0
Consequences of CEO Departures
and the Responsibility of the Board in
the Netherlands
MSc Finance
Abstract:
Name Student: S.M. Nanninga Student ID number: 1706276 Study Program: MSc Finance
Student E-mail: s1706276@student.rug.nl/sikko.nanninga@gmail.com
1 CONTENTS 1. INTRODUCTION………2 2. LITERATURE REVIEW………..……...4 3. METHODOLOGY……….………15 4. DATA………...19 5. EMPERICAL RESULTS...23
6. DISCUSSION AND CONCLUSION………..………..30
7. REFERENCES……….……..32
2
I. INTRODUCTION
“During the economic crisis, boards took a reactive approach to CEO turnover and postponed CEO transitions. Now they are actively planning CEO successions—companies are looking to build on the stability of a stronger economy and move forward with needed changes,” said Gary L. Neilson, Booz & Company senior partner and coauthor of the study. “Planned turnovers are at the highest rate ever, and insider CEOs make up the majority of new CEOs, indicating companies are taking a more thoughtful approach to ensure the right leaders are in place.” (Booz and Company Chief Executive Study Press Release, 2012)
3 far as CEO departures are concerned. Only Cools and van Praag (2007) tested the effect on stock prices of CEO departures of listed companies with an “insider” system. They used a sample consisting of Dutch companies, during the period between 1991 and 2000. Cools and van Praag (2007) investigated the consequences of forced departures and concluded that forced departures caused volatility in stock prices as shareholders value the monitoring task of the board. However, could it be that underlying reasons of forced departures, separately, could have a different impact on shareholder wealth and what would the impact be of other “unforced” reasons of CEO departures like retirement or the acceptance of a position in another company? Is shareholder wealth at stake in the event a CEO leaves for personal reasons? How do shareholders value one of the other important responsibilities of the board, namely the timely appointment of a successor? Could one argue that in the event that no successor is appointed the board has been taken by surprise and the CEO departure can be categorized as “uncontrolled”, whereas a concurrent appointment of a new CEO could be perceived by shareholders as an indication that the board is fully in control of such an event? This could indicate that shareholders do value a board managing the whole process of the replacement more than one that only manages the dismissal.
The objective of this research is to examine the impact of CEO departures on shareholder wealth, using a unique sample of executive turnovers of firms with an “insider” system listed on the Amsterdam Stock Exchange during the period of January 2000 until October 2013. Apart from this, this study will also take a closer look at the influence of the underlying reasons for CEO departures, since different reasons such as retirement, dismissal, external change of function, personal reasons, conflicts etc., may considerably vary from one and other. In addition, this study will investigate if shareholders value the monitoring task of the board by comparing the impact of forced and unforced departures on stock prices. Last but not least, this study will also investigate if shareholders value the hiring task of the board, by testing the impact of a concurrent appointment of a successor of the
departing CEO.
The following research questions were developed:
1. What is the effect of the various reasons of a CEO departure on shareholder value?
4
II. LITERATURE REVIEW
In the literature several authors have examined the contracts of CEO turnovers of large, listed companies. Evidence showed that their terms averaged eight to ten years and that term is generally not specified in advance except for a provision for mandatory retirement (Weisbach, 1995). It has been shown by many researchers that firm performance is generally worse than average prior to a CEO turnover. Therefore it can be said that executives are often fired because of poor performance. The hiring and firing of senior managers, such as the CEO, is one of the major functions of the board of directors, and often a consequence of monitoring by the board. Presumably, the board will appoint who they believe is the best person to run and lead the company as well as to represent the company with various stakeholder groups (Davidson et al., 2006). How the board makes these decisions and the related response of shareholders is an important subject in the event of a CEO departure. In the world different corporate governance structures exist. The next paragraphs will point out differences in board structures between Continental European countries and Anglo-Saxon countries, and will describe how board’s tasks related to CEO departures, are performed in different ways as a consequence of market differences in corporate control. In the end of this chapter an overview is given of empirical results of related studies.
2.1 Characteristics of Board Structures; One-Tier Board Structure vs Two-Tier Board Structure
5 and/or institutional investors. The management board (the lower layer) is usually composed of executive managing directors. It is normally not accepted by corporation laws that corporate statutes foresee in the possibility that directors combine the CEO and chairman roles in two-tier boards. Because the CEO has no seat in the supervisory board, its board leadership structure is formally independent from the executive function of the board. This is the case in two-tier boards in the Netherlands and Germany in particular (Maassen, 2002). In variants of the two-tier boards in these countries, executive managing directors are not entitled to have a position in the supervisory board of the corporation. In the two-tier board-model the main tasks of the supervisory board are to appoint and dismiss the members of the management board and to monitor them; the supervisory board acts in the interest of the shareholders (Jungmann, 2006). One could argue that shareholders in Anglo-Saxon countries are more directly involved in major board decisions like hiring and firing of CEOs, as in a Continental European system the supervisory board of directors plays a more proactive role in these matters. Therefore, one could say that an announcement of a CEO departure in an insider system may be more of a surprise to shareholders than in an outsider system.
2.2 The Monitoring Task of the Board
6 When CEOs are replaced due to monitoring, considering that monitoring maximizes the value of a corporation, shareholders can interpret this as a signal that the value of the particular corporation will increase in the near future. Therefore, a positive effect in abnormal returns, because of the announcement of a succession in the executive team, can be expected. Denis and Denis (1995) argue that if internal control mechanisms are effective, there should be i) a greater incidence of top management changes in poorly performing firms, and ii) improvements in firm performance following management changes. This consideration should be stronger for firms with a greater number of independent outsiders on the board of directors (Weisbach, 1988). A larger proportion of outsiders is expected to enhance the board’s ability to monitor management actions and performance as the probability of top management colluding and expropriating shareholder wealth might be lowered, and the viability of the board as a market-induced mechanism for low-cost internal transfer of control might be improved (Furtado and Kardan, 1990).
Besides a board with outsiders, economic theory points to four other parties disciplining the management of poorly performing firms: holders of large share blocks, acquirers of new blocks, bidders in takeovers, and investors during periods of financial distress (Franks et al., 2001). In any country of the world there are market differences in corporate governance and the way of monitoring. Franks and Mayer (2001) draw a distinction between “insider” and “outsider” systems. Anglo-Saxon countries have “outsider” systems of corporate control with large equity markets, dispersed ownership, one-tier board structures and active markets in corporate control. In contrast a majority of Continental European Capital Markets have “insider” systems with small numbers of quoted companies, concentrated share ownership, two-tier board structures and comparatively low levels of takeover activity (Franks and Mayer, 2001). There is an active takeover market in the USA and the UK, which affects CEO turnover. Top management’s concern with the constant threat of a takeover is evidenced by the numerous efforts taken by incumbent managers to protect target firms against takeover. CEO’s strive for high share prices to discourage a takeover and to safeguard their own position (Furtado and Karan, 1990). In countries with an insider system defense mechanisms make a takeover more difficult. For example, the Dutch situation offers companies numerous possibilities of defense many of which do not exist in the USA and the UK. These include (i) legal measures, such as the creation of structure companies; (ii) statuary measures such as issuing preferred defense shares, issuing priority shares, making binding appointments for directors and limiting voting power per shareholder; and (iii) non-statutory measures such as the issue of depository receipts of shares (Kabir et al., 1997). Although the Netherlands is modernizing its company law since 2004 with the acceptance of the “code-Tabaksblat”, defenses still occur (Quinn, 2004). In 2013, for example, the Mexican company América
Móvil could not acquire KPN as they were blocked by “Stichting Preferente Aandelen B KPN” who
7 external control mechanisms and shareholders are relatively well protected by the legal system; “insider” systems mainly rely on internal monitoring. The national market differences in corporate control and the way of monitoring might present differences in the reaction of shareholders to CEO departures between countries. Fewer defense mechanisms lead to more focus of a CEO on short-term value creation. In an outsider system, a new CEO will be highly motivated to optimize the shareholder value from day one, whereas new CEO in an insider system will be inclined to focus on value creation in a broader time frame.
2.3 The Appointing Task of the Board
It is the task of the board to decide when a turnover takes place and who the successor will be. In the literature several authors discussed the internal process of CEO successions. Most of them used the relay succession and horse race succession theory of Vancil (1987) as the basis for their study. According to Vancil (1987) in relay successions an inheritor serves a training period until it is the right time to step in and to become the new CEO. During a horse race succession there will be a competition between internal candidates, battling for the top spot (Vancil, 1987). In both, relay and horse race succession the successor is promoted from a position below the CEO in the corporate hierarchy (Davidson et al., 2006). Another type of succession occurs when a board hires a CEO from outside the company. Chan (1996) indicates that only when the need arises or when a clearly superior outside candidate is available, would succession-planning rules be violated and an outsider would be hired. Outside successions may take place in times of stress, such as poor performance, and the firm needs to show that it is changing directions and bringing in an outsider to initiate changes (Shen and Cannella, 2002). Succession rules depend on the situation and include various types of internal and external turnovers. In these cases boards have had time to plan the succession process and have chosen the most suitable candidate (Davidson et al., 2006). However, boards cannot always comply with succession rules, in situations requiring an immediate and unanticipated succession for example. When a CEO suddenly leaves and goes to another company or gets ill it may not be possible for a specific board to follow the normal rules of succession to search for either an internal or an external candidate. In these cases, the board may look for a candidate with the experience and knowledge to step into the position and assume the reins of command (Davidson et al, 2006).
8 into account the rules of successions, described above, one could say that when firms announce a CEO departure and publicize his or her successor at the same time, firms show shareholders that they have complied with the succession rules. In other words, the company signals the market that the board has the situation under control. Taking this into consideration, one could argue that a CEO departure is “controlled” by the board when the announcement includes the disclosure of the successor and a CEO departure is “uncontrolled” by the board when a successor is still not known.
2.4 Corporate Governance in the Netherlands
Schnyder (2012) suggests that the Netherlands has traditionally insider-dominated corporate governance regimes: control is exercised mainly by the insiders of the firm; capital market pressures on firms are low and outside shareholders’ interests are subordinated to other interests. Schnyder (2012) found an average ownership concentration of 24 per cent among the 250 largest Dutch companies. In addition, Schnyder (2012) found a blockholder holding 10 per cent or more of the capital in 46.2 per cent of the Dutch firms. Ownership concentration in the Netherlands is higher than in average Anglo-Saxon countries, but lower than other European Continental countries, such as Sweden and Switzerland (Schnyder, 2012). According to Hooghiemstra (2011) the Dutch corporate governance system has three important characteristics. Firstly, all Dutch public firms have a two-tier structure consisting of a management board and a separate and independently operating supervisory board. The second characteristic of the Dutch system is its stakeholder orientation, meaning that the supervisory board has to continuously weigh the interests of all stakeholders. The third and final characteristic is that investors’ rights are only weakly protected. The defence mechanisms in the Netherlands, as described earlier, limit the influence of shareholders. For example, while holders of certificates are entitled to receive dividends and can attend the annual general meeting of shareholders, they cannot vote at the meeting, as the voting rights are held by a separate foundation, which is known as the trust office. The trust office is loyal, in principle, to the company and its existing management (Hooghiemstra, 2011).
2.5 Various Reasons for CEO Departures
9 executive; his death; his departure to take up a post with another firm; and his dismissal by the board of directors (Dedman and Lin, 2002). In the literature several authors examined different reasons for CEO departures and defined those reasons in different ways. Cools and van Praag (2007) argue that if monitoring were serving shareholders’ interests, the market would value forced management departures, resulting in positive abnormal returns subsequent to the announcement of an executive’s dismissal. Cools and van Praag (2007) specified departures as forced whenever the announcement states a departure to be due to: (i) conflicts with other management board members, (ii) conflicts with the supervisory board causing the supervisory board to initiate the executive’s departure, (iii) bad performance or inadequate management, and (iv) a scandal, i.e. externally (often press) initiated “disgraceful” events (e.g. illegitimate insider trading). Moreover, they identified departures of which the business press and/or analyst reports speculated that they were forced, though the official, published motive was either “personal” or “not published”. In addition, Cools and van Praag (2007) considered their samples of departures due to health and death as a benchmark, since they argue that these departures are obviously not due to monitoring in any way. Dedman and Lin (2002) classify departures as being due to (i) retirement, (ii) succession, where a CEO succeeds to the position of Chairman or deputy Chairman of the board, (iii) dismissal, where press articles strongly suggest the CEO was forced to quit, (iv) resignation, where there was no information given, other than that the CEO had resigned, (v) split, where a joint CEO Chairman had relinquished the post of CEO (vi) death or illness of the CEO and (vii) unclassified, when it is impossible to find any articles pertaining to the event. Friedman and Singh (1989) stated that in general, the most common reason for CEO departures is pension or retirement.
2.6 The Relationship Between CEO Turnover and Stock Market Reaction
10 about the loss of valuable human specific capital. Their main prediction is that if the board members act in the shareholders’ interest then an announcement for a change in the top management team initiated by the board, through dismissal, should be met positively by the market (by an increase in stock price) or at least there should be no decrease in shareholder wealth. After conducting an event study on 323 US companies, they found that management turnover on average is associated with a small but significant increase in shareholder wealth (Furtado and Rozeff, 1987). Furthermore, they concluded that when a pressured dismissal is announced, shareholders wealth tends to increase (positive abnormal returns) around the announcement day, which they claim is consistent with the notion that the board acted in accordance to shareholder value maximization. According to Furtado and Rozeff (1987), the positive abnormal returns, immediately after an unexpected dismissal announcement, can also be explained by the fact that the market is already aware of the poor performance and this unexpected dismissal is seen as a positive sign.
Contrary to the study of Furtado and Rozeff (1987), Beatty and Zajac (1987), who analysed CEO changes in 209 US firms during the period between January 1979 and December 1980, found that CEO changes are typically associated with a reduction in firm value as reflected by the response from the stock market. Beatty and Zajac (1987) concluded that shareholders interpret the announcement of a CEO change as a signal of instability in an organization. According to Beatty and Zajac (1987) the announcement of a CEO departure should be seen and managed as a strategic event and firms should not only concentrate on smooth transition of management, but should also devote efforts to announce these intentions as clearly as possible: it is the investment community’s perception of the signal that is most important and not only the type of signal sent (Beatty & Zajac, 1987). A number of papers found no significant relationship between turnover and performance before using different specifications of the data samples. Warner et al. (1988), who looked at 351 US firms, found an inverse relation between the probability of a management change and a firm’s share performance. This relation could arise from monitoring by the board, other top managers or block holders. However, unless share performance is extremely good or bad, the logit models of Warner et al. (1988) had no predictive ability. In other words, Warner et al. (1988) could not predict the average stock price reaction due to the announcement of a top management change. Reinganum (1985), after investigating CEO turnovers by using a sample of US firms listed on the New York and American stock exchanges, did not find a statistically significant effect either. Furtado and Karan (1990) gave a possible explanation for these insignificant results. They claimed that a major problem of event studies is that the turnover announcements might be contaminated by other news, such as earnings performance, which might bias the results.
11 Cools and van Praag (2007) showed that forced departures of CEOs lead to an increase as well as a decrease in share prices. Cools and van Praag (2007) investigated the influence of 343 executives’ departures in the Netherlands during the period 1991 to 2000, announced by “Het Financieele Dagblad”. They proved that trading volume effects are great for forced departures and showed that events, which were unrelated to monitoring, like death and health related turnovers, did not generate significant abnormal trading volumes. The outcomes of Cools and van Praag (2007) were consistent with their expectations that shareholders react to a decision made by a monitoring party. However, Cools and van Praag (2007) did mention that measuring the announcement effects of forced departures is complicated since bad performance and the like is seldom mentioned in press releases as a reason for executive departure. In contrast to the findings of Cools and van Praag (2007) and Furtado and Roseff (1987), Deadman and Lin (2002) found evidence that market responses were negative when a dismissal of an executive operator is announced. They examined the market reaction to CEO departure announcements for UK firms included in the Financial Times All Share Index during 1990-1995. The different results of Cools and van Praag (2007), Furtado and Roseff (1987) and Deadman and Lin (2002), who investigated the announcement of forced CEO departures in different countries, namely the Netherlands, the United States and the United Kingdom respectively, indicate that the reaction of shareholders depends on the firm’s market. This may be due to differences in corporate control and the
way of monitoring.
A number of authors investigated various reasons for CEO departures. Johnson et al. (1985) examined the influence of a sudden death of a senior corporate executive on share price performance. Using a sample of 53 events, they found an insignificant stock price reaction to an unexpected death of a CEO. Friedman and Singh (1989) stated that CEO departures due to retirement are the most common. However, gathering responses from the senior-ranking human resources officer in Fortune 500 firms, they found that customary retirements are associated with no significant stockholder reaction. Friedman and Singh (1989) concluded that announcements of retirement were unlikely to create much disturbance. They argued that customary retirements were generally orderly, smooth transitions that involve successors well known to incumbent management. Neither the CEO nor the board takes a specific action to make a change in such a case; the level of discretion available to the successor is likely to be low, CEOs appointed after routine retirements arrive to find relatively inert and stable organizations (Friedman and Singh, 1989). Denis and Denis (1995) computed abnormal stock returns of US firms over a two-day period including the day of and the day prior to the turnover announcement. They found results consistent to Friedman and Singh (1989). Denis and Denis (1995) stated that normal retirements were not preceded by significant changes in operating performance, but are followed by small increases in operating income. Denis and Denis (1995) documented that these small increases were driven by firms that are subsequently acquired.
12 Mahajan and Lumer (1993), offered an alternative point of view on the impact of CEO turnovers on stock prices, adding voluntary resignation to their research. They analysed how the market accepts an announcement of a change of senior management of a company by observing the direction and magnitude of changes in stock prices. To do so, they divided their sample into five categories, depending on how an executive lost his position: firing, voluntary resignation, death, retirement or reshuffle. After conducting an event study they concluded that in general a succession does not have an effect on shareholder wealth. After splitting the sample in several subsamples depending on the reason for departure, they obtained results that suggest that when there is an announcement of a CEO departing from the firm after a voluntary resignation, the excess return is negative, but when the executive retains an advisory role (but has no discretionary power) there is a positive effect. This implies that CEO departure (after resignation) affects negatively firm value. The authors explained these results with the fact that when people with firm specific capital leave, this causes distress in the company. A CEO departure can create instability in the company, which affects stockholders’ wealth in a negative way.
Davidson et al. (2006) documented that in times when an immediate succession decision is needed because a key executive becomes ill, the normal rules of succession are not always followed and must be adapted to the situation. They examined a sample of 113 stories in the US, which indicated that a CEO or Chair had become ill or injured over the period between 1992 and 2002 and found that stock markets reacted positively to former-CEO succession announcement. Davidson et al. (2006) pointed out that none of the boards in their sample selected an outside candidate as successor. When CEOs become ill, they are more likely to be replaced by a board member that is a former CEO than would be the case under more normal succession decisions (Davidson et al., 2006). Thus, the horse-race succession processes are less likely to occur than in ordinary circumstances because the CEO’s illness makes an immediate decision necessary. In the end, Davidson et al. (2006) argued that the immediacy and urgency of the succession decision and the market reaction they found suggested that the market
responded positively to a board decision that is consistent with the functions of the rules.
2.7 Hypotheses
13 with an “outsider” system have mostly been researched; this study will make use of a sample of firms with an “insider” system, as defined by Franks and Mayer (2001). This study will examine departure announcements of CEOs of Dutch companies listed on the Amsterdam Stock Exchange. The effect on stock price performance due to announcements of CEO departures in general is hard to predict, because of the various results of different authors. Only Cools and van Praag (2007) examined executive successions in the Netherlands. As described earlier, they proved that trading volume effects are significant for forced departures, whereas CEO departures due to health and death, which are unrelated to monitoring, have no influence on stock prices. Furthermore, it is mentioned that both Friedman and Singh (1989) and Denis and Denis (1995) argued that announcements of CEO retirements are unlikely to create much disturbance. Since in general forced departures are in the minority, and retirements are the most common reason for CEO departures, the expectation of this study, using a unique sample of listed Dutch firms in the period from 2000 till 2013, is that an announcement of a CEO departure in general has no influence on shareholder value. This expectation is in line with the conclusions of Warner et al. (1988) and Reinganum (1985). The following hypothesis was devised:
H1: The announcement of a CEO departure of firms listed on the Amsterdam Stock Exchange has
no influence on shareholder wealth.
Differences in the results of related studies may also occur since different reasons of management departures can generate different price effects. This study will take a closer look at the various reasons for CEO departures in the Netherlands and will investigate if the various reasons for the exit of a CEO have a different influence on shareholder wealth. For example, Cools and van Praag (2007) identified CEO departures due to poor performance or a conflict as forced reasons to leave a company. However, those two reasons can, separately, have a different impact on share prices. On the other hand a CEO departure due to retirement can have a different effect on shareholder wealth than a departure due to an external change of function of the particular CEO, which is unrelated to the monitoring task of the board. Therefore, the sample of this research will be split into different reasons, based on the number of events, since the subsamples must be large enough to draw valid conclusions. The various reasons will be tested separately.
14 H2: The announcement of a CEO departure of firms listed on the Amsterdam Stock Exchange, due
to poor performance, has a positive influence on shareholder wealth.
H3: The announcement of a CEO departure of firms listed on the Amsterdam Stock Exchange, due to a conflict, has a positive influence on shareholder wealth.
Furthermore it can be expected, based on the results of Cools and van Praag (2007) that all other reasons of CEO departures have less influence on shareholder wealth. Cools and van Praag (2007) stated that CEO departures due to health problems or death, which in this study will be allocated to CEO departures due to personal reasons, have no impact on stock prices and both Friedman and Singh (1989) and Denis and Denis (1995) did not find significant changes in operating performance by the announcement of CEO departures due to retirement. As a CEO departure due to an external change of function is also unrelated to the monitoring task of the board, it is expected that the acceptance of a position outside the company will not lead to a significant change in stock price either. The following hypotheses will be tested:
H4: The announcement of a CEO departure of firms listed on the Amsterdam Stock Exchange, due
to retirement, has no influence on shareholder wealth.
H5: The announcement of a CEO departure of firms listed on the Amsterdam Stock Exchange, due
to an external change of function, has no influence on shareholder wealth. H6: The announcement of a CEO departure of firms listed on the Amsterdam Stock Exchange, due
to personal reason, has no influence on shareholder wealth.
Next to this, it will be tested if a difference exists between ‘’forced” reasons (conflict and poor performance), and “unforced reasons” (retirement, personal reasons, and external change of function) to test if shareholders in an “insider” system consider the monitoring task of the supervisory board as value relevant:
H7: Shareholders react differently to the announcement of forced departures in comparison to the announcement of unforced departures.
15 successors, into consideration as well. This study will test if “uncontrolled” CEO departures have a different effect on stock prices compared to “controlled” departures. Where an event will be classified as uncontrolled when the successor is unknown and will not be presented simultaneously with the announcement of a CEO departure and where the opposite applies for “controlled” departures, which means that the publication of a new CEO is announced at the same time as the departure of his or her predecessor. The impact of the announcement of a successor could be different for the various reasons of CEO departures. Therefore this study will test the impact of both uncontrolled departure and controlled departure per individual reason as well. Taken into consideration the findings of Davidson et al. (2006) it can be expected that shareholders react differently towards an uncontrolled event, since they could assume that the board is not managing the whole process, than towards a controlled event, as in this case the board should have control over the situation. Therefore, the last hypothesis is stated as follows:
H8: Shareholders react differently to the announcement of uncontrolled CEO departures in
comparison to the announcement of controlled CEO departures.
III. METHODOLOGY
To measure the effects of various CEO departures on stock prices, this study will perform univariate tests, using the event study methodology of Brown and Warner (1980, 1985) and MacKinlay (1997). For the total sample and each subsample the presence of abnormal returns will be tested in determined event windows. To examine these abnormal returns, Brown and Warner (1980, 1985) make use of three kinds of models: The Mean Adjusted Returns Model, The Market Adjusted Returns Model and the Market and Risk Adjusted Returns Model, which are described below.
To implement these three models, actual returns must be calculated first, using continuously compounding returns:
R
i,t=
LN(
P
i,t+1P
i,t)
(1)16 3.1 Mean Adjusted Model
The Mean Adjusted Model is consistent with the Capital Asset Pricing model; under the assumption that a security has constant systematic risk and that the efficient frontier is stationary (Brown and Warner, 1980). This model focuses on the returns to each sample security around the time of its event. It will examine whether or not the returns on the sample securities in the event window significantly differ from the returns on the securities in the time period of the estimation window (Brown and Warner, 1980). In this model, the abnormal returns are calculated by subtracting the predicted returns from the observed returns:
)
(
, , ,t it it iR
E
R
AR
(2)The predicted return can be determined by calculating the average of the estimation window.
3.2 Market Adjusted Model
The Market Adjusted Model is consistent with the Capital Asset Pricing Model only if all securities have systematic risk of unity (Brown and Warner, 1985). The model assumes that the actual returns move in the same way as the market wide movements. The abnormal returns, in this model, are calculated by deducting the actual market returns from the actual returns:
(3)
Where is the stock return of the particular market at the same time t as the sample firms experienced events.
3.3 Market and Risk Adjusted Model
The most accurate model of the three is the Market and Risk Adjusted model. This model calculates the abnormal return, using an estimation of the Alpha and Beta of the certain stock.
Beta displays the correlation between stock i and the market, and can be determined by dividing the covariance between stock i and the stock return of the market by the variance of stock i:
17 (4) Alpha estimates the volatility of stock i, which is caused by factors other than the market. Alpha can be calculated as follow:
(5)
Where is determined on the basis of the estimation window.
3.4 Student T-test, Non-Parametric Z-test and Cross-Regional Regression Analysis
When the abnormal returns of the different models are calculated, two tests can be used to define whether or not the abnormal returns are significant. Brown and Warner (1980, 1985) make a distinction between a parametric test and a non-parametric test. For the parametric test, the Student T-test can be performed. The Student T-T-test finds out if the Average Abnormal Returns (AAR’s), in the event window, significantly differ from zero. The “AAR’s” can be calculated by dividing the total amount of the Abnormal Returns by the number of events N:
(6)
The Student T-test can also be used to test the Cumulative Average Abnormal returns for significance. “CAAR’s” are calculated to look for persistency, to determine whether the effect of the announcement of a particular event can be recovered or to see if markets response before the announcement of an event is made, which means that information is drained.
CAAR’s can be calculated as follows:
(7)
The student’s t-test statistic relies on the assumption that the abnormal returns follow a normal distribution. According to Serra (2002) previous event studies have shown that abnormal returns distributions are not normally distributed, because the sample data shows fat tails and is right skewed. Parametric tests too often reject hypotheses when testing for positive abnormal performance and too seldom reject hypotheses when testing for negative abnormal performance. When the assumption of
i t i i Var Cov, ˆ
i i i
C
ˆ ˆ i
AARi,t ARi,t i1 N
N
18 normality of abnormal returns is violated, parametric tests are not well specified. Non-parametric tests are well specified and more powerful at detecting a false null hypothesis in abnormal returns (Serra, 2002). When the Jarque-Bera test of the sample data indicates values higher than 5,99 the sample data does not come from a normal distribution and a non-parametric test must be carried out (Brown and Warner, 1985). Brown and Warner (1985) indicate that the non-parametric Z-test is a valid test to investigate abnormal returns, when a sample deals with a high skewness or kurtosis. Therefore, in this study for each subsample that displays a Jarque-Bera value higher than the critical value a parametric Z-test will be performed.
In examining the Hypotheses, the outcomes of the parametric Student T-test or the non-parametric Z-test will be discussed to decide whether or not the hypotheses will be rejected. According to Brown and Warner (1980,1985) and McKinlay (1997) the outcomes can be tested at a significance level of 1%, 5% and 10%; where a significance level of 1% indicates the most reliable results.
To make univariate results more convincing a cross-regional regression analysis can be implemented to explain why CAAR’s of different companies after a similar event vary between each other. Thus, showing why, in this study, the impact of various CEO departures is greater for specific companies in comparison with other companies. A cross-regional regression will test if independent variables of stock i have influence on the dependent variable using the following equation:
(8) Where represents the constant term and symbolizes the error term of the regression equation. For this study the independent dummy variables (i) Forced Departures, (ii) Uncontrolled Departures will be conducted and will be used to test whether shareholders value both the monitoring task as well as the appointment task of successors of the board. As Weisbach (1988) and Furtado and Kardan (1990) considered that improvements in firm performance following management changes is related to firms with a greater number of independent outsiders on the board of directors, this study will use the number of members on the supervisory board as a control variable1.
19
IV. DATA
This study investigates the impact of CEO turnovers on shareholder wealth in the Netherlands. To receive valid outcomes it is necessary to have a large database. The study of Cools and van Praag (2007) is, in the literature, the only study which tested the influence of executive’s departures in the Netherlands. Their sample consists of the 100 largest firms listed on the Amsterdam Stock Exchange (AEX) at year-end 1999, thereby representing more than 80% of the market value of Dutch listed firms (Cools and van Praag, 2007). The annual reports of their sample revealed that 86 firms experienced at least one executive departure in the period between January 1, 1991 and January 1, 2000. In the end, their sample of events consisted of the 227 executive departures that these 86 firms experienced (Cools and van Praag, 2007). Cools and van Praag (2007) identified executive departures by comparing the names of board members in consecutive annual reports. They found exact announcement days in “Het Financieele Dagblad”, which is the Dutch equivalent of the Wall Street Journal.
20
Figure 1. Overview of Number of CEO Departures at Companies Listed on the Amsterdam Stock Exchange in the period 2000-2013
After eliminating (i) events in which contaminating information was announced during the event window, (ii) events of which the announcement date or reason of departure was not clear, (iii) events of which exchange rates of the particular firm could not be received by Datastream and (iv) events, which appeared because of a merger or acquisition, a total sample of 210 CEO departures in the Netherlands was achieved (see table 1.). Figure 1 displays the number of CEO departures per year. It appears that most CEOs have left their company in the years 2007, 2008 and 2009. This may be the result of the financial crisis, which reached its peak in these years. Obviously, the announcement of a CEO departure due to a merger or acquisition is not taken into account, since the merger or acquisition itself would have a huge impact on stock prices and a CEO departure is normal in these kinds of happenings.
Table 1. Total Sample CEO Departures in the Netherlands
Retirement 63
Poor Performance 44
Conflict 30
Personal Reasons 24
External Change of Function 23
Reorganization 9
Interim 9
Internal Change of Function 8
21 From table 1 it can be seen that, in this sample, retirement is the most common reason for a CEO to leave a company. This is in line with the findings of Vancil (1987), Denis and Denis (1995), Friedman and Singh (1989) and Cools and van Praag (2007). A closer look at the influence of underlying reasons of CEO departures on shareholder wealth will be taken by looking for abnormal returns after the announcement of a CEO departure due to (i) retirement, (ii) poor performance, (iii) conflict, (iv) personal reasons, and (v) external change of function. The last three reasons of table 1 are evaluated when looking at the total sample, but will not be investigated separately since the subsample size is too small to draw proper conclusions. Of the reasons, which will be tested separately, retirement, external change of function and personal reasons can be defined as “unforced reasons”, since a board has limited influence on these kind of departures. If a CEO retires or leaves a company for another position, most of the time this is the CEOs’ own choice. This study defines the reason for departure as personal whenever a CEO leaves the company because he has health problems, family problems, or because he died. Also in this case, a CEO does not leave the company as a consequence of a monitoring action of the board. The separate reasons; poor performance and conflict can be assigned to “forced” reasons of departure, since this is also the case in the study of Cools and van Praag (2007). In this study the “conflict” reasons for departure consist of the combination of CEO departures due to a conflict with other board members, a conflict with the supervisory board or a conflict with shareholders.
To test the effect of controlled and uncontrolled CEO departures on stock prices the total sample is divided into two categories (see table 2 and table 3). The controlled category is defined as the announcement of both the exit of a CEO and the appointment of its successor. An interim CEO is not regarded as a successor, since this can be seen as just a temporary solution. The uncontrolled category is defined as the announcement of CEO departure without the appointment of its successor at the same time.
Table 2. Controlled CEO Departures in the Netherlands Table 3. Uncontrolled CEO Departures in the Netherlands
Retirement 54 Poor Performance 19 Conflict 13 Personal Reasons 8
External Change of Function 12
Reorganization 6
Internal Change of Function 6
Total 118
Retirement 9
Poor Performance 25
Conflict 17
Personal Reasons 16
External Change of Function 11
Reorganization 3
Internal Change of Function 2
22 From table 2 and table 3 one can conclude that the board controls 59 percent of the CEO departures, since a successor was announced. Hence, 41 percent of the announcements of CEO exits did not include the appointment of a successor and are therefore defined as uncontrolled. It is striking to see that in the category “uncontrolled” exits, a substantial number consist of the unforced reasons; retirement, personal reasons, and external change of function, making this worthwhile to investigate further. When there is a decent relationship between the board and a CEO, one could expect that a board would be informed to identify and appoint a successor in time.
According to Mackinlay (1997) the most common choice to define the estimation window, when feasible, is using the period prior to the event window. In this research the parameters are estimated at over 200 days prior to the event window, which is in accordance with the study of Cools and van Praag (2007). Table 4 summarizes the event results of the sample of 210 CEO departures within the 200 days of the estimation period. From the table it can be derived that the average returns have the skewness and kurtosis matching a normal distribution for the total sample as well as most of the subsamples. It can be noticed that only for the subsample (i) poor performance, the Jarque-Bera value exceeds the critical value. Therefore, for this subsamples a non-parametric Z-test will be performed.
Table 4. Descriptive Analysis of Average Returns of the Total Sample and Subsamples within Estimation Window [-210;-10].
Variables Obs. Mean Median SD Minimum Maximum Kurtosis Skewness Jarque-Bera CEO Departures 210 0,0000 0,0000 0,0024 -0,0077 0,0064 0,3949 -0,1874 2,5932 Retirement 63 0,0000 0,0000 0,0026 -0,0083 0,0059 -0,1338 -0,1428 0,2612 Change of Job 23 0,0000 -0,0004 0,0070 -0,0204 0,0280 1,5352 0,3828 2,8201 Personal Reason 24 0,0000 0,0001 0,0081 -0,0212 0,0232 -0,1241 -0,1864 0,1544 Poor Performance 44 0,0000 0,0003 0,0060 -0,0357 0,0283 9,1962 -1,0011 7,6128 Conflict 30 0,0000 -0,0002 0,0063 -0,0180 0,0207 0,7573 0,1975 0,9119
The choice of event windows for this study is related to the study of Cools and van Praag (2007). Like Cools and van Praag (2007) this study will utilize four event windows to look for abnormal returns, namely:
1. [0]: Day “0” is the day of the announcement. Market efficiency suggests this to be the relevant event window.
2. [-10; -1]: This event window is chosen to trace possible information leakage before press releases.
23
V.
Results
In this study, the three models of Brown and Warner (1980,1985), namely the Mean Adjusted model, the Market Adjusted Model, and the Market and Risk Adjusted model, have been applied to determine abnormal returns during the event of an announcement of a CEO departure. Subsamples were made to investigate if the market reacted differently to the underlying reasons of a CEO departure. In addition, regression analyses have been made to establish the impact of the decision taken by the board to let a CEO go as well as the subsequent appointment of a successor. Outcomes of the Student T-test and non-parametric Z-test and the results of the regressions will be discussed in the next paragraphs.
5.1 Market Reaction to the Announcement of a CEO Departure in the Netherlands
Table 5 displays the calculated average abnormal returns as well as the cumulative average abnormal returns of the total sample in the event period [-10; 2]. As can be noticed the event day “0” is surrounded by negative returns. Because the three different methods of Brown and Warner (1980, 1985) all show similar results, the remaining research will focus on the Market and Risk Adjusted method only, since this method, as described in the methodology section, is regarded as the most accurate one.
Table 5. Average Abnormal Returns (AAR) and Cumulative Average Abnormal Returns (CAAR) in event period [-10; 2] for the Three Methods of Brown and Warner (1980, 1985)
Mean Adjusted Method Market Adjusted Method Market and Risk Adjusted Method
Cumulative Cumulative Cumulative
Average Average Average Average Average Average Abnormal Abnormal Abnormal Abnormal Abnormal Abnormal Event Day Return (AAR) Return (CAAR) Return (AAR) Return (CAAR) Return (AAR) Return (CAAR)
24 Table 6 displays the calculated average abnormal returns and the p-values of the total sample in the event period [-10; 2]. In addition, it shows the cumulative average abnormal returns with the associated p-values for the event windows [-10; -1], [-1; 0], [0] and [0; 2]. Although the data of the total sample have the skewness and kurtosis matching a normal distribution, the non-parametric Z-test was performed as an additional robustness check. Figure 2 shows a clear downward trend, resulting in a cumulative decrease of -2,5%. From table 6 it can be acknowledged that both event window [-10; -1] as well as event window [0; 2] display significant values. Event window [-10; 1] shows a cumulative average abnormal return of -1,52% at a significant level of 1%. This indicates that the stock market is reacting in advance of the announcement and that the upcoming announcement is already known to and anticipated by certain parties. Event window [0; 2] shows a cumulative abnormal return of -1,01% at a significant level of 10%, suggesting that the market is also reacting after the announcement has been made. Furtado and Rozeff (1987) found that management turnover on average is associated with a small but significant increase in shareholder wealth; Warner et al. (1988) and Reinganum (1985) did not detect any average stock price reaction as a result of the announcement of a top management change. The significant cumulative abnormal returns in the event windows [-10; -1] and [0; 2] show that announcements of CEO departures in the Netherlands do affect stock prices in a negative way. Therefore, Hypothesis 1, stating that the announcement of a CEO departure of firms listed on the Amsterdam Stock Exchange has no influence on shareholder wealth can be rejected. The outcomes displayed in table 6 are consistent with the study of Beatty and Zajac (1987) and can be explained by the fear that management changes may lead to temporary instability in organizations, which is detrimental to performance. The different outcome in comparison with the studies of Warner et al. (1988) and Reinganum (1985) may occur since different reasons for management departures can generate different price effects. The next paragraph will point out if the underlying reasons of CEO departures, separately, have a different impact on shareholder wealth.
Figure 2. Cumulative Average Abnormal Returns due to CEO Departures in event period [-10; 2): Total Sample
-0,03 -0,025 -0,02 -0,015 -0,01 -0,005 0 0,005 0,01 -10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2
25
Table 6. Average Abnormal Returns and Results of Student T-Test and Z-test of the Total Sample within Event Window [-10; 2]; Cumulative Average Abnormal Returns and Results of Student T-Test and Z-test of the Total Sample within Event Window [-10; 1], [-1; 0], and [0; 2]
Market And Risk Adjusted Method
Event Day AAR Student T-Test P-value Z-test P-value
-10 0,0032 0,0402 ** 0,0384 ** -9 -0,0060 0,0106 ** 0,0096 *** -8 -0,0013 0,4961 -0,5040 -7 0,0001 0,3391 0,2676 -6 0,0010 0,4985 0,5000 -5 -0,0045 0,0299 ** 0,0301 ** -4 -0,0001 0,4752 0,3859 -3 -0,0068 0,0022 *** 0,0064 *** -2 -0,0007 0,3884 0,3783 -1 -0,0027 0,1290 0,1112 0 -0,0029 0,1155 0,1539 1 -0,0015 0,2707 0,2810 2 -0,0058 0,0079 *** 0,0129 **
Event Period CAAR Student T-test P-value Z-test P-value [-10; -1] -0,0152 0,0073 *** 0,0069 ***
[-1; 0] -0,0056 0,1843 0,1841
[0] -0,0029 0,1155 0,1539
[0; 2] -0,0101 0,0515 * 0,0516 *
Note: ***= significant at 1%, **= significant at 5%, *= significant at 10%
5.2 Market Reaction to Various Reasons of CEO Departures in the Netherlands
To determine if the underlying reasons for CEO departures have a different impact on shareholder wealth subsamples of the total sample were made for CEO departures as a result of (i) retirements, (ii) external changes of function, (iii) personal reasons, (iv) poor performances, (v) conflicts. The cumulative average abnormal returns and the associated p-values of each subsample are shown in table 9 up to table 13 in section 8.1 of the appendix.
26 extent also applicable for retirements in the Dutch situation. Table 10 shows the average abnormal returns and p-values in the event a CEO departure found a new job. It can be concluded that the announcement of a CEO departure due to an external change of function has, as expected, no influence on share price performance as no significant returns in the various event windows are found. Therefore, Hypothesis 5 is accepted. Also, from table 11 it can be discerned that the announcement of a CEO departure due to personal reasons has no impact on stock prices. This means that hypothesis 6 is accepted as well. This outcome is consistent with the study of Johnson et al. (1985), who focused on CEO departures due to sudden death of a senior corporate executive, which is defined as a personal reason in this study. It appears that CEO departures due to both external change of function and personal reasons have no influences on stock prices as they are unrelated to the monitoring task of the board as stated by Cools and van Praag (2007). Table 12 and table 13 show the impact of CEO departures as a result of poor performances and conflicts. Although it was the expectation that shareholder wealth would be enhanced, the results in table 12 and table 13 show the opposite. The subsample “poor performance” shows a highly significant average return of -2,16% in event window [0], whereas the subsample “conflict” has a negative cumulative abnormal average return of 1,82% in the event window [0; 2] at a significance level of 10%. Therefore, Hypotheses 2 and 3 are both rejected. Assuming poor performance or a conflict often leads to the dismissal of a CEO, these findings are inconsistent with the results of Furtado and Rozeff (1987), who argue that if board members act in the shareholders’ interest then an announcement for a change in top management initiated by the board, through dismissal, should be met positively by the market. The difference exists, since the announcement of a dismissal for shareholders in an insider system comes more as a surprise as they are less involved in the decision making process of the board than shareholders in an outsider system. Another explanation for the different results is that shareholders in an outsider system react positively to a new CEO as short term value creation is expected, whereas in the Netherlands, in an insider system, value creation might take more time.
27 5.3 Market Reaction to Forced and Unforced CEO Departures in the Netherlands
Cools and van Praag (2007) concluded that monitoring by the supervisory board is valued by investors as they argued that forced departures of executive directors have a significant impact on stock prices. In the previous section the univariate results of the subsamples poor performance and conflict, which are assigned to forced departures, demonstrated highly significant negative p-values. Therefore, one could expect that forced departures cause the negative effect on the total sample. This is tested by performing a cross-sectional regression analysis, using forced departures as independent dummy variable. Results of the cross-sectional regression analysis shown in table 7 confirm the expectation. The independent variable forced departures has a negative coefficient of 2,13% at a significant level of 10%. It can be concluded that CEO departures, which are assigned to forced departures, differ materially from the total sample. Shareholders react differently to forced departures in comparison to unforced departures. Hence, Hypothesis 7 is accepted since shareholders consider the monitoring task of the board as value relevant.
Table 7. Regression Results for the Total Sample; Independend Dummy Variable Forced Departures and Control Variable Supervisory Board Members within Event Window [0]
Total Panel Observations: 210
Variable Coefficient Std. Error T-Statistic Prob.
FORCED -0,0213 0,0125 -1,7111 0,0886 *
SUPERVISORYBOARD -0,0032 0,0023 -1,3994 0,1632
C 0,0189 0,0133 1,4214 0,1567
R-squared 0,0190 Mean dependent var -0,0029 Adjusted R-squared 0,0095 S.D. dependent var 0,0844 S.E. of regression 0,0840 Akaike info criterion -2,1025 Sum squared resid 1,4595 Schwarz criterion -2,0547 Log likelihood 2,2377 Hannan-Quinn criter. -2,0832 F-statistic 2,0039 Durbin-Watson stat 0,0941 Prob(F-statistic) 0,1374
28 5.4 Market Reaction to Controlled and Uncontrolled CEO Departures in the Netherlands
To test the other responsibility of the board, next to dismissal, in the process of CEO turnovers, namely the timely appointment of a successor, another cross-sectional regression was performed using uncontrolled departures as independent dummy variable. Where an event will be classified as uncontrolled when the successor is unknown and was not presented in the announcement of a CEO departure. Results of the cross-sectional regression analysis are highlighted in Table 8. As can be noticed, the uncontrolled departures show a coefficient of -2,45% at a significance level of 5%. It can be concluded that uncontrolled CEO departures are negatively received by shareholders and have a different impact on shareholder wealth than controlled departures where the appointment of a new CEO is announced at the same time as the departure of his or her predecessor. Therefore, hypothesis 8
is accepted.
Table 8. Regression Results for the Total Sample; Independend Dummy Variable Uncontrolled Departures and Control Variable Supervisory Board Members within Event Window [0]
Total Panel Observations: 210
Variable Coefficient Std. Error T-Statistic Prob.
UNCONTROLLED -0,0245 0,0122 -2,0174 0,0449 ** SUPERVISORYBOARD -0,0034 0,0023 -1,4814 0,1400
C 0,0219 0,0135 1,6204 0,1067
R-squared 0,0243 Mean dependent var -0,0029 Adjusted R-squared 0,0149 S.D. dependent var 0,0844 S.E. of regression 0,0837 Akaike info criterion -2,1080 Sum squared resid 1,4516 Schwarz criterion -2,0602 Log likelihood 2,2434 Hannan-Quinn criter. -2,0886 F-statistic 2,5779 Durbin-Watson stat 0,2022 Prob(F-statistic) 0,0784 *
Note: ***= significant at 1%, **= significant at 5%, *= significant at 10%
29 found, a CEO departure will not have many influence on shareholder wealth in the short term, because shareholders may perceive that the situation is under control. In contrast, if a board has not appointed a successor yet, shareholders may think that there are instabilities within the organization, leading to a huge drop in stock prices.
Figure 3. Cumulative Average Abnormal Return due to Controlled CEO Departures in Event Period [-10; 2]
Figure 4. Cumulative Average Abnormal Return due to Uncontrolled CEO Departures in Event Period [-10; 2]
This study proves that shareholders value the appointing task of the board. To establish if the appointing task has a different impact per underlying reason of CEO departures, a cross-sectional regression analysis, using uncontrolled departures as independent dummy variable, was also done for the subsamples (i) retirement, (ii) poor performance and (iii) conflict. The subsamples external change of function as well as personal reason were not taken into account, since these reasons of departure did not influence stock prices as they did not show any significant results. Outcomes of the regressions are highlighted in section 8.2 of the appendix. From table 14 up to table 16 it can be seen that uncontrolled departures only affect departures due to retirement. A departure due to retirement without the appointment of a successor leads to a negative coefficient of 3,12% at a significance level of 5%. This negative impact is also confirmed by figure 5, which shows a negative cumulative average abnormal return of more than 5 percent in the event period [-10; -2]. It can be concluded that companies, that are
0 0,002 0,004 0,006 0,008 0,01 -10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2
Market and Risk Adjusted Method -0,07 -0,06 -0,05 -0,04 -0,03 -0,02 -0,01 0 0,01 -10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2
30 going to announce the retirement of a CEO should always announce the CEO’s successor at the same time to signal shareholders that they have complied with the rules of succession of Vancil (1987) and that the situation is under control. On the other hand, in the case of a CEO departure due to poor performance or a conflict the appointing of a successor is less relevant. Therefore, boards should not waste time and should dismiss the CEO immediately to get out of the downward spiral.
Figure 5. Cumulative Average Abnormal Return due to CEO Departure due to Retirement, without Appointed Successor
VI. DISCUSSION AND CONCLUSION
The objective of this research was to investigate the effect of CEO departures on shareholder wealth. Especially in the United States, CEO departures at firms with an outsider monitoring system, where corporate control is linked to large equity markets, dispersed ownership, unitary board structures and active markets, have been examined. This study focused on CEO departures at listed firms in the Netherlands, which have an insider system with small numbers of quoted companies, concentrated share ownership, two-tier board structures and comparatively low levels of takeover activity (Franks and Mayer, 2001). Furtado and Rozeff (1987) found that, on average, management turnover is associated with a small increase in shareholder wealth; Warner et al. (1988) and Reinganum (1985) did not detect average stock price reaction as a result of the announcement of a top management change. This study showed that the announcements of CEO departures in the Netherlands, in general, do affect stock prices in a negative way. This is explained by the fear that management changes may lead to temporary instability in organizations. The different outcome in comparison with related studies may occur since different reasons for management departures can generate different price effects. This study has taken a closer look at the various reasons for CEO departures in the Netherlands and has proved that the various reasons for the exit of a CEO have different influences on shareholder wealth. Of the different underlying reasons, departures due to external change in function and personal reasons have no significant impact on shareholder wealth. Departures due to poor
-0,06 -0,05 -0,04 -0,03 -0,02 -0,01 0 -10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2
31 performances and conflicts do impact stock value in a negative way, whereas departures due to retirement lead to a modest increase in stock price. This study has also investigated the role of the board in the event of a CEO departure taking into consideration both the monitoring task and the appointing task of the board. This study has proved that shareholders value the monitoring task of the board as it was found that shareholders react differently to the announcement of forced departures in comparison to the announcement of unforced departures. According to Furtado and Rozeff (1987) a change in top management initiated by the board, through dismissal, should be met positively by the market as the board acts in the shareholders’ interest. In this study stock prices significantly decrease in value as a result of forced departures. The different reaction to forced departures can be explained as this research is conducted in the Netherlands with a different corporate governance and a different corporate control environment. The announcement of a dismissal comes more as a surprise to shareholders of Dutch listed companies with an insider system as shareholders are less involved in the decision making process of the board than shareholders in an outsider system. Another explanation for the different result is that shareholders in an outsider system react positively to a new CEO as short term value creation is expected, whereas in an insider system value creation might take more time. Last but not least, this study investigated the appointing task of the board and proved that shareholders interpret a CEO departure without an appointment of a new successor as a negative signal, indicating that the supervisory board has not complied with the rules of succession of Vancil (1987) and, therefore, presume instability in an organization. Regression analysis done of each subsample has shown that companies that are about to announce the retirement of a CEO, should always announce the CEO’s successor at the same time to signal shareholders that the situation is under control. In the case of a CEO departure due to poor performance or a conflict the concurrent appointment of a successor is less relevant. Therefore, boards should not waste time and should dismiss the CEO immediately to curb the downward spiral.
32
VII. REFERENCES
- Bresser, R., Valle Thielle, R., Biedermann, A., & Ludeke, H. (2005). Performance Implications on CEO Dismissals: Evidence from a Stakeholder Environment. Zeitschrift fur Betriebwirtschaft, 75, 1165-1192.
- Beatty, R., & Zajac, E. (1987). CEO Change and Firm Performance in Large Corporations: Succession Effects and Manager Effects. Strategic Management Journal, 8(4), 305-317.
- Booz and Company. (2013). Booz & Company Chief Executive Study Finds Steep Rise in Planned CEO Turnovers as Companies Take Active Control of Their CEO Succession Planning.
- Brickley, J., Smith, C.& Zimmerman, J. (2003). Managerial Economics and Organizational Architecture McGraw-Hill, 3.
- Brown, S., & Warner, J. (1985). Using Daily Stock Returns: The Case of Event Studies. The Journal of Finance Economics, 14, 3-31.
- Chan, W. (1996). External Recruitment Versus Internal Promotion. Journal of Labor Economics, 14, 555-570.
- Cools, K., & van Praag, C.M. (2007). The Value Relevance of Top Executive Departures: Evidence from the Netherlands. The Journal of Corporate Finance, 13, 721-742.
- Davidson, W.M., Tong, S., Worrell, D.L., & Rowe, W. (2006). Ignoring Rules of Successions: How the Board React to CEO Illness Announcements. Journal of Business Strategies, 23(2), 94-113.
- Dedman, E., & Lin, S.W.J. (2002). Shareholder Wealth Effects of CEO Departures: Evidence from the UK. Journal of Corporate Finance, 8, 1-104.
