Are Dissenting Directors Rewarded?€¦ · circumstances and/or resignation letters must be...

Are Dissenting Directors Rewarded?*

Cassandra D. Marshall

Kelley School of Business Indiana University

[email protected]

First draft: July 22, 2010

This draft: August 28, 2010

Abstract

I explore whether directors who resign in dissent from their board are rewarded in the labor market for directors. Using a hand collected sample of 278 boardroom disputes reported in 8-K filings during 1995-2006, I show that firms which have disputes are small, highly levered, have poor profitability, and have boards dominated by management. I find that dissent is not rewarded. For all types of directors across all types of disputes, directors who resign in protest experience a net loss in board seats of 85% over the five year period following the dispute. This means dissenting directors are not able to recover the seat they give up by obtaining additional board seats at other public firms. For these dissenting directors it appears that dissent is its own reward.

* Contact: Cassandra D. Marshall, Finance Department, Kelley School of Business, Indiana University, 1309 E. Tenth St., Bloomington, IN 47405. I would like to thank my dissertation chair, Utpal Bhattacharya, for all of the helpful advice, encouragement, and comments on this paper as well as my dissertation committee members Daniel Beneish, Scott Smart, and Jun Yang for providing helpful comments and support while completing this paper. Additionally, I would like to thank Alex Borisov, Mike Faulkender, Stacey Jacobsen, and seminar participants at Indiana University for many helpful comments which have greatly improved this paper. All remaining errors are my own.

2

Are Dissenting Directors Rewarded? I. Introduction

What incentives do directors have to act in the interest of shareholders? Specifically, if a director

does not agree with a board that appears to go against shareholders’ best interest, is the director rewarded

for dissenting? In this paper, I explore this question by examining the future employment of directors

who resign from boards in dissent. I measure reward as the net gain (loss) in board seats obtained in other

public firms from the day before the resignation to five years after the resignation, and scale this number

by all current board seats held at the time of the dispute to obtain a percentage.1

I find that directors who dissent from a board are not rewarded in the market for external

directorships. On average, dissenting directors experience a net loss in board seats of 85% over the next

five years. Thus, dissenting directors are not able to fully recover the board seat that they resign from in

dispute. This finding is obtained whether the dissenting director is independent or not and whether the

board is dysfunctional, has a severe agency problem, or because of differences of opinion. For dissenting

directors it appears that dissent is its own reward.

A recent WSJ article highlights the issue of director dissent. In July of 2007, three of the thirteen

directors of a small local Tennessee bank resigned from their board in disagreement over recent pay raises

for the CEO and several top executives. One of the resigning directors claimed that the pay raises were

“crafted to provide excessive and retroactive compensation to executive managers.” This disagreement

over unwarranted pay raises arose while the bank was experiencing a very successful year. The bank

hired an outside compensation consultant who benchmarked the bank against much larger peers and

1 Future versions of this paper will explore alternative measures of reward by focusing on the quality of the

future board seats obtained by these dissenting directors. They will include measures of the prominence of the new board seat such as the new firm’s size, performance, and governance, measures of changes in the financial remuneration of the directors involved in the dispute events such as director and executive compensation as well as changes in the value of stock holdings held in the dispute firm and obtained at the firm providing the new directorship.

3

recommended large pay increases. One of the dissenting directors wrote in his resignation letter “I have

to question whether we are a board of directors – or a board of directed.”2

Directors have begun to speak out in the boardroom against decisions or actions with which they

are in disagreement. In 2004, the SEC passed a new law which requires firms to disclose if a director

chose to leave the board or not stand for re-election due to a disagreement with the board. The

circumstances and/or resignation letters must be included in the company’s 8-K filing. Previously, firms

were only required to disclose director departures due to disagreements if the resigning director requested

his resignation letter to be made public.

Directors have incentives to enhance their reputations to increase their value in the labor market

for outside directors (e.g. Fama and Jensen (1983)). Directors who resign in disagreement tradeoff

current financial compensation against an expected loss in reputation which can lower expected future

employment and compensation. Alternatively, the signal in dissent should increase the demand for a

director’s services by enhancing or protecting their reputation.

My results include the following. First, I show that firms which have disagreement events

significant enough to trigger a director resignation have some unique characteristics that likely contribute

to the dissension in the boardroom. Firms experiencing director dissent tend to be smaller firms with

poor operating performance in the year prior to the dispute event. They also tend to be younger firms

with more growth opportunities and are concentrated in certain industries such as business services and

the communication industry.

Second, I show that these firms also have specific board and CEO characteristics. They tend to

have smaller boards with a large presence of managers and are less likely to have independent

committees. The CEOs of these firms are typically younger and newer to the firm. The CEOs also

exhibit greater power in the boardroom by being more likely to be a member of the compensation

committee and are also the board’s Chairman in 50% of the firms. The boards and CEOs also own a large

amount of the firm’s stock, which gives them significant voting power and a large financial stake in the 2Thurm, Scott, “How Big Raises Undid an Insular Board,” The Wall Street Journal, August 8, 2007.

4

firm. Under these circumstances, power struggles may be more likely, and factions of directors can result

in certain members having to leave the board if a consensus cannot be reached.

Third, I show that directors who dissent from the board also tend to be younger and newer

members of the board and are less likely to be members of committees. Directors who are also attorneys

are less likely to dissent from the board, while directors who work in private equity or venture capital are

more likely to dissent. In looking at other forms of corporate leadership outside the firm, dissenting

directors are less likely to be current Chairmen of other firms and are more likely to be founders of other

firms. Directors with additional board seats outside of the firm are less likely to dissent.

Fourth, I classify dissent into four categories. The first category is the alleged presence of a

dysfunctional board. The reasons cited in the resignation letters range from complaints over standard

board operations such as lack of dissemination of information from management to directors or among

directors, lack of board meetings, lack of discussion of corporate events, or lack of cooperation from

management or other directors in carrying out their duties. The second category is alleged agency

problems. These include complaints about entrenched managers, unreasonable managerial compensation,

or poor corporate governance practices. The third category is differences of opinion. The differences of

opinion could be about strategic corporate decisions like mergers and acquisitions, new business or

product lines, or sources of finance. The fourth category is miscellaneous, where the reasons for

resignation – like differences in religious belief – cannot be categorized in the first three bins.

Fifth, I combine the above finding that certain types of directors are more prone to dispute and the

fact that there are different types of dissent and perform a multivariate analysis. I confirm the previous

findings – dissenting directors have shorter tenures, have fewer board seats outside of the firm, are more

likely to be founders of other companies, are less likely to be in a CEO position at another firm, are more

likely to work in private equity or venture capital, and less likely to be attorneys. Further analysis reveals

that independent directors, linked directors, and founders are more likely to leave because of

dysfunctional boards. Founders of other companies and directors who primarily work in venture capital

or private equity are more likely to leave over a disagreement due to agency problems. Attorneys are less

5

likely to leave due to agency problems or differences in opinion on corporate strategy and the future

direction of the company.

Sixth, and most importantly, I find that directors who dissent from boards in dispute are not

rewarded in the market for external directorships. The average dissenting directors experiences a net loss

in board seats of 85% over the next five years. This means that dissenting directors are not able to fully

recoup the board seat that they give up in leaving the board with the dispute. This outcome is seen for all

types of directors and in all types of disputes.

There is relatively little research examining director turnover.3 Research has studied director

turnover following rare corporate events such as director resignation following bankruptcy or financial

distress (Gilson (1990)), acquisitions (Harford (2003)), financial fraud (Fich and Shivdasani (2007)),

accounting restatements (Srinivasan (2005)), and CEO turnover (Farrell and Whidbee (2003)). A more

recent study by Fahlenbrach, Low, and Stulz (2010) looks at whether or not outside directors resign from

boards to either protect their reputation or to avoid a heavier workload when firms experience poor

performance or negative events and find some evidence to support this ‘dark side’ of outside directors’

incentives. My paper differs from this work because I follow the career path of the dissenter after the

resignation and consider all types of dissent.4

The closest paper to mine is that by Agrawal and Chen (2010). They also look at boardroom

disputes that lead to director departures. Their sample consists of 168 dispute episodes over the period

3 A large body of research in corporate governance has studied board composition, board size, committee structure, and director backgrounds (e.g. see Hermalin and Weisbach (1988, 1998) for studies on board composition, Yermack (1996) for a study on board size, and Klein (1998) for a study on committee structures). 4 Other studies have followed the career paths of directors, but have studied different events: the reaction to a change in law regarding the acceptance or rejection of antitakeover provisions for firms (Coles and Hoi (2003)) or how top executives perform in the external market for directors after their firms cut dividends (Kaplan and Reishus (1990)). Yermack (2004) looks at the financial incentives of new outside directors of Fortune 500 firms during the late nineties and shows that the incentive for directors to increase reputational capital which then increases the likelihood of receiving new seats is economically valuable to directors.

6

1995-2006.5 My paper, which covers the same period as the Agrawal and Chen (2010) paper, has a larger

sample size with 278 dispute episodes. I confirm many of their findings on the types of firms that have

dissent and who dissents, and I uncover some new findings such as how the type of director is related to

the reason for the dissent. The principal difference between our papers is that I focus on the

consequences for the dissenting director whereas their paper focuses on the consequences for the firm.

To my knowledge, this paper is the first to examine a unique sample of boardroom disputes to

determine how these event-specific disagreements affect the future reputations and opportunities of the

dissenting directors. My finding is that dissenting directors are not rewarded in the external market for

director services; they are not able to fully recoup the board seat that they lose from resigning from the

dispute firm.

The paper proceeds as follows. Section II describes the data as well as how the sample was

collected. Section III explores the where, who and why of disputes. Section IV, the main section of the

paper, analyzes whether dissenters are rewarded. Section V provides some concluding remarks.

II. Data Description

A. 8-K Disclosure Requirements and Dispute Events

The sample of director disputes was collected from SEC 8-K filings using the 10-K Wizard

covering the period 1994 through 2006. Directors who leave a board over a disagreement often provide

the firm with a letter describing their disagreement and final thoughts on the matter. Effective August 23,

2004, the SEC began requiring firms to disclose when a director resigns or refuses to stand for re-election

due to a disagreement relating to the company’s operations, policies, or practices. The company’s

disclosure must include 1) the date of resignation or refusal to stand for re-election, 2) any positions held

5 They find that these types of conflicts are more likely to occur in firms where the CEO is also the founder, where the CEO has a shorter tenure, where independent block-holdings are higher, and in firms with larger and less independent boards. They also find that directors involved in the disputes have shorter tenures and are more independent of management. Upon experiencing a dispute event, the stock price of the firm has an average abnormal return of -2.6% over the three-day announcement period [-1, +1], and the firm experiences poor operating performance and stock price performance in the years around the dispute event. These firms are also more likely to experience a shareholder class action lawsuit, proxy contests, asset divestitures, and stock market delisting after the dispute.

7

by the director within the company or any board committee membership at the time of the departure, 3)

and a brief description of the disagreement that caused the director to leave the firm. If the departing

director provides a letter to the company, it too must be included in the disclosure as an exhibit regardless

of whether the director wishes for it to be made public. The firm must then provide the director with a

copy of the disclosure and give the director an opportunity to provide correspondence stating whether

they are in agreement or disagreement with the company’s account of the dispute. Any additional

correspondence submitted by the director must be added to the original 8-K disclosure as an amendment

exhibit within two business days after it is received by the company. The firm has four business days to

disclose the director departure in an 8-K filing. Prior to August 2004, firms were not required to disclose

whether a director left the board due to a disagreement unless the director requested that the disagreement

be disclosed. In these cases, the firm had five business days to report the director resignation or refusal to

stand for re-election, and had to include the director’s letter of resignation as an exhibit.

Firms file many 8-Ks each year as is required by the SEC to provide a current report of certain

material corporate events that shareholders should know about. Due to the substantial number of 8-K

filings each year by public firms for many different reasons, the 10-K wizard was used to find the relevant

8-K filings using word search strings. Firms are supposed to file any director resignation letters as an

Exhibit 17, but it is often filed under the incorrect heading; so the word search strings are used to find all

possible resignation letters. The resulting 10-K wizard search provided several thousand potential

director resignations due to disputes. I read each disclosure to determine if it was in fact a director

departure due to a disagreement as stated by the company or director. The 10-K wizard allows for text

searches back to 1994 up through today.

The resulting sample of disagreements covers the period 1994 through 2006. In all, 278 director

dispute events were found over this 13 year period. The dispute events represent 269 unique firms,

because 9 firms had two distinct dispute events over this time period. Each dispute clearly states that the

director left due to a disagreement over the company’s operations, policies, or practices as described by

the Final Ruling of the SEC on director disputes. Table 1 displays the dispute events for each year in the

8

sample period. There were no disputes found in 1994. Due to the change in the disclosure requirements

in August of 2004, there is a clear jump in the number of dispute events that were disclosed starting in

2004 and going through 2006.

Table 1 also shows the number of directors who left each firm due to the disagreement. In many

cases more than one director left the board because of the same dispute. If there were multiple directors

leaving because of the same dispute, the date of the first dispute was recorded. The 278 director dispute

events caused 364 directors to resign or not stand for re-election over the time period.

B. Firm, Board, CEO, and Director Characteristics

Table 2 provides descriptive statistics for the firms’ size, performance, and other firm specific

characteristics. Firms in this sample are relatively small which explains why many firms are not found in

Compustat. Out of the 278 director disputes, 189 are from firms that are covered by Compustat. For the

remaining firms, accounting data and other firm specific variables were collected from 10-K filings or

Annual Reports during the fiscal year end just before the dispute took place. Accounting data was

collected to reconstruct measures of size, debt, and performance. The market value of equity was

collected from the introductory paragraph in the 10-K.

Table 3 provides descriptive statistics on the board of directors and CEOs for each of the 278

firm/years in which there were disputes. For every firm, I read proxy statements, 10-Ks, or other SEC

filings in order to construct each board of directors. I also read each director’s biography to determine

director specific characteristics and their role within the firm and on the board. General characteristics

were recorded such as age, length of tenure on the board, gender, individual stock ownership, and director

classification. I followed the standard classification of director independence where inside directors are

those who are current employees of the firm. Non-inside directors are determined to be linked if there are

disclosed conflicts of interest which include consulting contracts, family relationships with top

management, or significant business relationships with the firm that could be inferred to present a conflict

of interest. All non-inside directors without a disclosed link to the firm are considered to be independent

directors. Other director characteristics such as committee membership were used to determine the

9

independence and membership of the audit, nominating, and compensation committees of the board. I

gathered director, board, and director and officer stock ownership from the table of beneficial ownership

for the firm. Other board variables including the number of board meetings, audit meetings, and if the

board has staggered terms of election, were collected.

Table 4 provides descriptive statistics on detailed director characteristics. From the director

biographies I also collect the position of the inside directors such as Chief Executive Officer (CEO),

Chairman (COB), President, Chief Financial Officer (CFO), Chief Operating Officer (COO), and any

variation of Vice President (VP). I also recorded any primary occupation held outside of the firm for the

past five years for all directors as well as past leadership or director positions held outside of the firm. If

the director has previously or currently served as a CEO, chairman, or founder of another company this

would show evidence of past leadership ability and may be helpful in determining their level of skill in

managing or providing director services.

C. Dispute Characteristics

Table 5 provides classifications of the types of disputes. In the sample of 278 disagreement

events, the reasons for the dispute cover a wide range of events. To better analyze the disputes and the

repercussions of the dispute, the sample is divided into four main categories based on the reason for the

dispute and resignation. The categories are the same as those used to classify director disputes in

Agrawal and Chen (2010).

The first category includes director resignations due to a disagreement on how the board

functions as a governing body. The second category of disputes can be described as those where the

director leaves because of agency problems within the firm or corporate improprieties. The third category

includes disputes concerning a difference in opinion on the corporate strategy or future direction of the

company. All remaining disputes are placed in the fourth category which includes general disagreements

on operations, policies, and practices without further information and those reasons not falling into one of

the other three categories.

D. Measure of Reward

10

The data for the number of external board seats held by each director before their service at the

current firm, during their service at the dispute firm, and after the dispute event are collected from several

different sources. The 10-K wizard was used to search for each director in all SEC filings over the period

1994-2010. Director biographies from proxy statements or 10-Ks were read to determine outside board

seat membership in both public and private firms over the working years of each director. Each director’s

outside board seats were also collected from Boardex. Roughly 45% of the directors in the sample were

found in Boardex which provides detailed accounts of directors’ board services. Boardex covers many

public firms and begins data collection of many larger firms beginning in 2000. Boardex reports

directorships held at both public and private firms and the tenure of each appointment. Lastly, Compact

Disclosure was used as a third source of outside directorships held by directors in other public firms

during the period 1988-2005.

Three different measures of outside board seats are constructed depending on the time of service

on the outside board relative to the dispute event date. The first measure is the sum of all prior public

board seats held where service as a director ended before the disagreement event occurred at the current

firm. The second measure is the sum of all public board seats held by each director at the time of the

disagreement event. The last measure is the sum of all public board seats that were obtained within five

years after the dispute event date.

The reward is defined as the sum of all new board seats in other public firms received by each

director in the five year period following the disagreement event minus the loss of any current public

seats held at the time of the dispute within one year after the disagreement event (including the seat lost at

the firm with the dispute if the director dissents). The sum is then scaled by the current number of public

seats held at the time of the dispute (including the seat held at the firm with the dispute). A five year

period following the disagreement event is used based on the observation by Yermack (2004) that the

market for directors’ services takes time to assess the monitoring ability of new directors. If a director’s

reputation after being involved in a dispute were to put into question their ability to be a director at one of

their other boards, it is likely that the other outside boards would act fairly quickly in severing their

11

association with that director. Therefore, a one year window was chosen to measure the number of

outside board seats that are lost due to a change in director reputation.

III. Disputes – Where, Who, and Why?

A. Are firms which have a director dispute special?

Is there anything special about the type of firm that has a public dispute amongst its board

members? Table 2 presents information on the size, performance, industry, and other general

characteristics of firms which had a director disagreement leading to at least one director departure.

Compustat firms are used as a benchmark for each firm in the dispute sample in the same year. All

accounting measures are windsorized at the 1% and 99% level to remove severe outliers and obvious data

errors. The 278 dispute firm/year observations are compared to all Compustat firms with data (Sales and

Total Assets are not missing) between 1994 and 2006 which contains 127,257 firm/year observations. A

percentile rank is assigned to each measure for the dispute firm compared to the same measure for all

Compustat firms in the same year. I then compute the mean and median of the percentile rank for all

dispute firms and test whether the percentile rank is different from 50.

Panel A presents the firm characteristics. If I measure the size of the firm in assets or sales, Panel

A shows that firms which have a dispute are much smaller than firms in Compustat where there was not a

director dispute. The mean (median) percentile rank is 26.80 (18.00) for Total Assets and 28.575 (21.00)

for Sales. Both t-tests and sign tests show that the dispute firms are significantly smaller than the firms in

Compustat. The market value of equity for dispute firms has a mean (median) percentile rank of 34.79

(31.00) and firm value has a mean (median) percentile rank of 34.63 (29.5) which also confirm that the

dispute firms are significantly smaller than the mean (median) firm in Compustat. On average the dispute

firms are small, but the sample does contain several large or more prominent firms such as Wal-Mart,

American Insurance Group, General Motors, and Walt Disney.

Performance measures in Panel A show that the dispute firms have significantly lower operating

performance during the fiscal year end just prior to the dispute compared to Compustat firms with a mean

(median) percentile rank of 35.331 (31.00) for Cash Flow scaled by Firm Value, 33.54 (27.00) for Net

12

Income scaled by Firm Value, 46.723 (36.5) for ROE, and 26.022 (16.00) for ROA. The poor

performance of these firms likely puts a significant amount of pressure on managers and directors to make

some changes in the way the firm is operating, which could lead to differences of opinion that fuel

disagreements amongst the board members. It may also be a sign that new leadership is needed to bring

new ideas or ways to change the company’s direction for the future recovery of operating profits. The

fact that these firms are under financial distress contributes to the pressure put on the board of directors in

making necessary changes in the firms operations, policies, and practices. It makes sense that in these

difficult times directors are more likely to be getting into disagreements over the future direction of the

company or placing blame on management or other directors for not looking out for the best interest of

shareholders.

Other firm measures in Panel A show that dispute firms may have a little more debt than

Compustat firms which may fuel disagreements on sources of financing or pressure from creditors. These

firms also have significantly larger MTB values with mean (median) percentile ranks of 60.493 (67.00)

which show that they have more growth opportunities. A majority of dispute firms also trade on Over-

the-Counter markets which represent 66% of the sample’s trading venue. The next most prevalent trading

venue is the NASDAQ with 22% and the only 12% of firms trade on NYSE or AMEX. Dispute firms

have been incorporated for an average of 14 years.

Panel B shows the top ten industry groups by two-digit SIC code for the sample. There is a high

concentration of firms in the Business Services and Communication industries. A formal Chi-Square test

shows that the industry concentrations are statistically significant when compared to the industries that

represent the firms in Compustat. The test statistic has a value of 35.217 and is significant at the 1% level

showing that the distribution of industries across the two groups is different.

All of these measures together tell us that dispute firms are quite different from the average

Compustat firm. Dispute firms are smaller, have poorer operating performance, higher leverage, are

younger firms, and have significantly different industry concentration with more firms in the service

sectors.

13

B. Are the board and CEOs of firms which have director disputes special?

Table 3 presents characteristics of the boards of directors and CEOs of dispute firms. Panel A

presents the board characteristics. All boards of directors in Risk Metrics Directors database during the

years 1996 to 2006 are used as a benchmark to compare board and CEO characteristics. In Panel A we

can see that boards with disputes are significantly smaller in size with an average of 6.4 directors

compared to firms in the S&P 1500 covered by Risk Metrics which have 9.5 members on average. We

can also see by classification of directors that firms with director disputes have significantly fewer

independent directors and linked directors. This shows that these types of boards may be more controlled

by top managers within the firm which may contribute to the escalation of disagreements, especially with

outside directors. We can also see that dispute firms are less likely to have independent Audit,

Nominating, and Compensation committees. There is also a greater concentration of board ownership in

dispute firms which is likely due to the fact that they are so much smaller than S&P 1500 firms. This

does provide the board with significantly more voting power and a large financial stake within the firm.

Also, 20% of the dispute firms have staggered boards which allow for each director to remain on the

board for three years before coming up for re-election.

Panel B presents descriptive statistics on CEOs of dispute firms compared to CEOs of S&P 1500

firms. CEOs from dispute firms are 2 years younger on average and 4 years younger at the median

compared to CEO’s at S&P 1500 firms. This goes along with the fact that CEOs at dispute firms have a

much shorter tenure with the firm than the typical company. At the time of the dispute the average CEO

has been in their position for 5 years compared to almost 10 years for the average CEO in the control

group. The CEO of the dispute firm is more likely to be on the Compensation committee at 8% compared

to 1% for the benchmark firms, which means they are more likely to be able to influence the level of their

own compensation and highlights the potential for conflict. We also see that they have a much higher

fractional ownership of the firm which again is attributable to the size of the firms, but it still provides the

CEO with substantial voting power and a large financial stake. Other measures for the dispute sample

show that the CEO is also the chairman of the board in 50% of the firms, and the CEO is also one of the

14

founders of the company in 15% of the cases. Since these are young firms it is not surprising that the

founder is still the acting CEO.

The conclusion from Table 3 is that the boards of theses dispute firms as well as the CEOs are

quite different from the benchmark group.6 The boards tend to be smaller, more heavily controlled by

managers, have less independent committees, and greater board ownership. Likewise the CEOs are more

powerful in that they are more likely to be on the compensation committee, have higher fractional

ownership, and in 50% of the cases are also the board’s Chairman. The fact that inside directors play

such a dominant role on these boards may be contributing to these firms having such strong disagreement

events amongst the directors that lead to director departures.

C. Are dissenting directors special?

Table 4 provides some univariate statistics on characteristics of directors who leave the firm

because of the disagreement and those who decide to stay with the firm through the disagreement. There

are a total of 1789 directors of which 364 are dissenting directors and the other 1425 remain on the board.

We see that younger directors and those with shorter tenures are more likely to dissent. Members of the

nominating or compensation committee are less likely to dissent. Directors with primary occupations as

attorneys are less likely to dissent, probably because they are the company’s corporate counsel and so

have a large financial stake in remaining with the company. Directors who are in private equity or

venture capital are more likely to dissent. Founders of other companies are more likely to dissent.

Current or former CEOs and Chairmen of other companies are less likely to dissent, but the only

difference that is significant is for current Chairmen of other companies. Lastly, dissenting directors have

significantly fewer current outside board seats than those directors who decide to remain.

6 Admittedly, a more appropriate benchmark would be a matched sample of firms of similar size and industry. Results may change if this control sample is used. Since this comparison is not the focus of this paper, I leave this matched control sample for inclusion in future versions of this paper. Agrawal and Chen (2010) compare their sample of director disputes to a matched control sample of firms that did not experience a dispute and find that the average dispute firm has fewer independent directors, a larger board, are less likely to have independent audit or compensation committees, have higher independent blockholdings, CEOs with shorter tenures, and more CEOs belonging to the founding family.

15

D. Are disputes special?

In the sample of 278 disagreement events, the reasons for the dispute cover a wide range of

events. To better analyze the disputes and the repercussions of the dispute, the sample is divided into four

main categories based on the reason for the dispute and resignation.

Table 5 gives the four main categories for disputes. The first category includes director

resignations due to a disagreement on how the board functions as a governing body. Many of these

directors leave their boards because they do not feel they are able to perform their duties as a director due

to a lack of information about the company’s events or financials, not enough board meetings or lack of

notice for future meetings, not enough time to prepare for meetings, or the firm not carrying appropriate

director liability insurance. The second category of disputes can be described as those where the director

leaves because of agency problems within the firm or corporate improprieties. Examples of such

disagreements include the board being under the control of management, management acting without

regard to the board’s direction, accusations of wrongdoing by managers or other directors such as insider

trading or fraudulent accounting, a director unsuccessfully demanding the removal of a member of

management, unreasonable director or manager compensation, and other corporate governance issues that

do not represent the best interest of shareholders. The third category includes disputes concerning a

difference in opinion on the corporate strategy or future direction of the company. Disagreements involve

disapproval or approval of mergers and acquisitions, matters of insolvency or bankruptcy proceedings,

disagreements on changes necessary to bring the firm back to profitability, sources of financing, and new

directions for the company regarding new products, marketing plans, and staffing changes. All remaining

disputes are placed in the fourth category which includes general disagreements on operations, policies,

and practices without further information and those reasons not falling into one of the other three

categories.

Due to the very different nature of the kinds of disputes, it is necessary to separate the firms based

on the reason for the dispute to get a clearer idea of how different directors may be more likely to cause

certain types of disputes, and also to look at the repercussions or future reputational effects on the

16

directors who are involved in these specific types of disputes. Those leaving due to inefficient or

dysfunctional boards may have been trying to do their job and just were not provided with the means to

carry out their fiduciary duties. Those leaving due to agency problems exhibit a different kind of dissent

because here there are clear accusations of wrongdoing by other directors or managers. Those leaving

due to differences in corporate strategy reflect a third kind of dissent, because there may not be a clear

right or wrong decision in taking the company forward. Therefore, it is important to distinguish between

the different types of disputes.

E. What types of directors dissent for what types of reasons?

I now integrate the analysis of sections C and D to answer whether there are special combinations

of dissenter types and dissent types. Specifically, do some types of dissenters tend to have certain types

of disagreements?

Table 6 presents logistic regressions on directors of firms with a dispute event. The dependent

variable is equal to 1 if the director leaves the board due to the disagreement and 0 if the director remains

on the board. Given the different nature of the types of dispute events, I first run a regression on all

dispute events, and then I run regressions for each subgroup of dispute reason categories. Only results for

the first three types of disputes are presented. To recall, the three categories were: (1) a director left the

board due to a disagreement over the inefficiency of the board’s processes, (2) agency problems or

accusations of corporate wrongdoing, and (3) changes in corporate strategy and future direction of the

company.

Panel A of Table 6 presents the results for all directors. Marginal effects are reported and are

evaluated at the median value for each variable. It tells us that newer directors are more likely to leave in

disputes, and this is especially true if the dispute is over a break down in normal board functions or

changes in corporate strategy. Independent and linked directors are more likely to leave due to

dysfunctional board behavior as are Founders. We see that directors with higher stock ownership are less

likely to leave the firm if there is an agency problem. This is probably because such directors are

benefitting from the agency problem. Those with more external public board seats are less likely to leave

17

the firm. As we saw in the univariate results of Table 4, directors who are founders of other companies or

are in private equity or venture capital are more likely to dissent, especially if the dispute is about agency

problems. Attorneys are less likely to leave the board, and this is especially true if there are agency

problems. This is probably because attorneys, who are usually in-house legal counsel, benefit from this

agency problem. An interesting result is that after August 2004, when resignation letters had to be made

public, we see more dissent because of agency problems and less because of dysfunctional boards. That

may be because boards became more efficient, plus it was now compulsory to air agency problems.

Panel B shows the results for the same regressions but with only Independent Directors. The

results are similar to the results in Panel A, which implies that the results in Panel A (all directors) may be

driven by the effects of the Independent directors.

IV. Are dissenting directors rewarded?

Table 7 gives the main results of the paper. This table presents the results of OLS regressions

where the dependent variable is the reward for dissent. As discussed before, it is estimated as the net

change in the number of new board seats obtained in public firms up to five years after the disagreement

event net of any board seats that were held at the time of the disagreement and subsequently lost within

one year after the disagreement event (including the seat lost at the firm with the dispute if the director

dissents).

In all specifications, control variables that may affect the number of outside board seats are used

such as director age, independence, other current public directorships, as well as firm controls for

operating performance and size. A post SEC rule change indicator is used to control for the change in

SEC disclosure requirements after August 2004.7 Industry dummies and clustered standard errors8 at the

firm level are included where indicated.

7 The analysis was also done by comparing the rewards of dissenting director who left before the change in SEC rules to those of dissenting directors who left after the change in SEC rules but the differences were not significant. Likewise, the analysis was done by comparing the rewards of dissenting directors in cases where only one director left the board due to the dispute compared to cases where multiple directors left the board because of the same dispute and the differences were not significant. 8 Results are similar using White’s robust standard errors.

18

The question being answered is whether dissent, irrespective of who dissents, is rewarded. It

shows this for all types of dissenting directors, as well as for the three different categories of dissent. The

first row shows that dissenting directors are not rewarded. From the first column we can see that the

average director loses 85% of their current board seats in leaving the firm because of the disagreement.

This means they are not able to recoup the loss of the seat they give up in dissent by obtaining additional

directorships at other firms.

As for the control variables, the signs are as expected. In general, younger and independent

directors get more board seats. Directors from firms that are larger and have greater operating

performance receive a significantly greater number of new board opportunities.

The conclusion from Table 7 is that dissent is not rewarded on average. It does not answer

questions like whether the reward is different for certain types of directors dissenting for certain types of

reasons. This is an important issue, because in Table 6 we saw that independent directors, directors with

primary occupations in venture capital or private equity and founders of other companies were more

likely to leave due to certain types of disputes, whereas current CEOs of other companies, and attorneys

were less likely to leave due to certain other types of dispute. These interactions need further analysis to

see whether certain types of dissenting directors are rewarded differently for different types of disputes.

Table 8 presents the same OLS regressions as Table 7, but now includes interaction terms for

these types of directors. Table 8 focuses on dissenting directors who are also independent, venture

capitalists, CEOs of other companies, attorneys, or founders of other companies. In Panel A, I compare

the net change in the number of new board seats for independent dissenting directors and independent

directors who stay on the board. The coefficient on the interaction for dissenting independent directors

(b1) is negative and significant in all columns confirming that this type of dissenting director is not

rewarded on average. We can further see that when an independent director dissents due to a

dysfunctional board they are able to recoup a small portion of their loss in board seats and experience a

loss of only 69.4% compared to independent directors who dissent due to agency problems or due to

differences of opinion, who have a loss of 83.2% and 95.6% respectively. The difference between the

19

loss for independent directors who leave due to dysfunctional boards and those who leave due to agency

problems just misses statistical significance; however, the difference between the loss of those who leave

because of dysfunctional boards and those who leave due to differences in opinion is significant at the 5%

level. When testing whether non-independent dissenting directors are rewarded differently from non-

independent remaining directors we see again that dissenting non independent directors experience a net

loss in board seats.

In all other panels, we consistently see that dissenting directors of various types are unable to

replace the seat they give up at the dispute firm. In Panel B, we see some evidence that venture capitalists

who dissent due to dysfunctional boards experience a more severe loss of 110% of board seats relative to

venture capitalists who stay; however, when they leave for other reasons they have smaller losses. As

board seats are not a major source of income for venture capitalists, the above result should not be

exaggerated. In Panel C, CEOs of other companies who dissent do significantly worse than CEOs who

stay at the dispute firms; however, when they leave because of dysfunctional boards they only experience

a loss of 52% compared to 73% for all dissenting CEOs of other companies. In Panels D and E,

Attorneys and Founders of other companies who leave due to agency problems only experience a loss in

board seats of 78.4% and 75.9% respectively compared to the average attorney or founder of another

company who leave with a loss of 90.4% and 85.5% respectively.

V. Conclusion

This paper studies the nature of director disputes at public firms over the period 1995-2006.

All of the measures together tell us that dispute firms are quite different from the average firm. Dispute

firms are smaller, have poorer operating performance, higher leverage, are younger firms, and have

significantly different industry concentration than firms in Compustat.

The boards of dispute firms tend to be smaller, more heavily controlled by managers, have less

independent committees, and greater board ownership. Likewise the CEOs are more powerful in that

they are more likely to be on the compensation committee, have higher fractional ownership, and are also

the board’s Chairman in 50% of the firms. All of these measures show that the board and CEO of the

20

dispute firms have significant ownership stakes and voting rights. The fact that inside directors play such

a dominant role on these boards may be contributing to these firms having such strong disagreement

events amongst the directors that lead to director departures.

I also find that newer directors are more likely to leave in disputes especially when the dispute is

over a break down in normal board functions or due to differences in opinion. Independent and linked

directors are more likely to leave due to dysfunctional board behavior. We see that directors with higher

stock ownership are less likely to leave the firm due to agency problems. Directors who are founders of

other companies are more likely to dissent as well as directors who are in private equity or venture capital

when the disagreement is over agency problems or corporate wrongdoing. Attorneys are less likely to

leave because of agency problems or differences in opinion on corporate strategy.

In the five year period after the disagreement event, directors who dissent are not rewarded in the

external market for directorships and experience a loss of 85% of their current board seats. This means

that directors resigning from boards due to disputes are not able to recover the loss of the seat they give

up at the firm with the dispute by receiving additional board seats at other firms in the future. This

outcome is seen for all types of directors and in all types of disputes. For these dissenting directors

dissent is its own reward.

21

References:

Agrawal, Anup and Mark A. Chen (2010). Boardroom Brawls: Determinants and Consequences of Disputes Involving Directors, Working Paper.

Coles, Jeffrey L. and Chun-Keung Hoi (2003). “New Evidence on the Market for Directors: Board

Membership and Pennsylvania Senate Bill 1310,” The Journal of Finance 58, 197-230. Fahlenbrach, Rudiger, Angie Low, and Rene M. Stulz (2010). “The Dark Side of Outside Directors: Do

They Quit When They are Most Neded?,” NBER Working Paper Series. Fama, Eugene F. and Michael C. Jensen (1983). “Separation of Ownership and Control,” Journal of Law

and Economics 26, 301-325. Farrell, Kathleen A. and David A. Whidbee (2000). “The Consequences of Forced CEO Succession for

Outside Directors,” Journal of Business 73, 597-627. Fich, Eliezer M. and Anil Shivdasani (2007). “Financial Fraud, Director Reputation, and Shareholder

Wealth,” Journal of Financial Economics 86. 306-336. Gilson, Suart C. (1990). “Bankruptcy, Bards, Banks, and Bondholders: Evidence on Changes in

Corporate Ownership and Control when Firms Default,” Journal of Financial Economics 27, 355-388.

Harford, Jarrad (2003). “Takeover Bids and Target Directors’ Incentives: the Impact of a Bid on

Directors’ Wealth and Board Seats,” Journal of Financial Economics 69, 51-83. Hermalin, Benjamin E. and Michael S. Weisbach (1988). “The Determinants of Board Composition,”

RAND Journal of Economics 19, 589-606. Hermalin, Benjamin E. and Michael S. Weisbach (1998). “Endogenously Chosen Boards of Directors and

Their Monitoring of the CEO,” The American Economic Review 88, 96-118. Kaplan, Steven N., and David Reishus (1990). “Outside Directorships and Corporate Performance,”

Journal of Financial Economics 27, 389-410. Klein, April (1998). “Firm Performance and Board Committee Structure,” Journal of Law and Economics

27, 275-303. Petersen, Mitchell A. (2009). “Estimating Standard Errors in Finance Panel Data Sets: Comparing

Approaches,” Review of Financial Studies 22, 435-480. Srinivasan, Suraj (2005). “Consequences of Financial Reporting Failure for Outside Directors: Evidence

from Accounting Restatements and Audit Committee Members,” Journal of Accounting Research 43, 291-334.

Yermack, David (1996). “Higher Valuation of Companies with a Small Board of Directors,” Journal of

Financial Economics 40, 185-212. Yermack, David (2004). “Remuneration, Retention, and Reputation Incentives for Outside Directors,”

Journal of Finance 59, 2281-2308.

22

YearDispute Events

(269 unique firms)

One Dissenting Director

Two Dissenting Directors

Three Dissenting Directors

Four Dissenting Directors

Five Dissenting Directors

Total Director

Resignations1994 0 0 0 0 0 0 01995 5 3 1 1 0 0 81996 16 13 3 0 0 0 191997 23 19 3 1 0 0 281998 14 11 0 2 1 0 211999 18 12 4 2 0 0 262000 17 13 3 1 0 0 222001 22 19 2 1 0 0 262002 22 18 2 1 1 0 292003 16 13 3 0 0 0 192004 32 27 4 0 1 0 392005 43 32 5 2 3 1 652006 50 42 5 2 1 0 62Total 278 222 35 13 7 1 364

The sample consists of all director departures due to a disagreement on a company's operations, policies, orpractices from 1994 to 2006. The director departures are from 8-K filings submitted by the firm whichcontain the director's letter of resignation or refusal to stand for re-election. There are 278 dispute eventsover the 13 year period and 269 unique firms with at least one director departure meaning 9 firms had twoseparate dispute events over the time period. In some cases more than one director resigned because of thesame disagreement which is shown in the remaining columns. There were 364 directors that left theirboards because of one of the 278 disputes.

Table 1Director Disputes Per Year - Descriptive Statistics

23

Panel A: Firm Characteristics

Measures of Size Mean Median Std. Dev.(Sample)

(Sample/Benchmark)Total Assets ($ Millions) 487.000 14.167 1911.294 275Total Assets (Percentile Rank) 26.803 18.000 25.519 275/127257T-Test/Sign Test for Percentile Rank = 50 (0.000) (0.000)Sales ($ Millions) 1340.990 9.352 10248.500 275Sales (Percentile Rank) 28.575 21.000 24.455 275/127257T-Test/Sign Test for Percentile Rank = 50 (0.000) (0.000)Market Value of Equity ($ Millions) 709.156 17.350 4325.670 278Market Value (Percentile Rank) 34.799 31.000 24.370 278/110823T-Test/Sign Test for Percentile Rank = 50 (0.000) (0.000)Firm Value ($ Millions) 4277.119 34.937 31034.374 278Firm Value (Percentile Rank) 34.637 29.500 24.482 278/110662T-Test/Sign Test for Percentile Rank = 50 (0.000) (0.000)

Measures of Performance Mean Median Std. Dev.(Sample)

(Sample/Benchmark)Cash Flow / Firm Value -0.323 -0.046 0.593 278CF/FV (Percentile Rank) 35.331 31.000 25.969 278/107960T-Test/Sign Test for Percentile Rank = 50 (0.000) (0.000)Net Income / Firm Value -0.413 -0.104 0.713 278NI/FV (Percentile Rank) 33.540 27.000 25.432 278/110661T-Test/Sign Test for Percentile Rank = 50 (0.000) (0.000)ROE -0.558 0.039 6.294 278ROE (Percentile Rank) 46.723 36.500 37.443 278/126867T-Test/Sign Test for Percentile Rank = 50 (0.146) (0.010)ROA -2.654 -0.309 5.427 278ROA (Percentile Rank) 26.022 16.000 27.670 278/126828T-Test/Sign Test for Percentile Rank = 50 (0.000) (0.000)

Panel A: The sample consists of 278 dispute events leading to at least one director departure from 1994-2006. The firm level data for the 278 dispute events are from Compustat and 10-Ks, annual reports, andother SEC filings. All firm characteristics are measured in the fiscal year end just prior to the dispute anddirector resignation. The percentile rank is the rank of the sample firm to all firms in Compustat in the same year. Assets (AT), Net Sales (SALE) are measured in millions. Market value of equity (MKVALT) wascollected from the introductory paragraph of the 10-K if it was not found in Compustat and is in millions.Firm Value (MKVALT + LT) is market value of equity plus long term debt and is in millions. CashFlow/Firm Value ((OIBDP-TXT-TIE)/Firm Value) is operating income before depreciation minus incometaxes minus interest expense divided by Firm Value. Net Income/Firm Value is IB/Firm Value, ROE(IB/CEQ) is Return on Equity. ROA (IB/AT) is Return on Assets. Total Debt/Firm Value((DLTT+DLC)/Firm Value) is long term debt plus current liabilities divided by Firm Value. Big 6 Auditorsinclude: Ernst & Young, Deloitte & Touche, Arthur Anderson, Coopers & Lybrand, Peat Marwick, andPrice Waterhouse. All accounting variables were windsorized at 1% and 99%. T-tests and Sign tests bothtest the null hypothesis that the percentile rank is equal to 50. P-Values are in parenthesis.

Are firms which have a dispute special?Table 2

24

Other Firm Measures Mean Median Std. Dev.(Sample)

(Sample/Benchmark)Total Debt / Firm Value 0.173 0.085 0.204 251TD/FV (Percentile Rank) 54.948 53.000 26.847 251/110402T-Test/Sign Test for Percentile Rank = 50 (0.004) (0.528)MTB 8.594 1.811 20.048 278MTB (Percentile Rank) 60.493 67.000 31.789 278/110319T-Test/Sign Test for Percentile Rank = 50 (0.000) (0.000)

Firm Age (Years) 14.663 12.000 205Firm in Compustat 0.680 278Big 6 Auditor 0.402 272

Trading VenuesDissent Sample Compustat Observations

Fraction of Firms on NYSE 0.070 0.229 258Fraction of Firms on AMEX 0.047 0.047 258Fraction of Firms on Nasdaq 0.225 0.309 258Fraction of Firms traded OTC 0.659 0.271 258

25

Dissent SampleTwo-Digit SIC Code Description SIC Code % of FirmsBusiness Services 73 15.83%Chemicals and Allied Products 28 7.91%Instruments and Related Products 38 6.47%Electronic & Other Electric Equipment 36 6.47%Communication 48 6.47%Depository Institutions 60 6.12%Industrial Machinery and Equipment 35 3.96%Holding and Other Investment Offices 67 3.96%Health Services 80 3.96%Engineering & Management Services 87 3.60%Number of Firms 278

CompustatTwo-Digit SIC Code Description SIC Code % of FirmsBusiness Services 73 11.55%Chemicals and Allied Products 28 7.13%Instruments and Related Products 38 4.91%Electronic & Other Electric Equipment 36 6.33%Communication 48 3.85%Depository Institutions 60 8.18%Industrial Machinery and Equipment 35 4.60%Holding and Other Investment Offices 67 3.70%Health Services 80 1.36%Engineering & Management Services 87 1.73%Number of Firms 127,534Chi-Square Test Value P-Value 20.937 (0.0129)

Panel B: Industry Composition

Panel B: The top ten two-digit SIC code categories are provided forthe sample of 278 disagreement events resulting in a director departureover the period 1994-2006. These same ten two-digit SIC codes arealso proviced for Compustat firms during 1994-2006. The differencein SIC code distribution is tested between the two groups usingPearson's Chi-Square Test Statistic.

26

Panel A: Board CharacteristicsDissent Sample Risk Metrics Directors

Variable Mean MedianStd.

Deviation Obs. Mean MedianStd.

Deviation Obs.Mean

Difference P-ValueBoard Size 6.435 6.000 2.849 278 9.556 9.000 2.963 17411 -3.120 (0.000)Fraction of Independent Directors 0.509 0.556 0.250 278 0.642 0.667 0.183 17411 -0.133 (0.000)Fraction of Inside Directors 0.394 0.333 0.245 278 0.210 0.182 0.118 17411 0.184 (0.000)Fraction of Linked Directors 0.086 0.000 0.123 278 0.147 0.125 0.146 17411 -0.061 (0.000)Fraction of Independent Audit Committees 0.493 0.000 0.501 278 0.579 1.000 0.494 17411 -0.086 (0.004)Fraction of Independent Compensation Committees 0.367 0.000 0.483 278 0.602 1.000 0.489 17411 -0.235 (0.000)Fraction of Independent Nominating Committees 0.187 0.000 0.391 278 0.362 0.000 0.481 17411 -0.175 (0.000)Board Ownership (%) 0.289 0.233 0.274 266 0.110 0.043 0.164 10038 0.179 (0.000)Directors & Officers Ownership (%) 0.295 0.244 0.229 258Board Meetings 8.214 7.000 4.877 131Audit Committee Meetings 5.466 4.000 4.279 103Fraction of Staggered Boards 0.205 0.000 0.404 278

Panel B: CEO CharacteristicsDissent Sample Risk Metrics Directors

Variable Mean MedianStd.


Deviation Obs.Mean

Difference P-ValueCEO Age (years) 51.753 51.000 10.451 227 53.761 55.000 11.420 17410 -2.008 (0.004)CEO Tenure in Firm (years) 5.167 3.000 6.566 210 9.972 7.000 8.869 14466 -4.806 (0.000)Fraction of CEOs on Nominationg Committees 0.061 0.000 0.239 247 0.080 0.000 0.272 17411 -0.019 (0.207)Fraction of CEOs on Compensation Committees 0.085 0.000 0.279 247 0.012 0.000 0.108 17411 0.073 (0.000)CEO Stock Ownership (%) 0.135 0.061 0.180 232 0.039 0.011 0.075 10231 0.096 (0.000)Fraction of Firms where CEO is also COB 0.502 1.000 0.501 247Fraction of Firms where CEO is Company Founder 0.158 0.000 0.365 247

The sample consists of 278 disagreement events among directors on a company's operations, policies, or practices that lead to at least one director resignation over the period 1994-2006. The benchmark is all firms in Risk Metrics Directors database from 1996-2006. All board and CEO characteristics are measured as of the fiscal year end just prior to thedispute event. Board Ownership (Directors & Officers Ownership) is the beneficial ownership by the board of directors (Directors & Officers) as a percentage of the firm's sharesoutstanding as reported in the 10-K. Board Meetings and Audit Committee meetings are the number of respective meetings held during the fiscal year. Staggered board is equal to 1if the board elects directors for three year terms. The difference in means is tested and P-Values are reported.

Are the board of directors and CEO of firms which have disputes special?Table 3

27

Panel A

General Characteristics Mean MedianStd.


Deviation Obs.Mean

Difference

Mean Difference P-Value

Median Difference P-Value

Age (years) 53.755 54.000 10.895 314 54.876 56.000 11.306 1369 -1.121 (0.111) (0.002)Tenure on Board (years) 4.469 2.000 5.882 322 5.326 3.000 5.806 1356 -0.857 (0.018) (0.000)Individual Stock Ownership (%) 0.042 0.003 0.090 321 0.048 0.000 0.115 1425 -0.006 (0.337) (0.190)Fraction of Female 0.044 0.000 0.205 364 0.060 0.000 0.238 1425 -0.016 (0.189) (0.229)Fraction of Independent 0.525 1.000 0.500 364 0.580 1.000 0.494 1425 -0.055 (0.059) (0.059)Fraction of Employee 0.349 0.000 0.477 364 0.333 0.000 0.472 1425 0.016 (0.575) (0.575)Fraction of Linked 0.113 0.000 0.317 364 0.086 0.000 0.281 1425 0.026 (0.148) (0.121)

Panel B

Insider Executive Title Mean MedianStd.


Deviation Obs.Mean

Difference



Fraction of CEOs of the Disupte Firm 0.148 0.000 0.356 364 0.135 0.000 0.342 1425 0.013 (0.534) (0.524)Fraction of Former CEOs of the Dispute Firm 0.030 0.000 0.171 364 0.032 0.000 0.177 1425 -0.002 (0.839) (0.842)Fraction of COBs of the Dispute Firm 0.126 0.000 0.333 364 0.135 0.000 0.342 1425 -0.009 (0.650) (0.650)Fraction of Former COBs of the Dispute Firm 0.025 0.000 0.156 364 0.017 0.000 0.129 1425 0.008 (0.373) (0.319)Fraction of Founders of the Dispute Firm 0.044 0.000 0.205 364 0.046 0.000 0.210 1425 -0.002 (0.848) (0.848)Fraction of Presidents of the Dispute Firm 0.148 0.000 0.356 364 0.134 0.000 0.341 1425 0.014 (0.479) (0.478)Fraction of CFOs of the Dispute Firm 0.041 0.000 0.199 364 0.050 0.000 0.218 1425 -0.009 (0.493) (0.493)Fraction of COOs of the Dispute Firm 0.052 0.000 0.223 364 0.032 0.000 0.177 1425 0.020 (0.114) (0.070)Fraction of VPs of the Dispute Firm 0.030 0.000 0.171 364 0.056 0.000 0.230 1425 -0.026 (0.017) (0.016)

Dissenting Directors = 364 Remaining Directors = 1425

Are Dissenting Directors Special?Table 4

The sample consists of all directors of the firms which had the 278 dispute events leading to at least one director departure. The group is divided into the 364 directors who resigned from theirboards or refused to stand for re-election and the 1425 directors who remained on their boards through the dispute. Differences between the means and medians of all variables are tested and p-values are reported. Panel A reports general director characteristics. Linked directors are non-employee directors with disclosed conflicts of interst such as consulting contracts, family relationshipswith top management, or significant business relationships with the firm. Panel B reports the current titles or former positions held by Inside Directors of the firms as is reported in their directorbiographies in the firms' proxy statements. CEO is Chief Executive Officer, COB is Chairman of the Board, and COO is Chief Operating Officer. Panel C reports the directors' commiteememberships at the firm with the dispute. Panel D reports directors' primary occupations outside of the company within the last five years. Panel E reports past and present leadership roles held bydirectors in other firms such as CEO, Chairman, and Founder positions. Panel F presents directors' public board seats held before the dispute event occured, at the same time as the dispute event,and up to five years after the dispute event. The percentage net gain (loss) in public board seats after the dispute is the number of new seats gained up to five years after the dispute minus thenumber of seats held during the dispute event that were subsequently lost within one year of the dispute event (including the seat lost at the dispute firm) scaled by all public seats held at the time ofthe dispute.


28

Panel C

Board Experience Mean MedianStd.


Deviation Obs.Mean

Difference



Fraction of Nominating Commitee Chairs 0.014 0.000 0.117 364 0.017 0.000 0.129 1425 -0.003 (0.657) (0.676)Fraction of Nominating Committee Members 0.124 0.000 0.330 364 0.175 0.000 0.380 1425 -0.051 (0.011) (0.019)Fraction of Compensation Committee Chairs 0.027 0.000 0.164 364 0.041 0.000 0.198 1425 -0.013 (0.189) (0.239)Fraction of Compensation Committee Members 0.228 0.000 0.420 364 0.290 0.000 0.454 1425 -0.062 (0.014) (0.019)Fraction of Audit Committee Chairs 0.060 0.000 0.239 364 0.055 0.000 0.228 1425 0.006 (0.673) (0.673)Fraction of Audit Committee Members 0.275 0.000 0.447 364 0.278 0.000 0.448 1425 -0.003 (0.904) (0.904)Fraction of Financial Experts 0.019 0.000 0.138 364 0.018 0.000 0.134 1425 0.001 (0.901) (0.901)

Panel D

Primary Occupation Mean MedianStd.


Deviation Obs.Mean

Difference



Fraction of Academics 0.022 0.000 0.147 364 0.025 0.000 0.157 1425 -0.003 (0.718) (0.718)Fraction of Accountants 0.008 0.000 0.091 364 0.010 0.000 0.099 1425 -0.002 (0.770) (0.781)Fraction of Attorneys 0.036 0.000 0.186 364 0.064 0.000 0.245 1425 -0.028 (0.016) (0.041)Fraction of Consultants 0.093 0.000 0.291 364 0.084 0.000 0.278 1425 0.009 (0.577) (0.577)Fraction of Corporate Executives 0.464 0.000 0.499 364 0.480 0.000 0.500 1425 -0.016 (0.592) (0.592)Fraction of Engineers/Scientists 0.011 0.000 0.104 364 0.004 0.000 0.065 1425 0.007 (0.238) (0.122)Fraction of Investment Bankers 0.014 0.000 0.117 364 0.013 0.000 0.112 1425 0.001 (0.868) (0.867)Fraction of Medical Professionals 0.014 0.000 0.117 364 0.013 0.000 0.115 1425 0.000 (0.953) (0.953)Fraction of Others 0.181 0.000 0.386 364 0.171 0.000 0.377 1425 0.010 (0.650) (0.650)Fraction of Private Equity/Venture Capital 0.038 0.000 0.193 364 0.018 0.000 0.131 1425 0.021 (0.051) (0.015)Fraction of Retired 0.066 0.000 0.249 364 0.076 0.000 0.265 1425 -0.010 (0.521) (0.521)Fraction of Private Investors 0.014 0.000 0.117 364 0.020 0.000 0.141 1425 -0.007 (0.356) (0.410)



29

Panel E

External Corporate Leadership Mean MedianStd.


Deviation Obs.Mean

Difference



Fraction of Current CEOs of another company 0.091 0.000 0.288 364 0.116 0.000 0.320 1425 -0.025 (0.147) (0.173)Fraction of Former CEOs of another company 0.143 0.000 0.350 364 0.150 0.000 0.357 1425 -0.007 (0.723) (0.726)Fraction of Current COBs of another company 0.080 0.000 0.271 364 0.107 0.000 0.309 1425 -0.027 (0.100) (0.128)Fraction of Former COBs of another company 0.088 0.000 0.284 364 0.111 0.000 0.314 1425 -0.023 (0.178) (0.205)Fraction of Founders of another company 0.173 0.000 0.379 364 0.135 0.000 0.342 1425 0.038 (0.080) (0.062)

Panel F

Number of Public Board Seats Held Mean MedianStd.


Deviation Obs.Mean

Difference



Public Board Seats Held Prior to Dispute 0.479 0.000 1.178 363 0.486 0.000 1.150 1425 -0.007 (0.918) (0.971)Public Board Seats Held During Dispute 1.325 1.000 0.882 363 1.449 1.000 1.000 1425 -0.124 (0.020) (0.008)Public Board Seats Gained After Dispute 0.240 0.000 0.690 363 0.218 0.000 0.610 1425 0.022 (0.549) (0.974)Public Board Seats Lost Within One Year After Dispute 1.038 1.000 0.219 363 0.060 0.000 0.276 1425 0.978 (0.001) (0.001)Percent Net Gain (Loss) in Public Board Seats After Dispute -0.732 0.000 0.576 363 0.112 0.000 0.413 1425 -0.845 (0.001) (0.001)



30

Dispute Category Examples of Reasons for Resignation or Refusal to Stand for Re-election

Observations (% of

Sample)Dysfunctional or Inefficient • Not provided with requested or sufficient information on operations or financials to fulfill duties 66 (24%)

Board • Lack of consultation or poor preparation of SEC filings and press releases(Category 1) • Management does not keep board informed of company events

• Director & Officer Liability Insurance not renewed or provided• Not enough board meetings or lack of notification of board meetings• Special Committee did not share information with rest of the board• Board is not in compliance with independence standards• Disagrees with composition of the board or wants new members to join• Not compensated for board services

Agency Problems • Board is under management's control and rubber stamps decisions 96 (34%)Corporate Wrongdoing • Management acts without cosideration or approval of the board

(Category 2) • Management has conflicts of interest which impair their judgement• Accusations of insider trading, fraudulent accounting, or other improprieties by board or management• Lack of due diligence performed or breach of fiduciary duties• Wanted removal of a member of top management (CEO/COB/President)• Compensation of managers or directors is unreasonable• Corporate governance issues (poison pill adoption, shareholder rights plan)• Other directors or management are misrepresenting the company to shareholders

Differences in Opinion About • Disagrees with new direction of the company 82 (30%)Corporate Strategy and Future • Approval/Disapproval of Merger or Acquisition

Direction of the Company • Disagrees with how to deal with insolvency or bankruptyc proceedings(Category 3) • Different strategies for bringing the firm back to profitability

• Firm changed investment strategy• Disagrees with source of financing for the company• Disagreement regarding sale of assets or business units•Opposed new product development, marketing plan, staffing change• Opposed stock repurchase program, stock issuance plan, stock split• Company moved away from R&D focus at the detriment of shareholders• Believes the firm should go private

Other Miscellaneous Reasons • General disagreement with operations, policies, and practices 34 (12%)(Category 4) • Management does not promote diversity or leadership in the workplace

• Inadequate workplace for employees• Breach of employment agreement• Difference in religious beliefs

Table 5Are the disputes special?

All 278 disagreement events have been grouped into one of four categories based on the reason for the director resignation or refusal to stand for re-election. The cause of the resignation was determined by reading the description of the dispute presented by the firm and the dissenting director(s) inthe 8-K filing. Examples of disputes belonging to each category are provided.

31

Panel A: All Directors (marginal effects at the median)

All DisputesDysfunctional

Board Agency ProblemsDifference in

OpinionVariable 1 2 3 4Log (1 + Tenure Years) -0.045** -0.159*** 0.002 -0.052**

(0.014) (0.005) (0.958) (0.047)Independent -0.021 0.152** -0.079 -0.029

(0.590) (0.043) (0.309) (0.713)Linked 0.077 0.230* -0.006 0.097

(0.183) (0.066) (0.949) (0.364)Founder 0.021 0.342* -0.073 -0.030

(0.772) (0.075) (0.439) (0.863)Nominating Committee -0.023 0.075 -0.124* 0.053

(0.553) (0.544) (0.058) (0.471)Compensation Committee -0.023 -0.120 -0.032 -0.056

(0.475) (0.170) (0.595) (0.373)% Stock Ownership -0.002 0.001 -0.006** 0.000

(0.159) (0.899) (0.028) (0.877)Other Public Board Seats - Current -0.038** -0.025 -0.005 -0.101**

(0.050) (0.599) (0.854) (0.044)Founder of Other Company 0.066 -0.073 0.192** 0.024

(0.115) (0.412) (0.020) (0.746)Current CEO of Other Company -0.072* 0.021 -0.086 -0.152

(0.083) (0.861) (0.178) (0.151)Retired -0.004 -0.007 -0.092 0.139

(0.950) (0.968) (0.379) (0.149)Academic 0.002 -0.161 0.221

(0.984) (0.182) (0.113)Venture Capital/Private Equity 0.262*** 0.210 0.370*** 0.090

(0.004) (0.187) (0.002) (0.735)Attorney -0.112** -0.009 -0.178** -0.223**

(0.041) (0.963) (0.036) (0.042)Consultant 0.028 0.134 0.025 -0.035

(0.568) (0.176) (0.815) (0.703)Post SEC Rule Change 0.013 -0.097* 0.086* 0.025

(0.538) (0.093) (0.051) (0.572)Industry Dummies YES YES YES YESClustered S.E. YES YES YES YESObservations 1548 372 533 477R-Square 0.033 0.092 0.072 0.064

Table 6What types of directors dissent for different types of disputes?

The samples in the logistic regressions consists of all 1789 directors of the firms which had the 278 dispute events thatlead to at least one director departure. The dependent variable is equal to 1 if the director resigned from the board orrefused to stand for re-election due to a disagreement amongst the board members and a 0 if the director decided toremain on the board. Panel A presents results for All Directors and Panel B present results for only IndependentDirectors. The regressions are run for the entire group of directors in the first specification and then for each of the threerelevant subgroups of directors based on the reason for the dispute and director resignation as described in Table 5. Onlythe results from the first three categories are presented. An indicator for post SEC rule change is included because theSEC requires all director dipsutes to be reported begining in August 2004. Prior to August 2004 the disputes are onlyreported at the resigning director's request. Industry dummies are included in all specifications. Standard Errors areclustered at the firm level. P-Values are in parenthesis. Marginal effects evaluated at median values are reported insteadof coefficients.

32

Panel B: Independent Directors Only (marginal effects at the median)

All DirectorsDysfunctional


OpinionVariable 1 2 3 4Log (1 + Tenure Years) -0.040 -0.254*** 0.021 -0.043

(0.200) (0.001) (0.599) (0.444)Nominating Committee -0.038 0.170 -0.097** 0.089

(0.471) (0.308) (0.030) (0.345)Compensation Committee -0.017 -0.130 -0.046 -0.122

(0.729) (0.334) (0.299) (0.203)% Stock Ownership 0.001 0.003 -0.002 0.018

(0.789) (0.695) (0.741) (0.177)Other Public Seats - Current -0.026 0.001 0.012 -0.134*

(0.264) (0.986) (0.489) (0.062)Founder of Other -0.030 -0.066 0.003 -0.117

(0.650) (0.604) (0.967) (0.364)Current CEO of Other -0.150** -0.017 -0.052 -0.425***

(0.011) (0.902) (0.288) (0.008)Retired -0.017 0.061 0.032 0.067

(0.798) (0.768) (0.657) (0.526)Academic -0.056 -0.089 0.112

(0.567) (0.219) (0.396)Venture Capital/Private Equity 0.337*** 0.175 0.415*** 0.152

(0.001) (0.299) (0.006) (0.447)Attorney -0.195** 0.257 -0.436**

(0.022) (0.137) (0.025)Consultant 0.030 0.063 0.150 -0.186

(0.691) (0.685) (0.153) (0.209)Post SEC Rule Change 0.014 -0.309** 0.129* 0.013

(0.767) (0.015) (0.064) (0.884)Industry Dummies YES YES YES YESClustered S.E. YES YES YES YESObservations 864 204 292 287R-Square 0.057 0.153 0.131 0.171

33

All Dissenting Directors Vs. All Remaining Directors



OpinionVariable 1 2 3 4Dissenter -0.849*** -0.785*** -0.835*** -0.887***

0.000 0.000 0.000 0.000Log Age -0.182*** -0.128 -0.197* -0.145

0.000 (0.166) (0.054) (0.106)Independent 0.062** 0.061* 0.072 0.050

(0.010) (0.072) (0.130) (0.312)Linked 0.070 -0.036 -0.003 0.183*

(0.153) (0.372) (0.969) (0.080)Other Public Seats - Current 0.010 0.042 -0.004 -0.018

(0.340) (0.110) (0.879) (0.318)ROE 0.005** 0.007*** 0.002 -0.001

(0.012) (0.003) (0.630) (0.856)Log Assets 0.015*** -0.002 0.027*** 0.010

(0.001) (0.669) (0.003) (0.203)Post SEC Rule Change -0.026 -0.011 -0.074* 0.009

(0.299) (0.727) (0.077) (0.833)Intercept 0.751*** 0.509 1.057* 0.524

(0.001) (0.165) (0.054) (0.127)Industry Dummies Yes Yes Yes YesStandard Error Correction Clustered Clustered Clustered ClusteredObservations 1638 396 564 511R-Square 0.377 0.489 0.375 0.392

Table 7Are dissenting directors rewarded?

The sample in the OLS regressions consists of all 1789 directors of the firms which had the 278 dispute events that leadto at least one director departure. OLS regressions are run for the full sample and then for subgroups of directors basedon the reason for the disagreement and director resignation as described in Table 5. The three relevant categories ofdisagreements include those due to 1) Dysfunctional or inefficient boards of directors, 2) Agency problem or accusationsof corporate wrongdoing, and 3) Changes in corporate strategy or the future direction of the company. The dependentvariable is the net gain (or loss) in the number of new board seats after the dispute event has occurred. It is constructedto be the sum of new board seats in public firms gained over the five year period after the dispute event minus the sumof board seats held by the director at the time of the dispute that are subsequently lost within one year of the disputeevent (including the seat lost at the dispute firm) all divided by the number of current seats held at the time of thedispute (including the seat at the dispute firm). Industry dummies are for one-digit SIC codes. Standard Errors areclustered at the firm level where indicated. P-Values are in parenthesis.

34

Panel A: Dissenting Independent Director Interactions



OpinionVariable 1 2 3 4Dissent*Independent (b1) -0.830*** -0.694*** -0.832*** -0.956***

(0.000) (0.000) (0.000) (0.000)Non-Dissent*Non-Independent (b2) -0.038 -0.025 -0.071 -0.035

(0.110) (0.388) (0.155) (0.469)Dissent*Non-Independent (b3) -0.907*** -0.969*** -0.911*** -0.816***

(0.000) (0.000) (0.000) (0.000)Log Age -0.171*** -0.121 -0.197* -0.121

(0.001) (0.185) (0.054) (0.202)Other Public Seats - Current 0.012 0.042 -0.004 -0.013

(0.274) (0.103) (0.873) (0.488)ROE 0.004** 0.008*** 0.002 -0.001

(0.013) (0.001) (0.632) (0.789)Log Assets 0.015*** -0.002 0.027*** 0.012

(0.001) (0.723) (0.003) (0.126)Post SEC Rule Change -0.026 0.002 -0.074* 0.011

(0.310) (0.937) (0.081) (0.798)Intercept 0.760*** 0.506 1.131** 0.483

(0.001) (0.172) (0.040) (0.194)Industry Dummies Yes Yes Yes YesStandard Error Correction Clustered Clustered Clustered ClusteredObservations 1638 396 564 511R-Square 0.376 0.498 0.375 0.388Coefficient Tests: P-ValuesHo: Dissenting*Non-Independent = Non-Dissenting*Non-Independent(Test: b2=b3) (0.000) (0.000) (0.000) (0.000)

The sample in the OLS regressions consists of all 1789 directors of the firms which had the 278 dispute events that lead to at least one directordeparture. Results are for the full sample of directors and then subgroups based on the type of disagreement for each type of dispute withinteraction terms for dissenting Independent directors in Panel A, dissenting Venture Capital/Private Equity directors in Panel B, dissentingCEOs of other companies in Panel C, dissenting Attorneys in panel D, and dissenting Founders of other companies in Panel E. OLSregressions are run for the full sample in each panel and then for subgroups of directors based on the reason for the disagreement and directorresignation as described in Table 5. The three relevant categories of disagreements include those due to 1) dysfunctional or inefficient boardsof directors, 2) Agency problem or accusations of corporate wrongdoing, and 3) Changes in corporate strategy or the future direction of thecompany. The dependent variable is the net gain (or loss) in the number of new board seats after the dispute event has occurred. It isconstructed to be the sum of new board seats in public firms gained over the five year period after the dispute event minus the sum of boardseats held by the director at the time of the dispute that are subsequently lost within one year of the dispute event (including the seat lost at thedispute firm) all divided by the number of current seats held at the time of the dispute (including the seat at the dispute firm). Industrydummies are for one-digit SIC codes. Coefficient tests are at the bottom of the table to test for differences between respective interactionterms. Standard Errors are clustered at the firm level where indicated. P-Values are in parenthesis.

Table 8Are certain dissenting directors rewarded differently?

35

Panel B: Dissenting Venture Capital/Private Equity Director Interactions



OpinionVariable 1 2 3 4Dissent*VC/PE (b1) -0.844*** -1.110*** -0.765*** -0.593**

(0.000) (0.000) (0.000) (0.030)Non-Dissent*Non-VC/PE (b2) -0.003 -0.169 0.291*** -0.335

(0.981) (0.278) (0.000) (0.252)Dissent*Non-VC/PE (b3) -0.850*** -0.939*** -0.540*** -1.223***

(0.000) (0.000) (0.000) (0.000)Log Age -0.158*** -0.122 -0.216** -0.087

(0.002) (0.190) (0.037) (0.414)Other Public Seats - Current 0.015 0.042 0.013 -0.020

(0.160) (0.110) (0.568) (0.267)ROE 0.004** 0.007*** 0.002 -0.001

(0.015) (0.003) (0.480) (0.765)Log Assets 0.016*** -0.001 0.029*** 0.011


(0.307) (0.667) (0.094) (0.920)Intercept 0.694*** 0.690* 0.880 0.667*


Panel C: Dissenting CEOs of Other Company Interactions



OpinionVariable 1 2 3 4Dissent*CEO of Other Company (b1) -0.734*** -0.521** -0.756*** -0.433

(0.000) (0.013) (0.000) (0.137)Non-Dissent*Non-CEO of Other Company (b2) -0.063* -0.045 -0.046 -0.116

(0.095) (0.393) (0.383) (0.180)Dissent*Non-CEO of Other Company (b3) -0.921*** -0.861*** -0.893*** -1.013***

(0.000) (0.000) (0.000) (0.000)Log Age -0.152*** -0.122 -0.174* -0.117

(0.003) (0.166) (0.084) (0.217)Other Public Seats - Current 0.012 0.038* 0.002 -0.019

(0.263) (0.079) (0.926) (0.310)ROE 0.004** 0.007*** 0.002 -0.001

(0.014) (0.005) (0.512) (0.848)Log Assets 0.016*** -0.001 0.028*** 0.011

(0.000) (0.839) (0.002) (0.142)Post SEC Rule Change -0.023 -0.010 -0.069 0.012

(0.364) (0.745) (0.101) (0.780)Intercept 0.722*** 0.574 1.049* 0.543


36

Panel D: Dissenting Attorney Interactions



OpinionVariable 1 2 3 4Dissent*Attorney (b1) -0.903*** -0.955*** -0.787*** -1.020***

(0.000) (0.000) (0.007) (0.000)Non-Dissent*Non-Attorney (b2) 0.008 -0.090 -0.054 0.100**

(0.869) (0.435) (0.669) (0.048)Dissent*Non-Attorney (b3) -0.837*** -0.859*** -0.893*** -0.786***

(0.000) (0.000) (0.000) (0.000)Log Age -0.159*** -0.106 -0.187* -0.120


(0.168) (0.088) (0.890) (0.487)ROE 0.004** 0.007*** 0.002 0.000

(0.015) (0.003) (0.519) (0.920)Log Assets 0.016*** -0.001 0.029*** 0.011


(0.308) (0.756) (0.091) (0.831)Intercept 0.684*** 0.539 1.112* 0.368


Panel E: Dissenting Founder of Other Company Interactions



OpinionVariable 1 2 3 4Dissent*Founder of Other Company (b1) -0.854*** -0.890*** -0.763*** -0.840***

(0.000) (0.000) (0.000) (0.000)Non-Dissent*Non-Founder of Other Company (b2) 0.031 -0.065 0.033 0.079*

(0.270) (0.138) (0.601) (0.099)Dissent*Non-Founder of Other Company (b3) -0.813*** -0.821*** -0.821*** -0.808***

(0.000) (0.000) (0.000) (0.000)Log Age -0.162*** -0.106 -0.184* -0.123


(0.160) (0.074) (0.838) (0.529)ROE 0.004** 0.007*** 0.002 0.000

(0.015) (0.003) (0.492) (0.899)Log Assets 0.016

Are Dissenting Directors Rewarded?€¦ · circumstances and/or resignation letters must be...

Documents

Transcript of Are Dissenting Directors Rewarded?€¦ · circumstances and/or resignation letters must be...