directors BoardEx The percentage of directors on the board who are deemed independent according to the applicable NYSE or Nasdaq regulatory definitions. Adjusted stock return, months +x to +y CRSP Compounded monthly returns from the xth to the yth month after the award month, adjusted by the contemporaneous 2-digit-SIC industry median return. Large firm CRSP A firm which is larger than the median firm across a director s affiliated companies, measured by the market capitalization. Market capitalization CRSP The market value of equity, measured at the beginning of the fiscal year. Return on assets Compustat (Net income)/(total assets), measured at fiscal-year end Sales growth Compustat (Sales)/(last year s sales), measured at fiscal-year end Small firm CRSP A firm which is smaller than the median across a director s affiliated companies, measured by the market capitalization. 31

33 Table A2: Determinants of Director Awards This table reports coefficient estimates from a logit regression explaining the probability that a director wins an award in a given award month. The sample includes all directors in each month in which a director award was given. The dependent variable is a dummy variable equal to 1 if a director won an award. Data on directors board affiliations are drawn from BoardEx. Board affiliations are measured as of the end of the latest fiscal year preceding an award month. Longest director tenure is the log of a director s longest tenure across all boards with which he or she is affiliated. Number of board seats is the in the log form. Any board chair role, Any CEO role, Any lead director role, and Any committee chair role are dummy variables equal to 1 if a director served in the respective board role at any one of his or her board seats. Any AMAC award, Any BGB award, and Any BCC award are dummy variables equal to 1 if at least one company in the director s board affiliations won the respective firm award in the past 3 years. Other variables are measured as medians across companies associated with a director s board seats. The industry-adjusted 12- month stock return is the raw stock return minus the contemporaneous median 2-digit SIC stock return, both measured over the year ending in the month preceding the month of an award. Book-to-market and market capitalization are measured as of the beginning of the fiscal year. Market capitalization is in log form. ROA is net income divided by total assets. Other variables are as described in the Appendix. Z- statistics are reported in parentheses. *, **, and *** denote statistical significance at 10%, 5%, and 1%, respectively. 32

34 Table A2, Continued Logit Director age 0.024*** (3.66) Longest director tenure (-1.56) Director male 1.302*** (12.45) Number of board seats at quoted companies 1.743*** (17.78) Any board chair role 0.955*** (7.91) Any CEO role (-0.69) Any lead director role 0.629*** (4.72) Any committee chair 0.854*** (9.00) Median log market capitalization 0.516*** (15.10) Median board size (0.05) Median book-to-market 0.203*** (3.55) Median industry-adjusted ROA (-1.19) Median 12-month industry-adjusted stock return 0.131** (2.01) Any AMAC award 0.418*** (3.70) Any BGB award 0.533*** (4.44) Any BCC award 0.478*** (4.49) Award type dummies Year dummies Yes Yes Pseudo R Observations 355,620 33

35 Table A3: Covariate Balance within Propensity Score-Matched Sample Mean t-test Covariate Treated Control % bias t p > t Director age Longest director tenure Director male Number of board seats at quoted companies Any board chair role Any CEO role Any lead director role Any committee chair Median log market capitalization Median board size Median book-to-market Median industry-adjusted ROA Median 12-month industry-adjusted stock return Any AMAC award Any BGB award Any BCC award Award type dummies Year dummies Pseudo R 2 from logit regression on matched sample

36 Table A4: Construction of Firm Prestige Index This table summarizes the results of principal components analysis (PCA) to the four measures of positive firm desirability: Large Firm (indicator for whether a firm s market capitalization is in the 90 th percentile in a given year); AMAC (among the top 50 of America s Most Admired Companies according to Fortune Magazine); BGB (rated as one of the Best Global Brands according to Bloomberg/Interbrand); BCC (rated as one of the Best Corporate Citizens). To construct the index, the scores are applied as weights to the four prestige components, and the linear combination is standardized to have zero mean and unit standard deviation. Large Firm Most Admired Company Best Global Brand Best Corporate Citizen (2) (3) (4) (5) Component Component Component Component Score (1 st principal component) Eigenvalue

37 Appendix B1: Data on Director Awards NACD Awards: Founded in 1977 and headquartered in Washington, D.C., the National Association of Corporate Directors (NACD) is an independent, not-for-profit organization that focuses on advancing board leadership by providing information and insights to boardroom community. Currently, NACD has more than 16,000 director members from both the U.S. and overseas. Starting from 1987, NACD granted its Director of the Year award each year (except for 2001 and 2002) to one or two individual directors in recognition of his or her exemplary boardroom leadership. NACD rewarded only one individual as the Director of the Year for and announced the event in Therefore, we count this awardee as a winner for In 2007, NACD introduced new award titles in addition to Director of the Year, including the Kenneth West Lifetime Achievement, Private Company Director of the Year, Nonprofit Company Director of the Year, and the Directorship 100. Except for Directorship 100, each title honors one individual director in its area each year. As our focus is on directorships of public companies, we exclude from our sample the Private Company Director of the Year award. The Directorship 100 identifies 100 influential people in the boardroom community each year, including not only corporate directors, but also corporate governance experts, journalists, regulators, academics and counselors. Since categories for the latter group are not intended to recognize board service per se and do not have the same intent as the directorship reputation in our sample, we remove them from our data, only keeping those awardees who are corporate directors of public companies. In 2008, NACD introduced an award for the Corporate Governance Hall of Fame, which honors 2 to 5 persons every year. In 2008 (and only in that year), NACD released a list entitled Directorship Global 100, which recognized 100 directors serving in companies worldwide (including the U.S.). We include Global 100 directors in our sample if they serve on a board in the U.S. at any time during This results in 64 NACD awardees in our sample for The lists of NACD awardees are obtained from two sources: first, the NACD website ( for the year 2010 to 2013; and second, Directorship magazine published by NACD for 2009 and earlier. We use various sources to search for the announcement dates of NACD awards. Each year, winners of NACD awards are honored in a conference in December, but press releases in local media announce the winners beforehand, usually 3 to 6 months before the conference. Generally speaking, the Director of the Year, Kenneth West Lifetime Achievement, Private Company Director of the Year, and Nonprofit Company Director of the Year were announced on the same day. We use archived NACD websites and NACD press releases to obtain the announcement dates of these awards for 2007 to Although we are able to obtain the list of Director of the Year for , we do not find any indication that the list was publicly disclosed before the conference. Therefore, for pre-2007 years, we take the award announcement date to be the date of the NACD conference. The Directorship 100 list is announced later than the four individual award titles. To determine these announcement dates, we use NACD website archive, NACD press releases, and the date that the Directorship magazine cover story was posted on the NACD website. Sometimes, an individual award winner s company also releases news of an award. We search news for awardees one by one to ensure that their companies news releases were not earlier than the NACD news release. We obtained the press release for the Directorship Global 100 award, which was published on 10/1/2008. We use this date as the announcement date as extensive online searches uncovered no earlier news for the Global 100 award. 36

38 NACD Blue Ribbon Commission: Since 1993, with the exception of 1997 and 2006, NACD has convened a Blue Ribbon Commission (2 commissions in 2010) to study a critical area of board practice and then publish a report on important issues in this area. These commissions are usually led by 1 to 2 chairpersons, and the commissioners consist of experienced corporate directors and experts. Based on our observation, many of the Blue Ribbon Commissioners are awarded as distinguished directors by NACD or other organizations a few years later after serving in the commission. Therefore, we consider being named a Blue Ribbon Commissioner to be an important proxy for a director s reputation and include Blue Ribbon Commission (BRC) membership in our sample. Since 2011, NACD announced the formation of that year s Blue Ribbon Commission with a full list of commissioners on the NACD website. Each of these announcement dates is the earliest one that we are able to locate for that year s BRC list release in public news. In 2010, NACD formed two separate commissions: "Performance Metrics" and "The Audit Committee. We obtain the full list of commissioners for "Performance Metrics" from the NACD website release, but we do not find the list for "The Audit Committee". We obtain the announcement dates for the two commissions from an article in Directorship Magazine. Before 2010, the NACD website did not announce the formation of BRC but released a BRC report each year which contains the full list of commissioners. We collected these reports using various sources, such as online news release, Value Alliance website, and authorship information of BRC reports available on Amazon.com. However, we do not have any information about the announcement date of commissioners for the pre-2010 period. We use the available publication date of the BRC report as the announcement date. For the year 1998 and 2008, neither the list of commissioners nor the announcement date is available. FCD Director of the Year Award: Founded in 1991, the Forum for Corporate Directors (FCD) is a non-profit organization based in Orange County, CA, that aims to offer dynamic forums for corporate directors and conducts programs to directors and executives to help promote board leadership and corporate governance. The group had originally been a chapter of NACD. When the organization separated from the NACD, it became the International Forum for Corporate Directors and eventually changed its name to the Orange County Forum for Corporate Directors. Each year since 1995, the FCD honors 5 to 7 corporate directors of public U.S. companies who demonstrate excellence in the boardroom in several award categories. The core categories include Companies in Transition, Economic Value, Corporate Growth, Corporate Citizenship, Early Stage Companies, Corporate Governance, Chairman s Award, Corporate Leadership & Service, and Lifetime Achievement. Although some of the awardees have a board seat in Orange County, being located or affiliated in Orange County is not a requirement. The award is a national, not a regional award. The selection of awardees is performed in a two-step process. First, the nominating committee chooses a set of nominees and vets them. Second, a group of candidates is presented to the FCD board for final selection and approval. Winners are honored at a gala banquet in February, but press releases in local media announce the winners beforehand, usually between January and mid-february. We obtained the list of FCD honorees from the Hall of Fame section on the FCD website (fcdoc.org). After 2005, FCD publishes an annual news press release that announces all awardees. We use the date of the press releases as the announcement date. To make sure that these are the 37

39 earliest available dates, we used online resources such as Google Search, Factiva, and LexisNexis databases, as well as cached and archived web pages of the FCD website to check for earlier news releases by local media or company press releases. In the current version of the FCD website, press releases before 2005 are not available. Hence, we used other available sources to obtain the earliest announcement dates. We found no news before Using web archive search engines, we were able to obtain two news releases by FCD in 2000 and 2002 with the earliest announcements. In other years, if we were not able to obtain any earliest awardees announcements, we considered the date of a dinner event as the earliest date of the announcement of all awardees. In 1999 and 2003, two awardees were announced by local newspapers prior to a dinner event. In addition, we obtained the earliest announcement of the registration for a dinner event each year in the archived web pages. Since the links to these events have been deactivated, we were not able to confirm whether the list of the awardees was available at the time that registration was open for the Director of the Year dinner event. FT-ODX Outstanding Directors Award: The Financial Times-Outstanding Directors Exchange (FT-ODX) is a premier conference for corporate directors of public companies. This conference and awards program is based on the Outstanding Directors awards program, which began in 1998 under the ownership of Money Media, a New York-based company that publishes news and intelligence for professionals in the financial services industry and in corporate boardrooms. In 2008, Money Media merged into and became a division of Financial Times. The Outstanding Directors Award program was then operated by Financial Times Live, the global conferences and events arm of the Financial Times. Each year, the Outstanding Directors Award program honors 5 to 10 corporate directors for their extraordinary contributions to the companies on whose boards they serve. Awardees are selected after consulting with members of the Outstanding Directors Advisory Board sitting directors and CEOs from a cross-section of U.S. industry and completing extensive research into their companies and careers. We employ various sources to obtain honoree-lists and announcement dates of FT-ODX awards. First, from 1998 to 2002, the award news was released by a monthly magazine named Director s Alert, where we obtain the full list of awardees. For 1998 and 2002, the magazine release provides us specific announcement dates which are 10/1/1998 and 11/1/2002, respectively. Based on this information, we infer that the magazine is published at the first day each month and thus the award news is very possibly announced at the first day of a certain month. For 1999 and 2001, we identified announcement months from news reports and use the first day of the announced months as announcement dates. For the year 2000, only one relevant news article is found. This article states that Gordon, an individual honoree, was selected as an Outstanding Director in the summer of We chose 9/1/2000 as September is the last month of summer and, at the same time is close to announcement dates for other years (usually November or October). In 2003, the award list was released by the Board Alert service. We found the specific announcement date from Board Alert news release. For 2004 to 2006, we rely on the archived website of the Outstanding Director Award ( and news reports to obtain the awardee list. We then searched news reports for each awardee and used the earliest released date as the announcement date. For the year 2007, we do not obtain any information about announcement dates for the list of awardees. However, we find separate announcement dates for each of the 10 honorees from Agenda magazine and Agenda weekly news stories. The news stories suggest that, in 2007, awardees were not announced all at once, but rather one per week. Using these news stories, we obtain 10 different announcement dates for the 10 awardees. We find the 38

40 2008-list in the website archive of a one-page biographies of awardees released by FT-ODX. We checked all related websites but could only limit the announcement date in a window of Jan. 8 to Jan. 13. We finally use 1/13/2008 because the website archive in this date provides two outstanding directors names, while that in Jan 8 only provides a link to the Conference held on March 25-26, Since 2009, the ODX award has been released by Financial Times (U.S. edition) directly. The website of Financial Times Live ( Directors-Exchange-New-York) provides us the awardee list for the post-2009 period. We found specific announcement dates from the Financial Times (U.S. edition) e-paper for 2010, 2012 and For the year 2009, although there is a date from the e-paper, we found FT press released the awardees list even earlier. Thus, for 2009, we employ the FT press release date as the announcement date. For the year 2011, two individual awardees companies chose to announce the news of their being awarded before FT s announcement. Since the individual releases did not mention other awardees at all, we employ the companies release dates only for these two individuals, and we use the FT-OXD date for other awardees. 39

41 Appendix B2: Data on Company Awards Fortune Most Admired Companies Fortune America s Most Admired Companies ranking was published in 1983 for the first time in the Fortune magazine. To create this ranking Fortune in cooperation with Hay Group conducted Corporate Reputations Survey. The survey started with the Fortune 1000 largest American corporations which were sorted by revenue and industry. In 2013, the ranking was extended to the World s Admired Companies. Hay Group surveyed international companies and included an industry rank. To create an overall list of Most Admired Companies, executives, outside directors, and financial analysts were asked to select 10 most admired companies based on 8 attributes using a scale of 0 (poor) to 10 (excellent). The attributes were the following: the quality of management; quality of products or services; financial soundness; innovativeness; long-term investment; ability to attract, develop, and keep talented people; community and environmental responsibility; and use of corporate assets. Final score was received by averaging the scores on those attributes. We manually collected rankings from 1999 to 2016 from printed and online editions of the Fortuna magazine. Best Global Brands Best Global Brand ranking goes starts in 2000 and is published by Bloomberg BusinessWeek. At the first step, the share of the company s sales is estimated relative to the total sales. Next, the value of the brand is derived from the earnings that are directly derived from the brand itself net of the costs and relatively to the overall contribution to other company s intangible assets. At the final step, the NPV of the future earnings is estimated. To be included in the ranking, the brand needs to be global, transparent and visible with financial results. The key components of the ranking are based on the financial analysis, role of brand, and brand strength. Best Corporate Citizens The ranking of Best Corporate Citizens starts in 2000 and covers 100 best companies that fulfill corporate responsibility requirements. The ranking takes into account accountability and transparency of firms that is a vital criterion for investors, regulators, customers, suppliers, employees, and neighbors to make their investment decisions and business decisions. The CR Magazine takes into account 260 data elements among seven data categories and weights them according to the Methodology Committee. The categories include climate change, employee relations, environmental, financial, and governance issues, as well as human rights and community support. The committee uses company websites, sustainability reports, and company 10-Ks. The ranking is also available as a breakdown by industry and is published also within each category as indicated above. 40

42 Figure 1: Number of Director Awards, by Year and by Type This figure displays the number of recipients of director awards, by type, from 1999 to NACD represents the various director awards granted by the National Association of Corporate Directors. NACD BRC is the distinction of being named an NACD Blue Ribbon Commissioner. FCD is the Director of the Year Award granted by the Forum for Corporate Directors. FT-ODX is the Outstanding Directors Award granted by Financial Times-Outstanding Directors Exchange NACD NACD BRC FCD FT-ODX 41