1
An analysis of the announcement
effect of forced CEO resignation on
US stock returns during the crisis
Name: Berend Reeskamp Student Number: 10431659
Programme: Economie en Bedrijfskunde Track: Finance and Organization
Supervisor: drs. P.V. Trietsch MPhil Date: 29-06-2015
Abstract
This research examines the effect of announcements of forced CEO departures on stock returns in the United Sates during the global crisis period. A sample of 47 announcements from January 2007 to April 2013 is used. The influence of the financial crisis on the stock market reaction to forced CEO resignation announcements is the main focus of this study. By means of an event study of daily abnormal returns, a significant announcement effect of forced CEO changes is shown, which is negative for both the entire sample and the announcements during the financial crisis. Using cross-sectional regressions, the influence of several variables on the stock market reaction is shown. Evidence is presented for a negative influence of the financial crisis. In addition, an inverse relation between firm size and abnormal stock returns caused by the announcements is suggested. Moreover, a higher leverage ratio prior to announcements of forced CEO changes affects the announcement returns positively. No significant influences on stock returns are found for outside succession announcements and poor firm performance. An overall negative announcement effect of forced CEO resignations on stock returns is found in this research and this effect is amplified by the financial crisis.
2
Statement of Originality
This document is written by Student Berend Reeskamp who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
3
Content
1 INTRODUCTION ... 5
2 LITERATURE REVIEW ... 6
2.1EXPECTED SHARE AND FIRM PERFORMANCE SURROUNDING TOP MANAGEMENT CHANGES 6 2.1.1 Expected performance preceding top management changes in general ... 7
2.1.2 Expected performance following top management changes in general ... 7
2.1.3 Real effect and information effect of announcing top management changes in general ... 7
2.2FORCED CEO CHANGES ... 8
2.2.1 Firm performance and share performance preceding forced CEO changes... 8
2.2.2 Share price performance following forced CEO changes ... 8
2.2.3 How are forced CEO changes identified in literature? ... 9
2.3PREVIOUSLY FOUND EFFECTS OF ANNOUNCEMENTS OF TOP MANAGEMENT CHANGES ... 10
2.4CONTROLLING FOR THE ORGANIZATIONAL CONTEXT OF FORCED CEO CHANGES ... 10
2.4.1 Origin of the successor ... 11
2.4.2 How is the origin of the successor measured in the literature? ... 12
2.4.3 Firm performance ... 12
2.4.4 How is firm performance measured in the literature? ... 13
2.4.5 Firm size ... 13
2.4.6 How is firm size measured in the literature? ... 14
2.4.7 Leverage ratio ... 14
2.4.8 How is the leverage ratio measured in the literature ... 14
2.5FORCED CEO CHANGE ANNOUNCEMENTS DURING THE FINANCIAL CRISIS ... 14
2.5.1 Financial crisis variable and hypothesis ... 15
2.6SAMPLE SIZE AND METHODOLOGY OF EVENT STUDY USED IN THE LITERATURE ... 15
3 SAMPLE AND DATA ... 16
3.1SAMPLE OF FORCED CEO CHANGE ANNOUNCEMENTS... 16
3.2DATA SELECTION FOR THE EVENT STUDY ... 17
4 METHODOLOGY AND MODEL SET-UP ... 18
4.1SAMPLE PERIOD ... 18
4
4.3REGRESSION ... 19
4.3.1 Regression model ... 19
4.3.2 How are the variables included in the regression measured? ... 20
5 RESULTS ... 20
5.1SUMMARY STATISTICS OF THE INDEPENDENT VARIABLES ... 21
5.2EVENT STUDY RESULTS ... 22
5.2.1 Cumulative abnormal returns for main event window (-1;+1) ... 22
5.2.2 Cumulative average abnormal returns for different event windows ... 22
5.3REGRESSION RESULTS ... 24
5.4ROBUSTNESS OF REGRESSION RESULTS ... 27
6 CONCLUSION AND DISCUSSION ... 28
6.1CONCLUSION ... 28
6.2LIMITATIONS AND FUTURE RESEARCH ... 29
REFERENCE LIST ... 30
APPENDIX ... 32
A.1OVERVIEW OF TOTAL SAMPLE OF FORCED CEO CHANGE ANNOUNCEMENTS ... 32
A.2EVENT STUDY, TEST STATISTICS: ... 33
A.3AARTABLE FOR DAYS -20;+3 ... 35
A.4CAAR RESULTS ADDITIONAL EVENT WINDOWS ... 36
A.5ROBUSTNESS REGRESSION RESULTS ... 36
A.5.1 Scatter Plots (Linearity and detection of outliers) ... 36
A.5.2 Regression outcomes without leverage ratio outliers ... 38
5
1 Introduction
This study examines the effect of announcing forced chief executive officer (CEO) changes on the stock returns of firms in the United States during the global crisis period. Forcing a CEO out of a firm is a major decision of the board, with long-term implications for the operating, financing and investment strategies of a company and therefore for future performance (Huson et al., 2001).
A CEO resignation generally could have a significant effect on stock returns because shareholders react instantly to the announcement of a change on the stock market. This leads to either positive or negative abnormal stock returns. Positive abnormal stock returns should occur when the CEO change is in the interest of the shareholders (Bonnier and Bruner, 1989). Abnormal stock returns could be negative if the CEO change indicates that management performance is worse than the market had anticipated (Warner et al., 1988). Due to this ambiguity, the expected effect on stock returns of announcing CEO changes is not immediately clear.
Richard Fuld was ousted as CEO of Lehman Brothers in September 2008 after the firm suffered from huge losses caused by bad home loans and its shares had dropped 80 percent earlier that year. As in this extreme example, forced resignations are usually preceded by a period of poor management and low stock returns in existing literature (Denis and Denis, 1995). Shareholder wealth also increases when dismissals of key executives occur, implying that forced CEO changes lead to increases in stock returns (Furtado and Rozeff, 1987).
The effect of announcing forced CEO changes on shareholder value is researched during the global crisis period of January 2007 to April 2013. Abnormal stock returns caused by announcing forced CEO changes are calculated using event studies. Any cross-sectional variation in these abnormal returns could be explained by firm characteristics or situations. These variables could influence the stock market reaction to a forced CEO change and this is researched using cross-section regressions. The variables are described below.
The unique variable employed in this research is the financial crisis period dummy for forced CEO changes in the period from September 2007 to the end of 2008. The expectation is that a forced CEO change in the financial crisis amplifies the negative effect or weakens the positive effect on stock returns due to a significant drop in confidence on the stock market.
A variable for outside succession announced together with the forced CEO change is also included. An outside successor could change the company’s strategies and restore better firm performance, suggesting a larger positive effect on abnormal stock returns.
6 Other variables that could influence the stock market reaction to the announcements of forced changes are also examined. Firm performance preceding the forced CEO change is added as a variable. Moreover, this research controls for firm size. Finally, a leverage ratio variable, measuring the amount of debt used prior to the announcement, is included.
Thus, the main research question is: What is the effect of announcing forced CEO changes on stock returns of firms in the United Sates during the global crisis period?
To answer the research question, the following sub-questions concerning the influence of the variables on the announcement effect are addressed: In what way does announcing forced CEO resignation during the financial crisis influence the stock market reaction? How does announcing outside succession together with forced CEO resignation affect the
abnormal stock returns? In what way do firm performance, firm size and leverage ratio influence the abnormal stock returns following announcements forced CEO changes?
This research is also relevant due to the more recent period that is used. According to Huson et. Al (2001), the overall frequency of forced CEO departures increased significantly and therefore the relation between announcements of forced CEO leave and stock price behaviour might have changed. A recent study will show if the previously found results differ from those found in this research. The sample used in this study consists of 42 firms listed on the S&P 500 making a total of 47 forced CEO changes from January 2007 to April 2013.
In the following section the relevant literature and the hypotheses are discussed. In section 3, the sample and data selection process are described. Section 4 illustrates the methodology used in this research. The results and their analysis are presented in section 5. Finally, the discussion and conclusion are given in section 6.
2 Literature review
This chapter discusses relevant literature on the topic of CEO and top management changes. The previously found results of (forced) CEO changes are given. The variables employed in the literature and their influence on the stock market reaction to forced CEO changes are explained. By means of these explanations, the hypotheses that will be tested are drawn up.
2.1 Expected share and firm performance surrounding top management changes There are expectations in the literature about the performance before and after announcing a top management change in general. A change in general implies not only forced changes.
7 2.1.1 Expected performance preceding top management changes in general
There should be a negative association between the probability of top management changes and share prices if internal control is effective and share prices reflect the performance of a top-level manager, according to Warner et al. (1988). They also provide evidence in their research for this inverse relation. This implies that CEO departures and other top-level management changes should occur more frequently following poor share performance.
2.1.2 Expected performance following top management changes in general
When board actions regarding top management turnover are aligned with the interests of shareholders, stockholder wealth should not decline after the changes are announced (Furtado and Rozeff, 1987). According to Jensen and Ruback (1983), it is hard to find board actions related to corporate control that harm stockholders, meaning that board decisions are usually in the interests of shareholders.
Denis and Denis (1995) agree with Furtado and Rozeff and mention that when internal control is effective there should also be improvements in firm and share performance
following management changes. This is because the board can now appoint a better performing , more suitable manager than the manager that left the firm.
However, it is not certain that the board always makes the right decision regarding the CEO change and the CEO succession. The board might replace the CEO of a firm that is suffering from poor performance even if the executive is not responsible for the performance (Denis and Denis, 1995). In addition, there is no evidence that the board always finds a suitable and superior successor for the CEO that will be replaced (Denis and Denis, 1995).
2.1.3 Real effect and information effect of announcing top management changes in general There is a real effect of announcing a management change, which should be positive if the change is in the interest of shareholders (Bonnier and Bruner 1989). A possible effect on stock returns caused by the announcement of the change is measured by abnormal returns.
However, Bonnier and Bruner (1989) also mention the information effect, which entails that the effect on shareholder value could be negative if the announcement of the change implies that the company was performing worse than the market had realized.
The combination of the positive real effect and negative information effect results in a unclear expectation of the effect of top management changes on stock returns, since there will only be a significant effect when one effects is larger than the other (Warner et al., 1988).
8 Therefore, Bonnier and Bruner (1989) try to isolate the real effect of announcing a CEO change by only examining firms that were suffering from poor performance preceding the change, therefore expecting a positive shareholder reaction. They find that poorly
performing firms that announced top management changes experience a large positive effect on the stock returns, measured by an average abnormal return of 2.479%. Thus, these results provide evidence for the real effects and show that the information effect can be reduced.
2.2 Forced CEO changes
This research, however, focuses on the effect on stock returns of announcements of CEO resignations that are forced. Only the announcement effect of the forced change on stock returns is investigated. The shareholder reaction to the moment of the actual change is not included. The research is limited to forced CEO changes to examine whether the
announcement of a forced resignation leads to significant nonzero abnormal stock returns. In this section, the performance of a firm preceding and following an announcement of forced CEO resignation is illustrated shortly. Next, the criteria used in the literature for the
identification of forced CEO change announcements is clarified.
2.2.1 Firm performance and share performance preceding forced CEO changes
Several papers explain why firms would force their CEOs to resign. Warner et al. (1988), state that termination of a CEO or a top manager is usually a response to poor
management performance. They provide evidence by showing that firms that announce forced resignation have significant negative abnormal stock returns before the management change, meaning that a CEO change in most cases is preceded by declining stock prices.
In the sample of Denis and Denis (1995) 20 of the 73 forced CEO changes are preceded by poor firm performance. Other reasons for forced CEO resignation stated are pressure from blockholders, illegal conduct and unsuccessful takeover offers for the firm. The main results of Denis and Denis (1995) indicate that forced resignations of top managers often occur after the company suffered from negative stock returns and operating performance deterioration measured by changes in operating income. Thus, a forced resignation of a CEO is frequently preceded by a period of poor management.
2.2.2 Share price performance following forced CEO changes
The effect on the stock returns of forced CEO resignation found in the literature is usually positive. Furtado and Rozeff (1987) studied, for example, the effect of dismissals of
9 key executives and found that when dismissals occur, shareholder wealth increases, implying a significant positive effect on the firm’s stock returns.
In addition, larger positive abnormal returns were found by Denis and Denis (1995) for forced CEO resignations than for the normal changes. They find an abnormal return of 2.5% for the forced CEO changes. Finally, positive effects on stock returns, indicated by abnormal returns of 1.64%, were also found following announcements of forced CEO resignation by Borokhovich et al. (1996) in specific situations concerning the origin of the successor, which will be explained below.
The positive effects found around the announcement date of the forced resignation of CEOs and top management point to the main hypothesis of this research:
Hypothesis : An announcement of a forced resignation of a CEO will have a positive effect on the stock returns, creating shareholder value (1)
2.2.3 How are forced CEO changes identified in literature?
There is a problem concerning the identification of forced CEO changes, since few management changes are described as forced in press announcements (Warner et al., 1988). Firms would rather not disclose the information that a CEO was actually fired due to the negative information effect it causes for the firm (Bonnier and Bruner, 1989). Furtado and Rozeff (1987) show that dismissals of executives are infrequent and that firms prefer to engage in alternative methods such as a transfer, resignation or retirement.
It is important to find a way to identify distinguishing characteristics of forced CEO resignations, so that CEO changes that were forced can be classified as forced, even though they are not explicitly stated as such (Denis and Denis, 1995).
Several key papers employ criteria for identifying forced CEO changes. Denis and Denis (1995) compare management changes that were clearly forced to changes for which the reason was retirement and voluntarily resignation. They state that forced resignations will more frequently involve external appointments, the departing top manager leaving the firm and less often top level managers of normal retirement age.
Warner et al. (1988) conducted earlier a similar approach and identify forced CEO departures by examining CEO change-related Wall Street Journal articles. They inspect ages of the CEO changes indicated as retired to identify whether they really retired or that they were forced. Finally, Warner et al. (1988) also exclude CEOs that were leaving for another position and they exclude CEOs that were deceased had to quit due to health problems.
10 2.3 Previously found effects of announcements of top management changes
The effects of top management changes in general on stock returns shown in previous event studies are comparable in most cases. For instance, Furtado and Rozeff (1987), Bonnier and Bruner (1989), Weisbach (1988), Denis and Denis (1995) and Borokhovich et al. (1996) show positive announcement effects of CEO and top management changes on stock returns.
However, Reinganum (1985) and Warner et al. (1988) do not find significant excess returns when studying the effect of top management changes in general and Beatty and Zajac (1987) show a negative stock market reaction to CEO changes. The overall effect of the changes is measured by the cumulative average abnormal return (CAAR, table 1).
Table 1: Overview of the country, period, sample size, methodology, variables and results of previous studies..
The CAAR depicts the cumulative average abnormal return, which shows the overall effect on stock returns
2.4 Controlling for the organizational context of forced CEO changes
However, instead of studying the effect of changes in general, papers usually discuss and include important variables in their research that could explain cross-sectional variation in the abnormal stock returns surrounding the announcement (table 1). Reinganum (1985)
Paper Country Period CEO or other
(top management) Number of Observations Methods/ Models Forced (Yes/-) (control) variables Event Window Overall Outcome event study(CAAR) Reinganum (1985) US 1978-1979 Top management 353 Event study (AAR & CAAR) - Origin of successor (insider vs outsider) & firm
size -2;2 No significant nonzero abnormal returns Furtado & Rozeff (1987) US 1975-1982 Top Management 323 (62 forced)
Event study Yes Origin of successor & firm size -60;60 Positive effect: 0.95% Beatty & Zajac US 1979-1980 CEO 209 Event study/ cross section - Expected/ unexpected -10;10 Negative effect (No CAAR) Warner et al. (1988) US 1962-1978
CEO & top management 269 (64 forced) Logit models/ event study/cross section Yes Firm performance /forced & origin
of successor -1;0 No significant nonzero abnormal returns Weisbach (1988) US 1977-1980 CEO 286 Logit models/ event study/ cross section - Origin of successor -1;1 Positive effect (No CAAR) Bonnier & Bruner (1989) US 1969-1983 Top management and CEO 87 (distressed firms) Event study/ cross section Yes Origin of successor/ firm
size & title
-50;50 Positive effect: 2.479% Denis & Denis (1995) US 1985-1988 Top management
107 (forced) Event study Yes Performance/ forced -1;0 Positive effect: 0.63% Borokhovich et al. (1996) US 1970-1988 CEO 696 Probit model/ event study Yes Origin of successor/ forced -1;0 Positive effect (NO CAAR) Charitou et al (2010) US 1993-2005
CEO 158 (outside) Event study - Origin of
successor (Outsider)
-2;1 Positive effect: 2.42%
11 emphasizes this by stating that the organizational context of a the change is important in the analysis of the effect of the departure of an executive on stock returns. The variables used in this research are described next and for each variable a hypothesis is given (figure 1).
Figure 1: Conceptual model of the hypotheses, with arrows indicating what is tested by each hypothesis and + or – for the corresponding expected effect (positive or negative).The expected effects are explained below.
+
+ -
+
+
+
-2.4.1 Origin of the successor
The announcement of a forced CEO change is frequently paired with the announcement of the successor, which is either an insider that was promoted or someone hired from outside the firm.
Appointing an outsider could have a negative effect on firm performance and the abnormal stock returns, according to Bonnier and Bruner (1989). Inside appointments are not as disruptive as outside appointments and the board is familiar with the insider, making it usually a good appointment. Moreover, the insiders have more firm-specific knowledge and experience, and appointing an outsider could imply that the firm’s management is performing so poorly that an outsider has to be brought in, which suggests a negative effect on abnormal stock returns, due to a negative signal to shareholders (Bonnier and Bruner, 1987).
However, the alternative expectation is that a successor hired from outside the firm may be the person that could change the companies strategies and lead the company back to better performance. Bonnier and Bruner (1987) state that this should lead to a positive effect on the company’s abnormal stock returns.
The previously found effects of a CEO change on stock returns are usually positive when the successor is an outsider. For example, Borokhovich et al. (1996) state that there are more shareholder gains when an outsider is appointed CEO, especially when the departing
Shareholder Value Forced CEO changes (1)
Outsider (2) Poor Firm Performance (2) Firm Size (4)
Financial Crisis (5) Leverage Ratio (5)
12 CEO was forced to leave. Charitou et al. (2010) also conclude that there are positive abnormal returns around the announcement of an outside appointment of a CEO. Bonnier and Bruner (1989) also find a larger positive stockholder reaction for an outside succession of the CEO.
There is one exception to the positive effects of placing outsider, as the appointing an insider to the post of top manager leads to a larger positive reaction of the stock market, according to evidence from Furtado and Rozeff (1987).
However, due to the convincing amount of CEO change-related studies that suggest positive effects of outside CEO succession, the following hypothesis can be inferred:
Hypothesis: Announcing an outside succession of the CEO together with the forced resignation affects the abnormal stock returns positively (2)
2.4.2 How is the origin of the successor measured in the literature?
External appointments involve appointing an outsider, which is a person that was not
employed by the firm until recently (Denis and Denis, 1995). Internal appointments concern the appointing of an insider, who was already employed by the firm. In Bonnier and Bruner (1989), the origin the successor is measured by a dummy variable with a value of 1 in cases of external appointments , and 0 otherwise.
2.4.3 Firm performance
Multiple papers suggest that firm performance preceding a top management change is important for the analysis of the announcement effect of CEO changes. A performance
variable could isolate the abovementioned real effect of a forced CEO change on stock returns after poor performance and help to explain any cross-sectional variation in the stock returns. The isolated effect of forced changes on stock returns preceded by poor performance should be positive, because the change is in the interest of shareholders (Bonnier and Bruner, 1989). Firms suffering from poor performance usually experience positive abnormal stock returns surrounding a top management change. Friedman and Singh (1989) show that
shareholders react positively to forced CEO changes when performance is low preceding the announcement. In addition, the main result of Weisbach (1988) is that outside boards add the most firm value by replacing their bad CEOs when this change was preceded by poor firm performance. Finally, Denis and Denis (1995) find that forced resignations are often preceded by large declines in firm performance and losses in shareholder wealth and result in
13 Due to the positive previously found effects of CEO change announcements that were preceded by poor performance, the third hypothesis is introduced:
Hypothesis: Poor firm performance preceding the announcement of a forced CEO change influences the stock market reaction positively (3)
2.4.4 How is firm performance measured in the literature?
Firm performance is measured in different ways by existing literature. Denis and Denis (1995) use changes in the operating income before depreciation divided by the book value of the total assets (OIBD/TA) as the measure for firm performance. Friedman and Singh (1989) measure firm performance using the return on equity of the firm in the year before the change.
2.4.5 Firm size
Including firm size in the cross-sectional analysis helps to show the relation of firm size to the effect on stock returns caused by announcing forced CEO changes.
CEOs are considered to be the key strategic decision-makers, especially in large firms (Beatty and Zajac, 1987). In addition, Reinganum (1985) states that small firms have less complex control structures than large firms. Therefore, the effect of announcing a forced CEO change should be more meaningful in small companies than in large companies. If this is the case and if there are positive effects of management changes on the stock market, the
abnormal returns should be inversely related to firm size (Bonnier and Bruner, 1988). Previous results imply that the abnormal stock returns surrounding the announcement of a CEO change are usually inversely related to firm size. For instance, Reinganum (1985) only finds a significant improvement in stockholder wealth when the firm is small and the top management successor is an outsider. In addition, Furtado and Rozeff (1987), find that small firms show more positive effects of announcing forced changes on stock returns.
In contrast, Bonnier and Bruner (1989) find that size effects are positive as interaction effect with title and size, meaning that there are is a positive influence on excess stock returns when a large distressed company announces a change in their CEO.
However, due to the expectations and the majority of evidence for the influence of firm size on abnormal stock returns surrounding CEO change announcements, the fourth hypothesis is:
Hypothesis: The abnormal stock returns surrounding forced CEO changes are inversely related to firm size (4)
14 2.4.6 How is firm size measured in the literature?
Firm size is measured by Warner et al. (1998) as the sum of the market value of equity and the book value of long-term debt the natural logarithm of their measure of firm size is used in the regressions. Reinganum (1995) measures firm size by the market capitalization, which is calculated by multiplying the share price times by the number of shares outstanding.
2.4.7 Leverage ratio
Including a variable concerning the leverage ratio will provide insight on how the use of leverage influences the effect of a forced CEO resignation announcement on stock returns.
CEOs that implement risky strategies, tend to use more leverage. Firms can increase their operations and therefore their earnings by financing with debt. However, if the cost of debt becomes larger than the returns on activities financed with the debt, the company becomes distressed (Bonnier and Bruner, 1989). When the CEO of a firm that implemented risky strategies is forced to leave because of the these strategies, the stock market should respond positively to the announcement. This is because they notice that the change of the executive is in shareholders’ interest, since a new CEO is probably less likely to implement strategies financed with debt.
Distressed firms respond positively to CEO changes. This is shown by Bonnier and Bruner (1988), who show a large overall positive overall effect on stock returns of top management changes in firms that are distressed. Thus, more use of debt, measured by a higher leverage ratio, in the period before the announcement of a forced CEO change should have a positive on abnormal stock returns. This introduces the following hypothesis:
Hypothesis: A higher leverage ratio of firms announcing forced CEO changes affects the abnormal stock returns positively. (5)
2.4.8 How is the leverage ratio measured in the literature
There is no evidence on the use of the leverage ratio in the literature. However, Bonnier and Bruner (1989) use a sample consisting only of distressed firms.
2.5 Forced CEO change announcements during the financial crisis
The unique feature of this research is that the effect of announcing forced CEO
changes of US firms during the financial crisis period of September 2007 to the end of 2008 is studied. No research has been done on the influence of a financial crisis on the abnormal stock
15 returns caused by forced CEO leave announcements. The expectation is that the financial crisis, measured by a dummy variable, affects the stock market reaction to forced CEO change announcements negatively, which is illustrated below.
2.5.1 Financial crisis variable and hypothesis
The financial crisis had severe implications for the financial markets. Batram and Bodnar (2009) state that the world stock index was down 45.4% in February 2009, compared to the end of 2007. Chaudhury (2013) finds that the mean daily returns of S&P 500 firms worsened significantly during the financial crisis in the United States (Chaudhury, 2013). Thus, share prices and stock returns were low during the financial crisis.
It is not self-evident that the financial crisis will influence or have a different effect on stock returns than the previously found positive effects of announcing a forced CEO change. Forced CEO changes are often preceded by declining share prices and lower stock returns, implying that a crisis should not be much different from the situations in comparable studies.
However, the effects during the financial crisis could differ from earlier found
evidence on forced CEO changes, due to the stock market being evidently more pessimistic in a financial crisis than in relatively normal financial situations. The stock market might not react positively to a forced CEO resignation, because they do not believe the persisting poor performance can be turned around by a new CEO during a financial crisis, regardless of the abilities of the successor. A forced CEO change would therefore not be in the interest of shareholders, since they do not expect it will improve the firms performance or create shareholder value.
This would imply that only the information effect, explained by Bonnier and Bruner (1989), might be present, because the market realizes that the firm is performing even worse than they had already anticipated in a time of financial crisis. Therefore, the expectation is that the financial crisis influences the announcement effect of forced CEO resignation on stock returns negatively. This leads to the final hypothesis:
Hypothesis: The financial crisis affects the stock market reaction to the announcements of forced CEO changes negatively (6)
2.6 Sample size and methodology of event study used in the literature
Practically all the related key papers research the effect of the announcement and not the actual change of CEO and other top management changes on stock returns of US firms. Reinganum (1985), Warner et al. (1988), Weisbach (1988), Furtado and Rozeff (1987) and
16 Bonier and Bruner (1989) examine the effect of forced resignation around the actual
announcement date, since stock prices react immediately to announcements of any kind. The sample sizes of forced CEO resignations are not very large. For instance, the samples in Furtado & Rozeff (1987) and Warner et al. (1988) are 62 and 64, respectively.
The tests of the effect on stock returns of CEO and other top management change announcements are often carried out using an event study of daily abnormal returns (Furtado and Rozeff, 1987). Bonnier and Bruner (1989) estimate the expected returns with the market model using 100 trading days before the announcement of the CEO change, from day -200 to day -101. Warner et al. (1988) use a period of 150 days for their estimation and drop a company from the sample when there are fewer than 30 days available for estimation.
In existing literature short event windows are used for the announcements of a CEO change. For example, Weisbach (1988) uses an event window comprising days -1, day 0 and day 1 for calculating the abnormal returns caused by announcing the forced CEO change.
Once the abnormal returns have been calculated, the average of the sample for each day in the event window is computed and in addition, they are combined and averaged over the entire sample. This results in the average abnormal returns (AAR) and the cumulative average abnormal returns (CAAR), respectively. As in Denis and Denis (1995), the AARs and CAAR are tested for their significance, to determine the overall effect of announcing forced CEO changes on stock returns.
3 Sample and data
The sample and the data collection are illustrated in this section. First, the sample of forced CEO changes used in this study is described and it is explained how the sample was obtained. Next, an explanation of how the crucial data for the event study was collected is given. 3.1 Sample of forced CEO change announcements
The sample used in the research is limited to forced CEO changes in the United States. The sample consists of 42 firms making a total of 47 forced CEO changes.
A list in excel format of CEO changes of S&P 500 companies in the global crisis period was retrieved using the Compustat database of WRDS. Consulting this database resulted in 288 CEO changes from January 2007 to the end of April, 2013.
The nature of the CEO changes was determined to create a sample of announcements of forced CEO changes during the global crisis. The forced changes used in this research are
17 identified by reading CEO change related news articles from reliable news sources, such as the Wall Street Journal and the New York Times. However, it is sometimes not immediately clear in a news article whether a CEO was forced to resign or that that the executive left for another reason. For these cases, criteria similar to the literature will be used.
In short, these criteria entail that forced resignations, even though not explicitly stated as forced, frequently involve external appointments, a longer period of bad results, the departing CEO leaving the firm and less often CEOs of normal retirement age. CEO changes that were clearly announced as forced were compared to changes that might be forced, even though the reason indicated was, for example, retirement or resignation. Comparing changes and applying the several criteria helped to determine if a CEO change was forced.
In addition, forced CEO resignations caused by a bankruptcy are excluded. This is because the stock price data are not available for the specific event windows, since the stock becomes worthless after a bankruptcy. CEOs forced to leave their position in the firm due to takeovers and mergers were also excluded from the sample (Furtado and Rozeff, 1987). Finally, CEOs that were leaving for a position in another company were excluded. These criteria resulted in the exclusion of six announcements of forced CEO changes.
Applying the criteria resulted in 42 firms making a total of 47 forced CEO changes over the entire sample period. An overview of the forced CEOs, their companies and the corresponding announcement dates is provided in the appendix (A.1).
3.2 Data selection for the event study
An important part of carrying out the event study is determining the announcement date of each forced of these CEO changes, because the date is crucial for measuring the effect on stock prices and returns. The stock market responds to news facts virtually immediately and therefore the date of the announcement should be accurately estimated. The exact announcement date of each forced CEO resignation is identified by comparing the announcement dates mentioned in several reliable news sources, such as the Wall Street Journal.
After the announcement dates are identified, stock price data for each firm are needed in order to calculate the abnormal returns. The historical stock prices of every firm that announced a forced CEO departure are retrieved via DataStream, These daily stock prices are automatically adjusted for dividend payments and stock splits. The price-index of the S&P 500 for the entire crisis period is also retrieved using DataStream.
18
4 Methodology and model set-up
To determine the effect of forced CEO changes on stock returns, two methods will be used. The first method is an event study, which will, using the market model, show the possible abnormal stock returns surrounding an announcement of a forced CEO resignation. First, the sample period is illustrated and thereafter the methodology for the event study is described.
Controlling for the organizational context is important. Therefore, the relation between the variables and the effect on stock returns is researched using the second method, which involves cross-section regressions. The methodology for the regressions is described below.
4.1 Sample period
A specific sample period is used. Chaudhury (2013) uses information about notable events to identify September of 2007 as the beginning of the financial crisis. In the beginning of 2007, firms already suffered from declining stock returns, showing signs of an incipient financial crisis. Therefore, the first months of 2007 are also included in the sample period.
Because the global crisis had a long-lasting effect on share prices and firm performance, the total sample period will cover several years following the start of the
financial crisis in the United States. Thus, the period from January 2007 to April 2013 is used as the sample period to examine the effect of announcements of forced CEO resignations on US firms’ stock returns during the global crisis.
4.2 Event study methodology
Structured as an event study of daily excess returns, the abnormal stock returns for each firm surrounding the announcements of forced CEO resignations are calculated and tested for significance. This process is illustrated shorty below, together with the estimation period and the event window used for the event study. The full methodology of the event study, including the formulas used in this research, can be found in the appendix (A.2).
The expected stock returns for each firm are estimated for the event period using the market model (Fama, 1976). This model is used in multiple key papers that study the announcement effects of CEO and other top management changes.
Since the sample period concerns the crisis, the long estimation periods used in the literature for the expected stock returns of some CEO changes early in the sample period would partly cover the period before the crisis. This would lead to an unrepresentative
19 beta might not be representative for the risk and the stock price performance of the firm around the CEO departure. Therefore, a shorter estimation period of 50 trading days from day -110 through day-61 is used, where day 0 is the date of announcement of the CEO change. A forced CEO change is excluded from the sample if fewer than 30 days of stock price data are available for estimating the expected stock returns, to ensure a correct estimation.
Copying Weisbach (1988), this research uses a main event window comprising day -1, day 0 and day 1 for calculating the abnormal returns caused by the forced CEO change, where day 0 is the exact date of the announcement of the forced CEO departure. This event period covers the major stock market reaction to the announcements of the forced CEO departures, since the market reacts almost instantly to news facts.
The abnormal returns for each day in the event window are calculated by subtracting the expected returns from the actual return. Taking the average of the abnormal returns for the entire sample results in the average abnormal return(AAR) for each day in the event period. The overall effect for each firm is calculated by combining the abnormal returns for each firm, presenting the cumulative abnormal return (CAR ). The average of the CAR s is the cumulative average abnormal return (CAAR).
The CAAR is a measure of the abnormal stock returns over the entire event window (Reinganum, 1985). Finding a significant CAAR implies that there is in fact an announcement effect of forced CEO departures on the stock returns of US firms.
4.3 Regression
Controlling for the organizational context could is also important. Therefore, the influence of multiple variables on the effects on stock returns of announcing forced CEO changes are examined in a cross-section analysis. This section explains how the variables described in the literature review are included and measured in this research. The following variables are included: the financial crisis, the origin of the successor, firm performance preceding the CEO change, firm size and the leverage ratio.
4.3.1 Regression model
The relationship between the variables and the announcement effect of forced CEO leave will be tested using multiple linear regressions. The total regression model is depicted by:
= + ∗ + ∗ + ∗ +
20 The regression is run on the cumulative abnormal return, depicted by the . The represents the total effect on stock returns for each firm separately.
4.3.2 How are the variables included in the regression measured?
The origin of the successor is measured using a dummy variable, with a value of 1 in cases of external appointments , and 0 otherwise. The origin of the successor is retrieved by means of the articles used to identify the announcements of forced CEO changes. In these articles, it was also frequently mentioned whether there was an inside or outside succession. Inside interim CEOs that were placed after the CEO left were seen as inside replacements.
Firm performance is measured similarly as in Denis and Denis (1995). Changes in firm performance are measured as the change in operating income before depreciation divided by the book value of total assets (OIBD/TA) for the years -3 to -1, where year 0 is the fiscal year of the announcement. The change in firm performance is the difference in the OIBD/TA of year -1 and year -3. A negative change implies a worsened firm performance.
Firm size is measured as the book value of total assets. This is more or less comparable to Warner et al. (1988), who measure firm size as the sum of market value of equity and the book value of debt. In Stata, the logarithm of assets are used for this variable.
The leverage ratio is measured as the book value of total debt divided by the book value of total equity of a firm. The leverage ratio for the year prior to the CEO change is used.
The data for the firm performance, firm size and leverage ratio variables are collected for the entire sample using the Compustat database of WRDS.
Finally, the financial crisis also functions as a dummy variable, with a value of 1 if the announced CEO change occurred during the financial crisis, and 0 otherwise. Chaudhury (2013) identifies September of 2007 as the beginning of the financial crisis. Because in the end of 2008 governments and monetary authorities started to intervene (Chaudhury, 2013), the dummy ends in 2008 to study the stock price behaviour surrounding announcement of forced CEO departures before the interventions had their impact on the financial markets. Thus, the financial crisis period from September 01, 2007 to December 31, 2008 is used.
5 Results
A descriptive overview of the statistics of independent variables used in the regressions will be given and illustrated first. Next, the results of the event study are presented and explained for several event windows. This shows the effect of announcing a forced CEO resignation on
21 stock returns. Thereafter, the results of the regressions of on the variables are given. These results show which factors can explain cross-sectional variation in the abnormal stock returns. Finally, the testing of the results leads to possible rejections of the hypotheses.
5.1 Summary statistics of the independent variables
A summary of the independent variables included in the regressions are given in table 2. The statistics for each variable will be explained shortly below.
Table 2 shows that 28 percent of the forced CEO change announcement occurred in the financial crisis period. In only 13 percent of the changes in the sample, meaning in 6 cases, the successor announced was an outsider. Due to this small sample for outside succession, its influence on abnormal stock returns surrounding CEO changes could be difficult to interpret and might be less often significant.
The change in firm performance, measured by the changes in OIBD/TA over the years -3 to -1, is on average -0.0053. There is a relatively large variation in the changes of firm performance. Firm size is depicted by total assets and these are on average 25.2 million. Total assets for the smallest firm are 4,4 million and 1,9 billion for the largest firm. The large standard deviation confirms the image by implying a large variation in the firm sizes. In the regressions, firm size is used as the logarithm of assets. The same holds for leverage ratio; this ratio is on average 5.32 with a minimum of 0.176 and a maximum of 32. Thus, there is a large variation in the relative amount of debt used prior to announcements of forced CEO changes. Table 2: Summary statistics of variables financial crisis, outsider, firm performance, firm size and leverage ratio
Variable Mean
Standard
deviation Minimum Maximum
Financial Crisis Outsider .277 .128 .4522 .337 0 0 1 1 Firm performance Firm Size Leverage ratio -0.005 251537 5.32 .036 494401.5 7.161 -0.130 4380.4 .176 .081 1884318 32.525
The mean, standard deviation, minimum and maximum are given for the values of each variable using the entire sample. Financial crisis is a dummy variable, with a value of 1 if the announcement was during the financial crisis. Outsider is also a dummy variable, that represents the origin of the successor, with a value of 1 in case of an outsider. Firm performance is the change in the operating income before depreciation divided by the book value of totals assets. Firm size is measured by the total assets and the logarithm is used in the regressions. Finally, leverage ratio is the total equity divided by the total debt, using the book values.
22 5.2 Event study results
The announcement effect on stock returns is shown using the results of the event study. First, the cumulative average abnormal return (CAAR) found for the main event period of days -1, 0 and 1 are given and illustrated. The abnormal returns are calculated using the market model with an estimation period of 50 days, ending on day -61. Thereafter, other event windows used are introduced and the interpretation of the results for these windows is given. 5.2.1 Cumulative abnormal returns for main event window (-1;+1)
The cumulative average abnormal returns of the main window are presented below (table 3). First, the overall CAAR of forced resignation is given. In addition, the CAARs for
announcements of forced CEO changes in the financial crisis period and outside succession are presented. No significant CAARs are found using the -1;+1 window for the total sample, the announcements in the financial crisis and announcements involving outsider succession.
The CAAR of the total sample of forced resignations is -1.99%, which suggests a negative overall stock market reaction to the announcements of forced CEO changes. This contrasts with the evidence found in existing literature. However, due to the insignificance represented by a t-value of -0.757, no conclusions about a negative effect can be drawn.
The same holds for the forced CEO change announcements during the financial crisis. A negative effect on stock returns is indicated, depicted by a CAAR of -7.40%, but this result is insignificant, due to the t-value of -0.802. According to this event window, a negative effect on stock returns of announcing forced changes during the financial crisis cannot be inferred.
Finally, no conclusions on the effect of announcing a forced CEO change together with an outsider succession can be made, because of the insignificant CAAR of -13.90% with a t-value of -0.684. Although the CAAR is insignificant, the negative effect indicated is interesting, since this is different than the literature and the hypothesis suggests. Charitou et al (2010) find, for example, a CAAR of 2.26% for announcements of outside appointments.
5.2.2 Cumulative average abnormal returns for different event windows
The event window -1;+1 did not result in significant effects. Different event windows, used as robustness check, could possibly show different and significant effects. This is because of market anticipation or because the plans of the forced change sometimes leak before the actual announcement date. The announcement effect might also be delayed in some cases. Therefore, an asymmetric larger event window of day -20 to day +3 is used. Since the event window starts 20 days before the announcement, practically all of the anticipation of the
23 stock market is accounted for. Besides, because of the 3 days after the announcement of the forced resignation, the effect measured using this window will also cover the delayed part of the stock market reaction. The amount of event days following the announcement is
somewhat less than before the event date, to ensure that the stock market reaction to the actual execution of the CEO resignation is not included. In addition, the cumulative average
abnormal returns are also tested using the event windows -1;0 and 0;+1, as robustness checks. Several papers also use a longer event window to detect the actual effect of the
announcements of top management changes on stock returns. Beatty and Zajac (1987) use, for example, a longer period of days -10 until day 10.
The average abnormal returns for day -20 until +3 for the total sample can be found in the appendix (A.3). The average abnormal returns were predominantly negative for the 20 days preceding the announcement and significant for days -15, -7, -3 and-2. These results imply that a longer event period before the announcement is important for showing the actual announcement effect and that the market seems to anticipate to the announcement.
The event study using the event period-20;+3 did result in significant effects of forced resignation on stock returns for the total sample and for the financial crisis period. A negative insignificant effect was found for the outside succession CEO change announcements. The results for this window are presented in table 3 and illustrated below.
The event windows of -1;+0 and 0;+1 resulted in comparable and insignificant outcomes as the -1;+1 window, expect for the slightly significant negative CAAR for the financial crisis, which provides additional evidence for the negative announcement effect of forced changes during the financial crisis. These CAARs are shown in the appendix (A.4). Table 3: CAARs of total forced CEO changes, forced changes during the financial crisis and forced changes with outside successionof event windows (-1;+1) and (-20;+3)
N CAAR (-1;+1) CAAR (-20;+3) Total Financial crisis Outsider 47 13 6 -1.99% (-0.76) -7.40% (-0.80) -13.90% (-0.68) -5.66% (-2.15)** -16.44% (-2.46)** -14.92% (-1.01)
The N represents the amount of observations for the total sample, the financial crisis and outsider succession forced CEO change announcements. The t-values are shown in parentheses below the corresponding CAAR and denote the outcome for the cross-sectional t-test, which tests the significance of the CAAR. *,**, and *** denote statistical significance of the one-tailed tests at the 0.10, 0.05, and 0.01 significance levels, respectively. The degrees of freedom were taken into account in the evaluation of significance. Abnormal returns are calculated using the market model with an estimation period of 50 days ending on day -61.
24 Table 3 shows that the total sample led to a CAAR of -5.66%. This CAAR has a corresponding t-value of -2.514, meaning that it is significant at the 5% level. An
announcement of a forced CEO resignation leads to a significant negative effect on stock returns, according the -20;+3 event window, therefore reducing shareholder value. This large negative effect is in contrast to previous studies, which suggest a positive effect. Denis and Denis (1995) found, for example, a positive CAAR of 2.5% for forced CEO resignations.
In the case of a forced CEO resignation announced during the financial crisis period, a large CAAR of -16.44% is measured. This negative effect is significant at the 5% level with a t-value of -2.462. Measuring the announcement effect of forced CEO resignation during the financial crisis period using the -20;+3 window indicates a negative effect on stock returns. The relatively large CAARs could be explained by the increased share price volatility during the financial crisis (Ranjeeni, 2014).
The effect of the announcements of outside succession in combination with a forced CEO change is again negative and insignificant using the longer event window. This is depicted by a CAAR of -14.92%, with a t-value of -1.01.
5.3 Regression results
In this section, the cross-sectional variation of the stock returns surrounding announcement dates of forced CEO changes are examined. The results are produced by regressing the cumulative abnormal returns ( ) of each firm on the variables. The dependent variable used for the regressions are the s calculated for the -20;+3 and -1;+1 event windows. Table 5 shows the results of these multiple linear regressions that were conducted using Stata.
The outcomes for the -20;+3 window are the most important, because this window shows the most significant results in both the event study and the regressions. The outcomes for this window will be decisive for the eventual evaluation of the hypotheses. The outcomes of this model and the event study results leads to possible rejections of the hypotheses.
The financial crisis dummy variable has a negative and significant coefficient of -0.145, according to the regressions using the -20;+3 event window. After removing several non-significant variables, the coefficient remains negative and significant at the 5% level. A negative influence of the financial crisis on the abnormal stock returns caused by
announcements of forced CEO changes can therefore be inferred. This is in accordance with the expected influence of the sixth and last hypothesis. A comparable but smaller and insignificant negative coefficient is shown for the -1;+1 event window.
25 Table 4: Outcomes of the multiple linear regressions on the s of event window -20;+3 and -1;+1
Event window -20;+3 Event window -1;+1 Expected direction of coefficient Significant coefficients in literature (1) (2) (1) (2) - Financial Crisis -0.145 (2.62)*** -0.147 (2.29)** -0.059 (-1.05) Negative Outsider -0.081 (-0.62) -0.167 (-0.86) Positive 0.056 Firm Performance -0.636 (-1.09) -0.900 (-1.39)* -0.358 (-1.12) Negative - 0.163 Firm Size -0.039 (-1.64)* -0.037 (-1.66)** -0.044 (-1.51)* -0.030 (-1.36)* Negative -0.008 Leverage Ratio 0.007 (1.47)* 0.005 (1.81)** 0.009 (1.21) Positive - N 47 0.255 47 0.228 47 0.226 47 0.097
The regressions were done using robust standard errors. The t-values are shown in parentheses. *,**, and *** denote statistical significance of the one-tailed tests at the 0.10, 0.05, and 0.01 significance levels, respectively. The coefficients for the variables Financial Crisis, Outsider, Firm Performance, Firm Size and Leverage Ratio are given. The expected direction of coefficients (positive or negative) is also given. Coefficients found in the literature are presented to compare the direction of the coefficient. The Outsider coefficient is from Bonnier and Bruner (1989) and the Firm Performance and Firm Size coefficients are from Friedman and Singh (1989).
The outsider dummy variable shows a small and non-significant negative coefficient of -0.081 for the event window -20;+3.This insignificant coefficient cannot prove that an outside succession announced together with the forced CEO change leads to a positive influence on the shareholder reaction, as was expected in the second hypothesis. A negative influence is suggested, which contrasts with evidence on this coefficient of previous studies. The -1;+1 regressions resulted also in comparable insignificant negative coefficients.
Firm performance shows an insignificant coefficient of 0.636, due to the tvalue of -1.09. The direction of the coefficient suggests a positive influence on the stock market reaction to announcements of forced CEO resignation preceded by poor or worsened firm performance, but the influence is not significant. Therefore, the expected positive influence of the third hypothesis cannot be proven, but the direction suggested is comparable to the
literature. Using the event window -1;+1, the total regression model shows a significant coefficient in the first model. However, this coefficient also becomes insignificant after the
26 non-significant variables are excluded. A significant positive influence of poor performance cannot be inferred.
Firm size has a negative and significant coefficient of -0.039. It remains statistically significant when there is not controlled for the non-significant variables. Therefore, the statistical evidence suggests that the abnormal returns surrounding the announcement of the forced CEO leave are inversely related to firm size. The coefficient is somewhat larger than in literature, but the negative direction indicated is the same as suggested by previous studies and the fourth hypothesis. The -1;+1 event window also shows negative and significant coefficients, therefore confirming the influence indicated.
Finally, the leverage ratio included in the full regression model is represented by a small, but positive and significant coefficient of 0.005, with a t-value of -1.64. It remains significant after the non-significant variables have been excluded. The positive coefficient implies that the use of more debt, measured by a higher leverage ratio, prior to announcing a forced CEO change affects the stock market reaction positively. This provides evidence for the expected positive influence indicated in the fifth hypothesis. The -1;+1 event window regressions result in a comparable but insignificant positive coefficient of the leverage ratio.
Table 5 summarizes the hypotheses and its corresponding expected effect. In addition the event study and regression outcomes are represented. This leads to the evaluation of the hypotheses, which is shown in the last column.
Table 5: Results and evaluation of hypotheses
Hypotheses ( )
Expected announcement effect or influence on effect:
Outcomes Conclusion: Reject or do no reject / ( =no influence/effect) CAAR (-20;+3) Coefficient (-20;+3)
Total Sample Positive -5.66%** - H rejected
Financial Crisis Negative -16.44%** -0.147** H Rejected/H not rejected
Outsider Positive -14.92 -0.081 H not rejected
Poor Firm Performance Positive - -0.636 H not rejected
Firm Size Negative - -0,037** H rejected/H not rejected
Leverage Ratio Positive - 0,005** H rejected/H not rejected
For each variable the corresponding hypothesis is given. This shows the expected effect on stock returns or its expected influence on the abnormal stock returns caused by the announcement of forced CEO changes. The outcomes of the event study and the regression are given, which leads to a possible rejection of the hypotheses. H is the hypothesis described in the literature and H is the null hypothesis that indicates no expected effect.
27 5.4 Robustness of regression results
The linearity, possible outliers, multicollinearity and independence of the variables need to be examined in order to check the robustness and the validity of the regression results.
In the appendix (A.5.1), scatterplots of the s of event window -20;+3 and the variables are given, except for the dummy variables. The scatterplots of the s and leverage ratio, the s and firm performance (changes in OIBD/TA) and s and firm size (Assets_log) do not show a clear departure from linearity. Moreover, adding a quadratic variable to the model does not produce significant coefficients or change the regression outcomes. Therefore, the linearity assumption of this model is not explicitly violated.
There seem to be two outliers in the data of the leverage ratio variable, which can be seen in the scatter plot on the far right. Removing these observations from the regression resulted does not change the directions of the coefficients (Appendix, table A4). The sizes of the coefficients are also more or less comparable, but the firm size and leverage ratio
coefficients are not significant in the regressions of event window -20;+3. However, it is important that the large leverage ratio observations remain in the model. This is because the values of the leverage ratio and especially the corresponding are not too far outside the usual range of the data. Moreover, the leverage ratio is not caused by a data entry error. Therefore, the observations with large leverage ratios observations remain included in the regressions, as in table 4.
The multicollinearity is measured using VIF-values. The 1/VIF represents the
tolerance of the relation between the variables. A 1/VIF value lower than 0.1 implies that the variable could be a linear combination of the other variables. The 1/VIF values (Appendix, table A5) range from 0.524 to 0.907, which indicates that there is no multicollinearity.
Finally, there is no interdependence of the variables, since the correlations, presented in the appendix (Table A6), range from -0.221 to 0.419. All correlation levels are below 0.5, which indicates that the variables are relatively independent.
28
6 Conclusion and discussion
6.1 Conclusion
This study researches the announcement effect of forced CEO resignation on stock returns. Previous studies usually find positive effects of announcing forced CEO changes on stock returns. However, a negative effect is possible if the announcement suggests that the firm is performing worse than the market had realized. Therefore, the expected effect is not clear.
In this research, a recent sample period is used that covers the years of the financial crisis and the longer lasting global crisis, to identify the actual announcement effect during this specific period. The research question is: What is the effect of announcing forced CEO changes on stock returns of firms in the United Sates during the global crisis period?
Most related studies are conducted during relatively normal times. This research deviates from existing literature by examining forced CEO changes during the financial crisis.
The research is done using S&P 500 firms that announced forced CEO resignations. 42 firms making a total of 47 forced CEO change announcements were used to investigate the effect on stock returns. An event study was carried out first for determining this effect. This was done for the total sample, for firms announcing a forced CEO change in the financial crisis and for firms announcing an outside succession together with the forced CEO change.
The event study resulted in significant and large cumulative average abnormal returns (CAAR) of -5.66% for the total sample and -16.44% for the forced changes during the financial crisis. The results are found using the longer and asymmetric event window -20;+3. These results indicate a negative effect of announcing forced CEO changes on shareholder value, especially during the financial crisis. This evidence differs from the usually positive announcement effect of forced changes found in previous studies as Denis and Denis (1995).
Next, cross-sectional regressions were run to identify possible factors that affect the stock market reaction to announcements of forced CEO changes. Again, a negative influence of the financial crisis on the announcement effect of forced CEO changes is indicated. Statistical evidence also suggests that firm size is inversely related to abnormal stock returns caused by the announcements, which is in accordance with the literature. Finally, the
regression show that a higher leverage ratio prior to the announcement influences the
announcement effect of forced CEO changes positively. Shareholders seem to react positively to the forced resignation announcement of a CEO that implement debt-financed strategies.
No significant influences on stock returns are found for the announcements of outside succession or for poor firm performance preceding the announcement of forced CEO leave.
29 The main results of this study indicate that firms announcing a forced CEO departure in financially unstable times reduce shareholder value, which is in contrast to the literature. This suggests that companies announcing forced CEO departures should not expect to create shareholder value in poor financial times combined with a drop in confidence on the stock market. However, this research suggests that smaller firms and firms with a high leverage ratio prior to the announcement should experience less negative announcement effects.
6.2 Limitations and future research
This research focuses solely on forced CEO changes, not CEO changes in general. The effect on stock returns of announcing CEO or other top management changes in general during the crisis is a topic that could produce new and unknown results. In addition, the outside
succession announcement effect is not examined extensively in this research, due to the few observations of announcements of outside successions together with the forced CEO changes in this sample. Analysing its influence during the crisis is a research possibility. Finally, the announcement effect of forced changes in other countries or continents than the United States are not researched in this paper. The announcement effect on stock returns during the crisis outside the United States is a possible research topic and could function as a comparison.
30
Reference List
Bartram. S. M., Bodnar, G. M. (2009). No place to hide: The global crisis in equity markets in 2008/2009. Journal of International Money and Finance, 28, 1246-1292.
Beatty, R.P., & Zajac, E.J. (1987). CEO change and firm performance in large corporations: Succession effects and manager effects. Strategic Management Journal, 8, 305-317. Bonnier, K.A., & Bruner, R.F. (1989). An analysis of stock-price reaction to management
change in distressed firms. Journal of Accounting and Economics, 11, 95-106. Borokhovich, K.A., Parrino, R. &Trapani, T. (1996). Outside directors and CEO selection.
Journal of Financial and Quantitative Analysis, 31, 337-355.
Charitou, M., Patis, A. & Vlittis, A. (2010). The market reaction to the appointment of an outside CEO: An empirical investigation. Journal of Economics and International Finance, 2(11), 272-277.
Chaudhury, M. (2013). How did the financial crisis affect the daily stock returns? Journal of Investing, 23(3), 65-84.
Denis, D.J, & Denis, K.D. (1995). Performance changes following top management Dismissals. The Journal of Finance, 50(4), 1029-1057.
Fama, E. (1973). Foundations of Finance: Portfolio Decisions and Securities Prices. New York, NY: Basic Books.
Furtado, E.P., & Rozeff, M.S. (1987). The wealth effects of company initiated management Changes. Journal of Financial Economics, 18, 147-160.
Friedman, S.D. & Singh, H. (1989). CEO succession and stockholder reaction: The influence of organizational context and event content. Academy of Management Journal, 32, 718-744.
Huson, M.R., Parrino, R. & Starks, L.T. (2001). Internal monitoring mechanisms and CEO turnover: A long-term perspective. Journal of Finance, 56(6) 2265-2297.
Jensen, M.C. & R.S. Ruback. (1983). The market for corporate control: The scientific Evidence. Journal of Financial Economics,11, 1-50.
Ranjeeni, K. (2014). Sectoral and industrial performance during a stock market crisis. Economic Systems, 38, 178-193.
Reinganum, M.R. (1985). The effect of executive succession on stockholder wealth. Administrative Science Quarterly, 30, 46-60.
Warner, J.B., Watts, R.L. & Wruck, K.H. (1987). Stock prices and top management changes. Journal of Financial Economics, 20, 461-492.
31 Weisbach, M.S. (1988). Outside directors and CEO turnover. Journal of Financial
