Reward CEO Transition?
by
Khasy Forouzfar
University of Groningen
Faculty of Business and Economics
MSc Thesis BA Marketing
Does The Financial Community
Reward CEO Transition?
by
Khasy Forouzfar
University of Groningen
ABSTRACT
Despite of calls for changing the pressure from short-term results, critics consider that the Wall Street community (e.g., institutional investors, financial analysts) still has an inherent preference for short-term performance. In recent years this trend has emerged and is spread towards chief executive officer (CEO) succession policies of firms. The number of years a CEO remains in office is drastically decreasing. With this trend of CEOs who are being fired, this study tests and tries to find empirical support for the hypotheses that the financial community rewards firms for CEO succession. The results that this study finds are related to criteria set in recent studies. It suggests that the financial community rewards CEO transition with increased short-term analyst following. Further, the ‘openness to change’ of a newly chosen CEO significantly impacts long-term changes in investor following. Overall the results suggest that firms should carefully consider the CEO transition process, initiated by the board of directors. When firms decide to engage in CEO transition, a CEO who is more open-minded about change is the most beneficial for long-term firm visibility in the financial community.
Key words: Investor Following, Analyst Following, Investor Relations, CEO, Transition,
Succession, Firm Visibility.
TABLE OF CONTENTS
1. INTRODUCTION 5
2. THEORY 8
CEO Transition And Financial Community Reactions 11
The Moderating Role Of ‘CEO Openness To Change’ 15
3. METHODOLOGY 18
Data Collection 18
Dependent Variables 18
Independent Variables 20
4. MODEL AND ESTIMATION PROCEDURE 25
Analyst Following Models 25
Investor Following Models 26
Endogeneity 26
5. RESULTS 28
6. DISCUSSION AND IMPLICATIONS 32
7. LIMITATIONS AND FURTHER RESEARCH 34
APPENDIX 36
1. INTRODUCTION
Recently there have been calls for changing the pressure from short-term results to encourage CEOs to return their focus on long-term value maximization (Bushee, 2001). Despite of these calls, critics consider that the financial community (e.g., institutional investors, financial analysts) still has an inherent preference for short-term performance. In recent years this trend has emerged and is spread towards chief executive officer (CEO) succession policies of firms. As Justin Lahart, reporter of The Wall Street Journal, states: “Companies have become more focused on the short term as well, with managers
concentrating on hitting short-term targets, such as analysts' quarterly earnings estimates, and as a result often forgoing measures that promote long-term growth, such as research and development -- or even routine maintenance”.
This development has caused pressure on CEOs to deliver quick results. If CEOs do not meet their short-term objectives, they are succeeded in rapid succession. David Langstaff (former CEO of Veridian Corp.) argues: “Any CEO feels the pressure of meeting Wall Street expectations”. Ultimately, CEOs are pressured to make operational and accounting decisions that boost short-term earnings at the expense of long-term value (Jacobs, 1991; Porter, 1992; Laverty, 1996; Bushee, 1998). About a short-term oriented view Bushee (1998) argues that losing sight of long-term oriented activities (e.g. R&D or other long-term investments) can put the long-term survival of a firm at risk.
The selection of a CEO is a fundamental organizational decision that has consequences for a firm’s strategy and performance. Not unexpectedly, in the past thirty years there has been widespread research on topics related to CEO succession (Datta, Rajagopalan and Zhang, 2003; Le Breton-Miller and Miller, 2006). CEO succession has been discussed even more in the past decade with the increasing financial market pressure on firms, which is caused by the preference of analysts and investors for short-term results (Bushee, 2001). This has caused firms that feel pressure of the financial community to take preliminary actions to boost short-term financial results.
One measure that offers quick results to the board of directors is a change in CEO, either forced or voluntary. Often, the CEO is the first who is punished when financial
related to positive abnormal stock returns (Huson et al., 2004). Eventually, this has caused a vicious cycle of high pressure on firms and decreasing CEO tenure, which is caused by the inherent preference for short-term results of the financial market (Bushee, 2001). In fact, the average CEO tenures have decreased in the past 20 years from about 8 to less than 4 years (Le Breton-Miller and Miller, 2006). If this trend continues, firms would not be able to focus on long-term results, decreasing their chances of survival (Bushee, 2001).
With the increasing power of the financial community, firms are pursuing Investor Communication (IR) strategies to hamper the potential negative effect of financial
community pressure on stock prices and firm visibility (Bushee et al., 2010). For firms it is essential to be visible in the financial community since financial analysts and institutional investors play an important role in a firm’s survival. For example, a firm that has a large investor “base” (i.e., the number of investors that are aware of the firm’s existence) will reduce its cost of capital (Bushee, 1998). Moreover, with its widespread use, IR strategies are becoming an important activity for many firms; however, there has been limited academic research that has focused on the IR process (Brennan and Tamarowski 2000).
To help put the intended contribution of this study in context, previous empirical research on CEO succession and IR is discussed. Table 1 represents a comparison chart of previous research in this field. The first two columns of Table 1 discuss two proxies to measure IR Activities that have been used in recent years, namely, analyst following and
investor following (Bushee and Miller, 2010; Bushee et al., 2010). Since institutional
investors and financial analysts have large incentives in the results of a firm, they spend time and effort to follow management activities (Dikolli, Kulp and Sedatole, 2009). News events (e.g. earnings calls, conference presentations) provide CEOs an opportunity to explain the firm’s “story”, to develop credibility and trust with financial community with the ultimate goal of increasing analyst and investor following through direct interactions (Jackson, 2007). While firms are seeking methods to increase their analyst and investor following, there is no research that examines the effect of CEO transition on analyst and investor following.
underlying psychological orientation and knowledge base (Datta and Rajagopalan, 1998; Keisler and Sproull, 1982). The ‘openness to change’ of a newly chosen CEO is likely to be reflected in strategic decisions made during the post-succession period.
Interestingly, none of the above research streams examine the direct effects of CEO succession on subsequent financial community response. Yet, it can be reasonably argued that the CEO transition events are likely to play a more prominent role in the financial community.
Table 1
Comparison of empirical studies on the effect of CEO transition on the financial community
Analyst Following Investor Following CEO transition
‘CEO openness to change’
Datta, Rajagolopan and Zhang (2003)
Not researched Not researched Focus on newly
chosen CEOs
Effect on strategic persistence Huson, Malatesta
and Parrino (2004)
Not researched Not researched CEO succession is
associated with positive abnormal
returns
CEO age and tenure is mentioned but not related to performance
measures Dikolli, Kulp and
Sedatole (2009)
Not researched High percentage of
institutional investor ownership causes
CEOs to focus on short-term earnings
Not researched Not researched
Bushee, Jung and Miller (2010) Conference presentations impact long-term analyst coverage of firms. Conference presentations impact long-term institutional investor following of firms.
Not researched Not researched
Bushee and Miller (2010)
IR initiatives result in an increase of analyst
following and media coverage
IR initiatives result in an increase of institutional ownership
Not researched Not researched
This study Examining the
effect of CEO transition and ‘openness to change’ on analyst following Examining the effect of CEO transition and ‘openness to change’ on investor following Focus CEO transition The relationship between ‘CEO openness to change’
For one, CEO transitions provide an important mechanism overcoming declining financial performance and resistance to change current strategies (Ocasio, 1994; Datta et al., 2003; Huson et al., 2004). In addition, it can be argued that the level of ‘openness to change’ of a newly selected CEO should have a significant influence on the firm’s strategic direction and subsequent analyst and investor following (Datta et al., 2003). However, empirical evidence on the relationship between CEO transition and analyst and investor following lacks in recent literature.
In summary, this study is encouraged by the need to address an important gap in CEO succession and IR literature. This study intends to answer the question whether the financial community rewards CEO transition and what role ‘CEO openness to change’ plays in the subsequent changes in analyst and investor following.
The current study makes a number of contributions to the CEO succession and analyst and investor following literature. First, while prior work has primarily examined only the event of succession, this study develops a more holistic construct (‘CEO openness to change’) that integrates two distinct CEO demographic characteristics (i.e. age and
organizational tenure). This provides the opportunity to examine the combined effects of the event of CEO transition and the successor’s distinct demographic attributes. Second, for research in marketing and finance this is the first to examine the changes in analyst following and investor following caused by CEO transition.
The remainder of this paper is organized as follows: the next section elaborates on the current theory and background on CEO succession and investor and analyst following. Then, based on the theory, hypotheses are proposed. In the subsequent section the research method and data collection is discussed followed with the results. This paper concludes with implications and limitations of this research, a well as a discussion of further research.
2. THEORY
This section develops relationships between CEO succession and the successor’s characteristics, and financial community reactions. Further, section describes the key theoretical constructs used in this study and in the following section the specific research hypotheses are developed.
Zuckerman (1999) describes that analysts and investors play a very significant and distinct role in the financial community. Hirsch (1972) argued that because of their
financial community increases their chances of survival (Bushee et al., 2010).
Alike the significant role analysts play in the financial community, the important role of institutional investors also has been well established in finance literature. Gibson,
Safieddine, and Sonti (2004) find that firms with high levels of institutional holdings
outperform their benchmark portfolios significantly. Further, Chemmanur, He, and Hu (2009) find that firms with high institutional holdings enjoy higher stock returns than firms with low institutional holdings. In summary, analysts and investors play distinct roles in the financial community. While high analyst following is associated with increasing firm visibility, high investor following has more financial benefits for a firm (e.g. higher stock returns, lower cost of capital). What both analysts and investors have in common is that firm’s IR
communication strategies are directed towards them.
The motive for firms to engage in communication with the financial community is to increase their analyst and investor following. To stress the important effect of investors and analysts even more, Lehavy and Sloan (2008) suggest that having high analyst and investor following impacts stock prices even more than firm fundamentals. When a firm fails to become visible in the financial community, it has a negative impact on firm survival and stock prices (Zuckerman, 1999). A firm that is visible in the financial market generally has high levels of analyst and investor following (Bushee and Miller, 2010). Analyst following relates to the number of analysts who are covering a firm (i.e. issuing earnings forecasts and buy/hold/sell recommendations), while investor following relates to the percentage of a firm’s stock that is hold by institutions (Bushee and Miller, 2010).
Moreover, analyst and investor play distinctive roles in the financial community. Bushee and Miller (2010) argue that analyst and investor following are distinctive on the short-term and on the long-term. While firms may initially respond positively towards a certain news event (assed by the one-quarter changes in analyst and investor following), they might adjust their long-term reaction (the yearly changes in analyst and investor following). For example, when a firm discloses high earnings expectations, the initial (short-term) market reaction might be positive, which will result in an increase of a firm’s analyst and investor following in the following fiscal quarter. However, when the firm fails to live up to the earnings expectations, on the long-term, analysts and investors might punish the firm by decreasing their coverage and holdings in the firm.
commonly react negatively towards firms that report fewer earnings than expected (Bushee, 2001). Further, when a new CEO is appointed, analysts and investors have the opportunity to experience the new CEO (Bushee and Miller, 2010). Therefore, the characteristics of the newly chosen CEO play a prominent role in how the financial community responds on the long-term.
The characteristic of the newly chosen CEO is measured through the concept of ‘openness to change’, which integrates two CEO demographic characteristics that have been most extensively related to firm-level changes in prior research, namely, CEO firm tenure, education and age. In general, high levels of firm tenure, lower levels of education, and higher executive age have all been linked to low levels of propensity for change, greater levels of risk aversion, use of limited information sources etc. (Rajagopalan and Datta, 1996; Wiersema and Bantel, 1992). While these three demographic characteristics are distinct theoretical constructs, they appear to have similar effects in predicting CEO actions and organizational change. Using a integrative construct, such as ‘CEO openness to change’, provides to opportunity to capture a CEO’s underlying cognitive orientation. This construct has proven to be a reliable predictor of CEO’s willingness to change current strategies (Datta and Rajagopalan, 1998; Finkelstein and Hambrick, 1996). These studies find that higher ‘CEO openness to change’ is associated with lower firm tenure and age and high education level. In contrast, low openness to change is associated with lower education and greater firm tenure and age. Although this construct has proven to be reliable, this study will demonstrate later in this paper that age and tenure are empirically related to each other.
FIGURE 1
Conceptual Model
CEO Transition And Financial Community Reactions
This section starts by elaborating on the role analysts and investors separately play in the financial community. Next, hypotheses are developed to relate CEO transition to short-term and long-term financial community reactions, respectively. Finally, the moderating role of ‘CEO openness to change’ is discussed and subsequently, hypotheses are developed that address the moderator effect.
Analyst following. For a firm, an upturn in analyst following is associated with improved
visibility on the financial market, which in time will enhance the ability of firms to raise equity capital (Bushee and Miller, 2010). In fact, firms with high analyst coverage can benefit from the positive influence that analysts have on the financial market.
To be more specific, analysts influence the financial market primary through their
information acquisition and reporting activities, with the goal of improving the informational efficiency (Givoly and Lakonishok, 1979; Healy and Palepu, 2001; Barth and Hutton, 2004). Improved informational efficiency has two benefits. First, improved information reduces the overall information uncertainty about a firm’s expected cash flows, which in time decreases
the probability of default1 (Merton 1974). Second, improved market information also reduces
information asymmetry, which will narrow the differences between informed insiders and all of the other participants in the financial market. In time, a low level of information
asymmetry will lead to a decrease in information risk (Easley and O’Hara, 2004), which ultimately lowers a firm’s credit risk (Francis et al. 2005; Mansi et al. 2006).
Investor following. Several studies stress the importance for firms to engage in IR
activities directed towards attracting investors (Bushee and Miller, 2010; Bushee et al., 2010). Bushee and Miller (2010) define that the key drivers of the IR strategy should be focused on improving visibility in the financial community. As for the effectiveness of IR, Bushee and Miller (2010) illustrate that firms initiating IR programs experience significant improvements in market valuation (i.e., reductions in the book-to-price ratio). Comparable to analyst following, the key theoretical argument for increasing investor following is that it will improve the visibility of a firm in the financial market.
From a firm’s perspective, high visibility is beneficial because it can reduce risk. Bhojraj and Sengupta (2003) contend that a high level of investor following will reduce credit risk because of two reasons. First, since institutional investors monitor managerial performance, rating agencies believe that this will reduce agency costs. Agency theory suggests that a conflict of interest can arise between the goals of the board of directors and the CEO. The board is held to be interested in increasing shareholder value by maximizing stock prices, where a CEO is motivated by personal goals (Antia et al. 2010). When a situation occurs that short-term performance is declining, the board is faced with finding a way to ensure that the CEO will act in the shareholder’s benefits (i.e. agency costs). Second, institutional investors pressure firms to disclose as much as possible information to the market and thus reducing information risk.
While the benefits of investor following are similar to analyst following, one should consider these two participants of the financial market as dissimilar. However, an analyst’s
1 Probability of default refers to the degree of likelihood that the borrower of a loan or debt
decisions to follow a firm and an investor’s decisions to invest in the same firms are
somewhat jointly determined (O’Brien and Bhushan, 1990). On the hand side, analysts are held to follow firms with large institutional holdings because institutional investors are their primary customers because of their interest in analyst recommendations and forecasts. On the other hand side, institutional investors are held more likely to invest in firms with large analyst coverage because of the marketing of brokerage’s services.
Although analyst following and investor following are closely associated, both play a different role in the financial community. The main difference is that analysts are primarily focusing on providing information by producing earnings forecasts and stock
recommendations for firms. While in contrast, institutions invest in firm’s stocks and participate in equity ownership.
In sum, the link between firm benefits and high levels of analyst following and investor following has been established in prior literature. However, the link between analyst
following and investor following and CEO transition is less straightforward. This study hypothesizes that a CEO transition is associated with changes in analyst following and investor following. These hypothesized relationships are discussed in the subsequent sections.
CEO transition and short-term financial community reactions. This study expects
that a CEO transition will lead to a positive change in short-term analyst following and investor following. These hypotheses are based on the ‘improved management’ theory, which holds that CEO transition will increase managerial skills and consequently, future firm performance (Huson et al., 2004). In other words, quality is not directly observable and varies across managers. This suggests that if a firm’s performance is poor, the board of directors conjectures that the current CEO is of low quality and that the expected benefit of replacing him exceeds the expected cost. When another CEO is appointed, the expected quality exceeds that of his predecessor.
This is expected because both analysts and investors have a preference for short-term results (Bushee, 2001). Therefore, it is expected that analysts and investors will react positively when a new CEO is appointed because this gives beneficial perspective on improved short-term results. Consequently, more analysts follow a firm and investors will increase their holdings in the firm on the short-term.
future firm performance is to increase following a CEO transition. Formally, this leads to the following hypotheses:
H1a: A CEO transition will positively influence changes in short-term analyst following for
the firm.
H1b: A CEO transition will positively influence changes in short-term investor following for
the firm.
CEO transition and long-term financial community reactions. On the long-term, this
study expects that a CEO transition will lead to a negative change in analyst following and investor following.
The reason for a negative change in long-term analyst following and investor following is based on the ‘scapegoat’ theory (Shavell, 1979; Mirrlees, 1976). In contrast to the improved management hypothesis, the scapegoat theory holds that quality does not vary across managers. When the poor performance of a firm continues even when the CEO is replaced, the scapegoat theory explains that the poor performance is from chance alone rather than low managerial quality. In other words, the continuation of bad performance is not because of bad management but is the result of bad luck. Consequently, a manager who is fired for poor performance can be viewed as a scapegoat.
In the long-term this might negatively influence financial community reactions when analysts and investors come to the conclusion that the new CEO has the same managerial skills as the previous CEO. Therefore, the turnover phenomena itself does not increase managerial quality or expected firm performance. In summary, financial analysts, on the long run expect that future performance will not increase following the change in CEO. In a similar fashion, it is expected that investors, on the long-term, will decrease their holdings in a firm when they come to the conclusion that the new CEO is of the same quality of the previous CEO. Formally, this leads to the following hypotheses:
H2a: A CEO transition will negatively influence changes in long-term analyst following for
the firm.
H2b: A CEO transition will negatively influence changes in long-term investor following for
The Moderating Role Of ‘CEO Openness To Change’
Analysts and investors primary maintain a close relationship in order to retain access to close firm information. Once an analyst takes up reporting of a firm, or a institutional investor has holdings in a firm, a very real relationship exists between the analyst or investor and CEO (Zuckerman 1999, 2000).
It is expected that the CEO characteristic: ‘CEO openness to change’ has an effect on the decision of an analyst or investor to follow a firm. There are numerous situations where the financial community and CEOs can interact with each other. Interviews, conference calls, conference presentations and road shows are venues that CEOs frequently visit and interact with members of the financial community (Bushee and Miller, 2010). Multiple interactions between CEOs and the participants of the financial community also provide analysts and investors an opportunity to experience a newly chosen CEO, which can consequently help them to assess the ‘credibility and trust’ of a firm. In terms of this study, the evidence that interactions between CEOs and the financial community exist provides the ability to examine how CEO transitions affect the financial market response (i.e. changes in analyst and
investor following).
This study consequently expects that during these interactions, analysts and investors are influenced by a CEO’s openness to change. Analysts and investors could have more confidence in a newly chosen CEO who is open to change the current strategy of a firm instead of a CEO who is reluctant to change. Therefore, this study expects that a positive change in ‘CEO openness to change’ will result in a positive change in analyst and investor following.
‘CEO openness to change’ is a combined construct that takes effects between
individual demographic attributes into account. For example, while one might associate high firm tenure with lower openness to change, this effect may be counterweighted by a low age. CEO succession literature indicates that researchers use the construct of CEO ‘openness to change’ to examine individual demographic characteristics (Datta et al., 2003). The ‘CEO openness to change’ construct is based on the strategic choice paradigm (Child, 1972), which proposes that CEOs have significant control over an organization’s future direction. In line with this literature stream, Smith and White (1987) found systematic relationships between the characteristics of new CEOs (e.g., age, tenure, education background) and firm strategy.
directions (Finkelstein and Hambrick, 1996; Datta et al., 2003; Huson et al., 2004). CEOs with higher levels of organization tenure are likely more rooted with organizational routines and are not likely to change current strategies (Daft and Weick, 1984), executives with long tenures have invested time and effort in the current strategy and often have more to lose than benefit from strategic changes (Hambrick et al., 1993). Likewise, CEO age is associated with age low openness to change. Wiersema and Bantel (1992) argue that age expresses greater commitment to past strategies, limited exploration of new alternatives and ultimately lowering the likelihood of a CEO being open to change.
‘CEO openness to change’ and short-term financial community reactions. In line
with the ‘improved management’ theory, which holds that CEO transition will increase managerial skills and firm performance (Huson et al., 2004), it is expected that the market response is stronger when the new CEO is more ‘open to change’. This implies that when the newly chosen CEO is more open to change, there is a greater likelihood that the firm will not continue with the same strategy.
Since the principal reason for CEO transition is declining performance, a new strategy could result in better short-term firm performance, which could lead to a stronger positive change in short-term analyst and investor following. Therefore, it is expected that investors will react stronger when a new CEO is willing to change the current strategy. When a CEO is appointed who is more open to change can give beneficial perspective on improved term results. Consequently, investors will increase their holdings in the firm on the short-term and analyst will increasingly cover a firm.
In summary, ‘CEO openness to change’, as reflected in lower firm tenure and lower age should lead to stronger short-term market reactions (analyst and investor following). More specifically, it is expected that the short-term changes in analyst and investor following are stronger when the new CEO is ‘open to change’. This leads to the following hypothesis:
H3a: The positive impact of CEO transition on short-term analyst following and short-term
investor following is stronger, given a positive change in ‘CEO openness to change’.
‘CEO openness to change’ and long-term financial community reactions. The
However, the ‘openness to change’ of the newly chosen CEO can play a moderating role here.
Several studies have shown that CEOs display different tendencies for change, some might strive for change while others put in effort to maintain the status quo (Hambrick Geletkanycz and Fredrickson, 1993). In a similar fashion, there are CEOs who have a more open mind-set about change than others (Hambrick and Mason, 1984). For example, a newly chosen CEO might be committed more to change because of the belief that the firm can benefit from changing the current strategy. Therefore, when a firm’s strategy is focused on short-term results, a CEO with low ‘openness to change’ is not held likely to change the course of that strategy. On the other hand side, a newly chosen CEO who is more ‘open to change’ is held more likely to change the current strategy. The financial community might have more confidence in a CEO who is more open-minded about change, which could lead to a positive long-term market reaction since the financial community will give the new CEO time to change current strategy.
In summary, ‘CEO openness to change’, as reflected in lower firm tenure and lower age should lead to positive long-term market reactions (analyst and investor following). More specifically, it is expected that the long-term changes in analyst and investor following are positive (instead of negative) when the new CEO is ‘open to change’. This leads to the following hypothesis:
H3b: The negative impact of CEO transition on long-term analyst following and long-term
3. METHODOLOGY Data Collection
This study uses a random sample that consists of CEO successions in 100 of US firms. However, this study is restricted to common stocks found in the Center for Research in Security Prices (CRSP) file; this includes all stocks listed on the New York Stock Exchange (NYSE), American Stock Exchange (AMEX), and National Association of Security Dealers Automated Quotations (NASDAQ) National Market System. The data is collected from four different sources. CEO turnovers and characteristics are identified using the Standard & Poor’s (S&P) ExecuComp database. Throughout this study the terms "investors" and
"institutional investors” are used as synonyms for "an institution that files a 13F." The 13F of Thomson Reuters requires investors to report all stock over which they employ ownership. The institutional investor ownership data is consequently obtained from the Thomson Reuters Form 13F database. Analyst following data is accessed trough the Thomson Reuters I/B/E/S database and finally, data on firm characteristics used as control variables are obtained from the COMPUSTAT database.
The data is collected on quarterly basis during the first quarter of 2002 until the last quarter of 2009. The sample is then filtered according to the following criteria: each firm had to record at least one transition in the time period during 2002 – 2009. Further, each firm should have analyst coverage and shares owned by institutional investors.
After acquiring this data, the selection criteria where applied. First, firms must have sufficient data in the Thomson Reuters, I/B/E/S and COMPUSTAT databases. A satisfactory level of data is needed to compute new variables required to perform our tests, which are discussed in the next section. Firms that did not record a CEO transition and had no analyst or investor following where excluded from the sample. The Appendix of this study gives a detailed explanation of each step that the selection criteria where applied. Further, the Appendix also explains the merging process of the four different datasets used in this study.
This yielded a final sample of 129 CEO transitions in 70 firms between Q1 2002 and Q4 2009.
Dependent Variables
2001; Gompers and Metrick, 2001; Botosan, 1997). Institutional investor following is
measured trough institutional ownership, which is defined as the percentage of a company's stock that is owned by institutional investors (Gompers and Metrick, 2001). Bushee et al. (2010) examined the optimal longer-window changes in analyst and investor following. In line with Bushee et al. (2010), this study uses the four-quarter changes in analyst and
investor following to capture long-term market reaction. In addition, this study also examines shorter-term changes in analyst and investor following.
Analyst following. To ensure that this study fully captures the impact of ‘CEO openness to
change’ and CEO transition on analyst following, both one-quarter changes and four-quarter changes in analyst following are used (Bushee et al., 2010).
The short-term change in analyst following is computed in two steps. First, analyst following (LNANL) is computed as the log of one plus the number of analysts earnings forecasts released, with a forecast period indicator (FPI) of one year, during each fiscal quarter. The forecast period is limited to one year because the number analysts issuing a forecast appear to remain constant over FPI’s (Botosan, 1997). Second, the one-quarter changes in analyst following (LNANL1) are calculated as the difference between LNANL in quarter (t+1) and LNANL in quarter (t).
To measure long-term analyst following, the four-quarter changes in analyst and investor following is used (Bushee et al., 2010). The four-quarter changes in analyst following (LNANL4) are calculated as the difference between LNANL in quarter (t+4) and LNANL in quarter (t).
Investor following. To capture the impact of ‘CEO openness to change’ and transition on
investor following, one-quarter changes and four-quarter changes in investor following are used (Bushee et al., 2010).
Short-term investor following is calculated in two steps. First, the percentage
institutional holdings (PIH) is computed as the total shares owned by institutions divided by the total shares outstanding at the end of each fiscal quarter. Second, the one-quarter change in institutional following (PIH1) is calculated as the difference between PIH in quarter (t+1) and PIH in quarter (t).
Independent Variables
CEO transition. A measure of CEO transition (TRA) is created to examine the impact of
CEO analyst and investor following. The ExecuComp database provides detailed information about who the CEO is in each fiscal quarter. With this information, a new variable is created and is coded 1 if a CEO transition has taken place during a fiscal quarter and 0 if the CEO was still in office at the end of the fiscal quarter.
‘CEO openness to change’. The original ‘CEO openness to change’ (OTC) construct is a
composite measure of three demographic indicators, namely, age, organizational tenure and education level. However, this study has no access to information about the educational level of CEOs. Obviously, this is a limitation, which is addressed in the “Limitations and Further Research” section of this paper. However, the measure is still reliable according to Zajac and Westphal (1996) when only age and tenure are used. Zajac and Westphal (1996) find that both age and tenure are reliable predictors of ‘CEO openness to change’. The CEO age was measured as the present age at a fiscal quarter, and organizational tenure was measured as the number of years the CEO was in office prior to the fiscal quarter (Singh and Harianto, 1989).
As for the sample, the mean age of the CEOs is 60 years with an average tenure of 5 years. This is in line with earlier research that shows the decreasing trend of CEO tenure (Le Breton-Miller and Miller, 2006). Before the measure is actually used in this study, the validity the theoretical argument that these two demographic variables represent ‘CEO openness to change’ was measured trough principal components factor analysis.
TABLE 2
Factor Loadings And Correlations
Variables Factor Loadings
(Eigen value = 1.105) Correlations
‘CEO openness to
change’ Age Tenure
Age -0.847 -0.607* -
Tenure -0.847 -0.816* 0.435* -
Significance level: *p<0.01
To capture the ‘openness to change’ of a newly chosen CEO, an interaction between CEO transition and ‘CEO openness to change’ is calculated. In general, interaction effects represent the combined effects of predictor variables on the dependent variable. When an interaction effect is present, the impact of one variable depends on the level of the other variable. Pedhazur and Schmelkin (1991) highlight the importance of multiple effects in research because when interaction effects are present, it means that the interpretation of the individual variables may be misleading.
To capture this simultaneous effect, a new variable is created (OTC*TRA). The interaction variable is calculated by multiplying CEO transition with the change in ‘CEO openness to change’ measure. In sum, the interaction effect of this study expects that the reaction of the strength of the financial community reactions will depend on the ‘openness to change’ of a newly chosen CEO.
Control variables. The model also controls for a large number of firm characteristics that
prior literature finds are associated with financial market reaction and firm performance. Information about firm profitability and growth are associated with financial market reaction (Bushee et al., 2010). A number of measures for profitability and growth are used, including the earnings-price ratio (EP), dividend yield (DP), and the book-to-price ratio (BP).
Furthermore, literature in finance has also found evidence that firm risk is an important predictor of stock market reaction (O’Briend and Bhushan, 1990), hence the debt-to-assets leverage ratio (LEV) is used to proxy for firm risk. This paper also controls the effects of total assets (TA), return on assets (ROA), firm size (LMV) and liquidity (LIQ). Finance and
Further, this study expects that the financial community reactions have been impacted by financial crisis between 2007 and 2009. More specifically, the crisis began midway through 2007 and continued through the end of 2008. The crisis that almost caused a systemic collapse of the financial sector contributed in the collapse of a large number of
investment banks, including Bear Stearns, Lehman Brothers, Merrill Lynch and Wachovia.
These events caused a loss of confidence of investors in the banking system. Banks responded by tightening their lending standards, which instigated another wave of credit crisis. In turn, institutional investors tried to reduce their risk exposure by selling large numbers of assets at discount. In sum, several studies point out that the financial instability during this period affected the financial sector with investors being reticent (Barth and Landsman, 2010; Guo, Chen and Huang, 2011). When the financial community responds differently in the midst of a crisis, this study expects that analyst and investors will react differently during a crisis. To control for this phenomena, a dummy variable is created to
capture the effect of the financial crisis on analyst and investor following. The measure is
coded 1 during the financial crisis and 0 in the time period before/after the financial crisis.
This study also assumes different market reactions in each industry. This is expected because different industries display different patterns of disclosure and subsequent financial community reactions (Botosan, 1997). Industries tend to provide different levels of
disclosure. For example, while some industries might provide less disclosure to the financial community other industries prefer more and voluntary disclosure. When firms use different disclosure measures in different industries, it necessitates this study to take the industry differences into account. The firms that are used in this study’s sample are from six different divisions. Botosan (1997) argues that especially the finance division (i.e. insurances, real estate firms) is mainly different in disclosure activities than the rest of the divisions. To capture the industry effect, a dummy variable is included that is coded 1 if a firm is in the
finance division and 0 if not.
Table 3 illustrates all variables used in this study, including their
TABLE 3
Variable Definition and Literature Support (with WRDS variables
codes between brackets)
Source
Dependent Variables
Analyst Following (LNANL1)
The first step is to obtain LNANL, which is the log of 1 plus the number of analysts issuing earnings forecasts (numest) with a forecast period indicator of 1 year, during each fiscal quarter
(Bushee et al., 2010). To obtain LNANL1, the difference
between LNANL in quarter (t+1) and LNANL in quarter (t) is
calculated.
I/B/E/S
Institutional Investor Following (PIH1)
The first step is to obtain PIH, which is the total shares owned
by institutions at the end of a fiscal quarter (shares), divided
by the total shares outstanding at the end of a fiscal quarter (shrout2*100). To obtain PIH1, the difference between PIH in
quarter (t+1) and PIH in quarter (t) is calculated (Bushee et
al., 2010).
Thomson Reuters
Analyst Following (LNANL4)
To obtain LNANL4, the difference between LNANL in quarter
(t+4) and LNANL in quarter (t) is calculated (Bushee et al.,
2010).
I/B/E/S
Institutional Investor Following (PIH4)
To obtain PIH4, the difference between PIH in quarter (t+4)
and PIH in quarter (t) is calculated (Bushee et al., 2010).
Thomson Reuters
Independent Variables Definition and Literature Support Source
Transition (TRA) Proxy for the market reaction surrounding a conference call is
driven by a transition of a CEO, a variable is included (TRA) that equals 1 if there has been a CEO transition in a fiscal quarter.
ExecuComp
‘CEO openness to change’
Age (age) and Tenure (becameceo-year) of a CEO are
converted by subtracting each observation’s values from the highest values in the sample (Zajac and Westphal, 1996). Then, these two converted measures and CEO educational
level were standardized (mean=0, standard deviation=1) and summed to yield a composite measure of ‘CEO openness to change’.
Control Variables Definition and Literature Support Source
Earnings-price ratio (EP) EP is the earnings-to- price ratio at the fiscal quarter. Calculated as stock price (prccq) divided by earnings per share (actual) (Bushee et al., 2010).
COMPUSTAT
Dividend yield (DP) DP is the dividend-to-price ratio at the fiscal quarter. The
dividend yield is the return on investment for a stock.
Computed as dividends per share (dvtq) divided by price per
share (prrccq) (Bushee et al., 2010).
COMPUSTAT
Book-to-price ratio (BP) BP is the book-to-price ratio at the fiscal quarter. Calculated by
dividing the current closing stock price in a fiscal quarter (prccq) with the total assets (atq) minus intangible assets (intanq) and liabilities (Bushee et al., 2010).
COMPUSTAT
Firm size (LMV) Defined as the log of market value at the end of a fiscal
quarter (mkvaltq) (Bushee et al., 2010).
COMPUSTAT
Leverage Ratio (LEV) The total long-term debt (dlttq) to the sum of long-term debt
and market value (mkvaltq) (Hong and Sarkar, 2007).
COMPUSTAT
Total Asset (TA) The logged value of total assets of a firm (atq) (Beaver,
Kettler, and Scholes, 1970).
COMPUSTAT
Return On Assets (ROA) The ratio of operating income (oibdpq) to total assets (atq)
(Hong and Sarkar, 2007). COMPUSTAT
Liquidity (LIQ) The current liquidity ratio of a firm, which is the current assets
(actq) divided by current liability (ulcoq) (Beaver, Kettler, and Scholes, 1970).
4. MODEL AND ESTIMATION PROCEDURE
The objective is to assess the impact of CEO transition and openness to change on analyst following and investor following. Since the objective is to measure long-term reactions, this study measures, both one-quarter and four-quarter changes of analyst and investor
following. This leads to four regression models.
The four regression models examine the underlying differences in firms by means of a difference-to-difference approach. By comparing firm-specific (first order) difference in analyst following and investor following as well as the control variables, the models tend to control for a large number of potential remaining differences (Armstrong et al., 2010).
Further, differences in ‘CEO openness to change’ (ΔOTC) are incorporated in the model to capture the effect of ‘CEO openness to change’ of current CEOs on analyst and investor following. Finally, since the dependent variables are measured as a percentage (PIH) or a natural logarithm (LNANL), it is helpful to have other variables as percentages or differences (Gompers and Metrick, 2001). Thus, the independent variables use first
differences except for the dummy variables that measure CEO transition and the financial crisis.
Short-term Reactions Models
Two models are estimated to capture the effect of CEO transition and ‘CEO openness to change’ on short-term analyst following and investor following. Equation (1) captures the effect of the predictor variables on one-quarter changes in (short-term) analyst following. Equation (2) captures the effect of the predictor variables on one-quarter changes in (short-term) investor following.
1 !"#"!1!" = γ1 [Δ!"#!"] + γ2 !"#!" + γ3 !"#!"∗ Δ!"#!" + γ4 Δ!!" + γ5 DUM(FC)
+ γ6 DUM(FI) + !!"
2 !"#1!" = γ1 [Δ!"#!"] + γ2 !"#!" + γ3 !"#!"∗ Δ!"#!" + γ4 Δ!!" + γ5 DUM(FC)
+ γ6 DUM(FI) + !!"
Where;
TRA = CEO transition for firm i in quarter (t)
!!" = Control variables for firm i in quarter (t)
DUM(FC)= Dummy variable for financial crisis (FC) DUM FI = Dummy variable for finance industry (FI)
!!" = Random error term
Long-term Reactions Models
Likewise, two models are estimated to capture the effect of CEO transition and ‘CEO openness to change’ on long-term analyst following and investor following. Equation (3) captures the effect of the predictor variables on four-quarter changes in (long-term) analyst following. Equation (4) captures the effect of the predictor variables on four-quarter changes in (long-term) investor following
3 !"#"!4!" = γ1 [Δ!"#!"] + γ2 !"#!" + γ3 !"#!"∗ Δ!"#!" + γ4 Δ!!" + γ5 DUM(FC)
+ γ6 DUM(FI) + !!"
4 !"#4!" = γ1 [Δ!"#!"] + γ2 !"#!" + γ3 !"#!"∗ Δ!"#!" + γ4 Δ!!" + γ5 DUM(FC)
+ γ6 DUM(FI) + !!"
Where;
LNANL4 = The difference between LNANL for firm i in quarter (t+4) and LNANL in quarter (t) PIH4 = The difference between PIH for firm i in quarter (t+4) and PIH in quarter (t) OTC = ‘CEO openness to change’ for firm i in quarter (t)
TRA = CEO transition for firm i in quarter (t)
!!" = Control variables for firm i in quarter (t)
DUM(FC)= Dummy variable for financial crisis (FC) DUM FI = Dummy variable for finance industry (FI)
!!" = Random error term
Endogeneity
The reason of not including investor following as a predictor in equations (1) and (3), and
analyst following in equations (2) and (4) is the endogenous relationship between analysts
services. On the other side institutional investors are interested in firms who have high analyst coverage because of the services provided by analysts (e.g. earnings forecast, buying recommendations).
More direct evidence of this endogenous relationship is obtained from estimating a regression model with the inclusion of the endogenous variable. In the test, a regression model with the dependent variable term analyst following’ with the inclusion of ‘short-term investor following’ and other predictor variables was estimated. Leeflang et al. (2002) describe that endogeneity can be detected by identifying correlation between independent variables and the error term. Consequently, a correlation matrix was used to detect possible signs of endogeneity between analyst following and investor following. In support of the expectation, the correlation between the error term and the investor following was significant. This process was repeated to test for endogeneity with long-term analyst and investor following. This produced the same result as the initial test. Therefore, this study does not include investor following as a predictor in the analyst following models, and vice versa.
To address endogeneity, Leeflang et al. (2002) suggest a two-stage least square (2SLS) estimation technique. This technique used when the dependent variable’s error terms are correlated with an independent variable (i.e. analyst following and investor following). In 2SLS the basic assumption of the ordinary least square method that the value of the error terms is independent of predictor variables is relaxed. In OLS, when the assumption is broken, 2SLS solves the problem by assuming that there is a secondary predictor that is correlated to the problematic predictor but not with the error term. However, Antia et al. (2010) find that 2SLS estimation in measuring the effect of CEOs on firm performance show identical results as OLS regression, which suggests that the former analysis technique is unnecessary.
5. RESULTS
Descriptive statistics and correlations. Table 4 outlines the descriptive statistics and
correlations between the variables that are used in this study. The lack of high correlations between any of the independent or control variables indicates that multicollinearity is not an issue with the data. Unexpectedly the correlations between the key predictors and the dependent variables are not significant.
Although most of the correlations are not significant, as shown in Table 4, the
correlations of CEO transition are in the expected direction of long-term changes in investor following. Further, the correlations of CEO transition are also in the expected direction of short-term analyst following. Unfortunately, these relationships are not significant. According to Leeflang et al. (2002), one cannot draw any conclusions from a correlation that is in the expected direction but not statistically significant. Solely based on the correlations, there is a lack of evidence to support any of the hypotheses.
Not finding any significant correlations between variables does not automatically imply that there is no relationship between the variables. The lack of any significant correlations could be because of a limitation of the ‘Pearson’s correlation coefficient’, as shown in Table 4. This coefficient assumes that each pair of variables is bivariate normal and it is a measure of linear relationship. The variables of this study could be related, but since the relationship is not linear (e.g. LNANL1, TRA), the Pearson’s correlation coefficient alone is not an appropriate statistic for determining whether a relationship exists. The lack of any significant correlation between the dependent variables and key independent variables is further addressed in the “Limitations and Further Research” section.
TABLE 4 Correlations Matrix Variables N Mean S.D. 1 2 3 4 5 6 7 LNANL1 2074 0 0.33 1.00 PIH1 1846 0.84 10.71 -0.02 1.00 LNANL4 1886 0 0.41 0.45** 0.01 1.00 PIH4 1635 2.93 15.25 0.02 0.48** 0.03 1.00 TRA 1398 0.09 0.29 -0.03 0.03 -0.03 0.01 1.00 ΔOTC 135 1.50 1.58 0.05 0.08 0.13 -0.04 -0.27** 1.00 TRA*ΔOTC 1259 0.01 0.22 -0.03 0.06 -0.01 0.01 0.38** 0.31** 1.00
the results indicate that there is a significant association between one-quarter changes in analyst and investor following with CEO transition and ‘CEO openness to change’. Further, correlations between predictors were within acceptable ranges and VIFs (Variance Inflation Factor) were lower than 10, suggesting that multicollinearity was not a threat to the validity of the findings. The regressions adjusted R- square is 16% for the analyst following model and 14% for the investor following model.
TABLE 5
Short-term Reactions Models Short-term Analyst Following (LNANL1) Short-term Investor Following (PIH1)
TRA 0.14* 0.79
ΔOTC 0.04* 0.44
TRA*ΔOTC -0.15*** 2.37**
Δ Total Assets 0.00 0.00
Δ Liquidity 0.01 0.10
Δ Book-to-price Ratio 0.01 2.44
Δ Dividend Yield -8.09 33.31
Δ Earnings-price Ratio 0.01** -0.01
Δ Firm Size -0.42 -3.26
Δ Leverage Ratio -1.50 -22.34
Δ Return On Assets -2.50 -81.51
Crisis 0.02 -1.99
Finance Sector -0.01 3.35
Adjusted R-square 0.16 0.14
*** (p<0.01) ** (p<0.05) * (p<0.10)
The results of the control variables are unfortunately not significant. The only exception can be found in the short-term analyst following model. The results show that changes in earnings-to-price ratio have a positive impact on short-term analyst following (0.01, p < 0.05). This is in line with prior work in finance (Bushee and Miller, 2010)
Long-term Reactions Models. Table 6 provides regressions of the four-quarter changes in
investor following (LNANL4) and the four-quarter changes in investor following (PIH4).
Further, the correlations between predictors were within acceptable ranges and VIFs (Variance Inflation Factor) were lower than 10, implying that multicollinearity was not threat to the validity of the findings. The regressions adjusted R- square is 8% for the long-term
analyst following model (LNANL4) and 19% for the long-term investor following model
(PIH4).
TABLE 6
Long-term Market Reactions Long-term Analyst Following (LNANL4) Long-term Investor Following (PIH4) TRA -0.23* -1.69 ΔOTC -0.01 -0.58 TRA*ΔOTC -0.07 4.11** Δ Total Assets 0.00 0.00 Δ Liquidity 0.03 -4.25 Δ Book-to-price Ratio -0.34 -8.64 Δ Dividend Yield 19.84 317.98 Δ Earnings-price Ratio 0.00 0.05* Δ Firm Size -0.12 19.54 Δ Leverage Ratio -2.37 5.81 Δ Return On Assets -6.30 -206.39 Crisis 0.19 -10.18* Finance Sector 0.17 5.63 Adjusted R-square 0.08 0.19 *** (p<0.01) ** (p<0.05) * (p<0.10)
model does not explain the variation in the long-term analyst following data. The lack of exploratory power of the long-term analyst following model is further addressed in the “Limitations and Further Research” section.
The results of the long-term analyst following model are in support H2a, that is, a CEO transition will negatively influence changes in long-term analyst following for the firm (-0.24, p < 0.10). However, the overall regression is not significant and lacks explanatory power, this will be further addressed in the ‘Limitations And Further Research’ section. The results of the long-term investor following regression model are also not in line with the expectation of H2b, that is, a CEO transition will negatively influence changes in long-term investor following for the firm (-1.69, sig. 0.68).
The results of the control variables are unfortunately also not significant for the long-term analyst following and investor following models. The only exception can be found in the long-term investor following model. The results show that changes in earnings-to-price ratio have a positive impact on long-term analyst following (0.05, p < 0.10). Further, the financial crisis dummy also has a significant impact on long-term changes in investor following (-10.18, p < 0.01). This negative impact is in line with literature (Botosan, 1997), which indicates that investors are held less likely to increase their institutional ownership during a financial crisis.
Moderating role of ‘CEO openness to change’. The results only partially support
hypothesis H3a, that is, the positive impact of CEO transition on short-term analyst following and short-term investor following is stronger, given a positive change in ‘CEO openness to change’. The results of Table 5 indicate that CEO transition has a larger positive effect on investor following when the CEO is more ‘open to change’ (2.37, p<0.05). In comparison, the sole phenomenon of CEO transition has a regression coefficient of 0.79 as opposed to the regression coefficient of 2.37 of a CEO who is more ‘open to change’. In addition, the results of ‘CEO openness to change’ provide evidence that there is a positive relationship between short-term analyst following and ‘CEO openness to change’ (0.04, p <0.10).
6. DISCUSSION AND IMPLICATIONS
Recent literature suggests that a primary objective of IR strategies is to impact a firm’s visibility in the financial community by attracting financial analysts and investors, in particular institutional investors (Elgin, 1992; Byrne, 1999; Bushee and Miller, 2010). Additionally, IR strategies often tend to use financial analysts to increase firm visibility and attract investors (Bushee and Miller, 2010). In summary, the findings of recent literature suggest an
important role for IR in simultaneously addressing disclosure, visibility, analyst following, and investor following concerns for companies.
Not unpredictably, marketing and finance academics have turned their thought to relationships between a variety of characteristics of a firm’s IR strategy and firm visibility. This study was encouraged by the need to address an important gap in CEO transition literature. As Datta et al. (2003) indicate, there have been a number of studies that
investigate the relationship between CEO transition and firm performance. In contrast, very few have examined the financial community reactions of such transitions. Consequently, there is a small number of insights that relay CEO transition to analyst and investor following, an important metric of firm visibility in the financial community. This study examines the impact of CEO transition and the successor’s openness to change, an important indicator of a CEO’s willingness to change current strategy, on analyst and investor following.
The findings of this study offer support for theoretical predictions derived from literature. First, CEO transition increases a firm’s short-term analyst following. Specifically, the results suggest that a CEO transition leads to a positive change in a firm’s short-term analyst following. Combined with results from other studies on the effect of CEO transition on short-term market response (Huson et al., 2004), this leads to an interesting conjecture. Such a positive response from the financial community might cause firms to believe that CEO transition is rewarded by a short-term increase in analyst following. This might be hazardous to the already decreasing CEO tenures. It seems that on the short-term, a firm becomes more visible on the financial market when a change in CEO occurs. If this trend continues CEO tenures will keep decreasing which could endanger the long-term survival of firms.
Second, while CEO transition leads to a partial positive market response on the short-term (increasing analyst following), a change in CEO causes a decrease in a firm’s long-short-term investor following. This evidence favors the scapegoat hypothesis over the improved
improve future firm performance.
Third, investor following tends to increase stronger when newly chosen CEOs are more open-minded about changing firm strategy. The findings content that ‘CEO openness to change’ in the succession context will significantly moderate the influence of CEO transition on financial community response (investor following). This indicates that firms that want to replace their CEO enjoy more investor following when the newly chosen CEO is more open-minded about change. In other words, CEOs who are younger and have lower levels of firm tenure are more likely to challenge the status quo and move the firm in new strategic directions, which is ultimately rewarded by the financial community.
This study’s findings also have interesting implications from a managerial standpoint. It appears that firms can derive significant benefits from choosing new CEOs who are
younger and have lower firm tenure. And although there may be exceptions, in general,
CEOs shouldn’t be in office for more than a decade (Mulcahy, 2010). However, firms face with the tradeoff between keeping CEOs in office or choosing a new CEO. Obviously, CEOs shouldn’t be succeeded in a too excessive pace (Mulcahy, 2010). Further, aside the decision to replace the CEO, firms that wish to influence their visibility in the financial market are more likely to realize their goals through the selection of a CEO with demographic attributes associated with greater openness to change.
In terms of CEO succession and the negative long-term financial community response, one possible way to avoid the negative long-term reaction is that firms may discuss very early on who should succeed the current CEO. In some cases this happens as early as when the new CEO is just appointed (Antia et al., 2010). Mulcahy (2010) advocates
that the succession conversation between CEOs and their board of directors needs to start
very early on, preferably three to five years before the actual succession. CEO transitions have proven to affect market response and therefore, IR strategies should improve voluntary disclosure to hamper the possible negative market response. Bushee and Miller (2010) stress that voluntary disclosure activities can positively impact a firm’s analyst and investor
7. LIMITATIONS AND FURTHER RESEARCH
The study’s findings need to be regarded in the perspective of several limitations. Though, the findings also can provide some interesting opportunities for future research.
First, the nature of sample used in this study may limit the generalizability of the findings across all industries and organizations. Even a CEO, who is very open-minded about changing a firm’s current strategy, may be unable to pursue a new in an industry that offers limited opportunity for discretion (Hrebiniak and Joyce’s, 1980). Moreover, certain
circumstances of an industry can significantly pressure the ability of the new CEO to a new strategy. A CEO who is open to change in a conservative industry may experience high levels of obstruction because aggressive growth-oriented strategies may be restrained by a lack of prospect offered by the industry. However, this limitation also provides an opportunity for future research. Especially, researchers could examine the consequences of CEO transition in different organizational contexts.
Second, while this study controls for several important influences of CEO transitions on a firm’s analyst and investor following (Earnings-to-price ration, firm size, dividend yield), extant CEO transition literature indicates there are also organizational factors that have an important impact on the capability of the newly chosen CEO to initiate strategic changes. These include organizational structure (Ginsberg and Buchholtz, 1990), planning and control systems (Simons, 1994) and, organizational culture (Pettigrew, 1987). The use of secondary data sources prohibited the inclusion of this additional information, which is on firm level. However, future research that has uses primary data sources (e.g. questionnaire surveys or interviews) could examine how these organization-level factors are related to the new CEO’s ability to effect strategic change and how the financial community responds to it.
Third, the finding of a non-significant correlations and equation for CEO transition and long-term analyst following is notable and merits discussion. Not finding any significant correlations could have several reasons. According to Leeflang et al. (2002), when no significant relationships are found a possible solution is to increase the sample size. One might argue that the sample size of this study is too limited. Further, another possibility is that non-linear relationships exist between the dependent variables and the key predictor variables. A possible opportunity for further research lies in non-linear regression models (Leeflang et al., 2002).
the expectation. This could depend on the ‘CEO openness to change’ construct which is lacking another variable, namely CEO education. This is consequently the following limitation, which is addressed next.
Third, this study uses two demographic variables (age and tenure) as a proxy for the construct of ‘CEO openness to change’ while the original construct, uses CEO education in addition (Smith and White, 1987). Therefore, it could be that this study may not have fully captured the ‘openness to change’ construct.
Fourth, ‘CEO openness to change’ uses demographic variables to proxy for a CEO’s cognitions. While researchers such as Pfeffer (1983) and Datta et al. (2003) advocate that a reliance on demographic proxies is an assured measurement. However, future research that seeks to examine executive cognitions could also employ research methods that directly analyze underlying psychological mechanisms trough which CEOs influence the firm and the financial community. One might think of analyzing the direct communications of CEOs and analysts and investors, a noteworthy example of such research can be found in recent literature on CEO cognitions. For example, (Fanelli et al., 2008) use thematic text analysis of the initial letters to shareholders following a CEO succession event they analyze how CEO charismatic visions influence financial analyst’s recommendations and forecasts.
Future work can also involve greater use of thematic text analysis to analyze how different CEO cognitions impact financial community reactions.
APPENDIX
The data sources used for this study are access through Wharton Research Data Services (WRDS). Figure 1 gives a graphical overview of how the different dataset, collected from different sources, are merged in to a final dataset. The top row of Figure 1 shows the four different data sources that are used, namely, Thomson Reuters, I/B/E/S, COMPUSTAT and ExecuComp. The data collected from Thomson Reuters and I/B/E/S where aggregated on a monthly level. Since this study uses quarterly data, both datasets had to be aggregated from monthly level to quarterly level (portrayed by the second row in Figure 1). The data collected from COMPUSTAT and ExecuComp where both quarterly available. Next, the data had to be merged in to one dataset.
FIGURE 1
Graphical Overview Data Collection
I/B/E/S (Analyst Following) Thomson Reuters (Investor Following) ExecuComp (CEO Characteristics) COMPUSTAT (Control Variables)
Analyst & Investor Following
(Merged)
Firm & CEO Fundamentals
This was done in three steps. First, Thomson Reuters and I/B/E/S where merged on Ticker Symbol, Fiscal Year and Fiscal Quarter. Second, COMPUSTAT and ExecuComp where merged on Ticker Symbol, Fiscal Year and Fiscal Quarter. Finally, merging the analyst and investor following data with the firm financials created the final dataset. Table 1 shows for how many firms the data was available in each data source. For example, 81 out of the total sample of 100 firms had analyst coverage. Further, ExecuComp provided information about executives for only 70 firms. This also determined the final sample of 70 firms. The firms where no CEO transition information available had to be excluded from the final sample.
Table 1
Detailed Information Each Data Source
Data Source Firms Observations
