SEARCH

SEARCH BY CITATION

Keywords:

  • director resignations;
  • board independence

Abstract

  1. Top of page
  2. Abstract
  3. 1. INTRODUCTION
  4. 2. DATA
  5. 3. EVENT STUDY ANALYSIS
  6. 4. REGRESSION ANALYSIS
  7. 5. CONCLUSIONS
  8. REFERENCES

Abstract:  As is evident from recent changes in NYSE and NASDAQ listing requirements, board independence is assumed to be an important and effective governance mechanism. However, the empirical evidence regarding the value of board independence is mixed. We examine board member resignation announcements and their perceived importance in the context of firms' existing governance structures. We find that outside director resignations appear to send negative signals to market participants. However, this market reaction is less negative when the board is more independent before the departure and when institutional ownership is high, but is more negative for higher levels of officer and director ownership and CEO incentive compensation.


1. INTRODUCTION

  1. Top of page
  2. Abstract
  3. 1. INTRODUCTION
  4. 2. DATA
  5. 3. EVENT STUDY ANALYSIS
  6. 4. REGRESSION ANALYSIS
  7. 5. CONCLUSIONS
  8. REFERENCES

The Sarbanes-Oxley Act was enacted on July 30, of 2002 as a collection of requirements to help overcome many of the inadequacies related to corporate governance and financial reporting that led to the corporate scandals of the 1990s. Sarbanes-Oxley required greater independence between the major parties involved with the firm such as management, the board of directors, and auditors, and greatly narrowed the definition of ‘independent’ to mean truly autonomous from the firm. Shortly thereafter (in August 2002) the New York Stock Exchange (NYSE) filed with the Securities Exchange Commission (SEC) proposed changes in its listing requirements. The changes, aimed at ensuring board independence and improving corporate governance practices of listed firms, require ‘the board of directors of each listed company to consist of a majority of independent directors’.1 Similar changes filed with the SEC by Nasdaq went into effect on November 4, 2003.

Institutional investors such as the California Public Employee's Retirement System (CalPERS) and their associations like the Council of Institutional Investors (CII) adopted similar guidelines regarding board independence five years before such requirements were adopted by the major stock exchanges. Outside the United States board independence as a means of improving corporate governance has been recommended for a decade or more. For example, in 1990 the government in the United Kingdom appointed the Cadbury Committee that issued The Code of Best Practice that included a recommendation for having at least 3 non-executive directors on the board. Similarly, the 1994 Dey report submitted to the Ontario Exchange Commission led to changes in governance-related disclosure requirements for firms listed on the Toronto Stock Exchange (TSE). Firms listed on the TSE are required to disclose the association of each board member with management and the firm, and whether or not a majority of their directors are independent. Clearly, large and influential entities such as the SEC view board independence as desirable and effective. Yet, there is no consensus in the financial economics literature to suggest that board independence is a necessary and/or sufficient condition for improving firm governance. Given the far-reaching consequences of requiring independent boards for so many firms in the economy, more and better evidence regarding the importance and effectiveness of board independence is needed.

The primary purpose of this study is to use director resignations to examine investors' perceptions of the importance of changes in board independence. A number of studies have examined whether firm value and performance are affected by board structure. The evidence is mixed and often contradictory. Baysinger and Butler (1985) and Hermalin and Weisbach (1991) find that board composition and firm performance are not closely related. In contrast, Rosenstein and Wyatt (1990) find a positive stock price reaction when a new outside director is announced implying that firm value is affected by a firm's proportion of outside directors. Byrd and Hickman (1992) find that when firms make tender offer bids, firms where outside directors hold at least half of the board seats experience a higher announcement return than other bidders. Yermack (1996) documents an inverse relation between board size and firm value and performance. Director resignations provide an abrupt and often unanticipated change in board structure that can be used to gauge the market's perception of the resulting changes in board independence. Although director appointments have been examined previously in the literature (Rosenstein and Wyatt, 1990 and 1997; and Shivdasani and Yermack, 1999), we have a priori reasons to believe that director resignations may be more meaningful to shareholders than are appointments. For example, there is a wealth of evidence in the auditing literature that auditor switches (when the firm chooses to change auditors) has little impact on shareholder wealth (e.g., Nichols and Smith, 1983; Johnson and Lys, 1990; and Klock, 1994). Alternatively, studies of auditor resignations, which typically occur for firms that are in poor financial health, show that resignations have a negative impact on stock prices (e.g., Beneish et al., 2001; Shu, 2000; DeFond et al., 1997; and Wells and Loudder, 1997).

We examine investors' reactions to director resignation announcements and classify the impact on board independence in two ways: First, we assume that the type of director resignation dictates the perceived change in board independence. That is, if an inside director resigns, board independence increases and so on. Second, we classify an increase, a decrease, and no change in the percentage of independent directors by the actual alterations to the board following the resignation announcement. Because the NYSE, CalPERS and others recognize that there are governance mechanisms (e.g., ownership structure) that can substitute for board structure we also examine the resignation announcements' impacts in the context of firms' existing governance structures.2

We find, using a sample of 744 director resignation announcements occurring from 1990 – 2003, that outside director resignations result in an average 1.22% loss (significant at the 1% level) in market value over a three-day period surrounding the announcement. The average abnormal return experienced when insiders or gray directors resign is not statistically significant. Additionally, when there is a net decrease in board independence within a year of the announcement of the director resignation, the average response at announcement is –1.17% (significant at the 5% level), and net increases and no changes to board independence result in negative, but statistically insignificant abnormal returns. The significant, negative stock price reaction to outside director departures can be an indication that investors view such resignations as signals regarding firm quality, and that they value the resulting alterations in board independence (that can result in lower firm quality due to less effective monitoring). Our findings are analogous to studies like Rosenstein and Wyatt (1990) and Rosenstein and Wyatt (1997) that examine investor responses to outsider and insider director appointments, respectively. While these studies (and our initial approach) appear to establish that director appointments (resignations) are deemed to be important events to investors, Rosenstein and Wyatt (1990) do not investigate the market reaction in the context of firms' other governance characteristics or seek to examine the importance of board independence in addition to signaling implications.3 Our multiple regression analysis in addition to our event analysis attempts to shed light on whether board independence is valued, and if so under what circumstances.

We find evidence that investors value board independence (beyond the event analysis) when outside directors resign. Specifically, after controlling for firm performance we find that investors react more negatively to outside director resignations and net decreases in board independence (than to insider or gray resignations and net increases or no changes in independence) when the degree of board independence prior to the resignation is low and institutional ownership is low. However, we find that the market reaction to a loss in board independence is more negative for firms with higher percentages of inside ownership and incentive compensation for their CEOs.

2. DATA

  1. Top of page
  2. Abstract
  3. 1. INTRODUCTION
  4. 2. DATA
  5. 3. EVENT STUDY ANALYSIS
  6. 4. REGRESSION ANALYSIS
  7. 5. CONCLUSIONS
  8. REFERENCES

(i) Data Selection

We identify director resignation announcements between 1990 and 2001 using Lexis – Nexis.4 From 2002 through 2003 we use the Corporate Library (which has data for individual directors as well as governance data for S&P 1500 firms) to identify potential director resignations, then we use Lexis-Nexis to identify publicly announced director resignations. We begin the sample period in 1990 due to superior proxy statement availability beginning in the 1990s. We eliminate events due to confounding announcements occurring on days –3 through +1. Examples of significant confounding events include earnings announcements, dividend announcements, corporate fraud revelations, CEO turnover, and mergers and acquisitions. To be included in the sample, firms must have data available on CRSP and Compustat, and must have proxy statements available (as close as possible to the announcement date, but not exceeding one year prior to the event). These requirements result in a sample of 744 firm events.5

Table 1 reports the distribution of director resignation announcements by director type, by net change in board independence, and by announcement year. The final three years of the sample period contribute about 61% of the viable events reflecting greater Lexis-Nexis coverage, better proxy statement availability in recent years, and data availability (for some 2001 and all 2002-2003 events) from the Corporate Library. Directors who are employees of the respective sample firms are categorized as insiders. Gray directors are former employees of the firm, relatives of firm employees, bankers, accountants, consultants, or attorneys of the firm. Also, we classify directors with any business or other association with the firm that could compromise their ability to be completely independent as gray, even if they do not fit into the other gray director classification categories. All other directors are classified as outsiders. Using this classification scheme we identify 232 inside, 116 gray, and 369 outside director resignations. We also classify the change in independence as the net change (from before the resignation) in the percentage of independent board members reported in the proxy statements in the year following the resignation announcement. We find (Table 1) that there are 367 independence decreases, 91 no changes, and 286 increases to the boards of directors resulting from the resignations.6

Table 1.  Distribution of Sample Across Time
YearNumber ofDirectorResignationsNumber ofInside DirectorResignationsNumber ofGray DirectorResignationsNumber ofOutside DirectorResignationsNumber ofResignationsthat Lead to NetIncrease inBoard IndependenceNumber ofResignationsthat Lead to NoChange inBoard IndependenceNumber ofResignationsthat Lead toDecrease in BoardIndependence
  1. Notes:

  2. The sample includes the firms for which CRSP data, Compustat data, and proxy statements no more than one year prior to each resignation are available. For events before year 2001, we identify director resignation from news announcements on Lexis-Nexis Academic Universe. Thereafter we use the Corporate Library database for director resignations and then use Lexis-Nexis to obtain public announcements of the resignations. The events that are contaminated by other significant announcements about the firm over the 3-day period before, through the day after the announcement of director resignation are also excluded from the sample. The types of announcements used to identify contaminated events are about its earnings, dividends, corporate frauds, top management turnover, or merger and acquisition. Only Non-CEO director resignations are included in the sample. A director is classified as an insider if he/she is an employee of the firm. Gray directors are former employees of the firm, relatives of firm employees, bankers, accountants, consultants, or attorneys of the firm. All others are classified as outsiders. An event is classified as a board independence increasing (decreasing) event if the proxy statement after the resignation reports a higher (lower) percentage of outsiders on the board than before the event.

1990 15 7 3 511 1 3
1991 19 8 5 611 3 5
1992 13 3 4 6 6 2 5
1993 2410 5 911 5 8
1994 13 4 2 7 4 18
1995 31 8 91414 314
1996 3310101314 613
1997 3212 71319 5 8
1998 4212 72323 416
1999 3817 61519 315
2000 2910 41521 1 7
200117077 093911861
20021472116110 602067
2003138333867631956
Total744232 116 396 367 91286 

(ii) Sample Characteristics

The descriptive statistics and variable definitions for the sample (stated in 1995 dollars where applicable) are reported in Table 2.7 In this table we report the 5th and 95th percentiles, means, medians, and standard deviations of the firms' financial characteristics. The median age of the sample firms at the time of the director resignation is 20 years. All firms listed on CRSP during the sample period have a median age of six years revealing that our sample firms are older than the average CRSP firm. Additionally, the median total assets of the firms ($1,627 million) is larger than the median Compustat firm total assets during the same period of $106 million. The fact that our firms appear to be older and larger than CRSP and Compustat firms in general is not surprising given the sample selection criteria. Typically, larger and more well-established firms have proxy statement and Corporate Library data available.

Table 2.  Sample Firm Characteristics
VariableDefinition All DirectorResignationsInside DirectorResignationsGray DirectorResignationsOutside DirectorResignations
  1. Notes:

  2. There are 744 director resignations in the sample during the eleven-year 1990–2003 period. A director is classified as an insider if he/she is an employee of the firm. Gray directors are former employees of the firm, relatives of firm employees, bankers, accountants, consultants, or attorneys of the firm. All other directors are classified as outside directors. Annual Compustat data are collected at the end of the firm's fiscal year before the event date. Compustat data items are mentioned in parentheses.

NNumber of Observations 744232116396
Firm AgeNumber of years since the firm first apprears on CRSP to the date of director resignation announcementMean25.1724.5520.9626.76
Std20.1019.7917.1220.92
P53.003.003.003.00
Median20.0020.0016.0021.00
P9571.0065.0057.0075.00
SizeTotal Assets ($million) (Data6)Mean15033.7512897.188735.6618130.38
Std58393.6549687.7628984.9868567.86
P532.4332.4324.6039.50
Median1627.641495.411130.792001.30
P9549539.0047322.0840297.7155412.04
MVMarket value of firm ($million) = Market value of equity (Data199 * Data25) + Book value of debt (Data181 – Data35 +Data10)Mean27590.0721735.6713364.9335225.44
Std90622.4666018.9339307.62111050.02
P566.4563.1432.7395.06
Median3302.523157.501719.404203.93
P95124787.65109896.7560995.00195778.29
M/BRatio of market value of equity (Data199 * Data25) to book value of equity (Data60)Mean3.4943.5651.9133.915
Std9.926.534.8212.34
P50.540.550.490.55
Median2.0932.0971.6792.194
P9510.4412.697.6910.87
ROAindRatio of net income (Data172) to total assets (Data6), adjusted for 3-digit SIC industry median valuesMean0.0230.030−0.0270.034
Std0.240.220.260.24
P5−0.22−0.19−0.28−0.21
Median0.0140.017−0.0020.015
P950.360.360.1700.40
ROEindRatio of net income (Data172) to market equity (Data199 * Data25), adjusted for 3-digit SIC industry median valuesMean−0.199−0.046−0.074−0.325
Std4.440.720.286.06
P5−0.39−0.35−0.64−0.35
Median0.0140.0220.0040.017
P950.210.250.150.21
ROBindRatio of net income (Data172) to book equity (Data60), adjusted for 3-digit SIC industry median valuesMean−0.035−0.0990.253−0.077
Std3.631.542.304.68
P5−0.60−0.65−0.59−0.61
Median−0.0010.0240.0020.032
P950.410.420.480.40
CAR1yr1 Year CAR using equally weighted market indexMean−0.133−0.146−0.246−0.092
Std0.500.530.490.49
P5−0.98−0.95−0.95−0.85
Median−0.130−0.139−0.236−0.070
P950.600.680.480.59

The median market-to-book ratio of the sample is 2.09. This figure is comparable to the median market-to-book ratio of 1.72 for all Compustat firms over the sample period. Our sample firms under-perform their peers in terms of both accounting and market-based measures of performance. For example, the mean industry-adjusted return on assets is 2.30% (median of 1.40%). The one-year prior to resignation average market- adjusted cumulative abnormal returns for our firms is −13.30% (median of −13.00%).

Table 2 also provides a breakdown of firm characteristics by director type.8 Comparing medians across these sub-samples reveals that firms with outside director resignations are similar to firms whose outgoing directors are insider or gray. For example, the median size (total assets) of firms with insider resignations is $1,495.41 million, which is comparable to $1,130.79 million for firms with gray director resignations, but is smaller than the $2,001.20 million value for outsider resignations. The underperformance of the sample firms before the event (in terms of ROA, ROE and ROEind) is also similar across subsamples defined by the type of outgoing director, although there is enormous variability for these variables for the sample as a whole and across resignation subsamples.

We collect board structure, officer and director ownership, and CEO compensation data from Edgar proxy statements and the Corporate Library within one year prior to the announcement of the director resignations. Spectrum 13F filings are used for institutional ownership data. Table 3 presents the 5th and 95th percentiles, means, medians, standard deviations, and definitions for several governance characteristics of the sample firms. The typical board size is about ten and on average 64.60% of these are outside directors. Outside representation on the board for the sample is higher than the 45.6% reported by Shivdasani and Yermack (1999), but is closer to the percentages of independent directors found by Gillan et al. (2003) of 59%, by Boone et al. (2004) of 56.5%, and by Cheng (2008) of 64.9%. Approximately 75% of the boards in our sample have a majority (greater than 50%) of independent directors.

Table 3.  Sample Firm Governance Structure
VariableDefinition All DirectorResignationsInside DirectorResignationsGray DirectorResignationsOutside DirectorResignations
  1. Notes:

  2. There are 744 director resignations in the sample during the eleven-year 1990–2003 period. Only Non-CEO director resignations are included in the sample. A director is classified as an insider if he/she is an employee of the firm. Gray directors are former employees of the firm, relatives of firm employees, or bankers, accountants, consultants, or attorneys of the firm. All others are classified as outsiders. Firms without a nominating committee are classified as those with a CEO as a member of a nomination committee. Stock incentive compensation of CEO includes value of restricted stock and stock option awards. Industry takeover activity is defined as the fraction of firms acquired over the preceding 7 years that belong to the same two-digit SIC industry as the sample firm.

NNumber of Observations 744232116396
O&D OwnFraction of common stocks owned by officers and directors of the firmMean18.4919.4816.8318.39
Std18.3919.0119.3717.66
P50.918.700.060.98
Median12.8913.139.3014.00
P9554.1458.5060.3052.93
NonAffi-BlockOwnFraction of common stocks owned by Non-affiliated blockholders of the firmMean14.3813.5714.8014.74
Std14.6914.7215.8614.33
P50000
Median10.309.198.4011.22
P950.4342.3845.6843.45
InstiownFraction of common stocks owned by institutional investorsMean53.1351.8948.3355.27
Std24.1123.5125.2923.91
P56.465.756.030
Median56.2655.3451.1857.15
P9589.9186.0382.8464.14
Board SizeBoard size of the firmMean10.2010.589.5810.15
Std3.553.863.383.39
P56666
Median1010910
P9517181517
OutSideMajority= 1 if outside directors form more than 50% of the boardMean0.7450.6290.6200.851
Std0.4350.4840.4870.356
OutDir RatioFraction of the board that are outside directorsMean0.6460.5880.5620.705
Std0.1870.1890.1990.161
P50.2860.2720.2000.375
Median0.6920.6000.6000.727
P950.9000.8670.8570.909
CEO-Chair= 1 if the firm's CEO is also the Chairman of the board, 0 otherwiseMean0.6850.71106400.684
Std0.4650.4540.4820.465
CompTotal Compensation of CEOMean5,914,0934,183,0913,381,2267,670,167
Std22,742,6498,126,0396,696,46830,240,785
P5248,971208,127175,080296,211
Median2,004,4621588,9201,463,1132,480,794
P9517,288,37515,107,35011,823,41618,694,394
Incentive-CompRatio of CEO stock based compensation to his total compensationMean0.4160.3780.3720.451
Std0.3210.3180.3260.316
P50000
Median0.4360.3700.3310.509
P950.8920.8690.8990.905
D/VRatio of Long-term debt to total market value of the firmMean0.2290.2450.2310.220
Std0.2210.220.2360.215
P50000
Median0.1730.1960.1750.155
P950.8920.6450.6900.641
D/V ind-adjIndustry adjusted ratio of Long-term debt to market value of the firmMean0.0650.0700.0780.059
Std0.1790.1790.2260.163
P5−0.165−0.165−0.246−0.165
Median0.0180.0220.0000.022
P950.4070.4310.5660.352
TakeoverFraction of firms acquired over the preceding seven years that belong to the same two-digit SIC industry as the sample firmMean0.2820.2830.2540.290
Std0.0980.1040.0930.095
P50.1360.1310.1310.153
Median0.2730.2730.2500.274
P950.4370.4670.3930.437

We measure firm ownership structure using officer and director ownership, non-affiliated large block (greater then 5%) ownership, and institutional ownership. As shown in Table 3, the mean (median) officer and director ownership of the firms is 18.49% (12.89%), slightly lower than the 21.1% (14.4%) reported by Holderness et al. (1999). This difference in officer and director ownership can be explained by the fact that our firms are larger than those examined by Holderness et al. (1999), as larger firms tend to have more dispersed ownership. Non-affiliated block owners hold a mean of about 14.385% (median of 10.30%) and institutional owners hold a mean of about 58.13% (median of 56.26%) of the firms' shares. The distribution of these ownership variables is similar across sub-samples based on director type.

In Table 3 we also show that the majority of firms have CEOs as Chairmen of the board. A board's responsibilities include hiring and firing the top managers of the firm and structuring the compensation of the firm's top managers. When the chairman and CEO positions are held by the same person the board may not be as effective at monitoring the CEO, thus increasing the importance of monitoring by outside board members (Jensen, 1993). Likewise, when the CEO is a member of the board nominating committee, greater monitoring by outsiders can be needed. Shivdasani and Yermack (1999) report fewer nominations of independent directors when CEOs are on the nomination committee or when there is no formal nomination committee. We find that 68.50% of our firms have CEOs who serve as Board Chairmen, similar to Harley and Wiggins (2004) who find that 68.76% of their sample has CEO-Chair combinations. Additionally, we find that a majority of CEOs also serve on the board nominating committee.

Table 3 reveals that mean (median) CEO compensation is $5.914 million ($2.004 million) in 1995 US dollars. Total compensation is the sum of any grants of restricted stock or stock options, base salary, bonus, and any other long-term non-incentive-based compensation. We calculate the ratio of restricted stock and options-based compensation, Incentive_comp, to total compensation in the year prior to the announcement as a measure of incentive-based pay.9 The CEOs in our sample earn a mean of 41.60% (median of 43.60%) of their total compensation in the form of stock incentives. The percentage of incentive-based CEO pay is similar across the three sub-samples. Murphy and Hall (2002) report a rise in S&P 500 CEOs' incentive compensation from 21% of total compensation in 1992 to 47% in 1999.

Firms in industries with significant takeover activity can find that other governance mechanisms lose some of their importance. Following Agrawal and Knoeber (1996) we compute for each sample firm an industry-specific probability of takeover. The probability is calculated using 2-digit SIC codes and is defined as the fraction of firms acquired over the preceding seven years belonging to the same 2-digit SIC industry as the sample firm. We find that on average 28.2% of firms within our firms' 2-digit SIC industries are acquired over the seven year period before the event. Agrawal and Knoeber report a 27% probability of acquisition for NYSE firms from 1981-1987.

3. EVENT STUDY ANALYSIS

  1. Top of page
  2. Abstract
  3. 1. INTRODUCTION
  4. 2. DATA
  5. 3. EVENT STUDY ANALYSIS
  6. 4. REGRESSION ANALYSIS
  7. 5. CONCLUSIONS
  8. REFERENCES

(i) Event Study Methodology

We capture the market's perception of the importance of director resignations using standard event study methodology. The announcement date (AD) is defined as the date the announcement appears on the Lexis-Nexis database. We use the market model to obtain abnormal returns using a 180-day (AD – 201, AD – 20) estimation period and the CRSP value weighted market index. Firms must have returns for at least 150 trading days before the announcement day to be included in the sample. We also estimate CARs with different parameter estimation periods and using simple market-adjusted returns to control for biases introduced by market model parameter estimation. The results are robust to event study model specifications and to different parameter estimation periods, hence we report only those based on the market model.

To isolate investor reaction to an event, it is crucial to know when the relevant information becomes available to the market. The newswire sources do not capture the exact time of the resignation, and announcements are sometimes made after the close of the major stock exchanges. In order to overcome the ambiguity regarding the timing of the announcement we consider a three-day window (−1, 1) for the event study analysis.10

(ii) Event Study Results

Table 4 presents the mean CAR for all director resignations as well as for each type of resigning director and for the net change in board independence groups. Sample firms on average lose about 0.60% of their market value during the 3-day window around the day of announcement.

Table 4.  Announcement Period Abnormal Returns for Director Resignations
Panel A: Mean CAR
Event WindowEvents Classified by Type of Outgoing DirectorEvents Classified by Net Change in Board Independence
Mean CARPositive : NegativeMean CARPositive : Negative
(−1,+1)All director resignations (N= 744)All director resignations (N= 744)
  −0.60%***356:388  −0.60%***356:388
(0.01) (0.01) 
(−1,+1)Inside director resignations (N= 232)Net Increase (N= 367)
  0.08%113:119 −0.21%175:192
(0.77) (0.35) 
(−1,+1)Gray director resignations (N= 116)No Change (N= 91)
  0.14%62:54 −0.38%46:45
(0.88) (0.35) 
(−1,+1)Outside director resignations (N= 396)Net decrease (N= 286)
   −1.22%**181:215  −1.17%**135:151
(0.00) (0.01) 
Panel B: Test for Difference in Mean CAR
Event WindowEvents Classified by Type of Outgoing DirectorEvents Classified by Net Change in Board Independence
Outside – InsideOutside – GrayInside – GrayNet Decrease – Net IncreaseNet Decrease – No ChangeNet Increase – No Change
  1. Notes:

  2. Three-day cumulative abnormal returns associated with announcements of 744 director resignations over fourteen years (1990–2003). The events that are contaminated by other significant announcements about the firm over the 3-day period before, through the day after the announcement of director resignation are excluded from the sample. The types of announcements used to identify contaminated events are about its earnings, dividends, corporate frauds, top management turnover, or merger and acquisition. Only Non-CEO director resignations are included in the sample. A director is classified as an insider if he/she is an employee of the firm. Gray directors are former employees of the firm, relatives of firm employees, bankers, accountants, consultants, or attorneys of the firm. All others are classified as outsiders. An event is classified as board independence increasing if an insider or gray is replaced by an outsider, insider is replaced by gray, or outgoing director is an insider and identity of replaced director cannot be confirmed. An event is classified as a board independence increasing (decreasing) event if the proxy statement after the resignation reports a higher (lower) percentage of outsiders on the board than before the event. Market model parameters are calculated using 180 trading-days data prior to the event. p-values are reported in parentheses from two-tailed tests using z-scores for Mean CAR.

  3. ***, ** and * indicate statistical significance at the 1%, 5% or 10% levels, respectively, in two-tailed tests.

(−1,+1)  −1.30%** −1.36%**−0.06% −0.96%**−0.79%0.17%
(0.01)(0.02)(0.92)(0.05)(0.21)(0.77)

Also revealed in Table 4, the market reaction to director resignations is only significant when outside directors resign. Sample firms suffer a 1.22% decline in market value during the three-day (−1, 1) event window when outside directors resign. Rosenstein and Wyatt (1990) find a 0.22% positive market reaction to outside director appointments, suggesting that investors may be more concerned with outsider resignations than with their appointments. The market reaction to insider resignations is −0.08% and to gray director resignations is 0.14%, neither of which is statistically significant. We also find that the proportion of firms suffering negative market reactions is higher for outsider resignations than for the other kinds of director resignations. The results for the sample classified by the net change in board independence are very similar to those using the resigning director type to classify the event. That is, firms with decreases in the percentage of independent directors lose 1.17% of their market value, whereas the firms with increases or with no changes in the percentage of independent directors do not experience significant changes on average in their stock prices. We also show in Table 4 pair-wise comparisons between the market reactions to inside, gray, and outside director resignation announcements and between the net change groups. We find that outside director resignations and resignations that result in net decreases in board independence result in abnormal returns that are statistically significantly more negative than those for the inside, gray, or the net increase groups.

Although we find strong evidence that director resignation announcements result in negative wealth effects for shareholders, it is important to control for other governance mechanisms in place at the time of the resignation in order to determine the relative importance of the change in board independence. If other governance mechanisms are quite strong, investors may be less concerned over a loss in board independence, and vice versa. However, if investor reaction to director resignations is only due to a signal about a firm's deteriorating condition and not about the concerns regarding a loss or gain in board independence, the CARs should be unrelated to the existing governance structure including board independence.

4. REGRESSION ANALYSIS

  1. Top of page
  2. Abstract
  3. 1. INTRODUCTION
  4. 2. DATA
  5. 3. EVENT STUDY ANALYSIS
  6. 4. REGRESSION ANALYSIS
  7. 5. CONCLUSIONS
  8. REFERENCES

(i) Model Specifications

We perform a multiple regression analysis to evaluate the effect of firms' existing governance structures on market perceptions of changes in board independence. The following specifications are used for these tests:11

  • image(1)

where CARf is the event study three-day abnormal return around the director resignation announcement for firm f. Dout, and Din are indicator variables for outside and inside director resignation, respectively. Additionally, equation (1) is modified by replacing Dout and Din with Ddec and Dinc. GV is a set of gv corporate governance variables. CV is a set of c control variables. In this specification we control for firm age, size as measured by log of total assets, industry-adjusted return on equity, and past stock performance. In alternative specifications we use as control variables the reasons stated for director departure and director-specific characteristics (such as director age). The following tests are conducted to examine the effect of governance variables on the market reaction by type of director resignation:

  • image

Analogous tests for increases, decreases, and no changes in board independence are also conducted.

We control for firm age and size (as measured by log of total assets) because firms can require different levels of board oversight depending upon their ages and/or relative sizes (which can relate to age), but the predicted sign of these relations is ambiguous. Younger (smaller) firms can require more board oversight because they may not have other mechanisms in place. However, more mature (larger) firms tend to have more free cash flow, and thus can suffer from greater agency problems.

(ii) Regression Analysis Results

The OLS regression results showing the effect of each governance mechanism on CAR for each type of director resignation and their potential to substitute for board independence (equation (1)) are reported in Table 5.12 The effect of each governance mechanism on CAR for each type of director resignation is captured by interacting Din and Dout with each governance mechanism. The coefficients βgv, βingvgv, and βoutgvgv capture the influence of each governance variable on CAR when the outgoing director is a gray, insider, or outsider, respectively.13 The models reported in this table also control for other factors like firm age, size, and past performance that can influence investor reactions to director resignations.14 Model 1 reports coefficient estimates for all governance variables considered. The F-test for model 1 rejects the joint significance of the governance variables. Therefore, we include those variables in models 2, 3 and 4 that are statistically significant in some cases.

Table 5.  OLS Regression Estimates of Three-day CAR on Firm Characteristics
VariablesaModel 1Model 2Model 3
Co-effβingv + βgvβoutgv + βgvCo-effβingv + βgvβoutgv + βgvCo-effβingv + βgvβoutgv + βgv
  1. Notes:

  2. There are 744 director resignations in the sample during the fourteen-year 1990-2003 period. Only Non-CEO director resignations are included in the sample. A director is classified as an insider if he/she is an employee of the firm. Gray directors are former employees of the firm, relatives of firm employees, bankers, accountants, consultants, or attorneys of the firm. All others are classified as outsiders. The regression can be described in the following model:

    • image
  3. ***, ** and * indicate statistical significance at the 1%, 5% or 10% levels, respectively, in two-tailed tests.

  4. aThe variables are defined as follows:

  5.  CARf is three-day event return around director resignation announcement for firm f.

  6.  Din is indicator for inside director resignation

  7.  Dout is indicator for outside director resignation

  8.  GV is a set of firm's (GV) governance variables

  9.  CV is a set of firm's (C) control variables

  10. Governance Variables:

  11.  O&D own= Fraction of common stocks owned by officers and directors of the firm.

  12.  O&D own2= Squared (O&D own).

  13.  NonAffi-BlockOwn= Fraction of common stocks owned by non-affiliated blockholders of the firm.

  14.  NonAffi-BlockOwn2= Squared (NonAffi-BlockOwn).

  15.  Instiown= Fraction of common stocks owned by institutional investors

  16.  Instiown2= Squared (Instiown)

  17.  Inventive-Comp= Ratio of CEO stock based compensation to his total compensation.

  18.  OutDir Ratio= Fraction of the board that are outside directors.

  19.  CEO-Chair= 1 if the firm's CEO is also the Chairman of the board, 0 otherwise.

  20.  D/V ind-adj= Industry adjusted ratio of long-term debt to market value of the firm.

  21.  Takeover= Fraction of firms acquired over the preceding 7 years that belong to the same two-digit SIC industry as the sample firm.

  22. Control Variables:

  23.  Firm Age= Number of years since the firm first apprears on CRSP to the date of director resignation announcement.

  24.  Ln Size= Logarithm of total assets of the firm.

  25.  ROE ind-adj= 3-digit SIC Industry adjusted ratio of net income to market value of equity.

  26.  CAR1yr = One-year CAR using equally weighted market index

Constant−0.002   0.000   0.001  
Din−0.010−0.012  0.002 0.001  0.003 0.004 
Dout−0.078*** −0.08***−0.085** −0.085***−0.089*** −0.088***
O&D own 0.074−0.053−0.083* 0.078−0.063−0.077* 0.083−0.059−0.075***
O&D own2−0.118 0.061 0.169**−0.122 0.075 0.159**−0.120 0.075 0.159**
NonAffi-BlockOwn−0.002 0.001 0.001−0.002 0.001 0.001   
NonAffi-BlockOwn2 0.001 0.001 0.001 0.000 0.000 0.000   
Instiown−0.067−0.088 0.099**−0.056−0.080 0.091*−0.061−0.080 0.095***
Instiown2 0.029 0.100−0.078* 0.020 0.092−0.073 0.027 0.098−0.070**
Incentive-Comp 0.014−0.001−0.018* 0.015−0.001−0.017* 0.018 0.000−0.015**
OutDir Ratio−0.018−0.022 0.055***−0.020−0.016 0.048**−0.021−0.017 0.050***
CEO-Chair−0.006 0.009−0.008      
D/V ind-adj 0.006 0.014−0.02 0.009 0.017−0.018   
Takeover 0.024 0.029−0.029      
Firm Age−0.001   0.000     
Ln Size 0.004***   0.004***   0.003***  
ROE ind – adj−0.002   0.009     
CAR1yr 0.009*   0.000   0.007  
adj-R2 0.048   0.051   0.055  
Number of Obs  743    743    744  
Prob>F (test for 0.000   0.000   0.000  
joint significance)         

In Table 5 we show that officer and director ownership (O&D own), institutional ownership (Instiown), CEO incentive-compensation (Incentive-Comp), and the pre-announcement degree of board independence are related to CAR for outside director resignations. None of the governance variables shows any significant relation to CAR for insider or for gray director resignations. O&D own, and Instiown also show a non-linear relation with CAR. In particular, O&D own is negatively related with CAR and the square of O&D own is positively related with CAR. That is, investors react more negatively to outside director resignation when officers and directors have large ownership stakes in the firm; however, investor reaction is less negative for firms with higher levels of O&D ownership. This relation between O&D own and CAR suggests that board independence is less valuable for shareholders at higher levels of insider ownership.15 Our results are consistent with Bhagat and Black (2001) who use a simultaneous equation framework to examine the effect of board independence on firm performance while controlling for endogeniety. They find that board independence is likely to be low when CEO ownership is high. Our results are also consistent with Dahya and McConnell (2005) who find that firms with more independent boards are more likely to appoint outside CEOs (who initially have lower levels of ownership). These findings, consistent with our results, can be interpreted to mean that CEOs with high ownership are more entrenched and do not allow greater board independence. Therefore, a decrease in existing board independence when CEO ownership is high is viewed more negatively because the CEO is unlikely to allow for an independent replacement. Consistent with our findings, Denis and Sarin (1999) find a negative relation between board independence and inside ownership. Such a substitution effect between inside ownership and board independence is supported by the findings of Bathala and Rao (1995) and Berry, Fields and Wilkins (2006) as well.

The positive coefficient of Institown for outsider resignations is consistent with institutional ownership providing a monitoring function for the firm that can substitute for board independence. The CARs for outsider resignations are higher (or less negative) by 1% (from model 2 in Table 5) when there is a 10% increase in intitutional ownership. For very high levels of institutional ownership, this relationship turns negative, suggesting that at such high levels, the market does not perceive that increasing institutional ownership adds value as a corporate governance mechanism.

Table 5 also shows that firms that reward their CEOs with more stock-based compensation suffer a more negative market response when outsiders resign. A negative coefficient of −0.018 for Incentive-comp in model 2 indicates that for every 10% increase in incentive compensation, event returns for firms with outsider resignations are lower by 0.18%. This is contrary to the standard belief that higher incentive compensation for CEOs leads to better corporate governance. Recently, however, Erickson et al. (2004) show that the possibility of corporate fraud is an increasing function of the CEO's incentive compensation. The findings reported in Table 5 are consistent with incentive compensation at high levels exacerbating agency problems rather than being a potential mitigating factor.

In Table 5 we find that firms with higher pre-announcement percentages of board independence have less negative reactions to independent board member resignation. The negative relation between abnormal returns and existing board independence is consistent with the argument that shareholders view board independence as being important.16

The OLS regression results showing the effect of each governance mechanism on CAR for net increases, net decreases, and no change in board independence and their potential to substitute for board independence (modified equation (1)) are reported in Table 6. The effect of each governance mechanism on CAR for each board change category is captured by interacting Dinc and Ddec with each governance mechanism. The coefficients βgv, βincggv, and βdecgvgv capture the influence of each governance variable on CAR when there is no change, an increase in independence, or a decrease in independence, respectively.

Table 6.  OLS Regression Estimates of Three-day CAR on Firm Characteristics for Changes in Board Independence
VariablesaModel 1Model 2Model 3
Co-effβingv + βgvβoutgv + βgvCo-effβingv + βgvβoutgv + βgvCo-effβingv + βgvβoutgv + βgv
  1. Notes:

  2. There are 744 director resignations in the sample during the fourteen-year 1990–2003 period. An event is classified as a board independence increasing event if the proxy statement after the resignation reports a higher percentage of outsiders on the board than before the event. An event is classified as a board independence decreasing event if the proxy statement after the resignation reports a lower percentage of outsiders on the board than before the event. In case the proxy statement following the event is not available, we classify an outsider resignation event as independence decreasing, an insider resignation event as independence increasing and gray director resignation as no change in board independence. The regression can be described in the following model:

    • image
  3. ***, ** and * indicate statistical significance at the 1%, 5% or 10% levels, respectively, in two-tailed tests.

  4. aThe variables are defined as follows:

  5.  CARf is three-day event return around director resignation announcement for firm f.

  6.  Dinc is indicator if the director's resignation leads to an increase in board independence.

  7.  Ddec is indicator if the director's resignation leads to a decrease in board independence.

  8.  GV is a set of firm's (GV) governance variables

  9.  CV is a set of firm's (C) control variables

  10. Governance Variables:

  11.  O&D own= Fraction of common stocks owned by officers and directors of the firm.

  12.  O&D own2= Squared (O&D own).

  13.  NonAffi-BlockOwn= Fraction of common stocks owned by non-affiliated blockholders of the firm.

  14.  NonAffi-BlockOwn2= Squared (NonAffi-BlockOwn).

  15.  Instiown= Fraction of common stocks owned by institutional investors

  16.  Instiown2= Squared (Instiown)

  17.  Inventive-Comp= Ratio of CEO stock based compensation to his total compensation.

  18.  OutDir Ratio= Fraction of the board that are outside directors.

  19.  CEO-Chair= 1 if the firm's CEO is also the Chairman of the board, 0 otherwise.

  20.  D/V ind-adj= Industry adjusted ratio of long-term debt to market value of the firm.

  21.  Takeover= Fraction of firms acquired over the preceding 7 years that belong to the same two-digit SIC industry as the sample firm.

  22. Control Variables:

  23.  Firm Age= Number of years since the firm first apprears on CRSP to the date of director resignation announcement.

  24.  Ln Size= Logarithm of total assets of the firm.

  25.  ROE ind-adj= 3-digit SIC Industry adjusted ratio of net income to market value of equity.

  26.  CAR1yr = One-year CAR using equally weighted market index

Constant−0.015  −0.015  −0.010  
Dinc0.0170.002 0.0170.002 0.0130.003 
Ddec−0.074* −0.089***−0.076** −0.092***−0.086** −0.097***
O&D own0.182−0.050−0.103**0.183*−0.056−0.098**0.172−0.048 −0.103** 
O&D own2−0.350*0.0880.192***−0.351*0.0940.185***−0.313*0.084 0.189***
NonAffi-BlockOwn0.0010.0010.0000.0010.0010.000   
NonAffi-BlockOwn20.0000.0000.0000.0000.0000.000   
Instiown−0.088−0.0590.108*−0.087−0.0590.106*−0.104−0.062  0.116**
Instiown20.0700.043−0.0730.0700.042−0.0740.0980.055−0.082   
Incentive-Comp0.027−0.008−0.020*0.026−0.008−0.019*0.019−0.006−0.015   
OutDir Ratio−0.026−0.0130.063***−0.026−0.0130.060**−0.017−0.011  0.059**
CEO-Chair−0.0010.002−0.007      
D/V ind-adj0.0370.015−0.043*0.0370.016−0.041*   
Takeover−0.002−0.012−0.006      
Firm Age0.000  0.000     
Ln Size0.004**  0.004**  0.003**  
ROE ind – adj−0.001        
CAR1yr0.005  0.005  0.004  
adj-R20.042  0.050  0.046  
Number of Obs  743    743    744  
Prob>F (test for0.002  0.000  0.000  
joint significance)         

The results for models 1-4 in Table 6 are very similar to the results shown in Table 5. Specifically, we show that officer and director ownership (O&D own), to a lesser extent than shown in Table 4 institutional ownership (Instiown), CEO incentive-compensation (Incentive-Comp), and the pre-announcement board independence are related to CAR for decreases in board independence, but none of the governance variables is significant for increases or no change in independence. However, only O&D own shows a non-linear relation with CAR.

The results presented in Tables 5 and 6 are robust to controlling for firm age, size, and past performance. Firm size (measured by the log of total assets) and past stock performance (measured by one year cumulative abnormal returns before the event) are positively related to CAR suggesting that larger firms and firms that outperform to market face less erosion in value when outsiders resign. Firm age and accounting return appear to be unrelated to investor reactions to director resignations. Alternative specifications of the models above including control variables such as director age at departure were specified, but none was statistically significant.

Overall, our results suggest that investor reaction to decreases in board independence (whether measured by departures of outside directors or measured by the decrease in the percentage of independent directors) is affected by the strength of the firm's other governance mechanisms. Specifically, we find that institutional ownership can potentially serve as a substitute for board independence. We also find that the market reaction to a decrease in board independence is less negative for firms with greater board independence prior to the resignation, and is more negative for firms with higher inside ownership and CEO incentive compensation.

5. CONCLUSIONS

  1. Top of page
  2. Abstract
  3. 1. INTRODUCTION
  4. 2. DATA
  5. 3. EVENT STUDY ANALYSIS
  6. 4. REGRESSION ANALYSIS
  7. 5. CONCLUSIONS
  8. REFERENCES

We examine 744 board changes occurring between 1990 and 2003, and find that they are met with a statistically significant negative stock price response. Significant differences in the responses exist across director resignation types. Outside director resignations (or decreases in the percentage of independent directors) result in significant, negative abnormal returns. Shareholders do not, however, respond significantly to insider or gray director resignations (or to increases or no change in the percentage of independent directors).

We find that institutional ownership possibly substitutes for board independence. Our regression results show that as institutional ownership increases the negative reaction to a reduction in board independence is mitigated. Additionally, the market reaction to a loss in board independence is more negative for firms with higher percentages of incentive compensation for their CEOs and officer and director ownership, but is less negative for firms with higher pre-announcement percentages of independent directors. We do not find evidence of a relation between governance mechanisms and insider or gray resignations (or for increases or no change in board independence) in terms of the market's reaction to these events. In sum, our findings support the changes in listing requirements recently adopted by the NYSE and Nasdaq in that we find that investors do appear to value board independence.

Footnotes
  • 1

    Page 4, ‘NASD and NYSE Rulemaking: Relating to Corporate Governance’, available at http://www.sec.gov/rules/sro/34--48745.htm

  • 2

    The NYSE listing requirements allow exceptions to the ‘independent board’ rule for closely held companies, implying that ownership structure can substitute for board independence. Further, CalPERS recognizes that the effectiveness of an independent board can depend upon the existing governance structure and the business environment in which the firm operates. In their own words ‘CalPERS recognizes that some of these (guidelines) may not be appropriate for every company, due to differing development stages, ownership structure, competitive environment, or a myriad of other distinctions' (Page 3, ‘Corporate Governance Core Principles and Guidelines’ CalPERS (1998), available at http://www.calpers-governance.org/principles/domestic/US/page01.asp).

  • 3

    Rosenstein and Wyatt (1997) examine the relation between abnormal returns to inside director appointments and two governance variables (insider ownership and board composition). They find that abnormal returns are positive for moderate levels (5%-25%) of insider ownership, and that board composition is not significantly related to abnormal returns.

  • 4

    We identify director resignation by searching Lexis-Nexis database for keywords ‘director’ and (‘resignation’ or ‘quit’).

  • 5

    The sample of director resignations used in this study is smaller than the samples of appointments in previous studies (such as Rosenstein and Wyatt, 1990 and 1997) due to the stringent sample selection criteria we employ. The two primary reasons we lose sample points in our study are a lack of proxy statement data (not relevant to Rosenstein and Wyatt, 1990) and contaminating announcements close to the resignation announcements. Rosenstein and Wyatt require that there be no confounding announcements on the day of the appointment, while we require that there be no contaminating events from three days prior to one day following the event. In several cases we find that the resignations are part of the information released in the proxy statements where it states that the director is not being renominated. These observations are eliminated from the sample. Also, our sample of resignations could be smaller than the samples of appointments because there are quite simply more public announcements for appointments than there are for resignations.

  • 6

    Finding the true underlying reason for departure is difficult because often boiler plate statements such as the director ‘wants to spend more time with family’ or the director wants to ‘pursue other opportunities' mask an underlying acrimonious departure. Because we cannot determine which resignations are genuinely acrimonious and because we are concerned about the impact of the directors' departures on board independence (and how that impact is felt in the context of other governance characteristics) rather than on the signaling aspects of the departure, we do not attempt to determine resignation motives.

  • 7

    The industry distribution of the sample firms is identified using Compustat two-digit SIC codes. The majority of firms (46%) are considered to be manufacturing firms (SIC codes 2000 to 3999). Regulated industries such as banks and insurance companies comprise about 14% of the sample.

  • 8

    The summary characteristics for the sample classified based on net change in board independence (not reported here to conserve space) are very similar to those in Table 2.

  • 9

    The value of stock options is calculated using the Black and Scholes (1973) option pricing model adjusted for continuously paid dividends as Nt×[SteqTΦ (d1) −KerTΦ (d1−σT1/2)] where Nt is the number of options granted in year t at exercise price K, T is the number of years until expiration, r is the average monthly yield on 10-year treasury notes in year t, q is the firm's dividend yield in year t−1, St is the fiscal year closing price, and σ is the annualized standard deviation of stock returns over the previous sixty months. Φ(.) is the cumulative standard normal distribution, and d1 is defined as:

    • image

    If the expiration date is not specified in the proxy statement, we assume that there are ten years remaining until expiration. If more than one series of options is granted in a particular year, we take a weighted average of the strike prices of each of the option grants.

  • 10

    We also employ two-day windows such-as (−1, 0) and (0, 1). We find that our results are robust to alternative event window specifications.

  • 11

    The models are also estimated with and without a series of one-digit SIC code dummy variables and with and without a series of time dummy variables. The results are qualitatively unaffected by inclusion or exclusion of these dummy variables, therefore the results reported are those without dummies to prevent substantial losses in degrees of freedom. Similarly, we test for year effects by including year dummies, but the results are unaffected.

  • 12

    The results do not change when we exclude firms with multiple resignation announcements from the sample. There are 20 firms with multiple director resignations on the same announcement date. These firms average 2.25 resignations on the announcement date, and their CAR is –1.45%.

  • 13

    We also performed the analysis for each subgroup (insider, outside, and gray, as well as, increases, decreases, and no change in independence) separately. The results are qualitatively unchanged from those presented in Tables 5 and 6. Therefore, we present the combined analysis because doing so provides better estimation of the standard errors.

  • 14

    Our tests failed to reject the hypothesis that residual variance is homogenous (p-value of 0.89). Therefore, we conclude that heteroscedasticity is not a problem. Variance Inflation factors for regressions with and those without squared terms are less than 5, and therefore, our regression models do not suffer from multicollinearity.

  • 15

    To ensure that our results are not unduly influenced by extreme observations, we perform the analysis on winsorized data (eliminating observations in the top and bottom 1%). We find that the results are qualitatively unchanged. We report the results for the full (not winsorized) sample to preserve sample size.

  • 16

    We find using piecewise linear regression models the results diminish (for board independence) as the level of independence increases. In fact the most negative response is for boards with less than 50% independent directors. Between 50% and 75% the response is still negative but is of smaller magnitude. Finally, once the board exceeds 75% independent directors the response is close to zero.

REFERENCES

  1. Top of page
  2. Abstract
  3. 1. INTRODUCTION
  4. 2. DATA
  5. 3. EVENT STUDY ANALYSIS
  6. 4. REGRESSION ANALYSIS
  7. 5. CONCLUSIONS
  8. REFERENCES
  • Agrawal, A. and C. Knoeber (1996), ‘Firm Performance and Mechanism to Control Agency Problems Between Managers and Shareholders’, Journal of Financial and Quantitative Analysis, Vol. 31, No. 3, pp. 37797.
  • Bathala, C. and R. Rao (1995), ‘The Determinants of Board Composition: An Agency Theory Perspective’, Managerial and Decision Economics, Vol. 16, pp. 5969.
  • Baysinger, G. and H. Butler (1985), ‘Corporate Governance and the Board of Directors: Performance Effects of Changes in Board Composition’, Journal of Law, Economics, and Organizations, Vol. 1 (January/February), pp. 10124.
  • Beneish, M., P. Hopkins and I. Jansen (2001), ‘Do Auditor Resignations Convey Private Information About Continuing Audit Clients?’, Working Paper (Indiana University).
  • Berry, T., L. Fields and M. Wilkins (2006), ‘The Interaction Among Multiple Governance Mechanisms in Young Newly Public Firms’, Journal of Corporate Finance, Vol. 12, pp. 44966.
  • Black, F. and M. Scholes (1973), ‘The Pricing of Options and Corporate Liabilities’, Journal of Political Economy, Vol. 8, (May/June), pp. 63756.
  • Boone, A., L. Field, J. Karpoff and C. Raheja (2007), ‘The Determinants of Corporate Board Size and Composition’, Journal of Financial Economics, Vol. 85 (July), pp. 66101.
  • Byrd, J. and K. Hickman (1992), ‘Do Outside Directors Monitor Managers? Evidence from Tender Offer Bids’, Journal of Financial Economics, Vol. 32, pp. 195222.
  • Cheng, S. (2008), ‘Board Size and Variability of Corporate Performance’, Journal of Financial Economics, Vol. 87, pp. 15776.
  • Dahya, J. and J. McConnell (2005), ‘Outside Directors and Corporate Board Decisions’, Journal of Corporate Finance, Vol. 11, pp. 3760.
  • DeFond, M., M. Ettredge and D. Smith (1997), ‘An Investigation of Auditor Resignations’, Research in Accounting Regulation, pp. 2545.
  • Denis, D. and A. Sarin (1999), ‘Ownership and Board Structures in Publicly Traded Corporations’, Journal of Financial Economics, Vol. 52, pp. 187223.
  • Erickson, M., M. Hanlon and E. Maydew (2006), ‘Is There a Link between Executive Compensation and Accounting Fraud?’, Journal of Accounting Research, Vol. 44, No. 1, pp. 11343.
  • Gillan, S., J. Hartzel and L. Starks (2003), ‘Explaining Corporate Governance: Boards, Bylaws, and Charter Provisions’, Working Paper (University of Delaware).
  • Hall, B. and K. Murphy (2002), ‘Stock Options for Undiversified Executives’, Journal of Accounting and Economics, Vol. 33, No. 1, pp. 342.
  • Harley, E. and R. Wiggins (2004), ‘Who is in Whose Pocket? Director Compensation, Board Independence, and Barriers to Effective Monitoring’, Journal of Financial Economics, Vol. 73, pp. 497524.
  • Hermalin, B. and M. Weisbach (1991), ‘The Effects of Board Composition and Direct Incentives on Firm Performance’, Financial Management, Vol. 20, pp. 10112.
  • Holderness, C., R. Kroszner and D. Sheehan (1999), ‘Were the Good Old Days that Good? Changes in Managerial Stock Ownership Since the Great Depression’, Journal of Finance, Vol. 54, pp. 43570.
  • Jensen, M. (1993), ‘The Modern Industrial Revolution, Exit, and the Failure of Internal Control Systems’, Journal of Finance, Vol. 48, pp. 83180.
  • Johnson, W. and T. Lys (1990), ‘The Market for Audit Services: Evidence from Voluntary Auditor Changes’, Journal of Accounting and Economics, Vol. 12, pp. 281308.
  • Klock, M. (1994), ‘The Stock Market Reaction to a Change in Certifying Auditor’, Journal of Accounting, Auditing and Finance, Vol. 9, pp. 33047.
  • Nichols, D. and D. Smith (1983), ‘Auditor Credibility and Auditor Changes’, Journal of Accounting Research, pp. 53444.
  • Rosenstein, S. and J. Wyatt (1990), ‘Outside Directors, Board Independence, and Shareholder Wealth’, Journal of Financial Economics, Vol. 26, pp. 17591.
  • Rosenstein, S. and J. Wyatt (1997),‘Inside Directors, Board Effectiveness and Shareholder Wealth’, Journal of Financial Economics, Vol. 44, pp. 22950.
  • Shivdasani, A. and D. Yermack (1999), ‘CEO Involvement in the Selection of New Board Members: An Empirical Analysis’, Journal of Finance, Vol. 54, pp. 182953.
  • Shu, S. (2000), ‘Auditor Resignations: Clientele Effects and Legal Liability’, Journal of Accounting and Economics, Vol. 29, pp. 173205.
  • Wells, D. and M. Loudder (1997), ‘The Market Effects of Auditor Resignations’, Auditing: A Journal of Practice and Theory, Vol. 16, pp. 13844.
  • Yermack, D. (1996), ‘Higher Market Valuation for Firms with a Small Board of Directors’, Journal of Financial Economics, Vol. 40, pp. 185211.