IMPACT OF CHIEF EXECUTIVE OFFICER (CEO) SUCCESSION POLICY ON CEO TURNOVER ANNOUNCEMENT IN MALAYSIA

This paper presents a fresh perspective on chief executive officer (CEO) turnover, where the impact of CEO turnover on firm value is analysed based on whether the removal is planned or unplanned. A total of 146 announcements for ten years in Malaysia is examined using an event study method. The results indicate that, in general, CEO turnover announcements cause a significant reaction due to changes in the firm’s investment decisions. Specifically, a significant positive impact exists when CEO turnover occurs as planned. In a planned turnover, the negative news of the removal of the CEO is immediately minimised with the positive news of a CEO appointment, indicating the positive impact of establishing a CEO succession plan on firm value. This finding adds new knowledge to the current literature and allows policymakers to examine the establishment of a CEO succession policy.


INTRODUCTION
In the past, the efficient market hypothesis (EMH) has been widely used in empirical studies to measure announcement effects. Fama (1970) argues that "the market is efficient" and therefore, "all information is reflected in the share price". Despite this, Fama, Fisher, Jensen and Roll (1969) indicate that the share price reacts to new information that can increase a firm's future wealth. In addition, Redhead (2009) reports that an investor's decision to buy, hold or sell a share is influenced by elements such as "motives, knowledge, experience, feeling and other cognitive, emotional and social influence". A firm's decision to release new information to the market to influence an investor's decision is known as "signalling theory". Spence (1973) suggest that a firm takes advantage of exclusive new information from the chief executive officer (CEO) transition event to increase firm wealth. With reference to signalling theory, this study examines how a firm uses CEO transition events to control share price movement.
Over time, the issue of CEO transition has received considerable attention in the academic literature and consulting reports. Even though investigations of CEO transition have been conducted since the 1950s, but a heighten focus on CEO transition emerged in the 1980s (Mehrabani & Mohamad, 2011). CEO transition is a continuous process of changing leadership that comprises both CEO turnover and CEO appointment. A change of CEO is expected to have an impact on firm performances (Ishak, Ismail, & Abdullah, 2013). Often, the appointment of a new CEO is more crucial to a firm than the removal of former CEO. A new CEO has an impact on a firm's sustainability and future growth. Despite this, most findings on CEO transition focus on CEO turnover because past studies have indicated that the sudden removal of a CEO influences the sustainability of the firm's future performance, which is reflected in the volatility of the share price.
Statistics from PricewaterhouseCoopers (PWC, 2018) reveal the increasing trend of CEO turnover globally. Based on Table 1, in a period of 19 years, the percentage of CEO turnover globally increased by 4.6 percentage point to 17.5%, in 2018, compared with 12.9% in 2000.
Similarly, during the same period, the planned CEO turnover rate for most of the region also shows an increasing trend. Even though the highest CEO turnover rate in 2018 is observed for Western Europe and other mature economies at 19.8% (increased of 9.6 percentage points from 10.2% in 2000) and 19.7% (increased of 9.7 percentage points from 10% in 2000), the percentage change is more significant for Brazil, Russia, India and China (BRIC countries) and other emerging economies. Overall CEO turnover rate for BRIC increased 13.5 percentage points (from 4% in 2000 to 17.5% in 2018); meanwhile other emerging economies increased 14.8 percentage points (from 1.8% in 2000 to 16.6% in 2018). Despite this, only the United States (U.S.)/Canada show a declining trend at 3.2 percentage points (from 17.9% in 2000 to 14.7% in 2018). However, this decline in CEO turnover may be related to the measures the U.S. government has taken to include a CEO transition plan as part of the corporate governance framework. The U.S. Securities and Exchange Commission (SEC) requires all listed firms to disclose their CEO transition plans in their corporate governance framework. On 29 October 2009, the SEC issued a statement prohibiting firms from using Rule 14a-8(i)(7) to avoid submission of a CEO transition plan. In recent times, Malaysia's corporate world has been flooded with a series of news reports on the sudden exit of various CEOs. The prominent exit news has included coverage of the sudden death of YTL CEO in 2017 (Lopez, 2017), the sudden resignation of two Malaysian Airlines System (MAS) CEOs in 2016 and 2017 (Tay, 2016;Yunus, 2017), the departure of Telekom Malaysia Berhad (TM) CEO in 2018 (Zainul, 2018), D'nonce Technology Berhad (D'nonce) in 2019 (Zainul, 2019) and Malaysia Airport Holding Berhad (MAHB) in 2020 (Yusof, 2020), the suspension of FGV Holdings Berhad (FGV) CEO in 2018 after the investigation of the firm's financial loss (Dhesi, 2018), as well as the removal of several CEOs due to political change in Malaysia, such as Permodalan Nasional Berhad, Khazanah Nasional Berhad and Armed Forces Fund (LTAT) (Kaur & Tan, 2018). All these announcements had an immediate impact on the firms' share price. For example, the share price of TM rose 26 sen to RM3.99 (Murugiah, 2018) while; FGV shares closed four sen or 2.58% lower at RM1.51 (Dhesi, 2018). D'nonce shares traded down three sen or 7.32% at 38 sen (Zainul, 2019) and MAHB shares dropped 2.91% to RM7.34 despite the appointment of an interim CEO (Yusof, 2020). To minimise these impacts on a firm, the board of directors takes major responsibility in selecting a CEO who matches the current performance objectives of the firm as well as the firm's competitive surroundings (Khazanah Nasional Berhad, 2006).
Even though the Malaysian government through the Putrajaya Committee for GLC High Performance launched the green book (Enhancing Board Effectiveness) under the GLC Transformation Programme to address the CEO transition plan and the issue of sudden CEO removal (Khazanah Nasional Berhad, 2006), this policy is not mandatory and applies only to governmentlinked companies. In addition to the green book, the Ministry of Finance has also established a guideline for the board appointed by the Ministry of Finance Incorporated (MOF Inc), to recommend a CEO transition plan to the board. Nevertheless, this guideline is limited to MOF Inc companies. The Minority Shareholder Watch Group (MWSG, 2017) has reported that only one-fifth of the top 100 firms listed in Bursa Malaysia in the year 2016 have adopted a proper CEO transition policy. Meanwhile, the Associated Chinese Chambers of Commerce (ACCCIM) Family Business Survey Report 2018 (ACCCIM, 2019) shows that only 13% of family-owned firms have a robust and documented transition plan, while 24% do not think having such a plan as a priority. About 63% of the respondents claim to have established an informal plan or to be thinking about doing so but have not documented the plan. Failing to establish a CEO transition plans causes a delay in the CEO selection process, which leads to business risk (O'Brien & Ferris, 2010), as well as court battles, especially in family-owned businesses (Lopez, 2017). Since a change in CEO influences the firm's future performance, the firm should give high priority to the succession planning process as highlighted by Ishak et al. (2013).
Thus, using 146 announcements from Bursa Malaysia from 2007 to 2016, this study examines the relationship between announcements' effects and changes in the share price in the emerging market of Bursa Malaysia. This study investigates investors' reaction based on the planned or unplanned CEO turnover signal that the firm send. The result is expected to add knowledge to prior findings on the CEO turnover and make way for policymakers to re-examine the establishment of a CEO succession policy, especially in an emerging market.

LITERATURE REVIEW
According to Ahmad, Hassan and Jaffar (2016), based on 105 announcements of CEO turnover in Malaysia during 2008-2014, CEO turnover has no effect on share price, especially for the short event window. This finding is consistent with Warner, Watts and Wruck's (1988) findings on 269 firms in the U.S. from 1963-1978. According to the study, CEO transition has no impact on share price unless the CEO is removed by force. Removal by force is an unanticipated event that sends a negative signal to the market. According to Denis and Denis (1995), forced resignation causes a large and significant decline in a firm's operating performances.
Most often, removal of a CEO by force occurs due to a corporate scandal or a firm's financial condition (Ertugrul & Krishnan, 2011). However, Jenter and Kanaan (2015) explain that the reason to terminate a CEO after bad firm performance involves factors beyond the firm's control. According to Denis and Denis (1995), forced removal (by resignation) is rare and occurs due to external factors such as block-holder pressure and takeover attempts rather than normal board monitoring. Based on Adams and Mansi's (2009) observation of CEO turnover for 1973-2000, removal of a CEO results in lower bondholder value depending on the type of turnover, the successor's origin as well as the riskiness of the firm's debt. If removal occurs due to wrongdoing, Pukthuanthong, Ullah, Walker and Wu (2017) indicate that a delay in removing the CEO causes a negative abnormal return on the day of the announcement because the delay in resignation causes an investor to lose confidence in the firm due to ineffective corporate governance and supervision mechanisms. Therefore, with respect to the announcement of CEO turnover between 1990 and 1995, Dedman and Lin (2002) highlight that firms in the United Kingdom generally do not announce the departure of top executives, especially if they are dismissed or leave to take another position because such a departure represents a financial risk to a firm; it may also cause a higher chance of failure in the future.
Apart from the reason for turnover, some scholars have found that CEO turnover also has a negative relationship with firm performance for a familyowned business. According to Tsai, Hung, Kuo and Kuo (2006), based on 304 listed firms in Taiwan, CEO turnover is lower in a family-owned business, and CEO removal has a negative impact on firm performance. In some cases, CEO removal causes a mixed reaction. According to Ishak and Latif (2012), based on 247 announcements in Malaysia throughout 2008-2010, the share price reacts positively prior to (0.91%, 0.7%) and on the day of announcements (0.56%, 0.157%) but negatively on the day after announcements (-0.32%, -0.6%) using Market Model (MM) and capital asset pricing model (CAPM) estimations, respectively. According to Suchard, Singh and Barr (2001), a positive reaction before the announcement occurs because the removal of an inefficient CEO in a poor-performing firm can potentially improve the firm's future performance, while a negative reaction during the post-transition period implies a higher probability of strategic changes and uncertainty regarding the firm's future cash.
Meanwhile, some scholars have viewed CEO turnover as having a positive return. For example, based on 135 announcements in Malaysia throughout 2002-2008, Hassan, Jaffar and Rosly (2016 found that CEO turnover had a positive impact on the stock market before (2.10%), during (2.41%) and after (3.11%) the announcement of CEO turnover. Hassan et al.'s (2016) study is consistent with Cools and Praag's (2007) study on 343 CEO turnovers in the Netherlands in 1991-2000. The authors clarify that CEO turnover has a positive impact before (0.01%), during (0.21%) and after (0.26%) the event announcement.
A positive reaction is also observed for planned CEO turnover. Based on findings by Lambertides (2009) on 202 firms in the United States for December 2001 and February 2006, removal of the CEO through retirement, as opposed to an exit caused by a sudden death or illness, seems to have no direct effect on the long-term performance. This finding is similar to that of Denis and Denis (1995), who report that the removal of a CEO through retirement is usually followed by small increases in operating income but a slightly higher incidence in postturnover corporate control activity.
Apart from the removal of a CEO through retirement, a positive reaction is also observed when the removal of a CEO is unanticipated. According to Rhim, Peluchette and Song (2006), the market responds positively to an unanticipated turnover (0.9241%) compared to an anticipated turnover (about 0.2249%). Kang and Shivdasani (1996) observe that CEO transition increased from 0.52% to 1.02% when turnover was by force compared to voluntary turnover at 0.40%. This is similar to Huson, Parrino and Starks' (2001) findings, which indicated that abnormal returns increased from 0.50% to 2.02% when removal was by force.
As explained earlier, CEO transition is a process of change in leadership, involving both the appointment and turnover of a CEO. However, CEO transition is mostly referred to as CEO turnover. Warner et al. (1988) describe CEO transition as an event that occurs due to "redesignation, retirement, no reason, control change, death, and poor performance, take other position, policy difference, health, fired and others". Since the definition of CEO turnover is oversimplified, Farrell and Whidbee (2003) and Allgood and Farrell (2000) classify CEO turnover as an event that occurs as "voluntary" or "by force". This definition is driven by the EMH, which emphasises the timing of the information released. A surprising event is expected to result in a more significant reaction. According to Farrell and Whidbee (2003) and Allgood and Farrell (2000), voluntary events occur due to "retirement, normal management succession, death, or illness, or those involving CEO's departure for a prestigious position elsewhere, meanwhile, forced removal occurs due to resignations, pressure from the board of directors, pressure from outside block-holders, pressure from bank lenders, policy or personality disagreements, demotion, being fired, scandal, poor performance, bankruptcy, and reorganisation".
Similarly, Lambertides (2009) classifies CEO turnover as an "initiated" event. An event initiated by the CEO is similar to a voluntary event, while an event initiated by a board of directors is a forced event. However, some events classified as voluntary occur unexpectedly, beyond the CEO's control. Thus, Rhim et al. (2006) reclassify CEO turnover as "anticipated" or "unanticipated". The "anticipated" classification includes events such as "retirement, meanwhile unanticipated classification includes event caused by death, health, the poor performance of a firm, legal problems, forced resignation and other personal reasons of the predecessor, merger/acquisition-related, and restructuring/ reorganisation of a firm" (Rhim et al., 2006).
In the past, many scholars have used this classification as a point of reference to investigate the impact of CEO transition on the share price. The different classification of the CEO turnover will influence the findings. The different reason for CEO turnover also has a different implication for the share price. For example, Farrell and Whidbee (2003) and Allgood and Farrell (2000) classify retirement as a "voluntary" event while Rhim et al. (2006) classify retirement as "anticipated". However, unless this retirement is based on a CEO transition plan, removal of the present CEO will have an impact on the firm performance if the firm fails to appoint a successor. Thus, this study will consider the planned CEO transition as any removal of a CEO that occurs simultaneously with the appointment of new CEO because past findings have indicated that investors react positively when the announcement of CEO turnover is immediately followed with the announcement of a new CEO.
Based on Hambrick and Mason's (1984) model for upper echelon characteristics, CEO characteristics do not have an impact on share price unless moderated with policy measures. By adopting a proper CEO transition plan, the impact of the CEO's sudden removal can be minimised. However, if a firm has established an informal CEO transition plan but fails to identify a successor, not announcing the changes simultaneously may lead to a reduction in the firm's future wealth. Based on the above arguments, in general, the following hypothesis is tested: Hypothesis: A significant positive relationship exists between a planned CEO turnover and share price reaction.

DATA AND METHODOLOGY
The overall data for CEO transition includes 354 announcements, with turnover accounting for 146; after removal of outliers, the data include 124 announcements. Based on Table 2, Panel A, the highest percentage of CEO removal is in 2008 and 2015 at 13.7% each. In Panel B, the highest percentage of CEO turnover occurs in the trading and services sectors at 26.6%. Meanwhile, when the turnover is observed by classification, planned turnover accounts for 74.2% while unplanned turnover is 25.8%. For unplanned CEO turnover, the data for CEO turnover are reduced from 109 announcements to 92 due to significant outliers. Meanwhile, for planned CEO turnover, the data are extracted from the simultaneous announcement and reduced from 74 announcements to 32.
This study follows the usual event study method, where the reaction of investors is based on the movement of the share price, which is calculated using the abnormal return. The abnormal return for firm i on day t is formulated based on CAPM following Nthoesane and Kruger (2014) as in Equation (1): where, AR i,t = abnormal return of firm i on day t; R i,t = daily return for firm i on day t; E(R i,t ) = expected return firm i on day t, formulated based on Equation (2).
where, α i and β i are parameters using estimation period; rf = risk-free rate (based on three (3) months Treasury bill); R m,t = change in the market price for firm i at the end of day t-1. The data are collected based on a 21-day event window, including 10 days prior to the announcement and 10 days after it. To obtain a clearer understanding of the effect of the announcement on share price, three event windows, (-10, +10), (-10, 0) and (0, +10), are observed. A longer event window is chosen to view the investor's reaction as the investor may take time to analyse and react to the information regarding CEO characteristics.
This study comprises dependent, independent, moderating and dummy variables. The dependent variable is measured by CAAR for the event window (-10, +10), (-100) and (0, +10), while the independent variables are measured by the CEO characteristics of origin, prior experience, education, stock ownership, age and gender. A CEO who is younger than 45 is classified as a young CEO, 45-55 years as a middle-aged CEO and older than 55 as an old CEO. Since this observation focuses on the middle-aged CEO, CEO age (AGE) is measured by a dummy variable using a binary scale (1 middle-aged CEO and 0 otherwise). Origin is categorised as an internal or external selection. An internal selection is the appointment of a CEO from within the firm, while an external selection is appointment of the CEO from industry. Origin (ORG) is measured by a dummy variable using a binary scale (1 internal and 0 otherwise). This study also measures the impact of CEO prior experience (EXP) on the firm. Thus, this study measures experience as prior CEO experience from industry or none based on a dummy variable using a binary scale (1 EXP and 0 otherwise). To investigate the impact of the CEO's educational background on share price, CEO education is based on higher education, which is categorised as Sijil Pelajaran Malaysia (SPM) leavers, bachelor's degree and master's degree and above. Since the focus is to measure the impact of a highly educated CEO, CEO education (EDU) is measured by a dummy variable using a binary scale (1 master's or above and 0 otherwise). Stock ownership (SO) is measured by a dummy variable using a binary scale (1 own stock and 0 otherwise). Gender is categorised as male or female. The focus of this study is to examine the impact of a female CEO on share price. Therefore, gender (GEN) is measured by a dummy variable, using a binary scale (1 female and 0 otherwise).
CEO transition includes independent and moderating variables, which are measured by a planned or unplanned transition. The CEO transition plan (CT) is measured by a dummy variable using a binary scale (1 planned transition and 0 otherwise). Meanwhile, the control variables are measured by firm characteristics, which include return on equity (ROE), earnings per share (EPS), size, leverage and sector. All the elements under the control variable have been widely tested prior to this and shown to have an impact on the investigation. To mitigate potential biases in this finding, these elements are proposed to function as a control variable. Financial measurement is therefore based on valuation ratios such as EPS (measured by net income/average outstanding shares), ROE (measured by net income/total assets), firm debt (measured by leverage [total liabilities/total assets] and firm size (SIZ), measured by the natural log of total assets. For industry, only two out of eleven industries are observed, which are trading and services and industrial product. Trading/services (TS) are measured by a dummy variable using a binary scale (1 if TS and 0 otherwise) as in Equation 3. Meanwhile, industrial product (IP) is measured by a dummy variable using a binary scale (1 if IP and 0 otherwise).
The regression analysis is measured using four steps.
Step 1 is to examine the impact on the control variable, which includes the firm characteristics as in Equation 3; Step 2 includes the type of turnover as an independent variable as in Equation 4 and Step 3 adds in CEO characteristics as an additional independent variable as in Equation 5.

Average Abnormal Return
The average abnormal return (AAR) is examined based on planned CEO turnover and unplanned CEO turnover as in Table 3. For planned CEO turnover, the exit of the CEO has a positive but insignificant impact on the stock market, before the announcement day at 0.68%, on day (-1) and on the event day at 0.39%, but a positive and significant impact at 1.6% on day (+1). The increase of AAR for the event window (-1, +1) observed in the plotted graph is shown in Figure 1. This finding follows findings by Hassan et al. (2016) and Cools and Praag (2007). Apart from day (+1), AAR for planned CEO turnover is also significant and positive at 1.41% on day (+6) but negative and significant at 5.59% on day (-3) and -1.53% on day (+5).
Meanwhile, for an unplanned CEO turnover, the CEO exit has mixed results. The unplanned CEO turnover is negative and insignificant before the announcement at 0.17% on day (-1) and 0.1% on the event day, but positive and insignificant after the announcement at 0.17% on day (-1). This finding is consistent with Rhim et al. (2006), Kang and Shivdasani (1996) and Huson et al. (2001), who argue that removal of a CEO by force has a positive impact in the post-announcement period. For an unplanned CEO turnover, AAR is positive and significant at 0.51% on day (-5) before the event day and 0.65% on day (+3) after the event day.
Generally, investors are expected to react negatively to CEO removal, especially when it is sudden and unanticipated. However, as shown in Figure 1, a higher gain is obtained in a shorter event window for a planned CEO turnover. The higher gain obtained in a planned CEO turnover is the turnover based on the CEO transition plan, where the announcement of the CEO removal is immediately followed by the announcement of the appointment of a new CEO. Thus, negative news of the CEO removal is minimised with positive news of the appointment of the new CEO. Note: * significant at 0.1; ** significant at 0.05; *** significant at 0.01. Figure 1. AAR for event window date (-10, +10) using CAPM estimation

Cumulative Average Abnormal Return
The AAR is cumulated to observe the share price reaction in a longer event window. Based on Table 4, positive and significant CAAR is obtained at 3.8% and 3.95% for the event window (0, +10) and (-10, +10), respectively, for the planned announcement. Meanwhile, CAAR for the announcement of an unplanned CEO turnover is negative but not significant for any of the event windows. When observing for a 21-day event window (-10, +10) as in Figure 2, CAAR for the announcement of planned and unplanned CEO turnover moves in an increasing trend. However, CAAR for planned turnover is higher than for unplanned turnover across the event window. The higher gain obtained at the removal of the CEO is followed by the positive news of the appointment of the successor. Thus, this signal influences the investor's long-term investment decision.   Further analysis is conducted using Pearson correlation to examine the relationships among CEO characteristics. Based on the Pearson correlation matrix table (see Table 6), EPS is highly correlated with ROE, with a coefficient value of 0.483, and type of turnover. Both predictors measure the firm's financial performance.
Moreover, the financial performance has a positive relationship with the planned turnover. Firm size is positively correlated with firm performance, ROE and leverages with coefficient values of 0.331 and 0.334, respectively. The larger firm has high leverage, but also high risk. Meanwhile, being female is positively correlated with industrial product and origin with a coefficient value of 0.275 and 0.245, respectively. Being female also has a positive relationship with origin and industrial sector. It is possible that most females appointed as CEOs are insiders in a family-owned business.

Regression Analysis
For robustness, CAAR is regressed against the type of turnover dummy and CEO characteristics. The regression analysis is conducted in two parts. In the first part, based on Table 7, the regression analysis is conducted separately based on planned CEO turnover and unplanned CEO turnover. In the planned CEO turnover, the announcement of CEO turnover and CEO appointment coincides. When the model is tested in Step 1 for the impact of firm characteristics, the model is significant for the event window (−10, 0) and (0, +10). However, when the analysis is extended to examine the impact on CEO characteristics in Step 2, the model is significant only for the event window (−10, 0). The significant value is contributed by the significant value in the several predictors; for example, origin, education and stock ownership are significant. All three predictors (i.e., origin, education and stock ownership) have a significant negative relationship with share price at −4.740 (t < 0.01), -6.026 (t < 0.05) and 4.221 (t < 0.01), respectively.
To better understand the planned CEO transition, it is moderated based on the CEO characteristics. As seen in Table 8, when the model is tested for the impact of firm characteristics on CEO turnover in Step 1, the model is not significant for any event window. However, the EPS and log of the assets have a positive relationship with firm value while leverage has an inverse relationship with firm value, though none of the predictors is significant. In Step 2, when the independent variable is included in the regression model, the model is significant only in event window (−10, +10). Table 6 Correlation analysis for CEO turnover (n = 124)

1.00
Notes: ** Correlation is significant at the 0.01 level (2-tailed); * Correlation is significant at the 0.05 level (2-tailed) Table 7 Regression analysis for planned CEO turnover for the event window (-10, +10), (-10,0) and (0, +10) (−10, +10) (−10, 0) (0, +10) S1 S2 S1 S2 S1 S2 Coeff Note: * significant at 0.1; ** significant at 0.05, *** significant at 0.01. Dependent variables are CAARs for event window (-1,+1). The positive and negative sign indicates prior and posts announcement. The model would be tested in two steps. In Step 1 CAAR is regressed with different control variables and in Step 2, when CAAR is regressed with different control variables and independent variables. Independent variables are CEO characteristics such as age (measured by dummy variables, 1 age between 45-55; and 0 otherwise); origin (measured by dummy variables, 1 internal; and 0 otherwise); education (measured by dummy variables, 1 master's degree and above; and 0 Otherwise); experience (measured by dummy variables, 1 ex-CEO; and 0 otherwise); stock ownership (measured by 1 own stock; and 0 otherwise) and gender (measured by dummy variables, 1 female; and 0 otherwise). Control variables are firm performance (measured by EPS, ROE; firm size (measured by log of an asset), leverage and industry type dummy: trading/services (trad-serv) and manufacturing (ind-prod). This regression model is tested for multicollinearity, and all variance inflation factor is below 10. This model also is adjusted for heteroscedasticity using Andrew Hayes. The total number of the announcement is adjusted to 64 due to the large outliers especially on ROE.
Moreover, a planned CEO transition shows a significant positive coefficient value of 8.271 (t < 0.05) and 8.534 (t < 0.01) for event windows (−10, +10) and (0, +10), respectively. The significant positive coefficient value indicates that CEO turnover that co-occurs with CEO appointment has a positive impact on firm value. Investors view the planned removal of a CEO positively, which is in line with prior findings by Lambertides (2009), Cools andPraag (2007) and Denis and Denis (1995). For CEO characteristics, only gender is significantly positive with a coefficient value of 6.832 (t < 0.05) and 4.065 (t < 0.1), respectively, for event window (−10, +10) and (−10, 0). For CEO turnover, a positive coefficient value is viewed as a negative signal from the investor.

In
Step 3, the CEO transition plan is moderated based on the CEO characteristics. The findings indicate that none of the models is significant. However, the type of turnover is significantly positive with a coefficient value of 6.382 (t < 0.1) and 6.382 (t < 0.1), respectively, for event windows (−10, +10) and (−10, 0). Moreover, the predictors are all positive for event window (0, +10), which indicates that in a planned CEO turnover, all characteristics have a positive impact on the share price. The positive reactions are reflected from the simultaneous announcement of the CEO appointment.  Note: * significant at 0.1, ** significant at 0.05; *** significant at 0.01. Dependent variables are CAARs for event window (−1, +1). The positive and negative sign indicates prior and posts announcement. The model is tested in three Step: Step 1 CAAR is regressed with different control variables and in Step 2 when CAAR is regressed with different control variables and independent variables, and in Step 3 CAAR is regressed with different control variables, independent variables and moderating variable. Independent variables are CEO characteristics such as age (measured by dummy variables, 1 age between 45-55; and 0 otherwise); origin (measured by dummy variables, 1 internal; and 0 otherwise); education (measured by dummy variables, 1 master's degree and above; and 0 otherwise); experience (measured by dummy variables, 1 ex-CEO; and 0 otherwise); stock ownership (measured by 1 own stock; and 0 otherwise) and gender (measured by dummy variables, 1 female; and 0 otherwise). Moderating variable is CEO transition (measured by a dummy variable, 1 planned appointment and 0 otherwise). Control variables are firm performance (measured by EPS, ROE; firm size (measured by log of an asset), leverage and industry type dummy: trading/services (trad-serv) and manufacturing (ind-prod). This regression model is tested for multicollinearity, and all variance inflation factor is below 10. This model is tested for correlation using Spearman correlation. This model also is adjusted for heteroscedasticity using Andrew Hayes. The total number of the announcement is adjusted to 124, a combination from planned (32 announcements) and unplanned (92 announcements). For the other event window, education is significantly negative with a coefficient value of (−7.464) (t < 0.1) for event window (−10, +10); meanwhile, origin is significantly positive at 9.748 (t < 0.1) for event window (−10, 0). The negative reaction for education indicates that investors are not delighted with the removal of an educated CEO. The investor may foresee further destruction of the firm's financial position, as in event window (−10, +10) with poor firm performance. However, origin has a positive relationship in that removal of an insider incites share price. The positive reaction in a planned CEO turnover indicates that investors feel more favourably towards the removal of an insider in a planned CEO turnover than in an unplanned turnover. In a planned CEO turnover, an insider is replaced with an eligible candidate who has been groomed to take over the positive smoothly. However, for the unplanned turnover, investors may react negatively to the removal of an insider because the insider may be a founder, owner or long-serving officer who has a good reputation in the industry and whose exit causes investors to lose confidence in the firm.

CONCLUSION
Past findings have mostly shown that a forced CEO turnover positively affects firm performance (Lassoued & Attia, 2013;Dedman & Lin, 2002;Lambertides, 2009;Huson et al., 2001;Kang & Shivdasani, 1996;Denis & Denis, 1995). However, this study provides evidence that a planned CEO turnover has a positive impact on share price when compared to an unplanned CEO turnover. The gain from a planned CEO turnover is higher than from an unplanned CEO turnover. Although the EMH suggests that information that is known has a lesser impact on share price, Malaysia is a semi-strong market that reacts to asymmetric information. This finding is consistent with Lambertides' (2009) findings on the planned removal of a CEO. A CEO succession plan is essential as it extends the business's lifespan and ensures sustainable growth (Lim, 2019). According to a report by Minority Shareholder Watch Group (MSWG, 2017), only 1/5 of the 100 firms listed in Bursa Malaysia has a proper CEO succession policy. Firms and policymakers are not convinced of the need to implement a proper succession plan because a considerable gap exists in the literature on CEO transition and its impact on the stock market. Thus, policymakers should re-examine the policy on CEO succession established under GLCs and make it mandatory for firms listed in Bursa Malaysia. By establishing a CEO succession plan and announcing the removal and appointment of a CEO simultaneously, any negative signal from the CEO removal is negligible because all negative signals are minimised with the appointment of a successor. Further analysis is proposed for a short event window to view investors' immediate reactions. Moreover, further investigation into the planned CEO appointment, especially the appointment of female CEO, is needed to obtain a more comprehensive view of investors' reaction to a succession plan.