1
CEO’s term of office, Governance Structure and
Acquisition Performance
--
Based on North American market
Master Thesis Finance
Xiaoqian Sun
10839801
Supervisor: Tolga Caskurlu
July, 2015
2
Statement of Originality
This document is written by Student [Xiaoqian SUN] who declares to
take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is
original and that no sources other than those mentioned in the text and
its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the
supervision of completion of the work, not for the contents.
3
Content
Abstract ... 4
Section I: Introduction ... 5
Section II: Literature Review ... 7
2.1 CEO’s term of office ... 7
2.2 Governance factors ... 10
Section III: Data &Methodology ... 13
3.1 Sample selection ... 15 3.2 Data analysis ... 15 3.2.1 Dependent Variables ... 15 3.2.2 Independent Variables ... 16 3.2.3 Control Variables ... 17 3.3 Methodology ... 17
3.3.1 Part one: CEO tenure & likelihood of the acquisition ... 17
3.3.2 Part two: CEO’s term of office after the acquisition ... 18
Section IV: Empirical Test & Analysis ... 18
4.1 Regression result ... 18
4.1.1 Part one: CEO tenure & likelihood of the acquisition ... 18
4.1.2 Part two: CEO tenure after the acquisition ... 21
4.2 Robustness checks ... 27
Section V: Conclusion ... 29
Section VI: References ... 33
4
Abstract
Using cumulated abnormal return as the measure of acquisition performance, the paper researched the CEO’s term of office, acquisition performance as well as the corporate governance factors. Specifically, the study finds that the longer CEO tenure before the acquisition would increase the likelihood of acquisition completion; the bidders who complete the deals show more significant negative relationships between acquisition performance and CEO after-‐acquisition turnover comparing to the bidders who cancel the deal; CEO who comes with a longer tenure before the acquisition decreases the likelihood of being dismissed in case of a bad acquisition. Moreover, the governance factors (number of independent shareholder; the shares of single largest shareholder; board size; E-‐index; CEO-‐Chairman duality) influence the likelihood of the deals and after-‐acquisition turnovers at different extents.
Key word: CEO tenure; CEO turnover; acquisition performance; corporate
governance structure
5 Section I:
Introduction
CEO’s term of office has been paid increasing attention by scholars and investors since it does not simply reflect to the governance efficiency but also crucial to the corporate performance (Hermalin and Weisbach, 2001; Murphy and Zimmerman, 1993; Huson, Parrino and Starks, 2001). There are large numbers of current research that focus on CEO’s term with the governance structure (Weisbach, 1988; Goyal and Park, 2002; Fisman et al, 2013) and CEO’s turnover with corporate performance (Lehn and Zhao, 2006; Kaplan and Minton, 2012), but there is very few paper that focusing on CEO’s term of office and acquisition performance from a governance perspective. The central research question of the paper is, how does the acquisition behavior affect the CEO’s term of office on the bidder side? Around the main research question, the following questions will be discussed as well: Does a longer CEO tenure before the acquisition lead to a larger possibility of success? Does the bidders who complete and withdraw the deal differs in CEO turnover afterwards? Does a longer CEO tenure before the acquisition reduce its turnover rate in case of a bad acquisition? Furthermore, how does the different governance structure influence the CEO turnover for the change?
The research sample is constructed from seven databases as well as hand collection. Based on North American market, the paper study the CEO’s term of office, corporate governance structure and acquisition performance from 2002 to 2011. The study illustrates CEO’s term of office in two parts: before and after the acquisition. The result shows that the bidder’s CEO tenure is positively correlates with the likelihood of the acquisition. In addition, bidder’s CEO that comes with a longer tenure before the acquisition decreases the likelihood of being dismissed in case of a bad acquisition. The paper would contribute to the existing literatures in multiple ways. Most of the current literatures mainly
6
focused on the CEO tenure and corporate performance, or on the acquisition performance and the CEO turnover. There is very few available paper to study the relationships between CEO’s term of office and the corporate performance from an acquisition perspective. More specifically, CEO’s term of office before and after the acquisition will be separately discussed, which include bidder’s CEO tenure, CEO turnover and acquisition performance. Different from the previous studies (Haleblian and Finkelstein, 1999; Lehn and Zhao, 2006), this paper use Fama-‐French three-‐factor model to calculate the abnormal returns as acquisition performance. In addition, the governance mechanism factors will be further illustrated together with the tenure and turnover in order to fill in the vacant.
Merger and acquisition (M&A) has become a popular way of enlarge the corporation and increase its competitive power due to the rapid growing of global economics. From the history, we can still see the long procedure of the acquisition made by AOL (America Online) and Warner Bros., which gained win-‐win strategy and built the great Time Warner today. However, the media giant 21st Century Fox withdrew the deals of acquiring Time Warner in 2014, which caused a huge loss for the company. From the story, we can see that M&A is not only meaningful to the strategy of corporate development but also crucial to the corporate performance.
There is an increasing debate of the CEO’s term of office and the corporate performance. The resource-‐based scholars argues that the longer the CEO stays within the company, the more specific experiences he has and it would be beneficial to the company (Beigh, 2001). While the Upper Echelons Perspective suggest that, rather than the short CEO retention, the longer tenure goes against to improve the corporate performance due to the power concentration and over confidence (Hambrick and Fukutomi, 1991; Miller and Shamsie, 2001). In this paper, I would study the CEO’s term of office and the acquisition
7
performance since M&A is an significant measurement for corporate performance. Moreover, it is the CEO who takes charge and who is mainly responsible for the acquisition, and a successful acquisition is an individual achievement of the bidder’s CEO ( El-‐Khatib, Fogel and Jandik, 2015).
The study proceeds as follows: Section 2 presents the literature review of the related contributions; Section 3 describes the data and research methodology; Section 4 discusses the empirical results and robustness checks and Section 5 presents the conclusion.
Section II:
Literature Review
2.1 CEO’s term of office
CEO tenure and the corporate performance has been widely discussed in the field of enterprise strategic management (Miller, 1991), and it attracts increasingly attention during the recent years. With the separation of ownership and control, to some extent, the board of directors is the nominal highest decision maker while CEO is the actual controller within the company. According to Xu and Li (2007), CEO can potentially construct his own ‘team' with those who obey his order. It might probably result in a situation that, even though, there are numbers of independent directors, the monitoring is not perfect due to the information asymmetry, and the independent directors may not take too much time for getting to know the company. As a result, the board of directors have to concede and give the highest decision maker position to the CEO for series of investment decisions (Rosenstein, 1988). CEO is crucial to the enterprise, as it reflects to the supreme administrator. Therefore, the CEO tenure plays a critical role in affecting the investment decision and thus
8
influences the corporate performance (Hambrick and Mason, 1984)
However, there are debates regarding to the CEO tenure and corporate performance. The resource-‐based scholars suggest that, the longer the CEO stays within the company, the more specific experiences he has. Thus, it is beneficial to increase the corporate performance and the competitive power (Beigh, 2001). In addition, as a rule of survival of the fittest, the CEO who comes with an inadequate performance faces a bigger chance of turnover comparing to the one with a better work. Hence, the long tenure also reflects a preferable corporate performance as well as a mark for individual success (Boeker, 1992). Furthermore, Beigh (2001) assumes that the longer CEO retention correlates with a better performance in the acquisition. In turn, the shorter tenure associates with a higher probability of being acquired by others in the future. As to Chikh and Filbien (2011), they find that CEOs with stronger social ties usually have a lower possibility to cancel the deal even if they met negative market responses. As the CEO has a long tenure before acquisition, he might comes with stronger management power, tighter social ties and more experiences. Thus, we might argue that a longer CEO tenure might be more likely to lead to an successful deal.
On the other hand, the Upper Echelons Perspective insist that, comparing with the short CEO retention, the longer tenure goes against to improve the corporate performance (Hambrick and Fukutomi, 1991; Miller and Shamsie, 2001). They regard CEO tenure as the management lifecycle, CEO differs in the degree of the mind, the information and the attempt across different periods. Accordingly, CEO might come with different management performances in each period. At the beginning of the management lifecycle, CEO is comparably open minded, with huge interests and develops numerous methods of acquiring information to adapt the new environment and make suitable strategies. Therefore, within a short tenure, CEO usually comes with a positive influence
9
on the corporate performance. While as the tenure getting longer, the former success might become traps for today. They become over-‐confidence as well as conceit. Thus, they refuse to make innovations and stay on the old governance. As a result, the corporate is going far away with the society and lead to a bad performance. Moreover, the longer tenure might associates a stronger management power, which increase its social ties but might neutralise the monitoring power from board members (Fracassi and Tate, 2012; El-‐Khatib, Fogel & Jandik, 2015). Thus, it is meaningful to study the CEO tenure and the likelihood of acquisition.
The successful acquisition is an individual achievement for good CEO performance on the bidder side (El-‐Khatib, Fogel and Jandik, 2015). According to Lehn and Zhao (2006), although merger and acquisition activities are approved by the entire board, CEO is the core person who launched and who is responsible for the whole procedure. Moreover, CEO faces various considerable decisions during the M&A activities, which every little factor matters for the acquiring valuation. As soon as the acquisition decision is being made, the CEO also faces various market reactions and pressures. It is therefore he might come across another decision to continue or cancel the deal. Once the deal is completed, the corporation might still face a favourable or dissatisfying return. Thus, the CEO might confront with a dismiss in case of a bad situation. Similar to Allgood and Farrell (2000), the paper also study the CEO turnover with the corporate performance. The study will also include the forced turnover only, which exclude the retirement, death, illness or no reasons reported on the database (Allgood and Farrell, 2000).
In perfect markets, the stock price reflects the corporate performance (Morck, Yeung and Yu, 2000). Thus, the stock price after the acquisition announcement could reflect the bidder's performance straightforwardly. As a consequence, the ‘good bidders' will associate with a higher abnormal stock return, while the
10
‘bad bidders' will probably face a negative return and their CEOs face disciplines (Lehn and Zhao, 2006). As to Lehn and Zhao (2006), they tested the bidder returns and the CEO turnover and find there is an inverse relationship between them. Furthermore, they conclude that there is no significant relationship of returns between cash and stock acquisitions, which is opposite to Shleifer and Vishny’s (2003) ‘stock market driven’ theories (Lehn and Zhao, 2006). Based on their research, the paper will also study the relationship between CEO’s term of office after the acquisition and its acquisition performance. The study will adopt Fama-‐French three-‐factor model to calculate the abnormal returns rather than CAMP model. I also assume the bad acquisition performance associates with a relatively short term of office afterwards. Moreover, if the CEO experienced a long tenure before the acquisition, his turnover possibility might be much more lower in case of a bad acquisition than the CEO with a short tenure. Because of the accumulated experiences and management power, the turnover rate might be not easily affected by a single acquisition. As to Dikolli, Mayew and Nanda (2014), they study the CEO’s turnover and find it declines with its tenure. In addition, the study will also include the governance variables based on Lehn and Zhao’s (2006) research, which only focus on the governance variables and the cumulative abnormal returns, but does not figure the associations between the governance variables and the CEO turnover.
2.2 Governance factors
Corporate governance has been paid attention by scholars and investors increasingly since it does not simply influence the firm value but also crucial to the whole economy (Bebchuk & Hamdani, 2009). Hermalin and Weisbach (2001) examined its importance and find that the board structure (e.g. scale or construction) influences the behavior of the board of directors (e.g. CEO turnover). In addition, besides of the discipline from the external market, the
11
internal governance also plays a role in affecting the CEO tenure (Walters, Kroll and Wright, 2007). Hence, it is important to include the governance structure in studying the CEO tenure concerning the acquisition.
Independent board directors. Baysinger and Butler (1985), Wei et al. (2007)
suggest that the percentage of independent director shows a significantly positive relationship with the governance performance. They base on the agency theory and argue that the independent board directors are more objective in monitoring the governance (Fama and Jensen, 1983). Hence, we can expect that the CEO is better monitored in a company with more independent board directors. Thus, once the bidder's return does not perform good, the CEO faces discipline also from the independent board directors. On the other hand, the other scholars insist the independent board directors and the corporate performance are inversely correlated due to the potential information cost (e.g. Fosberg, 1989). Thus, there will be one probability that there is no significant relationship between the independent directors and the corporate performance if we put these benefits and disadvantages together.
Single largest shareholder. Grossman and Hart (1980) examine that the block
holder or ownership concentration within a company is beneficial to the corporate governance. He argue that the small shareholder has to pay more for the monitoring comparing to their management income under the diffuse ownership, such that they usually lack motivations for the governance supervision. Thus, one might argue that the shares of the single largest shareholder is positively correlated with the monitoring and might lead to a higher chance for CEO turnover in case of being a ‘bad bidder'. However, Demsetz and Lehn (1985), Shleifer and Vishny (1997) demonstrate that the majority shareholders might harm the minority shareholders' interests in order to increase their own benefits. It would lead to a negative relationship between the single largest shareholder and the governance performance.
12
CEO-‐chair duality. With the introduction of the Stewardship theory by Davis,
Schoorman and Donaldson (1997), they argue that the CEO-‐chair duality could reduce the power conflicts between the two positions, such that it reduces the agency cost and increases governance efficiency. They assume the duality position might increase the probability of CEO’s serving more sincerely for the company and maximize shareholder's benefit. On the contrary, Fama and Jensen (1983) hold the opposite opinion. Under the duality, the monitoring function reduces since the CEO is monitoring himself. As to Jensen (1993), the CEO might not implement the monitoring which is away from his interest. He suggests separating the CEO and chairman in order to make the governance more efficient. In addition, the duality would make the CEO have significant power in making important decisions with a higher probability to lead into a dictatorship. Therefore, it might positively affect his tenure even though after a bad acquisition performance.
Board size. According to Jensen (1993), companies with smaller board size have
higher governance efficiency and corporate performance. He suggests that the board should keep the numbers of directors within seven or eight people. Similarly, Lipton and Lorsch (1992) also argue that the board members should not exceed ten. Otherwise the large number might influence the internal communication and affect corporate performance. From that perspective of view, CEO is better monitored in corporations with smaller board size and faces a higher probability of discipline in case of the bad performance. However, an alternative view is that large board size might include complementary knowledge in management and experiences, which could generate more information and resources (Ocasio, 1994). Thus, by averaging the points of view of the two perspectives, the relationship between CEO tenure and the board size is full of uncertain.
13
E-‐index. Based on the GIM index from Gompers, Ishii and Metrick (2003),
Bebchuk, Cohen and Ferrell (2006) study the E-‐index governance entrenchments in evaluating the governance performance of the companies. They show an index consist of six provisions (staggered boards, limits to shareholder bylaw amendments, poison pills, golden parachutes, and supermajority requirements for mergers and charter amendments) that inversely correlate with the performance. Bebchuk, Cohen and Ferrell (2006) show that the low index reflect a 'Democracy' governance and the high index lead to the ‘Dictatorship'. The index also reflects to the ability of shareholder to replace the management team (Dikolli, Mayew and Nanda, 2014). Therefore, one might expect that the CEO is more likely to be taken over under a low E-‐index if it acts as a ‘bad bidder'.
Section III
Data & Methodology
3.1 Sample selection
The database is constructed from seven databases. It is initially based on the North American companies that are available from Thomson One Securities Data Company (SDC) Merge and Acquisitions database. The study includes the deal ‘complete’ and ‘withdraw’ from 2002 to 2011, which provides the sample large and newly enough for the current research. Furthermore, the initial database is further merged with Compustat, since Compustat and CRSP are the main sources of the financial data and stock prices. Following the previous research (e.g. Lehn and Zhao, 2006) for representing large investment and avoiding of potential errors, the sample excludes the target firms that are no more than 5% value of the bidders. The CEO status are collected from ExecuComp, which include the date of being a CEO, date for leaving and the
14
leaving reasons. The governance data is obtained from Field-‐Ritter dataset of company founding dates, Thomson Reuters and Institutional Shareholder Services (ISS). The study also hand-‐collected missing data from company
Table 1
Code Variable Source
Deal Acquisition status Thomson One Securities
Data Company (SDC)
Turnover Whether bidder’s CEO is replaced
within 3 years Execucomp
Ten0 Years of bidder’s CEO tenure before
the acquisition Execucomp
Ten1 Years of bidder’s CEO tenure after the
acquisition Execucomp
CAR Bidder’s cumulated abnormal return
around announcement date CRSP
Indep_board Percentage of independent board
members Thomson Reuters
Sing_largest Single largest shareholder Thomson Reuters
Boardsize Board size Thomson Reuters
Eindex Governance provisions index Institutional Shareholder Services (ISS)
Duality CEO-‐ Chair duality Thomson Reuters
DEratio Debt-‐equity ratio Compustat
Tobinq Market to book value Compustat
Ceo_age CEO’s age Institutional Shareholder
Services (ISS)
Ceo_share CEO’s share holding Institutional Shareholder
Services (ISS)
Firm_age Firm age Field-‐Ritter dataset
Firm_size Firm size Compustat
Industry Whether the bidder and target are
from the same industry
Thomson One Securities Data Company (SDC)
Table 1 presents a summary of all the variables which provides brief descriptions and data sources of all variables from the sample.
15
websites. Unfortunately, there are unavoidable missing values in the dataset. Nevertheless, the observations in the sample are convincing enough to support the study results.
3.2 Data analysis
Table 1 presents a summary of all the variables which provides brief descriptions and data sources of all variables from the sample.
3.2.1 Dependent variable
Deal (0 or 1): It shows the status of an acquisition. All the sample companies
declared the announcement of acquisitions, while some companies completed and someone withdrew. Hence, the Deal takes 1 if ‘complete’ and 0 if ‘withdraw’. There are 335 out of 428 companies which completed the deal in the sample as the result of further merges with the other databases. For the same CEO companies that completed more than one deal, the study will only keep the first deal for each company and each CEO.
Turnover (0 or 1): Following with Parrino (1997) and Lehn & Zhao (2006), the
turnover is initially defined that as the CEO is being replaced due to internal governance, takeovers or bankruptcy. It except the retirement, death, illness or no reasons reported on the database (Allgood and Farrell, 2000). It takes 1 if the CEO is replaced within three years after acquisition. That is one of the reasons for the data selection period from 2002 till 2011. There are 131 sample companies replace their CEO within three years after the acquisition regardless of the retirement. In addition, 43 out of 131 are the ones who withdraw the acquisition after the announcement.
16
Due to the sample limitation, the dataset only covers the deals after 2002. Thus, the maximum of the CEO’s tenure after the acquisition is approximately 15 years in the sample. The average CEO’s term of office after the acquisition is around five years among the sample.
3.2.2 Independent variable
Tenu0: It measures the number of years which CEO’s tenure before the
acquisition. The average CEO’s tenure in the sample is 7.5 years, the shortest tenure is less than 3 months and the longest is more than 35 years.
Cumulative abnormal return (CAR): Unlike the previous literature adopting
CAMP model (Haleblian and Finkelstein, 1999; Lehn and Zhao, 2006), the abnormal return in this study is calculated by using Fama-‐French three-‐factor model that includes size and book-‐to-‐market-‐equity to illustrate the acquisition performance more precisely. The acquirer’s cumulative abnormal returns will be computed by using a three-‐day and a five-‐day event window.
Corporate Governance factors: As mentioned above, there are some include
variables which study the relationship between the governance structure and the CEO tenure for the acquisition. The percentage of independent board
members (Indep_board) measures the corporate monitoring power; Single largest shareholder (Sing_largest) shows the ownership concentration; CEO-‐ chair duality (Due) measures the CEO power; Board size (Boardsize) presents
the number of board members in the corporation and E-‐index (EIndex) measures the corporate governance provisions, which is based on Bebchuk, Cohen and Ferrell (2009). Our database is constructed from 2002 since the E-‐index database is only available from 2002.
17 3.2.3 Control variable
All of the control variables are collected from the year before acquisition.
Leverage (DE_ratio) measures the capital structure of the company. Debt to
equity ratio is frequently used in the empirical research (Jensen and Meckling, 1977), which shows the financial structure and default risks.; Tobin’s Q (Tobinq) presents the value of the firm, following the calculation by Gompers, Ishii and Metrick (2003), the ratio is calculated by the market value to book value. As to Claessens et al. (2002), an elder (Firm_age) and larger firm (Firm_size) has less distress probabilities. The firm size is calculated by the logarithm of the total asset, while the age of the company is measured as the difference between current year (2015) and the founding year.
Moreover, the variables like CEO ownership (Ceo_share) and CEO age (Ceo_age) will also be controlled in the study. As Hambrick and Mason (1984) argue, CEO with a large holding percentage might reduce the agency problems. While an elder CEO might comes with more experiences in corporation management (Xu and Li, 2007). Furthermore, the study also control the business relativeness
(Industry) with the bidders and the targets. It equals to one if the two parties
from the same industry (based on SIC code) and zero otherwise.
3.3 Methodology in empirical test
The study will be mainly discussed in two parts: before and after the acquisition. Part one presents the result of the CEO tenure and the likelihood of the acquisition; part two discusses the CEO tenure after the acquisition as well as the CEO turnover.
3.3.1 Part one: CEO tenure & likelihood of the acquisition
18
In order to research the CEO tenure and the likelihood of the acquisition, a Probit model will be applied in this research. Two regressions will be analysed separately with and without corporate governance factors:
A1: Prob(Deal)= 𝛼!+ 𝛽!𝑇𝑒𝑛𝑢!+ 𝑐𝑜𝑛𝑡𝑟𝑜𝑙 𝑣𝑎𝑟𝑠
A2: Prob(Deal)= 𝛼!+ 𝛽!𝑇𝑒𝑛𝑢!+ 𝛽!𝐼𝑛𝑑𝑒𝑝 + 𝛽!𝑆𝑖𝑛 + 𝛽!𝐷𝑢𝑒 + 𝛽!𝐵𝑠𝑖𝑧𝑒
+𝛽!𝐸𝐼𝑛𝑑𝑒𝑥 + 𝑐𝑜𝑛𝑡𝑟𝑜𝑙 𝑣𝑎𝑟𝑠
3.3.2 Part two: CEO’s term of office after the acquisition
To examine the CEO tenure after the acquisition, the study analyse the cumulated abnormal returns on the 428 U.S. bidders using event study. Unlike the previous research, the study treats the cumulated abnormal return (CARs) as the independent variable, which measures the acquisition performance in the CEO tenure afterwards. The main models that will be used in Part two are:
B: Turnover=𝛼!+ 𝛽!CAR+ 𝛽!𝐼𝑛𝑑𝑒𝑝 + 𝛽!𝑆𝑖𝑛 + 𝛽!𝐷𝑢𝑒 + 𝛽!𝐵𝑠𝑖𝑧𝑒 +𝛽!𝐸𝐼𝑛𝑑𝑒𝑥 + 𝑐𝑜𝑛𝑡𝑟𝑜𝑙 𝑣𝑎𝑟𝑠 C: Turnover=𝛼!+ 𝛽!𝑇𝑒𝑛𝑢!+ 𝛽!𝐼𝑛𝑑𝑒𝑝 + 𝛽!𝑆𝑖𝑛 + 𝛽!𝐷𝑢𝑒 + 𝛽!𝐵𝑠𝑖𝑧𝑒 +𝛽!𝐸𝐼𝑛𝑑𝑒𝑥 + 𝑐𝑜𝑛𝑡𝑟𝑜𝑙 𝑣𝑎𝑟𝑠 D: Turnover=𝛼!+ 𝛽!CAR + 𝛽!𝑇𝑒𝑛𝑢!+ 𝛽!𝐼𝑛𝑑𝑒𝑝 + 𝛽!𝑆𝑖𝑛 + 𝛽!𝐷𝑢𝑒 +𝛽!𝐵𝑠𝑖𝑧𝑒 + 𝛽!𝐸𝐼𝑛𝑑𝑒𝑥 + 𝑐𝑜𝑛𝑡𝑟𝑜𝑙 𝑣𝑎𝑟𝑠 Section IV
Empirical test & analysis
4.1 Regression result
19
Table 2
Prob. of the DealModel A1 Model A2
Ten0 0.0047b (2.06) 0.0057b (2.50) 0.0083a (3.47) 0.0093a (4.19) 0.0080a (3.54) 0.0075a (3.33) 0.0094a (3.87) Indep_board 0.0050a (3.35) 0.0041a (2.71) Sing_largest 0.0039c (1.87) 0.0089b (2.11) Boardsize 0.0245a (5.66) 0.0266a (6.06) Eindex 0.0256a (2.86) 0.0177 (1.59) Duality 0.0479b (1.89) 0.0543b (2.19) DEratio -‐0.0129a (-‐2.59) -‐0.0119b (-‐2.35) -‐0.0097c (-‐1.74) -‐0.0133b (-‐2.54) -‐0.0167a (-‐3.16) -‐0.0139 (-‐2.66) -‐0.0080 (-‐1.61) Tobinq 0.0335a (2.66) 0.0416a (3.27) 0.0403a (3.01) 0.0499a (3.96) 0.0397a (3.12) 0.0417a (3.26) 0.0400a (3.06) Ceo_age 0.0164a (8.19) 0.0142a (7.16) 0.0156a (7.48) 0.0155a (7.57) 0.0150a (7.18) 0.0139a (6.72) 0.0163a (7.62) Ceo_share 0.092a (3.91) 0.0627a (8.99) 0.0640a (10.01) 0.0601a (10.93) 0.0676a (10.22) 0.0669a (10.19) 0.0579a (9.32) Firm_age 0.0002 (0.68) 0.0005b (1.94) 0.0003c (1.70) 0.0006b (2.39) 0.0005b (2.02) 0.0003 (1.60) 0.0005c (1.89) Firm_size 0.0313b (3.75) 0.0197b (2.26) 0.0150c (1.73) 0.0157c (1.72) 0.0127 (1.65) 0.0201b (2.24) 0.0205b (2.22) Industry 0.1032a (4.36) 0.1186a (4.09) 0.1241a (4.64) 0.1148c (4.52) 0.1542a (6.18) 0.1412a (5.27) 0.0074c (1.86) Observation 428 423 423 423 412 419 412 Overall R2 0.116 0.108 0.105 0.119 0.110 0.102 0.098
Table 2 presents the result of the CEO tenure with the likelihood of the acquisition completion. Model A1 shows
that the relationship between CEO tenure and the likelihood of the deal completion without adding governance factors; Model A2 shows that the relationship between CEO tenure and the likelihood of the deal completion
together with the governance factors; In Model A2 , each of the governance factor is tested individually, and then
all factors are regressed together with the probability of the deal completion. Significant levels are indicated by a, b and c for 1%, 5% and 10% respectively.
20
As Table 2 presents, Model A1 shows the result that CEO tenure before
acquisition is positively correlated with the likelihood of the acquisition. In Model A2, each of the governance factor is tested individually, and then all
factors are regressed together with the probability of the deal completion. The results reflect that as the CEO comes with a longer tenure and more management experiences, he might has a lower probability of cancelling the deal. The result is even more significant at 1% level in all of the regressions in model A2 as we assumed before. In Model A1, as the CEO’s tenure increase one
year longer before the acquisition, the probability of the cancelling the acquisition deal will decrease by 0.0047. As we can also see from the table, the larger firm that has a higher market value is also correlates with a lower probability of cancelling the deal. Moreover, the bidders and targets from the same industry are also less likely to cancel the deal rather than the two parties from different areas.
The corporate governance factors are included in the regressions of Model A2.
The increase of one independent shareholder will increase the probability of deal completion by 0.005. While after including all the other factors into the regression, the coefficient would slightly change to 0.0041 but still at 1% significant level. Furthermore, companies with a larger blockholder are also positively correlates with the probability of deal completion. As the single largest shareholder increase by one percent share, the likelihood of the deal completion will increase by 0.0039. The result is also slightly changes after putting all the other governance factors into the regression, and the significance changed from 1% to 5% level. Moreover, the board size has a positive relationship with the deal completion as well. Both of the two regressions showed 1% significant level for the board size. The last regression of Model A2 reflects that the probability of cancelling the deal would decrease
by 0.0266 as the board size increase by one percent. The reason might be that the shareholders would get more information and have more social ties as the
21
board becomes larger. Thus, they would be more determined to face the various market reactions and continue the acquisition procedure. In addition, the CEO-‐Chairman duality shows positive correlations with the deal probability at 5% significant level in both of the two regressions, which reflects that the power concentration would also decrease the cancelling probabilities. Unlike all the governance factors that illustrated above, E-‐index shows positive relationship with the deal completion probability at 1% significant level in the individual test but at a low significance level in the total one.
4.1.2 Part two: CEO tenure after the acquisition
To better describe the data for further regressions, Table 3 summaries the cumulated abnormal returns for three-‐day and five-‐day event windows. Comparing to the CARs around announcement date with a five-‐day event window, the three-‐day event widow has lower CARs. It is probably due to the various strong market reactions close to the announcement date would lead to a greater difference with the market predictions. The mean of CARs (-‐3,+3) sample is -‐0.698, which is 0.367 lower than CARs (-‐5,+5) event time. Moreover, the mean of CARs (-‐5,+5) is -‐0.381, which reflects that the cumulated abnormal return in the United States close to the announcement date might increase shortly at a rapid speed.
In order to illustrate the difference between the bidders who complete and
Table 3
Sample Mean (Std. Dev.) Median
CAR(-‐3,+3) 428 -‐0.698 (0.190) -‐0.681
CAR(-‐5,+5) 419 -‐0.381 (0.276) -‐0.305
Table 2 presents the summary of cumulated abnormal returns for three-‐day and five-‐day event windows.
22
withdraw the deals in CEO turnover afterwards, model B is applied into the test. Table 4 shows the regression results of CARs on CEO turnovers after the acquisitions. In addition, the different deal status and different event windows
Table 4
CEO Turnover
Event time(-‐5,+5) Event time (-‐3,+3)
Deal=0 Deal=1 Deal=0 Deal=1
CAR -‐0.0870 (-‐0.96) -‐0.3092c (-‐1.88) -‐0.2991 (-‐1.50) -‐0.3324 (-‐1.76) Indep_board -‐0.0136 (-‐1.74) -‐0.0137c (-‐1.83) -‐0.0133b (-‐2.09) -‐0.0132c (-‐1.90) Sing_largest 0.0023 (0.29) 0.0035 (0.82) -‐0.0031 (-‐0.42) -‐0.1033c (-‐1.94) Boardsize -‐0.0238 (-‐0.98) -‐0.0242 (-‐1.02) -‐0.0142 (-‐0.85) -‐0.0106 (-‐0.99) Eindex -‐0.0034 (-‐0.31) -‐0.0031 (-‐0.70) -‐0.0431 (-‐1.07) -‐0.0405 (-‐1.55) Duality 0.1130 (1.42) -‐0.1043c (-‐1.80) 0.0499 (1.60) -‐0.0349 (-‐0.65) DEratio 0.0142 (0.65) 0.0303 (1.43) 0.0085 (0.60) 0.0099 (1.41) Tobinq 0.0934 (1.41) 0.0921 (1.67) 0.0921 (1.60) 0.0492c (1.83) Ceo_age -‐0.0402a (-‐2.66) -‐0.0402a (-‐3.06) -‐0.0512a (3.61) -‐0.0464a (-‐3.88) Ceo_share -‐0.0074 (-‐0.22) -‐0.0079 (-‐0.70) -‐0.0192 (-‐1.09) 0.0193 (0.58) Firm_age 0.0019 (1.57) 0.0025c (1.85) 0.0015 (1.66) 0.0011c (1.89) Firm_size 0.0113 (0.44) 0.0108 (0.38) 0.0029 (0.69) 0.0038 (0.97) Industry -‐0.2144 (-‐1.41) -‐0.2143 (-‐1.54) -‐0.0733 (-‐1.54) -‐0.0699 (-‐1.21) Observation 90 329 93 335 Overall R2 0.1683 0.1042 0.1805 0.0998
Table 2 presents the difference between the bidders who complete and withdraw the deals in CEO turnover afterwards. Two CARs event windows are individually regressed with the CEO turnover using Model B. The Deal takes 1 if complete and 0 if withdraw. Significant levels are indicated by a, b and c for 1%, 5% and 10% respectively.
23
are individually presented. In general, the CARs are negatively correlated with the CEO turnover. In CARs (-‐5, +5) event window, the cumulated abnormal return is negatively correlated with the CEO turnover at 10% significant level in completed deals. It is less significant in CARs (-‐3, +3) comparing to the five-‐day event window in completed deals. The increase in cumulated abnormal return would lead to a lower chance of CEO turnover in completed deals generally, but the relationship is relatively weak in the cancelled deals. Thus, the study will only focus on the completed deals in following regressions.
Besides of cumulated abnormal return, the percentage of independent shareholders also play an important role in effecting the CEO turnover. Except for the cancelled deals in CARs (-‐5, +5), all the other three lines show significant negatively relationship with the turnover. It shows that with an increase of one independent shareholder, the turnover would correspondingly decrease. Moreover, it means that as the number of independent shareholder increases, the board monitoring power would increase as well (Fama and Jensen, 1983). As a result of a public monitoring, the CEO has to be extremely careful at every decision and avoid of making mistakes. At the meantime, the board with more independent shareholders might also increase the social ties of the corporation. As being mentioned before, the stronger social ties might get hold of larger information, which would lead to a less probability of cancelling the deal (Chikh and Filbien, 2011). Furthermore, the single largest ownership from the completed deal in CARs (-‐3, +3) also shows a negative relationship with turnover at 10% significance. It means that as the corporate shareholding gets more centralized, the single largest shareholder has a larger power in decision making and monitoring, which would also lead to a lower CEO turnover probability as the independent shareholder increases. As we can see from Table 4, the cancelled deals show less significant relationship between
24
acquisition performance and CEO turnover. Thus, the study will focus on the completed deals in further analysis.
Table 5 presents a summary of sample tenure before (Ten0) and after the
acquisition(Ten1: till Jan 2015) only for completed deals. With an average of
34 CEO samples in each of the ten years, the mean of CEO tenure before the acquisition is approximately seven years in the sample. The minimum CEO tenure before acquisition is less than three months while the largest is about 35 years. In addition, the standard deviation of the Ten0 mean has a low
confidence level due to the large variance CEO tenure of each company. By comparing the data year by year, it could kindly reduce the year effect that the acquisitions took place in different years. However, the Ten1 and Ten0 are still
incomparable since the sample period are too close from now and there is very few years to be counted after the acquisition. Nevertheless, we can still see from the table that the CEO tenure after the acquisition is relatively less than the years before the acquisitions by computing the extreme values and the means.
Table 5
Year Sample
Ten0 Ten1 (till Jan2015)
Min Max Mean Median Min Max Mean Median 2002 28 0.13 23.16 7.52 5.85 0.09 12.89 5.40 5.63 2003 36 0.56 35.63 7.45 6.07 0.03 11.80 6.10 5.96 2004 32 0.02 31.11 8.23 6.22 0.19 10.95 6.55 6.77 2005 38 0.68 29.97 5.55 5.10 0.07 9.94 5.62 5.49 2006 35 0.31 23.92 5.24 4.90 0.13 8.95 5.56 6.43 2007 40 0.16 22.87 5.84 3.55 0.10 7.90 5.47 6.42 2008 28 0.94 22.05 5.99 3.87 0.25 6.88 4.60 4.33 2009 36 0.04 32.39 7.48 5.48 0.37 5.93 4.69 4.88 2010 34 0.99 35.41 9.53 9.16 0.28 4.99 3.95 3.59 2011 35 0.17 23.02 9.01 8.14 0.15 3.93 2.92 3.23 Total 342 0.04 35.63 7.26 5.82
25 Table 6 CEO Turnover CAR(-‐3, +3) CAR(-‐5,+5)
Model B Model C Model D Model B Model C Model D
CAR -‐0.3324c (-‐1.76) -‐0.2528 (-‐1.65) -‐0.3092c (-‐1.88) -‐0.2407 (-‐1.70) Ten0 -‐0.1170a (-‐2.93) -‐0.0912b (-‐2.01) -‐0.1291a (-‐3.02) -‐0.1037b (-‐2.38) Indep_board -‐0.0132c (-‐1.90) -‐0.0111a (-‐3.89) -‐0.0140c (-‐1.80) -‐0.0137c (-‐1.83) -‐0.0111a (-‐4.40) -‐0.0154c (-‐1.71) Sing_largest -‐0.1033c (-‐1.94) -‐0.0212b (-‐2.69) -‐0.0039c (-‐1.83) 0.0035 (0.82) -‐0.0811a (-‐3.06) -‐0.0085c (-‐1.78) Boardsize -‐0.0106 (-‐0.99) -‐0.0088 (-‐1.47) -‐0.0093 (-‐1,13) -‐0.0242 (-‐1.02) -‐0.0081 (-‐1.50) -‐0.0196 (-‐0.86) Eindex -‐0.0405 (-‐1.55) -‐0.0007 (-‐0.30) -‐0.0394 (-‐1.21) -‐0.0031 (-‐0.70) -‐0.0023 (-‐0.17) -‐0.0061 (-‐0.84) Duality -‐0.0349 (-‐0.65) -‐0.0231 (-‐0.30) -‐0.0459 (-‐0.77) -‐0.1043c (-‐1.80) -‐0.0339 (-‐0.61) -‐0.0909 (-‐1.34) DEratio 0.0099 (1.41) -‐0.0100 (-‐0.93) 0.0009 (0.90) 0.0303 (1.43) -‐0.0086 (-‐0.99) 0.0191 (0.77) Tobinq 0.0492c (1.83) 0.0499b (2.08) 0.0581c (1.79) 0.0921 (1.67) 0.0499a (2.55) 0.0659c (1.81) Ceo_age -‐0.0464a (-‐3.88) -‐0.0489a (-‐6.46) -‐0.0491a (-‐3.90) -‐0.0402a (-‐3.06) -‐0.0482a (-‐3.88) -‐0.0457a (-‐3.11) Ceo_share 0.0193 (0.58) 0.0140 (0.37) 0.0205 (0.67) -‐0.0079 (-‐0.70) -‐0.0126 (-‐1.55) -‐0.0099 (-‐0.92) Firm_age 0.0011c (1.89) 0.0003 (0.60) 0.0009c (1.78) 0.0025c (1.85) 0.0003 (0.73) 0.0008c (1.82) Firm_size 0.0038 (0.97) -‐0.0009 (-‐0.33) -‐0.0059 (-‐0.59) 0.0108 (0.38) -‐0.0029 (-‐0.30) 0.0044 (0.49) Industry -‐0.0699 (-‐1.21) -‐0.0269 (-‐1.00) -‐0.0412 (-‐1.42) -‐0.2143 (-‐1.54) -‐0.0405 (-‐0.94) -‐0.1078 (-‐1.33) Observation 335 335 335 329 329 329 Overall R2 0.0998 0.1101 0.1294 0.1040 0.1092 0.1309
Table 6 presents the result of whether the CEO’s term of office before the acquisition matters for the CEO turnover afterwards in the completed deals. Model B shows the relationship between the CARs and the forced CEO turnover after the acquisition; Model C shows the relationship between the CEO tenure before the acquisition and the forced CEO turnover afterwards; Model D regressed CARs and CEO tenure together with the forced CEO turnover after the acquisition. Significant levels are indicated by a, b and c for 1%, 5% and 10% respectively.