The effect of the former CEO on CEO turnover

The effect of the Former CEO

on CEO Turnover

Student name: Mehmet Sahin Karip

Thesis Title: The effect of the Former CEO on CEO Turnover Student number: 10867767

Date of final version: 20-06-2015 Supervisor: Dr. P. Kroos Word count: 9741

Education: MSc Accountancy and Control, Accountancy and Control track

Amsterdam Business School

2 Statement of Originality

This document is written by student Mehmet Sahin Karip 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 Abstract: I study how the decision to retain the former chief executive officer (CEO) on the board influences the sensitivity of CEO turnover to firm performance. The findings suggest that when the former CEO is retained on the board, for at least 2 years, independent of how the CEO performs, it is less likely that the new CEO will be replaced within the first 3 years of employment. These findings can be supported with the view that the former CEOs, who are retained on the board, have more substantial power compared to regular board members, and that the firms who retain their CEOs are more likely to select new CEOs who are younger and have no prior experience and therefore more easily influenced by the former CEOs. (Evans, Nagarajan, & Schloetzer, 2010).

4

Content

2 Literature review and hypothesis ... 8

2.1 Agency theory ... 8

2.2.1 Theoretical perspectives on the performance- CEO turnover relationship ... 9

2.2.2 Empirical studies on the sensitivity of performance to CEO turnover ... 10

2.2.3 Moderating variables on relationship between performance and CEO turnover ... 12

2.3 Retaining former CEOs on the board ... 14

2.4 Hypotheses ... 15 3 Research methodology ... 17 3.1 Sample ... 17 3.2 Empirical models ... 18 3.2.1 Performance measures ... 19 3.2.2 Control variables... 20 4 Descriptive statistics ... 21 4.1 Descriptive statistics ... 21 4.2 Main analyses ... 23 4.3 Robustness test ... 25 5 Conclusion ... 27 6 Appendix ... 29

5

1. Introduction

1.1 Background

Over the past two decades empirical research has produced a large body of evidence on the various aspects of the relationship between firm performance and CEO turnover. The first studies that focused on the relationship between CEO turnover and firm performance were performed in the 1980s (e.g., (Coughlan & Schimdt, 1985; Warner, Watts, & Wruck, 1988; Weisbach, 1988). Coughlan and Schmidt (1985) suggest that the board has the power to affect the change in management, so the CEO turnover. They predict that if the boards discipline the CEOs for their actions or results that harm the shareholders, the stock price performance will be a predictor of changes in CEOs. They found significant evidence that stock price performance and the probability of a change in CEO are inversely related. These findings show that poor stock performance in a given fiscal year is associated with an increase in the probability of a subsequent CEO turnover. Warner et al. (1988) have

investigated the relation between a firm’s stock price performance and subsequent changes in its top management. The top management change is defined as any change in the set of individuals holding the title of CEO, president or the chairman of the board. The findings of this study are consistent with the findings of Coughlan and Schmidt (1985). The findings also indicate an inverse relation between the probability of a top management change and stock performance; this means that top management is more quickly replaced when the stock market performance decreases. This relation was also consistent with the joint hypothesis in this study that; 1) information about management performance is reflected in stock returns and 2) such information is used in evaluating the top management.

Weisbach (1988) examined how monitoring CEOs is influenced by the quality of board corporate governance. In this study, CEO turnover is predicted by using stock returns and earnings changes as measures of prior performance. The findings of this study show that a poor stock return increases the probability of a CEO losing his job. Weisbach (1988) also states that this result replicates the results of Coughlan and Schmidt (1985) and Warner, Watts and Wruck (1988). Weisbach (1988) expects also that inside and outside directors behave differently in their decisions to remove the top management. He exploits the wide variation across firms in the composition of the board of directors to study how the relation

6 between poor performance and CEO turnover varies with the composition of the board. The findings suggest that firms with outsider dominated boards are more likely than firms with insider dominated boards to remove the CEO on the basis of poor performance. The evidence suggests that the sensitivity of top management turnover increases with the fraction of outside board members.

Since the 1980s, researchers have produced a large number of empirical papers focusing on various aspects of the relationship between CEO turnover and firm performance. Goyal and Park (2002) studied whether the combination of CEO and board chairman duties affects the decision of the board to dismiss an ineffective and poor performing CEO. The findings of Goyal and Park (2002) show that the sensitivity of CEO turnover to firm performance is significantly lower (declines with almost one-half) when the CEO and the chairman duties are vested in the same individual. The findings of Goyal and Park (2002) strongly supports

Jensen`s (1993) view that independent leadership, separate positions for the CEO and the chairman of the board, is important for effective monitoring of top management by the board of directors. Denis et al. (1997) report that the ownership structure of the company significantly affects the likelihood of CEO turnover. The findings of this study indicate that the CEO turnover is significantly greater in poorly performing companies with low

managerial ownership than in poorly performing companies with higher managerial ownership. In addition, the findings indicate that CEO turnover is more sensitive to poor performance in companies with outside blockholders. The findings of Denis et al. (1997) suggest that higher managerial ownership insulates managers from internal monitoring efforts and that outside blockholders may perform a monitoring function within these companies.

An important fact that has received little attention is that CEOs often join the board of directors of their own firm after they retire (Lee, 2011). Prior research has shown that better prior performance increases the likelihood that the former CEO is retained on the board (Brickley, Linck, & Coles, 1999). However no research to my knowledge has examined how the decision to retain a former CEO on the board influences the likelihood of the dismissal of the new CEO following poor performance.

7

1.2 Research question

The aim of this paper is to investigate the relationship between poor firm performance and CEO turnover when the former CEO is retained on the board of directors. Therefore the research question is stated as follows: How is the relationship between poor performance and probability of CEO turnover affected when the retired CEO remains on the board of that particular firm?

1.3 Motivation and Contribution

This study contributes to the existing literature on the role of the board of directors in monitoring and replacing the CEOs. Previous work has, amongst others, focused on how board composition (Weisbach, 1988), insider and institutional ownership (Denis, Denis, & Sarin, 1997), main bank ties (Kang & Shivdasani, 1995), product market competition (DeFond & Park, 1999), and combining the CEO and chairman positions (Goyal & Park , 2002) affect the sensitivity of CEO turnover to firm performance. There is no prior research yet which examines how the decision to retain the former CEO on the board influences the relation between poor performance and CEO turnover. It is important to focus on the boards because board governance plays an important role in corporate governance, particularly in monitoring the CEO of the company. Given that the board is the shareholders` first line of defense against the incompetent and underperforming CEO (Weisbach, 1988), addressing this research question also has practical relevance. In addition, a limited number of papers have looked into the topic of former CEOs that are retained on the board. While these studies have strongly focused on the determinants, my study is focused on the

consequences of that decision. Specifically, how the decision to retain the former CEO on the board influences the sensitivity of CEO turnover to firm performance.

1.4 Structure

The remainder of this paper will be structured as followed. Section two discusses the prior literature and the hypothesis development. The research methodology will be discussed in section three. This will include a description of the sample selection, empirical models, and variable measurement. Section four provides an overview of the empirical findings. Finally, section five will present the conclusion, research limitations and the future implications.

8

2 Literature review and hypothesis

2.1 Agency theory

Agency theory refers to the agency relationship due to the separation between ownership and control. In an owner-managed firm, the owner makes management decisions and has also a claim to the profits of the firm. In a publicly traded company characterized by a separation of ownership and control, the shareholders own the residual claims but lack direct control over management decision making. Subsequently, the manager has the control but possesses relatively small (if any) residual claims. There are benefits attached to separating ownership and control. First, under certain conditions, hierarchical decision making may be more efficient than market allocation (Marks, 1999). Second, due to economies of scale in both decision making and production, the optimal firm size can be large (Marks, 1999). Third, optimal investment strategy requires investors to be able to diversify and pool and to be able to change their allocations in response to changing market conditions (Marks, 1999). Besides benefits of separation of ownership and control, there are also costs. Adam Smith (1776) considered the separation of ownership and control to be problematic, because the managers of such firms would lack the incentives to operate the firm in the same manner as owner-managers. Jensen and Meckling (1976) characterized the separation of ownership and control within a firm as an agency problem. As developed in prior literature (Fama & Jensen, 1983; Jensen & Meckling, 1976; Ross, 1973), agency theory has been primarily concerned with the relationship between the principals (shareholders) and the agents (managers). Jensen and Meckling (1976) define the agency relationship as a relationship in which one or more persons (the principal(s)) hires another person (the agent) to perform a service on their behalf which involves delegating some decision making

authority to the agent. The cornerstone of the agency theory is the assumption that the interests of the principals and the agents diverge (Hill & Thomas, 1992).

Jensen and Meckling (1976) suggest that if both parties are utility maximizers, there is a great chance that the agent will not always act in the best interest of the principal. Because the interests of both parties diverge and the agent has more information

(information asymmetry), the principal cannot directly ensure that the agent will always act in the best interest of the firm. The agents have the ability to operate in their own

self-9 interest rather than in the best interests of the principals of the firm because of asymmetric information, the agents have more information than the shareholders whether they are capable of meeting the shareholders` objectives. The agents are able to use this information advantage out the expense of the shareholders. Information asymmetry contributes to two problems namely moral hazard and adverse selection. First, the managers don’t have the incentives to select those actions aligned with the interests of the principals; this is called the moral hazard problem. Moral hazard refers to the lack of effort on the part of the manager (Hill & Thomas, 1992). The argument here is that the manager does not put forth the agreed-upon effort, so the manager is shirking. The second is that the managers don’t have the ability to conform to the interests of the principals, because the managers may be incompetent. This is called the adverse selection problem (Ayres & Crampton, 1994; Smith G. , 1996). The principal can use performance measurement as a means to address the moral hazard and adverse selection problem, as realized performance represents a (noisy) signal of the effect choices and abilities of the respective agent.

2.2.1 Theoretical perspectives on the performance- CEO turnover relationship

Huson et al. (2004) developed two theoretical perspectives on the relationship between CEO turnover and firm performance; these are the improved management hypothesis and the scapegoat hypotheses. The improved management hypothesis (that refers to adverse selection problem) suggests that forced CEO turnover tends to increase managerial quality and therefore the expected firm performance. This hypothesis suggests that the quality varies across the CEOs, even though it is not directly observable. The boards of directors tend to infer the CEO quality from the realized firm performance. When the realized firm performance is sufficiently poor, the boards of directors conclude that the incumbent CEO is of low quality and that the expected benefits of replacing him exceeds the expected cost. The board of directors will choose for another CEO whose expected quality exceeds that of his predecessor. Furthermore, poor firm performance tends to co-occur with bad luck as well as low CEO quality. Therefore, the future firm performance is expected to increase following the change in CEO for two reasons: the CEO luck is expected to return to average and the expected increment in CEO quality is positive.

The scapegoat hypothesis, according Huson et al. (2004), is based on the moral hazard models of Mirrlees (1976), Holmström (1979) and Shavell (1979). The scapegoat hypothesis

10 suggests that quality does not vary across CEOs, which is in contrast with the improved management hypothesis. Under the scapegoat hypothesis poor CEO performance arises from chance alone and does not arise from low CEO quality. Thus, poor CEO performance results from bad luck, not bad management. In general, CEOs are effort-averse so they should be threatened with dismissal when the performance is poor. According to this hypothesis, only the CEOs who have bad luck are replaced by another CEO. Boards of directors understand that all CEOs have the same quality, but must fire the CEOs when the firm is performing poorly to induce other managers to provide desired level of effort. Since the replacement candidates are of the same quality as the outgoing CEO, the CEO turnover itself will not increase the managerial quality or the expected firm performance. Thus, the CEO who is fired for low quality can be seen as a scapegoat. According to this hypothesis the expected change in the firm performance following the CEO turnover will be positive, even though the turnover itself does not increase managerial quality. This is because the turnover is triggered by bad luck which reverses over time.

2.2.2 Empirical studies on the sensitivity of performance to CEO turnover

In this section we will look to the empirical studies which focus on the sensitivity of firm performance to CEO turnover. Coughlan and Schmidt (1985) suggest that the board has the power to affect the change in management, so the CEO turnover. They predict that if the boards discipline the CEOs for their actions or results that harm the shareholders, the stock price performance will be a predictor of changes in CEOs. In their research they admit that the empirical confirmation of such relation between poor stock performance and CEO turnover is complicated, because there exist also other possible reasons for a change in CEO like; normal retirement, death or promotion towards a better paying position. In their study they excluded all of these cases. The regression results in this study are consistent with the hypothesis that stock price performance and the probability of a change in CEO are inversely related. The results show that poor stock performance in a given fiscal year is associated with an increase in the probability of a subsequent CEO turnover.

Warner et al. (1988) have investigated the relation between a firm’s stock price performance and subsequent changes in its top management. The top management change is in this study defined as any change in the set of individuals holding the title of CEO, president or chairman of the board. One of the hypotheses, in this study, is that the probability of a top

11 management change is inversely related to stock price performance. This study examines the association between stock price performance and top management changes for 269 NYSE and AMEX firms in the period 1963-1978. The results suggest an inverse relation between the probability of a top management change and stock performance. This relation is also consistent with the joint hypothesis in this study that; 1) information about

management performance is reflected in stock returns and 2) such information is used in evaluating the top management. Weisbach (1988) examined the relation between monitoring of CEOs by inside and outside directors and CEO turnover. In his study, CEO turnover is predicted by using stock returns and earnings changes as measures of prior performance. The results show that a poor stock return increases the probability of a CEO losing his job. Weisbach (1988) also states that this result replicates the results of Coughlan and Schmidt (1985) and Warner et al. (1988). The results of Weisbach (1988) also show that there is a relation between changes in earnings and the probability that the CEO will be replaced. The findings of Huson et al. (2004) show that firm financial performance tends to decline prior to top management turnover, this evidence is statistically significant and robust in their sample. They state that this evidence is consistent with the findings of Warner et al. (1988) and Weisbach (1988). Farrell and Whidbee (2003) examined CEO turnover and replacement decisions from a different perspective by examining the role of performance expectations. They argue similar to Puffer and Weintrop (1991), that 1-year analyst forecast errors, the deviation of realized earnings from expected earnings, provide additional

information to the board regarding the performance of the CEO. “To the extent that earnings forecasts proxy for the board of director’s earnings expectations, forecast errors may capture the component of firm performance that the board attributes, in large part, to CEO

performance” (Farrel & Whidbee, 2003). The results of this study suggest that CEOs are not simply held accountable for the firm performance, but that the boards of directors make use of firm performance expectations as part of their criteria for evaluating the performance of the CEO. Further, the results suggest that CEOs with negative forecast errors face a greater likelihood of turnover. Consistent with their hypothesis, the results also suggest that when the signals from the forecasted error are more precise, it is assigned more weight in the CEO turnover decision.

12 2.2.3 Moderating variables on relationship between performance and CEO turnover

In this section I will review the literature on moderating variables that influence the relation between CEO performance and CEO turnover. Weisbach (1988) states in his article that boards of directors play an important role in corporate governance, particularly in

monitoring the Chief Executive Officer (CEO). The board members are supposed to supervise the actions and the performance of the management and provide advice and veto poor decisions. Weisbach (1988) states that the board is the shareholders` first line of defense against the incompetent CEO. The most important duty of the boards of directors is to evaluate the CEO. The board is responsible for evaluating the CEO of the firm and replacing him if he fails to perform well (Weisbach, 1988). Weisbach (1988) states that this task is to be executed by the outside directors, because inside directors` careers are tied to the CEOs and they are often unable or unwilling to remove the incumbent CEOs of the firm.

Borokhovich et al. (1996) report also that outside directors are more likely than inside directors to dismiss a poorly performing CEO and to replace him with an executive who will increase the firm value. The results of Mace (1971) show that boards of directors typically are not involved in important corporate decisions like selecting other directors. Mace (1971) states that the boards of directors serve as valuable source of advice for many CEOs and are also responsible for removing the CEO in crisis situations. Mace (1971) emphasized that outside directors took more initiative than the inside directors to remove the bad

performing incumbent CEOs. The results of Weisbach (1988) are consistent with the results found by Mace (1971). Weisbach (1988) found stronger association between prior

performance and the probability of CEO turnover for firms with outsider-dominated boards than for firms with insider-dominated boards (these results was not a function of ownership effects, size effects, or industry effects). The obvious interpretation of these results is that the outside directors have a stronger monitoring role within firms.

Goyal and Park (2002) studied whether bestowing CEO and board chairman duties on one individual affects a boards` decision to dismiss an ineffective and underperforming CEO. They assume that the lack of independent leadership in a firm with a single CEO-Chairman reduces monitoring by the board and makes it difficult for the board to replace a poorly performing CEO. Goyal and Park (2002) suggest that the probability of CEO turnover is likely to be less sensitive to performance in a firm with a combined CEO and chairman position than in firms with two separate positions. Their findings show that the sensitivity of CEO

13 turnover to firm performance declines by almost one-half when the same individual holds both the CEO and the chairman position. The findings of Goyal and Park (2002) strongly support Jensen`s view that when the CEO also holds the position of the chairman of the board, internal control systems fail because the board cannot effectively perform its key functions including those of evaluating and firing CEOs. DeFond and Park (1999) are stating that relative performance evaluation (RPE) is likely to improve boards of directors ability to identify unfit CEOs, because RPE provides a more precise measurement of CEO performance. RPE is only beneficial when a peer group faces common economic factors and when any given CEO`s actions cannot affect other CEO`s output, and in addition the benefits of RPE increases with the number of CEOs in the peer group (Holmstrom, 1982). DeFond and Park (1999) are stating that these conditions are more likely to occur in competitive industries where large numbers of firms operate in similar kind of environment, and where the actions and decisions of CEOs do not affect each other’s output. Their findings show that an RPE-based accounting measure is significantly associated with CEO turnover in high competition industries but not in industries with low competition. Denis et al. (1997) examine the impact of ownership structure on internal monitoring efforts by documenting the rate of non-routine CEO turnover in 1394 firms over the period 1985-1988. Their results suggest that ownership structure plays an important role in determining the effectiveness of internal control mechanisms within firms. The findings show that the likelihood of CEO turnover is significantly greater in poorly performing firms with low managerial ownership than in poorly performing firms with higher managerial ownership. In addition, the findings show that CEO turnover is more sensitive to poor performance in firms with outside blockholders than in firms without blockholders. These findings of Denis et al. (1997) suggest that higher equity ownership insulates mangers from internal monitoring efforts and that outside blockholders may perform a monitoring function within firms. Shleifer and Vishny (1986) are pointing out that large stockholders are likely to have greater incentives to monitor the management than do small stockholders. This is because the benefits that large stockholders could receive from monitoring the CEOs are more likely to exceed the costs they bear. The findings of Denis and Serano (1996) suggest that outside blockholders are instrumental in removing poorly performing CEOs. Denis et al. (1997) show that the probability of CEO turnover in general is positively related to the presence of such blockholders. Many prior research, including Pound (1992) and Black

14 (1992), have claimed that institutional shareholders perform a monitoring function similar to that of blockholders. These studies assume that institutions increase the pressure on boards to make management replacement decisions that serve the interests of the stockholders, so the quality of internal monitoring would be positively related to institutional shareholders. Engel et al. (2003) examine how the relation between accounting performance measures and CEO turnover is affected by properties of the firm`s accounting system. Their findings show that accounting information appears to receive greater weight in CEO turnover decisions when the accounting-based measures are more precise and sensitive. Engel et al. (2003) found also evidence that market-based performance measures receive less weight in turnover decisions when accounting-based measures are more sensitive or market returns are more volatile (susceptible to noise).

2.3 Retaining former CEOs on the board

Brickley et al. (1999) examine the CEO incentives (1989-1993), more specifically the concerns of CEOs about post-retirement board service. They argue that CEO`s career concerns need not end at retirement and suggest that managers remain active during retirement, through serving on corporate boards. Brickley et al. (1999) found a strong relation between the likelihood of post-retirement board service of the CEOs and their performance while on the job. Their findings suggest that especially stock-market performance and to lesser extent accounting performance during the CEO`s career have substantial power to explain the likelihood of a CEO serving on his own board after retirement. More specifically, Brickley et al. (1999) found evidence that retiring CEOs who stay on their own board generate 10.9% higher annual abnormal stock returns, and 2.0% higher annual accounting returns, over their last year serving as a CEO than CEOs who do not stay on their own board. The findings also suggest that the probability that a CEO serves on his own board two years after exiting as a CEO is 0.43 for former CEOs with prior stock performance at the bottom quartile, in contrast to 0.56 for CEOs performing at the top quartile. Additionally, the probability that a CEO stays on his own board is 0.45 when accounting performance is at the bottom quartile, in contrast to 0.52 CEOs performing at the top quartile. Lee (2011) analyzes the directorships held by CEOs who retired during the periods 1989-1993, 1995-1999 (before the Sarbanes-Oxley Act) and 2001-2005 (after the Sarbanes-Oxley Act). Lee (2011) compares the outcomes of his

15 study to that of Brickley et al. (1999), and analyzes what has happened over time to the selection of retired CEOs as board members. Lee (2011) found evidence that abnormal stock returns strongly explain the probability of a CEO`s serving as chairman or insider director on his own board in the 1989-1992 and 1995-1999 sample, however this positive effect is declining for the 2001-2005 sample. In particular, the findings suggest that a 21.2% increase in abnormal stock returns of the firm, produces an 11.2% increase in the probability of holding a chairman or inside directorship after retirement in the 1989-1993 sample, and 13.1% in the 1995-1999 sample, compared to 8.1% in the period after SOX. Evans et al. (2010) are also focusing on the option in which the former CEOs remain on the board of directors. They label this option as Retention Light, because the CEOs are retained in the board but with reduced decision rights. The former CEO continues to provide services for an extended period of time, but now as a board member rather than as chief executive. Evans et al. (2010) found evidence that CEOs who are powerful individuals and whose firms achieve relatively better preturnover performance are more likely to remain on the firm’s board after leaving office.

2.4 Hypotheses

To my knowledge, no research has examined how the decision to retain a former CEO on the board of directors influences the likelihood of the dismissal of the new CEO following poor firm performance. Evans et al. (2010) suggest that the former CEO who is retained on the board of directors as chairman or director, which I call Retention CEO from now on, is likely to have substantial power compared to regular board members. This is because the CEO who is retained on the board is more likely to have been the chairman of the board or the firm`s founder. The findings of Evans et al. (2010) show that compared to firms whose CEO was dismissed, firms who retained their CEO on their board are more likely to select new CEOs who are younger, have no prior CEO experience and may have family ties to the

Retention CEO. According to Evans et al. (2010), these firms are more likely to select the new CEOs with those characteristics because these incoming CEOs are more easily influenced by the Retention CEO. Consistent with entrenchment, the former CEO within the board of directors can continue exert its power and affect the firm, with associated positive or

16 and Lee (2011) show how Retention CEOs have superior pre-turnover performance

indicative of the firm-specific knowledge that those CEOs accumulated over time.

Based on the evidence found by prior studies, one may expect that retaining a former CEO on the board, so the Retention CEO, will have a positive impact on the likelihood of the dismissal of the new CEO following poor firm performance. I expect that the Retention CEO will use his accumulated firm-specific knowledge and power to provide valuable advice to the incoming CEO and to stringently monitor the new CEO, in order to ensure the firm is performing well. It is also possible that these Retention CEOs hold a fraction of firm stock and are therefore attached to the firm’s success, which in turn creates more incentives for them to monitor the incoming CEO. This more stringent monitoring by the Retention CEO may translate into a greater likelihood of dismissal after poor performance. In addition, Evans et al. (2010) described that these firms select new CEOs who are younger and have no prior experience. This could also result in a greater likelihood of dismissal after poor

performance given that the new CEO has relatively little power to insulate herself against dismissal after poor performance (compared to the pressure that may be exercised by the powerful Retention CEO). However, may also be possible that, retaining a former CEO on the board will have a negative impact on the likelihood of the dismissal of the new CEO following poor firm performance. That is, The Retention CEO may want to retain its influence on the corporate decision making by using his power and influence to insulate the new CEO against dismissal. That is, the new less CEO enables the Retention CEO to exert its influence relative to a new CEO successor that may want to operate more independently from the Retention CEO. Here, I assume that there will be Retention CEOs who do not care so much about the firm performance because they lack pecuniary or nonpecuniary attachment to the firm. Thus, it may be expected that when the incoming CEO is performing poorly, the Retention CEO will use his power to convince the board members to not to dismiss the new CEO, by the fact the Retention CEO still can influence the new CEO (with low power) for his own personal benefits. Based on the aforementioned, I developed the following non-directional hypothesis;

Hypothesis: There is an association between retaining a former CEO on the board and the likelihood of dismissal of the new CEO following poor firm performance.

17

3 Research methodology

3.1 Sample

This paper`s sample will be based on the U.S market from 1992 to 2010. The U.S market is chosen because of the easy access to the accounting information on U.S firms. I start with an initial sample of ExecuComp firms from 1992 to 2010. The coverage of the ExecuComp database roughly corresponds with the S&P 1500; this combines three leading indices, the S&P 500, The S&P MidCap 400 and the S&P Smallcap 600. These indices cover approximately 90 percent of the United States market capitalization. This coverage of ExecuComp yields an initial sample of 32.038 observations. From this data, I focus on the first 3 years of

employment of the incoming CEOs. This yields a sample of 6.323 observations. The

computation of market performance, accounting performance and data on control variables, requires additional data from the Compustat database. After merging Compustat data with ExecuComp data, I dropped all the observations for which no match has been found. This yields a sample of 5.399 observations. Finally, after dropping all incomplete observations, the sample size is reduced to 2.599 observations. For example, to assess whether the former CEO remains on the board, I compare the data when the former CEO left office with the date when the former CEO left the firm. Missing data on each of these variables leads to the exclusion of the observation. My final sample is composed of 2.599 observations.

18

3.2 Empirical models

My hypothesis is tested by means of the following empirical model. Consistent with prior literature focusing on the relationship between performance and CEO turnover, a logit regression model approach will be used. The logit regression model is as follows;

CEO_Turnover= β0 + β1ACC_Perf+ β2ACC_Perf *D_Stay + β3Market_Perf +β4Market_Perf*D_Stay + β5D_Stay + Controls + Ɛ

Where,

Turnover= An indicator variable which is equal to one if there is a CEO

change in the subsequent year.

ACC_Perf= Return on Assets (see section 3.2.1).

Market_Perf= Percentual change in market performance for that fiscal year (see section 3.2.1).

D_Stay= An dummy variable which takes the value of one if the previous

CEO is retained for at least 2 years on the board of directors and a value of zero when the previous CEO is retained less than

2 years on the board of directors (or not at all).

Controls= Several variables are included in the empirical model to control for differences that are systematically associated with CEO turnover. Including CEO age, Leverage, Market-to-book ratio and Firm size (see (see section 3.2.2).

The coefficients β1 and β3 give the relationship between respectively accounting

performance and market performance on the one hand and the likelihood of CEO turnover on the other hand for the CEOs whose predecessor has been retained less than 2 years on the board of directors. The sum of coefficients (β1+β2) and (β3+β4) give the relationship between respectively accounting performance and market performance on the one hand and the likelihood of CEO turnover on the other hand for the CEOs whose predecessor has been retained on the board of directors for at least 2 years. Therefore, the coefficients β2 and β4 represent the difference in the relationship between accounting and market

19 retained less than 2 years and the CEOs whose predecessor has been retained longer than 2 years on the board of directors. On the basis of my non-directional hypothesis, I expect that β2≠0 and β4≠0.

3.2.1 Performance measures

Consistent with prior literature, I use both accounting earnings and stock returns as measures of firm and CEO performance in determining the likelihood of CEO turnover (Rosen, 1990; Milgrom & Roberts, 1992). Some prior literature found that CEO turnover is significantly related to stock returns (Warner, Watts, & Wruck, 1988; Weisbach, 1988; Murphy, 1999; Murphy & Zimmerman, 1993; Kaplan, 1994; Coughlan & Schimdt, 1985), but it is not clear yet whether stock return measures are more informative than earnings in measuring the performance of the CEOs. Kaplan (1994) suggests in his paper that stock return measures also reflect changes in discount rates and therefore accounting earnings measures may be more informative. The findings of Murphy and Zimmerman (1993) show a strong relationship between CEO turnover and accounting earnings measures. One large advantage of accounting earnings over stock returns for measuring the performance of the CEO is that accounting earnings data measure short-term profits. The stock price reflects the present discounted value of the expected future cash flow of the firm (Weisbach, 1988). As we can see, the empirical evidence is ambiguous about this issue. To ensure the robustness of the results, I will use both performance measures; market performance and accounting performance. The first measure of firm and CEO performance is the market performance. This is calculated by; ((market value t1 + total dividends t1 –

market value t-1)/(market value t-1))*100%. This indicates whether the market performance of the firm is better or worse compared with the market performance of the previous year. The second measure is an accounting-based performance measure, Return on Assets, this is calculated by dividing net income by total assets of the firm for each fiscal year, it is

displayed as a percentage. This performance measure is an indicator of how profitable a company is relative to its total assets. It gives an idea as to how efficient a CEO is at using its assets to generate earnings for the firm.

20 3.2.2 Control variables

Several variables will be included in the empirical model to control for differences that are systematically associated with CEO turnover. First, CEO AGE will be included to control the impact of age on CEO turnover. Weisbach (1988) and Murphy and Zimmerman (1993) found strong relationship between CEO age and CEO turnover, 30% of the CEOs were leaving the firm at the age 64 to 65. Other literature also suggest that CEO turnover around the ages 63 to 65 are more likely due to normal retirements (routine CEO turnover) than forced

departures (non-routine CEO turnovers) (Murphy, 1999). Because of this strong relationship between CEO age and CEO turnover, I will include CEO age as a control variable in the analysis. Second, with respect to firm size, larger firms are more complex, which can lead to higher CEO turnovers. Prior literature shows that firms with forced CEO turnover are larger than firms with no turnover (Hazarika, Karpoff, & Nahata, 2012). To account for the effects of firm size on CEO turnover, the control variable SIZE is used, which is defined as the natural logarithm of book value of the total assets. Third, leverage will be added to the empirical models because this might affect the volatility of performance variables (Adams, Almeida , & Ferreira, 2005). The findings of prior literature suggest that firms with forced CEO turnover are more highly levered than firms with no turnover (Hazarika, Karpoff, & Nahata, 2012). To account for this, the control variable LEV is used, the total liabilities are divided by book value of total assets. Finally, market-to-book ratio will be added to the empirical models. Prior literature suggests that the mean of market-to-book ratio is significantly lower for firms with forced turnover than the means for firms with voluntary turnover and no turnover (Hazarika, Karpoff, & Nahata, 2012). The finding that firms with forced CEO turnover have higher risk and lover market-book ratios suggests that the probability of forced turnover could reflect firms` operating environments. To account for these effects the variable MTB is used as a control variable. This is defined as market value of equity divided by the book value of equity for the firm.

21

4 Descriptive statistics

4.1 Descriptive statistics

Table 1 reports descriptive statistics on variables used in the empirical model, for the full sample. It states some centrality measures like mean, and also some dispersion measures, such as the standard deviation. Table 1 shows that 16% of the incoming CEOs are replaced within the first 3 years of their employment. Firm performance is measured by accounting and market performance. More specifically, the accounting performance of the CEOs has an average ratio of 1.9% for this sample. The average market performance for the whole sample is about 19.5%. The Table 1 also shows that about 2.7% of the predecessors of the incoming CEOs are retained for at least 2 years on the board of directors.

For the control variables, the table shows that the firms have an average leverage ratio of 56.78. So, more than the half of their assets are financed through debt. Size (in assets) has an average value of 7.69. The market-to-book ratio has an average value of 3.01, which suggests a fairly high availability of growth to the sample firms. Finally, Table 1 shows that the average CEO age is around 55 for the whole sample, where 80% of my sample ranges between 46 and 62 years.

Table 2 gives an overview of the Pearson correlations (***, **, * denotes 1%, 5%, and 10% significance levels respectively). Table 2 shows a negative correlation between CEO turnover and accounting performance, consistent with the intuition that lower accounting

performance increases the likelihood of CEO turnover. It also shows a negative correlation between retention of the former CEO and CEO turnover. This implies that retaining a CEO on the board is associated with a lower likelihood of CEO turnover. Further, this table shows a positive significant correlation between CEO age and CEO turnover, which indicates that older CEOs are more likely to leave the firm because of retirement. It shows also a positive significant relation between market performance and accounting performance. For more correlations, I refer to Table 2.

22

Table 1 reports descriptive statistics for variables used in the analyses

stats N mean sd p10 p25 p50 p75 p90 CEO_turnover 2599 0.164 0.370 0 0 0 0 1 ACC_Perf 2599 1.901 17.410 -7.070 0.793 4.253 8.147 12.547 Market_Perf 2599 19.510 82.326 -41.048 -17.099 7.331 35.587 76.164 D_Stay 2599 0.027 0.163 0 0 0 0 0 LEV 2599 56.779 26.534 27.110 41.390 56.193 70.167 80.448 SIZE 2599 7.695 1.650 5.652 6.552 7.622 8.794 10.063 MTB 2599 3.019 23.490 0.845 1.313 1.969 3.310 5.827 AGE 2599 54.570 6.229 46 51 55 59 62

Table 2: Pearson Correlation Matrix; ***, **, * denotes 1%, 5%, and 10% significance levels respectively.

CEO_turnover ACC_Perf Market_Perf D_Stay LEV SIZE MTB

CEO_turnover 1 ACC_Perf -0.08*** 1 Market_Perf -0.01 0.1*** 1 D_Stay -0.06*** 0.02 -0.02 1 LEV -0.04 -0.09*** -0.01 -0.06*** 1 SIZE -0.08*** 0.24* -0.03* 0 0.21*** 1 MTB 0.03 0.03 0.02 0 0.01 0.02 1 AGE 0.1*** 0.05*** 0.04* 0.03 0.01 0.11*** 0.02

23

4.2 Main analyses

Table 3 (Logistic regression) reports regression estimates for the empirical model (see section 3.2). It examines the non-directional hypothesis that predicts that there is an association between retaining a former CEO on the board of directors and the likelihood of dismissal of the incoming CEO following poor firm performance. Table 3 shows that the independent variable accounting performance is significantly negative correlated to CEO turnover. So, this implies that when the former CEO is not retained on the board, accounting performance is inversely related with CEO turnover. Lower accounting performance

increases the likelihood of CEO turnover. The coefficient on market performance is insignificant, indicating that when the former CEO is not retained on the board, a lower market performance is not associated with the likelihood of CEO turnover. So these findings imply that when the new CEO is performing poorly, calculated as ROA ratio, then it is more likely that he or she will be replaced within the first 3 years of his or her employment, which is consistent with the findings of Murphy and Zimmerman (1993).

My hypothesis is about the interaction terms ACC_STAY (ACC_Perf*D_Stay) and

Market_STAY (Market_Perf*D_Stay). The independent variable ACC_STAY indicates the accounting performance of the new CEO whose predecessor is retained on the board of directors. The independent variable Market_STAY indicates the market performance of the new CEO whose predecessor is retained on the board of directors. Table 3 shows no

significant coefficients for the relationship between the independent variables ACC_STAY and Market_Perf on the one hand and CEO turnover on the other hand. So, I have to reject my hypothesis that retaining a former CEO on the board would influence the relationship between (poor) performance and the (greater) likelihood of CEO turnover.

However, interesting is that the independent variable D_Stay is significant and negatively related to CEO turnover. This finding implies that when the former CEO is retained on the board of directors, for at least 2 years, independent of how the CEO performs, it is less likely that the new CEO will be replaced within the first 3 years of employment. This finding can be supported with the findings of Evans et al. (2010) suggesting that the former CEO who is retained on the board of directors as chairman or director, is likely to have substantial power compared to regular board members. As described earlier, Evans show that compared to firms whose CEO leave the firm, firms who retained their CEO on their board are more likely to select new CEOs who are younger and have no prior experience. These firms are more

24 likely to select the new CEO s with those characteristics because these incoming CEOs are more easily influenced by the Retention CEOs. Regardless, how the new CEO performs, in the first 3 years of his employment, the Retention CEO will use his power to convince the board members to not to dismiss the new CEO, by the fact the Retention CEO still can influence the new CEO (with low power) for his own personal benefits.

Several variables are included in the empirical model to control for differences that are systematically associated with CEO turnover. First, Table 3 shows a significantly negative relation between firm size and CEO turnover. This implies that the new CEOs are more likely to be replaced, within 3 years, in smaller firms compared to larger firms. Second, Table 3 shows a positive significant relation between CEO age and CEO turnover. This implies that new CEOs who are older are more likely to leave office than the CEOs who are younger. Finally, the model is significant as indicated by the F value.

Table 3 reports regression estimates for the empirical model. The p-values are reported in parentheses. ***, **, * denotes 1%, 5%, and 10% significance levels respectively.

CEO turnover Variables Pred. ACC_Perf -0.008*** (0.003) ACC_STAY β2≠0 -0.016 (0.672) Market_Perf 0 (0.53) Market_STAY β4≠0 -0.032 (0.219) D_STAY -1.843** (0.02) LEV -0.003 (0.12) SIZE -0.114*** (0.001) MTB 0.003 (0.159) AGE 0.05*** (0) R2 0.033 Prob > F 0*** Observations 2599

25

4.3 Robustness test

To test the robustness of the findings described in the previous section, I take an additional test. Table 4 (Logistic regression) reports regression estimates for the empirical model (see section 3.2). It examines, again, the non-directional hypothesis that predicts that there is an association between retaining a former CEO on the board of directors and the likelihood of dismissal of the incoming CEO following poor firm performance.

For this additional test, I drop all the incoming CEOs who are older than 63 years. This yields a sample of 2433 observations. I only keep this group of incoming CEOs to control for the impact of age on CEO turnover found by various literatures. Weisbach (1988) and Murphy and Zimmerman (1993) found strong relationship between CEO age and CEO turnover, 30% of the CEOs were leaving the firm at the age 64 to 65. The findings of other literature also suggest that CEO turnover around the ages 63 to 65 are more likely due to normal

retirements (routine CEO turnovers) than forced departures (non-routine CEO turnovers) (Murphy, 1999). The main inferences from Table 3 are not affected. Still, retaining a former CEO on the board does not influence the relationship between performance and the likelihood of CEO turnover. Retaining a former CEO on the board still decreases the

likelihood of CEO turnover regardless the performance. Furthermore, when the former CEO is not retained on the board, lower accounting performance still increases the likelihood of CEO turnover. The only main difference is that now market performance is also negative and significant. This indicates that for those cases where the former CEO leaves the firm, the likelihood that the new CEO leaves the firm within 3 years becomes bigger when either the accounting performance and the stock market performance is lower. Finally, the model is significant as indicated by the F value.

26

Table 4 reports the regression estimates.

The p-values are reported in parentheses. ***, **, * denotes 1%, 5%, and 10% significance levels respectively.

CEO turnover Variables Pred. ACC_Perf -0.008*** (0.005) ACC_STAY β2≠0 -0.012 (0.735) Market_Perf -0.002* (0.071) Market_STAY β4≠0 -0.028 (0.254) D_STAY -1.646** (0.035) LEV -0.003 (0.193) SIZE -0.118*** (0.002) MTB 0.004 (0.153) AGE 0.022** (0.037) R2 0.024 Prob > F 0*** Observations 2433

27

5 Conclusion

Over the past two decades empirical research has produced a large body of evidence on the various aspects of the relationship between firm performance and CEO turnover.

An important fact that has received little attention is that CEOs often join the board of directors of their own firm after they retire (Lee, 2011). Only a limited number of papers have looked into this topic where former CEOs are retained on the board of directors (Brickley, Linck, & Coles, 1999; Evans, Nagarajan, & Schloetzer, 2010). While these studies have strongly focused on the determinants, my study is focused on the consequences of the decision to retain a former CEO on the board of directors. The findings of this study imply that for those cases where the former CEO leaves the firm, the likelihood that the new CEO leaves the firm within 3 years becomes bigger when the accounting performance is lower, consistent with the findings of Murphy and Zimmerman (1993). Furthermore, the findings of the robustness test (see section 4.3) suggest that for those cases where the former CEO is not retained on the board, the likelihood that the new CEO leaves the firm within the first 3 years of employment, becomes bigger when either the accounting performance and the stock market performance is lower, the latter is consistent with evidence found by prior literature (Coughlan & Schimdt, 1985; Weisbach, 1988).

There is no evidence found that retaining a former CEO on the board influences the relationship between (poor) performance and the (greater) likelihood of CEO turnover. However, the findings of this paper imply that when the former CEO is retained on the board, for at least 2 years, independent of how the (new) CEO performs, it is less likely that the new CEO will be replaced within the first 3 years of employment. These findings can be supported with the findings of Evans et al. (2010), suggesting that firms who retained their CEO on their board are more likely to select new CEOs who are younger and have no prior experience. These firms are more likely to select the new CEO s with those characteristics, because these new CEOs are more easily influenced by the former CEO.

The former CEO, who is retained on the board, may want to retain its influence on the corporate decision making by using his power and influence, to insulate the new CEO against dismissal independent of his performance by the fact he still can influence the new CEO (with low power) for his own personal benefits. That is, the new CEO (with low power) enables the former CEO to exert its influence relative to a new CEO successor that may want to operate more independently from the former CEO.

28 The findings of my study are subject to at least one important limitation. I only focused on the first 3 years of employment of the incoming CEOs. Future research can improve this study by focusing on a longer period of time, and examine whether the evidence found in this study still remains the same.

29

6 Appendix

Adams, R., Almeida , H., & Ferreira, D. (2005). Powerful CEOs and Their Impact on Corporate Performance. Oxford University Press, 1403-1432.

Ayres, I., & Crampton, P. (1994). Relational Investing and Agency Theory. Cardozo Law Review. Black, B. (1992). Institutional investors and corporate governance: the case for institutional voice.

Journal of Applied Corporate Finance, 19-32.

Blackwell, D., Brickley, J., & Weisbach, M. (1994). Accounting information and internal performance evaluation: evidence from Texas banks. Journal of Accounting and Economics, 331-358. Borokhovic, K., Parrino, R., & Trapani, T. (1996). Outside directors and CEO selection. Journal of

Financial and Quantitative Analysis, 337-355.

Brickley, J. (2003). Emperical research on CEO turnover and firm-performance: a discussion. Journal

of Accounting and Economics, 227-233.

Brickley, J., & Van Horn, R. (2002). Managerial incentives in nonprofit organizations: evidence from hospitals. Journal of Law and Economics, 225-249.

Brickley, J., Linck, J., & Coles, J. (1999). What happens to CEOs after they retire? New evidence on career concerns, horzion problems, and CEO incentives. Journal of Financial Economics, 341-377.

Butler, S., & Newman, H. (1989). Agency control mechanisms, effectiveness and decision making in an executive's final year with the firm. Journal of Institutional and Theoretical Economics, 451-464.

Coughlan, A., & Schimdt, R. (1985). Executive compensation, management turnover, and firm performance: an emperical investigation. Journal of Accounting and Economics, 43-66. Dechow, P., & Sloan , R. (1991). Executive incentives and the horizon problem: Some empirical

evidence. Journal of Accounting and Economics, 51-89.

DeFond, M., & Park, C. (1999). The effect of competition on CEO turnover. Journal of Accounting and

Economics, 35-56.

Denis, D., & Serano, J. (1996). Active investors and management turnover following unsuccessful control contests. Journal of Financial Economics, 239-266.

Denis, D., Denis, D., & Sarin, A. (1997). Ownership structure and top executive turnover. Journal of

Financial Economics, 193-224.

Denis, D., Denis, D., & Sarin, A. (1997). Ownership structure and top executive turnover. Journal of

30 Engel, E., Hayes, R., & Want, X. (2003). CEO turnover and properties of accounting information.

Journal of Accounting and Economics.

Evans, J., Nagarajan, N., & Schloetzer, J. (2010). CEO Turnover and Retention Light: Retaining Former CEOs on the Boards. Journal of Accounting, 1015-1047.

Fama, E., & Jensen, M. (1983). Seperation of ownership and control. Journal of Law and Economics, 301-325.

Farrel, K., & Whidbee, D. (2003). The impact of firm performance expectations on CEO turnover and replacement decisions. Journal of Accounting and Economics .

Ghosh, A., Gu, Z., & Moon, D. (2007). “CEO Tenure and Asymmetric Informativeness of Earnings.”.

Working Paper, Carnegie Mellon University.

Goyal , V., & Park , C. (2002). Board leadership structure and CEO turnover. Journal of Corporate

Finance, 49-66.

Hazarika, S., Karpoff, J., & Nahata, R. (2012). Internal corporate governance, CEO turnover, and earnings management. Journal of Financial Economics, 44-69.

Hill, C. W., & Thomas, M. J. (1992). STAKEHOLDER - AGENCY THEORY. Journal of Management

Studies, 131-154.

Hill, C., & Phan , P. (1991). CEO tenure as a determinant of CEO pay. Academy of Management

Journal, 707-717.

Holmström, B. (1979). Moral hazard and observability. The Bell Journal of Economics, 74-91. Holmstrom, B. (1982). Moral hazard in teams. Bell Journal of Economics, 324-340.

Holmström, B., & Kaplan, S. (2003). The state of U.S. corporate governance: what's right and what's wrong? Finance, 8-20.

Huson , M., Malatesta, P., & Parrino, R. (2004). Managerial succession and firm performance. Journal

of Financial Economics, 237-275.

Jensen, M. (1993). Presidential Address: the modern industrial revolution, exit and the failure of internal control systems. Journal of Finance, 831-880.

Jensen, M., & Meckling, W. (1976). Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure. 1-78.

Kang , J.-K., & Shivdasani, A. (1995). Firm performance, corporate governance, and top executive turnover in Japan. Journal of Financial Economics, 29-58.

Kaplan, S. (1994). Top executive rewards and firm performance: a comparison of Japan and the United States. Journal of Political Economy, 510-546.

Kimberly, J., & Zajac, E. (1988). The dynamics of CEO/Board relations. In D. C. Hambrick (ed.), The

31 Lee, C. (2011). New evidence on what happens to CEOs after they retire. Journal of Corporate

Finance, 474-482.

Mace, M. (1971). Directors, myth and reaiity. Harvard Business School Press, Boston, MA. Marks, S. (1999). THE SEPARATION OF OWNERSHIP AND CONTROL. 692-724.

Milgrom, P., & Roberts, J. (1992). Economics, Organization, and Management. Prentice-Hall,

Englewood Cliffs, NJ.

Mirrlees, J. (1976). The optimal structure of incentives and authority within an organization. The Bell

Journal of Economics, 105-131.

Mizruchi, M. (1983). Who controls whom? An examination of the relation between management and boards of directors in large corporations. Acadamy of Management, 426-435.

Murphy , K., & Zimmerman, J. (1993). Financial performance surrounding CEO turnover. Journal of

Accounting and Economics, 273-316.

Murphy. (1999). executive compensation In: O. Ashenfelter and D. Card (Eds.), Handbook of Labor.

Elsevier, 2485-2563.

Parrino, R. (1997). CEO turnover and outside succession: a cross-sectional analysis. Journal of

Financial Economics, 165-197.

Pearce, J., & Zahra , S. (1991). The relative power of CEOs and boards of directors: Associations with corporate performance. Strategic Management Journal, 135-153.

Pound, J. (1992). Raiders, targets, and politics: the history and future of American corporate control.

Journal of Applied Corporate Finance, 6-18.

Puffer, S., & Weintrop , J. (1991). Corporate performance and CEO turnover: the role of performance expectation. Administrative Science Quarterly, 1-19.

Rosen. (1990). Contracts and the market for executives. Working Paper No. 3542. NBER, Cambridge,

MA, December.

Ross, S. (1973). The Economic Theory of Agency: The Principal's Problem . The American Economic

Review, 134-139.

Salancik, G., & Meindl, J. (1984). Corporate attributions as strategic illusions of management control.

Administrative Science Quarterly, 238-254.

Shavell, S. (1979). Risk sharing and incentives in the principal and agent relationship. The Bell Journal

of Economics, 55-73.

Shleifer, A., & Vishny, R. (1986). Large shareholders and corporate control. Journal of Political

Economy, 461-488.

32 Smith, G. (1996). Corporate Governance and Managerial Incompetence: Lessons From Kmart. North

Carolina Law Review.

Sonnenfield, J. (1988). The Hero's Farewell. Oxford university Press, Oxford.

Vancil, R. (1987). Passing the Baton: Managing the Process of CEO Succession. Harvard Business

School Press, Boston, MA.

Villalonga, B., & Amit, R. (2006). How Do Family Ownership, Control and Management Affect Firm Value? Journal of Financial Economics, 385-417.

Walsh, J., & Seward, J. (1990). On the efficiency of internal and external corporate control mechanisms. Academy of Management Review, 421-458.

Warner, J., Watts, R., & Wruck, K. (1988). Stock prices and top management changes. Journal of

Financial Economics, 461-492.

Weisbach, M. (1988). Outside directors and CEO turnover. Journal of Financial Economics, 432-460. Zald, M. (1969). The power and functions of board of directors: A theoretical synthesis. American