MASTER THESIS
CORPORATE SOCIAL RESPONSIBILITY, CEO
POWER & TURNOVER:
How does prior Corporate Social Responsibility (CSR)
investment explain CEO turnover with the moderating role
of CEO power on the U.S. market?
Written by: Fatimah Ahmad Munawar Student Number: 10864644
Supervised by: Dr. R. Perez Ribas Master Programme: Finance Track: Corporate Finance Date: 01/07/2018
2
Statement of Originality
This document is written by student, Fatimah Ahmad Munawar who declares to take full responsibility for the contents of this document.
I declare the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
3 Abstract
This paper investigates the impact of prior Corporate Social Responsibility (CSR) investment on chief executive officer (CEO) turnover with the moderating role of CEO power in the United States. To test this relation, the present study employs the non-linear logistic
regression on a sample of 1,068 public firms during the period of 2007 to 2017. The dataset includes all forms of turnovers (i.e. voluntary and involuntary). Results show that prior CSR investment has a negative but insignificant influence on CEO turnover. However, this predicted negative relationship becomes significant when moderated by CEO power. It suggests that powerful CEOs are less likely to be penalized for undertaking investment strategies that may be value-destroying such as overinvesting in CSR activities. In other words, firms that engage in higher levels of CSR and led by powerful CEOs are more likely to have weak shareholder power and poor corporate governance, consistent with the agency theory.
4 Table of Contents
1 Introduction ... 5
2 Literature review ... 7
2.1 The Concept of Corporate Social Responsibility... 8
2.2 Corporate Social Responsibility and Agency Theory ... 9
2.3 Corporate Social Responsibility, CEO Power and CEO Turnover... 10
3 Data ... 11
3.1 Sample and Data Collection... 11
3.2 Variables ... 14
3.3 Descriptive Statistics ... 15
4 Empirical Method ... 16
5 Results ... 18
5.1 The Impact of Aggregate CSR and CEO Power on Turnover ... 18
5.2 The Impact of Individual CSR Components and CEO Power on Turnover ... 19
5.3 Robustness Tests ... 19
6 Discussion... 20
6.1 Interpretation and Implication... 20
6.2 Limitations ... 21
7 Conclusion ... 22
References ... 23
5 1 Introduction
Over the recent years, the concept of Corporate Social Responsibility (CSR) has continued to grow in importance and significance at the global level. In 2016, 64% of CEOs are increasing investments in CSR activities and view such investments as being core to their business strategy (PwC, 2016). The widespread application of CSR makes it important to determine its implications for different stakeholders (Aguinis and Glavas, 2012). Among these groups, employees, specifically, top managers are crucial to any discussion of the consequences of CSR. However, most CSR research primarily focuses at the macro level (i.e. organizational level) and on external stakeholders such as shareholders and consumers rather than on the underlying mechanisms through which CSR influences manager outcomes. Accordingly, this study contributes to the gap by exploring the relationship between CSR and CEO turnover.
The effectiveness of CEO effort is generally assessed by a firm’s performance, thereby, it is also a strong indicator of CEO dismissal. Researchers have commonly found that the probability of managerial turnover increases with poor firm performance in both developed and developing economies including the U.S. (Huson, Parrino and Starks, 2001), Germany (Brunello, Graziano and Parigi, 2003), Japan (Kaplan & Minton, 1994) and Thailand (Rachpradit, Tang and Ba Kang, 2012). Although this conclusion has gathered consistent support over decades of research, it does not fully explain the variance in CEO turnover. In the context of agency theory, I argue that CEOs can overinvest in CSR activities that may be value-destroying as an effective entrenchment tool by meeting the wants and needs of other stakeholders such as local communities and activists. Thus, the main research question of this study is: “Controlling for firm performance, does prior CSR investment negatively affect CEO turnover?”
I examine the impact of investing in CSR activities on the likelihood of CEO
dismissal using logistic regressions on a sample of 1,068 public firms in the U.S. from 2007 to 2017. Additionally, CEO turnover is regressed on various dimensions of CSR, specifically, corporate governance, social and environmental sustainability. Lags are added to the models, which is known as the lag identification strategy to reduce potential endogeneity issues. A composite CSR score is determined based upon firm level ratings of corporate social performance from the Kinder, Lydenberg and Domini Research and Analytics (KLD) database, which has been widely used in prior research that investigate CSR (for i.e. Adams, Almeida and Ferreira, 2005; Bebchuk, Cremers and Peyer, 2011).
6
To further develop our understanding of the relationship between CSR investment and CEO turnover, it is critical to consider other factors that may impact this association. An overlooked factor in this relation is CEO power. The level of CEO power has been shown to influence CSR level (Borghesi, Houston and Naranjo, 2014) as well as having an influence on CEO turnover (Huson, Parrino and Starks, 2001; Goyal and Park, 2002). Thus, it is reasonable to suggest that the extent of CEO power may have a moderating effect on the CSR-turnover relation. Therefore, the second research question of this study is: “How does CEO power moderate the relationship between prior CSR investment and the likelihood of CEO turnover?” To answer this research question, logit models are used by regressing CEO turnover on the interaction term of CSR and CEO power, which is constructed based on three proxies: CEO ownership, duality and tenure.
Research on CSR and CEO turnover is relatively nascent, but there are a few studies (for i.e. Cooper, 2016; Barrios, Fasan and Nanda, 2014) that establish the moderating role of CSR on the relation between firm financial performance and managerial turnover. Using a large sample of firms over a 21-year period from 1992 to 2003, Cooper (2016) finds that CSR is positively related to CEO turnover. Her finding suggests that firms with better CSR
performance have higher rates of replacing their underperforming CEOs. Additionally, in a study of 1,111 public U.S. companies from 1996 to 2005, Barrios et al. (2014) find that the probability of terminating a CEO due to poor firm performance increases with a firm’s overall CSR performance. In summary, their results are inconsistent with the agency theory of CSR and instead suggest that underperforming CEOs who invest in higher levels of CSR are exposed to greater risk of termination.
Unlike prior research, my results corroborate the agency theory, which states that powerful CEOs can entrench themselves by engaging in CSR projects that can potentially be costly for shareholders. An explanation for the conflicting finding is that previous CSR-turnover research examines CSR as a moderator on the firm performance-CSR-turnover relation, while this study investigates the direct effect of investing in CSR activities on managerial turnover. The empirical findings of this paper are summarized as follows. First, on average, there is no significant effect of CSR on CEO turnover. In other words, controlling for firm performance, CSR investment by itself does not have a direct impact on the probability of replacing a CEO. In addition to the main finding, I find that the predicted negative
relationship between CSR and CEO turnover becomes significant with the interaction of CEO power. It suggests that firms led by powerful CEOs (i.e. firms with weak corporate
7
CEOs associated with high CSR investments can protect themselves from the disciplinary actions of shareholders. In line with the agency theory, powerful CEOs are more likely to take on strategies that are personally beneficial for them.
The present study contributes to the existing literature in a few ways. First, this thesis adds to prior research which has attempted to prove the relationship between CSR and CEO turnover by adding a further layer of analysis in terms of CEO power. By incorporating power as a moderator, it provides insights into how corporate governance as driven by CEO power influences the underlying processes under which CSR leads to specific outcomes, specifically, a CEO’s career. Additionally, this paper extends the reach of the agency theory providing a rationalization for the link between CSR and CEO turnover. The study of
managerial power (entrenchment) in agency theory ensures that the findings of this paper will offer insightful results, thereby providing a baseline for further research and discussion in CSR at a micro-level.
The remainder of the paper is organized as follows. Section 2 summarizes the most relevant literature and develops the hypotheses. Section 3 describes the sample and variables. Section 4 is methodological and describes the empirical models to be tested. The results on the relation between CSR and CEO turnover, as well as the moderating effect of CEO power are presented in Section 5. Section 6 discusses the implications of the main empirical
findings, its limitations and potential avenues for further research. Lastly, the final section provides the main conclusions of this paper.
2 Literature Review
2.1 The Concept of Corporate Social Responsibility
CSR is a very broad concept with many definitions and practices. The first significant
scholarly definition of CSR was proposed by Bowen and Johnson (1953) who states that “the obligations of a business is to pursue policies that are desirable in terms of the objectives and the values of the society.” While there is no single, universally accepted definition of CSR, it generally concerns the link between a company’s business strategy and its compliance to profits, legislative requirements, social demands and ethical values (Garriga and Melé, 2004). In line with this definition, McWilliams and Siegel (2001) regard CSR as voluntary social actions performed by the management of a company that appears to further the promotion of some social good, beyond the interests of the firm (shareholders) and beyond minimum legal requirements. A large body of research (for i.e. Fabrizi, Mallin and Michelon, 2013; Borghesi
8
et al., 2014) suggest that top managers, specifically, CEOs play an important role in the direction of their firms’ corporate strategy and are responsible for discretionary decisions such as engaging in CSR. Thus, the propensity of firms to engage in CSR may be affected by chief executives’ characteristics, preferences and values.
2.2. Corporate Social Responsibility and Agency Theory
According to agency theory, the separation between ownership and control in a company can lead to conflicting interests between the principals such as shareholders and the agents such as the managers (Jensen and Meckling, 1976). Previous research (for i.e. Cassel, Huang, Manuel Sanchez and Stuart, 2012; Dudley, 2012) find that agency problems may occur when managers make investment choices that are against the best interests of
shareholders. In other words, investing in projects that are non-value-maximizing. According to Borghesi et al. (2014) and Bhandari and Javakhadze (2017), engaging in CSR is regarded as a form of investment. In the context of CSR, the classical agency problem may arise when managers over-invest in CSR activities for their own private benefits such as improving their professional and/or personal reputations even if they are not necessarily profitable for the shareholders (Friedman, 1970).
Empirically, Borghesi et al. (2014) find that CEOs who donate to both major US political parties and CEOs with higher media coverage are more likely to invest in CSR. Their findings imply that CEOs can privately benefit from investing in CSR levels that are not necessarily value enhancing because they believe that CSR is an effective strategy for personal reputation-building. Additionally, overinvestment in CSR can be driven by impure altruistic motives of the managers (Barnea and Rubin, 2010; Petrenko, Aime, Ridge and Hill, 2015). Petrenko et al. (2015) find that narcissistic CEOs are more likely to engage in CSR activities. Furthermore, their results show that firm financial performance is lower for firms with more narcissistic CEOs than for firms led by less narcissistic CEOs when CSR
engagement is higher. The authors suggest that investing in CSR may be less strategic in terms of firm financial performance, but instead is a result of managers’ personal need for attention and reputation building. Based on these empirical findings, it can be concluded that agency conflicts between shareholders and managers of a company may manifest from investing in CSR initiatives.
9
2.3 Corporate Social Responsibility, CEO Power and CEO Turnover
In line with the agency view, Shleifer and Vishny (1989) find that managers can entrench themselves by making manager-specific investments in a way that any potential replacement is costly for the shareholders. Manager-specific investments such as CSR made by the managers can curtail his or her probability of being dismissed, extract higher rents and larger perks from shareholders. Following this concept, Cespa and Cestone (2007) provide a theoretical model in which ineffective CEOs often exploit investment in CSR initiatives as a defensive strategy to avoid negative consequences of managerial discretion by “buying off” the company’s stakeholders. Specifically, incumbent CEOs are more likely to maintain a close relationship with stakeholder representatives when he or she is under the threat of being replaced and when the effectiveness of social activism is increased, thereby reducing the likelihood of managerial turnover. Similarly, Pagano and Volpin (2005) theorize that high-level managers and lower high-level employees can collude together to entrench managers and avoid takeover threats. Particularly, top managers can gain the support of workers by increasing employee benefits in the event of takeover threats. Thus, reducing the likelihood of managerial dismissal.
Garcia-Meca, Garcia-Sanchez and Martinez-Ferrero (2015) confirm the stated theories of CSR-turnover relation in their international study on managerial discretion. They find that firms that are engaging in poor accounting practices strengthen CSR engagement to obscure the negative consequences of this unethical behaviour. Therefore, CSR can be considered as an entrenchment strategy to conceal disreputable managerial discretion. Conversely, Harjoto and Jo (2011) find that firms engaging in more CSR activities are more likely to experience higher CEO turnover. Inconsistent with the agency view, an implication of their finding is that firms engaging in a higher level of CSR are more likely to have an effective governance structure (Harjoto and Jo, 2014). In their view, managers of better governing firms are more likely to invest in CSR, which reduces agency problems that may arise between shareholders and other stakeholder groups. In summary, the agency theory argues that CEOs may utilize CSR as a strategic tool to meet the wants and needs of non-shareholder stakeholders, then the incumbent CEO can reduce the probability of his/her replacement. In line with this, I hypothesize that the relationship between the level of CSR investment and the likelihood of CEO turnover is negative.
In addition to the CSR-CEO turnover relation, several studies (for i.e. Jensen and Ruback, 1983; Cooper, 2016) suggest that other factors such as CEO power can influence this relation in the context of agency theory. Since a CEO is regarded as the corporate leader i.e.
10
the most powerful member of an organization, any action taken by or against him or her will have an impact on a company’s direction, policies and future outcomes. Jensen and Ruback (1983) suggest that the main theoretical underpinning the rationalization of CEO power-turnover relationship is agency theory. The authors assert that CEOs can exercise their power to extract private rents by undertaking value-destroying projects and achieve larger
advantages from shareholders such as higher pay, longer tenure and greater job security. This implies that the existence of power amongst CEOs is an indication of weak shareholder power and poorly governed firms.
Following this theory, empirical studies support the negative CEO power-turnover relation. In a sample of 455 CEO turnovers from 1992 to 1996, Goyal and Park (2002) find that CEO turnover-performance sensitivity is weaker for firms where the CEOs hold greater structural power, particularly, CEOs who are also the chairman of the board. Their findings are consistent with the view that combining CEO and chairman positions are more likely to reduce the effectiveness of monitoring by the board. Thus, single CEO-Chairman role makes it difficult for the board to dismiss underperforming managers. A similar conclusion is drawn by Denis, Denis and Sarin (1997). In their study of 1,394 distinct firms over the period 1985 to 1988, Denis et al. (1997) find that the greater is the CEO’s ownership power, the lower is the likelihood of his or her turnover. They conclude that CEO ownership power holds an important influence on internal monitoring mechanisms and that this influence is attributed to the CEO turnover relation. In light of the agency theory and the negative CEO power-turnover relation, I further hypothesize that powerful CEOs who hold greater control over corporate boards are more likely to undertake CSR projects as an effective entrenchment mechanism that may be value-destroying and can thereby, reduce the probability of being replaced.
3 Data
3.1 Sample and data collection
The initial sample consists of 3,534 publicly listed U.S. firms that are observed during the period of 2007 to 2017. Multiple data sources are used. Data on CEO turnover and other CEO characteristics are obtained from Compustat Execucomp. CSR score data is collected from KLD (i.e. MSCI ESG) database. CEO duality, which is one of the variable constructing the composite measure for CEO power is extracted from the Institutional Shareholder
11
collected from Compustat North America. From the full sample of firms, I exclude financial firms (i.e. firms with SIC codes 6000-6999). I further remove all the firm-year observations with missing assets or sales. Financial ratios are winsorized at the 1 percent level and 99 percent level. After combining data from these databases, the final sample consists of an unbalanced panel of 4,434 firm-year observations for 1,068 unique non-financial listed firms.
3.2 Variables
Table 1 summarizes the definition, measurement and data sources for each variable used in this research. The dependent variable in this study is CEO turnover, which takes the value of 1 if the company has replaced its CEO with respect to the previous year and 0 otherwise. The final sample has 354 turnover events, which account for 8 percent of the total firm-year observations. These turnovers include both voluntary and involuntary forms such as
retirement, resignation, dismissal, contract expiration, personal reasons, promotion to board positions and legal disputes.
Table 1. Dependent, Independent and Control Variables
Variable Measurement Sources
Dependent variable
CEO Turnover Binary variable for which Turnover = 1 if the CEO is replaced in a given year, 0 otherwise
Execucomp
Independent variables
CSR Total strengths of all issue areas minus total concern of all issue areas
KLD Social Net score of community, human rights, workforce diversity
and employee minus total
KLD
Environmental Net score of environment KLD
Corporate Governance Net score of corporate governance and product quality KLD CEO Ownership The ratio of number of shares owned by the CEO to total
shares outstanding
Execucomp CEO Duality Binary variable for which Duality = 1 if the CEO is the
chairman of the company in a given year, 0 otherwise
ISS
CEO Tenure The number of years since becoming CEO Execucomp
Control variables
ROA Return on assets (in percentage points) = Earnings before interests, tax, depreciation and amortisation (EBITDA) divided by the book value of assets
Compustat
Firm size Natural logarithm of book value of total assets Compustat Leverage The ratio of long-term debt to total assets Compustat R&D The ratio of R&D expense to total operating expenses Compustat CAPEX The ratio of capital expenditures to total PP&E expenditures Compustat
CEO Age The age of the CEO Execucomp
CEO Gender Binary variable for which Gender = 1 if the CEO is a woman, 0 otherwise
Execucomp CEO Compensation Natural logarithm of total current compensation Execucomp
12
CSR scores and CEO power are the explanatory variables in this analysis. In line with previous empirical research (Ng and Rezaee, 2015; Bhandari and Javakhadze, 2017), CSR is collected from the KLD database, which is widely regarded as the most reliable data source available to measure CSR. The database includes the strengths and concerns of seven CSR-related areas: the community, human rights, corporate governance, workforce diversity, employee relations, the environment and product quality. For each individual key area, a binary (0/1) rating is assigned to the strength and concern. Thus, aggregate CSR investment is measured as the total CSR strengths of all issue areas minus the total CSR concerns of all issue areas. Additionally, the aggregate CSR measure is decomposed into three individual components: corporate governance, social and environmental sustainability. Corporate governance score includes key areas of corporate governance and product quality. Social score is constructed based on the areas of the community, human rights, workforce diversity and employee relations, while environment sustainability includes the key issue of the environment.
Power refers to the legitimate authority of CEOs and their extensive knowledge of the firm they serve. Finkelstein (1992) classifies CEO power into four dimensions: structural, ownership, expert and prestige power. Structural power derives from holding formal
organization positions, which gives the CEO legitimate authority to conduct and control over the company’s activities and outcomes (Finkelstein, 1992). Ownership power such as holding greater stock ownership increases the CEO’s power over both internal and external
constituencies (Mintzberg, 1983). Expert power reflects on the CEO’s ability and organizational knowledge, which allows the CEO to effectively handle organization and environmental contingencies (Finkelstein, 1992). Lastly, prestige power refers to the level of a CEO’s reputation or status in the institutional environment. This present study focuses mostly on a CEO’s structural and ownership power. In this thesis, the level of CEO power is a composite index constructed by three CEO characteristics: CEO ownership, duality and tenure. Information on CEO ownership and tenure are obtained from Execucomp, while data on CEO duality is retrieved from the ISS database. The measure for CEO ownership power is included as the percentage of shares owned by the CEO to total shares outstanding in the firm at the end of the fiscal year. CEO duality is a binary variable that takes the value of 1 if the CEO is also the chairman of the board and CEO tenure is defined as the number of years that a CEO has been in his/her managerial position.
In addition to the variables above, I also construct the following covariates to capture other factors that may influence the likelihood of CEO turnovers. These covariates are
13
categorized into two aspects, firm characteristics and CEO characteristics. In total, eight control variables have been added into the regression models. Firm-specific characteristics used as control variables are the return on assets (ROA), firm size, leverage, research and development (R&D) expenses and capital expenditures (CAPEX). These variables are collected from Compustat. With reference to the measurement of firm performance, this study uses the accounting-based performance measure. ROA represents firm performance obtained by the ratio between the earnings before interest, tax, depreciation, and amortisation (EBITDA) on the lagged value of total assets. Instead of net income, EBITDA is chosen as it is a good measure in comparing companies within and across industries (Fiodelisi and Ricci, 2012). Firm size is measured as the natural logarithm of book value of total assets, to
consider greater expectations for CEOs at larger firms. Leverage is measured as long-term debt divided by total assets. To capture differences in the firms’ operating strategies, R&D and CAPEX intensity are included. R&D intensity is calculated as the ratio of R&D expense to total lagged value of operating expenses and CAPEX intensity is calculated as capital expenditures scaled by total lagged value of plant, property and equipment (PP&E) expenditures.
CEO-specific characteristics used as control variables are age, gender and log of total compensation including salary, bonus, stock options and other forms of compensation. The reasons for controlling for these CEO-specific characteristics are the following. Previous research finds that CEO age is positively related to turnover given that as a CEO becomes older and is approaching retirement age, he or she would be more likely to decide to leave the firm or be dismissed by the firm (Brickley, 2013; Liu, 2014). Additionally, Cooper (2016) finds that female CEOs have a lower risk of being replaced relative to male CEOs in poor performing companies. Moreover, Chakraborty, Sheikh and Subramanian (2009) find that incentive compensation is positively related to CEO turnover, especially in firms that are underperforming relative to its peer firms. Given that these variables may influence the relation between CSR and CEO turnover through their association with turnover, I control for such variables in the analysis.
3.3 Descriptive statistics
Summary statistics are presented in Table 2. The average CEO turnover rate in the sample is 8 percent. The average CSR score for firms in the sample is 0.36 indicating high social performers with more strengths than concerns across the three key issues. The median is 0, which suggests that the strengths and concerns in key issues are approximately equal. This
14
indicates that most firms in the sample are neutral, in which they have no strengths or concerns on a given key issue. The minimum and maximum CSR score are -9 and 20
respectively. With respect to the individual CSR components, environment and social scores are, on average, positive except for corporate governance score. Consistent with the overall CSR score, the median for each key issue is 0, which suggests that strengths of key CSR issues balance its concerns. The percentage of shares owned to total shares outstanding by the CEO is on average 281 percent and the median is 83 percent. On average, 55 percent of the sample indicates that the CEO of a given company is also the chairperson of the board. The average (median) tenure of a CEO in the sample is 13.2 years (10 years).
Table 2. Summary statistics
Mean SD Min. Median Max. n
Dependent variable CEO Turnover 0.08 0.27 0 0 1 4,434 Independent variables CSR Score 0.36 3.16 -9 0 20 4,434 Social Score 0.38 2.39 -7 0 15 4,434 Environment Score 0.26 1.01 -5 0 5 4,434
Corporate Governance Score -0.29 0.72 -4 0 2 4,434
CEO Ownership 2.81 6.13 0 0.83 87.60 4,434 CEO Duality 0.55 0.50 0 1 1 4,434 CEO Tenure 13.2 31.39 1 10 61 4,434 Control variables ROA (%) 0.14 0.10 -1.69 0.14 0.95 4,434 Firm Size 7.75 1.57 3.99 7.56 13.57 4,434 Leverage 0.18 0.16 0 0.17 1.49 4,434 R&D 0.05 0.29 0 0 17.44 4,434 CAPEX 0.08 0.14 0 0.04 2.29 4,434 CEO Age 57.69 7.17 35 57 96 4,434 CEO Gender 0.04 0.18 0 0 1 4,434 CEO Compensation 8.01 1.20 0 8.08 11.77 4,434
Note: This table presents summary statistics for the variables used in this analysis.
In this sample, average and median ROA for firms in the sample is 14 percent and 10 percent. Firms have an average log asset of 7.75 and a median of 7.56. The mean firm has long-term debt of 18 percent of its total assets. On average, expenses on R&D scaled by total operating expenses count for 5 percent. The average (median) capital expenditure scaled by total PP&E expenses is 8 percent (4 percent). Overall, the majority of CEOs in the full
sample of firm-years is male. In total, 96 percent of the sample is male while the remaining is 4 percent is female. The average age of a CEO in the sample is approximately 58 years. On average, CEOs receive log compensation of 8.01, which is $4.5million in absolute terms.
15 4 Empirical Method
4.1 Regression Models
Following the literature (Fiordelisi and Ricci, 2014; Gao, Harford and Li, 2012), the analysis of the impact of CSR on the probability of a CEO’s dismissal is conducted by logistic regressions as the dependent variable, CEO turnover is binary. I estimate two models to test my hypotheses regarding the relation between CSR and CEO turnover and the moderating effects of CEO power on the CSR-turnover relation. The aim of the first regression is to test whether there is a negative relationship between prior CSR
investment and CEO turnover probability:
Pr(CEO Turnoverit = 1) = α0 + β1CSRt-1 + β2(Controlst-1) +Industry Fixed Effects + Year
Fixed Effects + εit (1)
The aim of the second model is to test whether the negative relationship between prior CSR investment and CEO turnover probability will be strengthened when the level of CEO power is high:
Pr(CEO Turnoverit = 1) = α0 + β1CSRt-1 + β2CEO Power+ β3 (CSRt-1*CEO Power)+
β4(Controlst-1)+Industry Fixed Effects + Year Fixed Effects + εit (2)
The variables of interest in these models are CSR and the interaction term between CSR and CEO power, CSR*CEO Power. Therefore, these coefficients indicate the impact of CSR on CEO turnover and the moderating effect of CEO power on the probability of a CEO being replaced by a firm. The first hypothesis of the present study predicts a negative
correlation between CSR and CEO turnover, namely I expect that β1<0 in regression (1). The
second hypothesis predicts that the negative CSR-turnover is stronger for powerful CEOs, namely β3<0 in regression (2). Additionally, individual components of CSR score (corporate
governance, social and environmental sustainability) are regressed against managerial turnover to further investigate the relationship between CSR and CEO turnover and to determine its variability.
Endogeneity problem has been researched extensively in corporate finance. Many solutions to endogeneity problem has been introduced including control variables, fixed effects and instrumental variables. In this study, control variables, robust standard errors clustered by firm as well as year and industry fixed effects are included in each model to
16
address potential endogeneity issues. Furthermore, following prior research (Brick, Palmon and Wald, 2006), CSR investment and control variables are lagged one year to reduce potential biases from reverse and double causality between the independent variables and CEO turnover. Using this empirical method called lag identification, CSR investment and control variables are expected to be exogenous when using lagged values. However, it is important to note that using lagged variables does not fully resolve the potential issue of endogeneity and simultaneous causality.
5 Results
5.1 The Impact of Aggregate CSR Score and Power on CEO Turnover
The logistic regression results examining the direct impact of the aggregate lagged CSR investment on CEO turnover are reported in Table 3. The main variables of interest here are the aggregate CSR investment and the interaction terms between the explanatory variables, lagged value of CSR and CEO power. Table 3 reports the marginal effects, while Appendix 2 reports the logit coefficient estimates. Models (3) and (4) of Table 3 employ an industry-adjusted CSR, which is measured based on four-digit SIC codes. The reason for running the main regressions on industry-adjusted CSR is that when a company decides to engage in CSR practices, it may compare the firm’s CSR investment to other peer firms in the industry. Consequently, this may affect the relation between prior CSR investment and CEO turnover.
Table 3. Marginal Effects of Aggregate CSR and CEO Power on CEO Turnover
Dependent variable: CEO turnover
(1) (2) (3) (4) CSRt-1 -0.002 -0.002 -0.001 -0.003 (0.002) (0.002) (0.002) (0.029) CEO Power -0.051*** -0.051*** (0.011) (0.012) CSRt-1 x CEO Power -0.002* -0.002* (0.003) (0.004)
Industry-adjusted CSR No No Yes Yes
Controls Yes Yes Yes Yes
Year and Industry Fixed Yes Yes Yes Yes
Wald Test 112.51 142.81 141.41 145.20
Observations 2,321 2,321 2,321 2,321
Note: This table reports the results of logit regression models where the dependent variable is CEO turnover during the period of 2007 – 2017. Each number represent the marginal estimate of the corresponding variable. Clustered robust standard errors are shown in parentheses. Financial ratios are winsorized at the 1% and 99% level. Models (1) and (2) is run over firm-specific CSR. Models (3) and (4) is run over industry-adjusted CSR. * denotes significance at the 10 percent level; ** denotes significance at the 5 percent level; *** denotes significance at the 1 percent level.
17
Model (1) presents the regression results from equation (1). We show that past year’s CSR is negatively related to CEO turnover, however, it is not statistically significant. Since this coefficient is insignificant, this implies that prior CSR investment does not influence the personal consequences that CEOs face when making a non-value-maximizing investment. Model (3) shows similar results when regressed using industry-adjusted CSR. The negative and insignificant coefficients of firm-specific CSR and industry-adjusted CSR are
inconsistent with the expectation of the first hypothesis, rejecting the agency view that
investments in CSR, per se, can be used by the CEOs to entrench themselves in the company. The second hypothesis states that according to the agency view, there is a negative association between CSR and CEO turnover for firms with powerful CEOs. Holding all else constant, the coefficient estimates on CEO power are negative and significant at the 1 percent level. The greater is the extent of CEO power, the lower is the turnover probability of the CEO. The same patterns can be observed when interacting with the explanatory variable CSRt-1, which implies that the level of power a CEO holds does influence the relation
between CSR investment and CEO turnover. Thus, the data support hypothesis 2. Model (2) shows that CEO power has a significant moderating effect of the negative CSR-CEO
turnover relationship at the 10 percent level. The findings in Table 3 indicate that the change in the probability of CEO turnover for a powerful CEO decreases by 0.2 percent when they engage in CSR activities. Similarly, Model (4) reports that CEO power moderates the negative relationship between CSR and CEO turnover when industry-adjusted CSR is used.
Conclusively, my results report that there is no relation between prior CSR investment and CEO turnover, but this relation becomes negative when moderated by CEO power. To better understand the insignificance of the main effect of CSR on managerial turnover and the significant negative interaction effect of CSR and CEO power on CEO turnover, I further investigate the relation between CEO power and prior CSR investment. Appendix 2 presents the simple OLS regression coefficient estimates on the impact of CEO power and its proxies (ownership, duality and tenure) on the lagged value of CSR, including the control variables from the main analysis. Based on the simple OLS analysis, the relation between CEO power and past year’s CSR investment is positive and significant at the 1 percent level. The results indicate that firms led by more powerful CEOs are more likely to engage in CSR by 2.6 percent in comparison to less powerful CEOs. Overall, my results suggest that powerful CEOs who hold greater control over the company are more likely to overinvest in CSR. Consistent with the agency theory, overinvestment in CSR is a form of job shield to protect powerful CEOs from the threat of being terminated.
18
5.2 The Impact of Individual CSR Components and Power on CEO Turnover
The same analysis described above is performed to further test the hypotheses in which the individual CSR components are used instead of the aggregate CSR score. Table 4 presents the marginal effects of individual lagged CSR components and its interactive terms with CEO power on CEO turnover while Appendix 3 reports the logit coefficient estimates of the
regression models. Table 4 reports that corporate governance and environmental aspects of CSR are negatively related to CEO turnover. However, models (3) and (4) shows that social CSR score is positively related to managerial turnover. Without taking into account the statistical significance of the logit estimates, it suggests that firms with a stronger social stance are more likely to fire their CEOs for overinvesting in CSR. Nonetheless, the
coefficients are insignificant. Consistent with the findings of Table 3, my results indicate that the level of CSR investments in the respective aspects of social, environmental and corporate governance made by a CEO does not influence his or her replacement.
Table 4. Marginal Effect of Individual CSR Components, Firm Performance and CEO Power on CEO Turnover
Dependent Variable: CEO Turnover (1) Corporate Governance (2) Corporate Governance (3) Social (4) Social (5) Environmental (6) Environmental CSRt-1 -0.014 -0.014 0.005 0.006 -0.001 -0.005 (0.009) (0.009) (0.003) (0.003) (0.007) (0.007) CEO Power -0.052*** -0.051*** -0.048*** (0.012) (0.012) (0.011) CSRt-1 x CEO Power -0.004* -0.001* -0.017** (0.012) (0.004) (0.008)
Controls Yes Yes Yes Yes Yes Yes
Year and Industry Fixed
Yes Yes Yes Yes Yes Yes
Wald Test 115.50 147.28 112.93 142.23 112.50 145.56 Observations 2,321 2,321 2,321 2,321 2,321 2,321 Note: This table reports the results of logit regression models where the dependent variable is CEO turnover during the period of 2007 – 2017. Each number represent the marginal estimate of the corresponding variable. Clustered robust standard errors are shown in parentheses. Financial ratios are winsorized at the 1% and 99% level. Models (1) and (2) is run over corporate governance aspect of CSR. Models (3) and (4) is social aspect of CSR. Models (5) and (6) is run over environmental aspect of CSR. * denotes significance at the 10 percent level; ** denotes significance at the 5 percent level; *** denotes significance at the 1 percent level.
Additionally, Table 4 shows that the negative relation between prior CSR investments and CEO turnover is moderated by CEO power. The coefficients of CEO Power are negative and significant at the 1 percent level, which reinforces the entrenchment theory stating that
19
powerful CEOs based higher level of ownership in the firm, holding both CEO and chairman position, as well as longer services, reduce the probability of the CEO being replaced. The interactive term between each CSR dimension and CEO power suggests that for each aspect of CSR, CEO power has an influence on the relationship between CSR and CEO turnover. Results of Tables 3 and 4 suggest that the greater power a CEO has the more likely that he or she will engage in CSR and the less likely he or she will be dismissed for overinvesting in such initiatives.
5.3 Robustness Tests
Three sensitivity analyses are conducted to assess the robustness of my results. First, I rerun the main analysis of Table 3 by including firm governance factors as additional controls: percentage of independent directors on board, percentage of common stock owned by all directors and percentage of common stock owned by institutional shareholders. These
covariates are collected from the ISS database. As shown in Appendix 5, the marginal effects of lagged value of firm-specific CSR (β = -0.001, p = 0.905), CEO power (β = -0.071, p = 0.001) and the interactive term (β = -0.003, p = 0.073) are similar to the main results from the primary analysis. Results indicate that the primary model is robust to the inclusion of firm governance factors.
Second, I differentiate between high and low CSR. High CSR is net CSR score greater than zero and low CSR is net CSR score lower than zero. I rerun the main analysis of Table 3 using this classification to further test whether higher levels of CSR are associated with lower probability of turnover. The results are reported in Appendix 6. Without
considering the statistical significance of the coefficients, results show that firms engaging in higher levels of CSR (i.e. firms with net CSR score greater than zero) are associated with lower turnover probability, while firms that engage in lower levels of CSR (i.e. firms with net CSR score lower than zero) are associated with higher turnover probability. Furthermore, the interaction effect of prior CSR investment and CEO power do not have any significant impact on managerial turnover in firms with net CSR score lower than zero. However, this
interaction effect is significant in firms that engage in higher levels of CSR (β = -0.026, p = 0.017). It suggests that turnover probability is lower in firms that engage in higher levels of CSR and in firms that are led by powerful managers.
Third, I further examine whether firms with high CSR levels (i.e. firms with net CSR score greater than zero) are associated with poor governance structures. Using two sample t- test analysis, I test whether there is a difference between firms with net CSR score greater
20
than zero and firms with net CSR score lower than zero in terms of firm governance characteristics: percentage of shares held by CEO, duality, tenure, compensation, tenure, percentage of independent directors, percentage of common stock held by all directors and percentage of common stock owned by institutional shareholders. Results in Appendix 6 show that firms that engage in high levels of CSR are attributed with higher CEO ownership, longer tenure and lower percentage of independent directors. It suggests that high CSR firms are more likely to be attributed to weak shareholder power and poor governance.
Collectively, both main and supplemental analyses provide consistent evidence suggesting that CEO power moderates the negative relationship between CSR and CEO turnover.
6 Discussion
6.1 Interpretation and Implication
This present study establishes that there is no relation between the level of CSR investment and the likelihood of CEO dismissal. This thesis distinguishes three dimensions of CSR and finds that none of the dimensions have an impact on managerial turnover. The finding that there is a negative but, insignificant relation between CSR and CEO turnover is inconsistent with studies such as Barrios et al. (2014) and Cooper (2016). They find that CSR is positively related to CEO turnover in underperforming firms. One reason for the differences in our findings is due to the methodological approach. Their research primarily investigates the moderating role of CSR on firm financial performance and CEO turnover relation, while this thesis examines the direct impact of CSR on the likelihood of replacing a CEO. Combining the findings of this paper and that of previous research, it can be concluded that the board and owners of a company do not observe prior CSR investment as a predictor in evaluating a CEO’s effort, but it is merely an investment that contributes to the bottom line of the
company. The fact that CSR does not directly influence managerial turnover could imply that investing in CSR activities are unlikely to be induced by individual managers itself but is instead a reflection of the collective decision made by the firm’s owners.
In addition to the main finding, this thesis provides evidence for the moderating effect of CEO power on the CSR-CEO turnover relation. I find that the level of CSR investment is negatively related to CEO turnover when taking into account the power a CEO has in the company in terms of shares ownership, duality and tenure. This finding is in line with the classic agency view. In other words, overinvesting in CSR reduces the likelihood of CEO turnover only when the CEO holds greater authority and control over the firm (i.e. the CEO
21
owns a higher portion of company shares, the CEO is the chairman of the board and the CEO has been in service for longer periods). An explanation for this finding is that power gives CEOs the ability to engage more in investments such as CSR for their private benefits potentially at the cost of shareholders. Consequently, the greater power a CEO has, the more entrenched he or she is in the company, thus making the decision to dismiss the CEO more difficult. Conclusively, the empirical evidence of this paper suggests that high powered CEOs are able to extract advantages such as personal reputation-building and job security through CSR engagement.
6.2 Limitations and Further Research
As with each study, this one also has its limitations. First, this study only focuses on CEOs in the U.S market. Thus, the findings of this paper may not fully represent the CSR-turnover landscape of other markets due to the differences in corporate culture or governance
structures. Hence, further studies in other countries are encouraged to conclude the impact of CSR on CEO turnover. Additionally, given the fact that the area of managerial turnover is fluid and has many different dimensions in the decision-making process, other factors that have not been addressed in this paper may be important to the CSR-turnover. For example, CEO’s gender, board’s structure and executive compensation can have moderating roles that may influence the relationships presented in this study. Thus, further studies could include other variables that may moderate the relationship between CSR and CEO turnover.
Furthermore, although this study applies the lag identification strategy to address endogeneity concerns, it does not fully resolve these potential issues. Further research on the
CSR-turnover relation could apply a different methodology to improve the empirical model of this paper. Another limitation of this study is that the analyses conducted in this paper include both voluntary and involuntary turnovers (i.e. resignation, dismissal, death). Relatedly, further study could explore how CSR and CEO power affect different forms of turnover by investigating the impact of CSR and CEO power on CEO turnover through the differentiation between forced and voluntary turnovers.
7 Conclusion
Research on CSR has critical strategic implications for firms, as investing in a socially responsible manner is primarily focused on establishing and maintaining relationships with primary stakeholders. Consequently, a growing body of research have attempted to identify
22
key drivers of CSR. Using CEO data on the U.S. market over the period of 2007 through 2017, results of this paper suggest that investing in CSR influences career outcomes for CEOs, but that this influence is dependent upon the governance structure of the firm. My findings shed insights on the role of CEO power on the CSR-turnover relation. There is limited evidence that CSR investment, per se, directly affects executive turnover. However, when moderated by CEO power, this relation becomes negative as predicted. Furthermore, I also find that firms with high levels of CSR are associated with poor-quality governance. The results provide important implications from an agency theory perspective. My study indicates that powerful CEOs are more likely to engage in CSR activities and are faced with a lower risk of termination for overinvesting in such initiatives. This may be explained by the existing literature that CEOs can gain power via ownership, control or tenure and consequently, these accumulated powers will make the managers become more entrenched in their company. In line with the agency view, high powered CEOs can utilize CSR as a form of entrenchment strategy to gain the support of non-shareholder stakeholders and can thereby, avoid the consequences of disciplinary actions by the board and shareholders.
23 References
Adams, R. B., Almeida, H., & Ferreira, D. (2005). Powerful CEOs and their impact on corporate performance. The Review of Financial Studies, 18(4), 1403-1432.
Aguinis, H., & Glavas, A. (2012). What we know and don’t know about corporate social responsibility: A review and research agenda. Journal of management, 38(4), 932-968.
Amici, A., Fiordelisi, F., Masala, F., Ricci, O., & Sist, F. (2012). Strategic alliances and joint venture results in banking. Journal of Banking and Finance, 37, 1386-1396.
Andreoni, J. (1990). Impure altruism and donations to public goods: A theory of warm-glow giving. The economic journal, 100(401), 464-477.
Barnea, A., & Rubin, A. (2010). Corporate social responsibility as a conflict between shareholders. Journal of business ethics, 97(1), 71-86.
Barrios, J., Fasan, M., & Nanda, D. (2014). Is Corporate Social Responsibility an Agency Problem? Evidence from CEO Turnovers.
Bebchuk, L. A., Cremers, K. M., & Peyer, U. C. (2011). The CEO pay slice. Journal of Financial Economics, 102(1), 199-221.
Bhandari, A., & Javakhadze, D. (2017). Corporate social responsibility and capital allocation efficiency. Journal of Corporate Finance, 43, 354-377.
Borghesi, R., Houston, J. F., & Naranjo, A. (2014). Corporate socially responsible investments: CEO altruism, reputation, and shareholder interests. Journal of Corporate Finance, 26, 164-181.
Bowen, H. R., & Johnson, F. E. (1953). Social responsibility of the businessman. Harper. Brick, I. E., Palmon, O., & Wald, J. K. (2006). CEO compensation, director compensation, and firm performance: Evidence of cronyism?. Journal of Corporate Finance, 12(3), 403-423.
Brickley, J. A. (2003). Empirical research on CEO turnover and firm-performance: A discussion. Journal of Accounting and Economics, 36(1), 227-233.
Brunello, G., Graziano, C., & Parigi, B. M. (2003). CEO turnover in insider-dominated boards: The Italian case. Journal of Banking & Finance, 27(6), 1027-1051.
Cespa, G., & Cestone, G. (2007). Corporate social responsibility and managerial entrenchment. Journal of Economics & Management Strategy, 16(3), 741-771.
Chakraborty, A., Sheikh, S., & Subramanian, N. (2009). The relationship between incentive compensation and performance related CEO turnover. Journal of Economics and
Business, 61(4), 295-311.
Cooper, E. (2017). Corporate social responsibility, gender, and CEO turnover. Managerial Finance, 43(5), 528-544.
24
Denis, D. J., Denis, D. K., & Sarin, A. (1997). Ownership structure and top executive turnover. Journal of financial economics, 45(2), 193-221.
Fabrizi, M., Mallin, C., & Michelon, G. (2014). The role of CEO’s personal incentives in driving corporate social responsibility. Journal of Business Ethics, 124(2), 311-326.
Finkelstein, S. (1992). Power in top management teams: Dimensions, measurement, and validation. Academy of Management journal, 35(3), 505-538.
Freeman, R. E. (2010). Strategic management: A stakeholder approach. Cambridge university press.
Friedman, M. (1970). The social responsibility of business is to increase its profits. New York Times, September 13: 122-126.
Gao, H., Harford, J., & Li, K. (2012). CEO pay cuts and forced turnover: Their causes and consequences. Journal of Corporate Finance, 18(2), 291-310.
García-Meca, E., Garcia-Sanchez, I. M., & Martínez-Ferrero, J. (2015). Board diversity and its effects on bank performance: An international analysis. Journal of Banking & Finance, 53, 202-214.
Garriga, E., & Melé, D. (2004). Corporate social responsibility theories: Mapping the territory. Journal of business ethics, 53(1-2), 51-71.
Goyal, V. K., & Park, C. W. (2002). Board leadership structure and CEO turnover. Journal of Corporate Finance, 8(1), 49-66.
Harjoto, M. A., & Jo, H. (2011). Corporate governance and CSR nexus. Journal of Business Ethics, 100(1), 45-67.
Hazarika, S., Karpoff, J. M., & Nahata, R. (2012). Internal corporate governance, CEO turnover, and earnings management. Journal of Financial Economics, 104(1), 44-69.
Huson, M. R., Parrino, R., & Starks, L. T. (2001). Internal monitoring mechanisms and CEO turnover: A long‐term perspective. The Journal of Finance, 56(6), 2265-2297.
Jensen, M. C., & Meckling, W. H. (1976). Theory of the firm: Managerial behavior, agency costs and ownership structure. Journal of financial economics, 3(4), 305-360.
Jensen, M. C., & Ruback, R. S. (1983). The market for corporate control: The scientific evidence. Journal of Financial economics, 11(1-4), 5-50.
Kaplan, S. N., & Minton, B. A. (1994). Appointments of outsiders to Japanese boards: Determinants and implications for managers. Journal of Financial Economics, 36(2), 225-258.
Kitzmueller, M., & Shimshack, J. (2012). Economic perspectives on corporate social responsibility. Journal of Economic Literature, 50(1), 51-84.
25
Liu, Y. (2014). Outside options and CEO turnover: The network effect. Journal of Corporate Finance, 28, 201-217.
McWilliams, A., & Siegel, D. (2001). Corporate social responsibility: A theory of the firm perspective. Academy of management review, 26(1), 117-127.
Mintzberg, H. (1983). The case for corporate social responsibility. Journal of Business Strategy, 4(2), 3-15.
Mitnick, B. M. (2015). Agency theory. Wiley Encyclopedia of Management.
Murphy, K. J. (1985). Corporate performance and managerial remuneration: An empirical analysis. Journal of accounting and economics, 7(1-3), 11-42.
Ng, A. C., & Rezaee, Z. (2015). Business sustainability performance and cost of equity capital. Journal of Corporate Finance, 34, 128-149.
Pagano, M., & Volpin, P. F. (2005). Managers, workers, and corporate control. The journal of finance, 60(2), 841-868.
Petrenko, O. V., Aime, F., Ridge, J., & Hill, A. (2016). Corporate social responsibility or CEO narcissism? CSR motivations and organizational performance. Strategic Management Journal, 37(2), 262-279.
PwC. (2016). Redifining business success in a changing world. CEO Survey. Retrieved from https://www.pwc.com/gx/en/ceo-survey/2016/landing-page/pwc-19th-annual-global-ceo-survey.pdf
Rachpradit, P., Tang, J. C., & Ba Khang, D. (2012). CEO turnover and firm performance, evidence from Thailand. Corporate Governance: The international journal of business in society, 12(2), 164-178.
Ross, S. A. (1973). The economic theory of agency: The principal's problem. The American Economic Review, 63(2), 134-139.
Shleifer, A., & Vishny, R. W. (1989). Management entrenchment: The case of manager-specific investments. Journal of financial economics, 25(1), 123-139.
Videras, J. R., & Owen, A. L. (2006). Public goods provision and well-being: empirical evidence consistent with the warm glow theory. The BE Journal of Economic Analysis & Policy, 5(1).
26 Appendix
Appendix 1. Logit Estimates of Aggregate CSR and CEO Power on CEO Turnover
Dependent variable: CEO turnover
(1) (2) (3) (4) CSRt-1 -0.002 -0.001 -0.002 -0.004 (0.028) (0.002) (0.028) (0.029) CEO Power -0.58*** -0.58*** (0.128) (0.128) CSRt-1 x CEO Power -0.018* -0.02* (0.035) (0.042) Constant -4.571*** -5.697*** -4.571*** -5.717*** (1.511) (1.506) (1.51) (1.506)
Industry-adjusted CSR No No Yes Yes
Controls Yes Yes Yes Yes
Year and Industry Fixed Yes Yes Yes Yes
Wald Test 112.51 146.21 112.51 144.31
Observations 2,321 2,321 2,321 2,321
Note: This table reports the results of logit regression models where the dependent variable is CEO turnover. Each number represent the logit coefficient estimate of the corresponding variable. Clustered robust standard errors are shown in parentheses. Financial ratios are winsorized at the 1% and 99% level. Models (1) and (2) is run over firm-specific CSR. Models (3) and (4) is run over industry-adjusted CSR. * denotes significance at the 10 percent level; ** denotes significance at the 5 percent level; *** denotes significance at the 1 percent level.
Appendix 2. OLS Estimates of Aggregate CSR and CEO Power on CEO Turnover
Dependent variable: CSRt-1 (1) (2) (3) (4) CEO Power 0.026*** (0.065) CEO Ownership 0.009* (0.007) CEO Duality 0.038*** (0.101) CEO Tenure -0.008* (0.006) Constant -6.756*** -6.848*** -6.807*** -6.989*** (1.021) (0.576) (0.581) (0.596)
Controls Yes Yes Yes Yes
R-squared 0.21 0.21 0.21 0.21
Observations 3,226 3,226 3,226 3,226
Note: This table reports the results of simple OLS regression models where the dependent variable is lagged value of CSR during the period of 2007 - 2017. Each number represent the coefficient estimate of the corresponding variable. Clustered robust standard errors are shown in parentheses. Financial ratios are winsorized at the 1% and 99% level. * denotes significance at the 10 percent level; ** denotes significance at the 5 percent level; *** denotes significance at the 1 percent level.
27
Appendix 3. Logit Estimates of Individual CSR Components and CEO Power on CEO Turnover
Dependent Variable: CEO Turnover (1) Corporate Governance (2) Corporate Governance (3) Social (4) Social (5) Environmental (6) Environmental CSRt-1 -0.151 -0.157 0.037 0.043 -0.051 -0.085 (0.098) (0.102) (0.04) (0.041) (0.079) (0.08) CEO Power -0.602*** -0.587*** -0.559*** (0.131) (0.012) (0.128) CSRt-1 x CEO Power -0.061* 0.011 -0.173* (0.141) (0.048) (0.099) Constant -4.658*** -5.818*** -4.323*** -5.465*** -4.689*** -5.815*** (1.497) (1.493) (1.507) (1.504) (1.504) (1.493)
Controls Yes Yes Yes Yes Yes Yes
Year and Industry Fixed
Yes Yes Yes Yes Yes Yes
Wald Test 115.06 148.51 112.55 143.00 112.04 145.56
Observations 2,321 2,321 2,321 2,321 2,321 2,321
Note: This table reports the results of logit regression models where the dependent variable is CEO turnover during the period of 2007 to 2017. Each number represent the logit coefficient estimate of the corresponding variable. Clustered robust standard errors are shown in parentheses. Models (1) and (2) is run over corporate governance aspect of CSR. Models (3) and (4) is social aspect of CSR. Models (5) and (6) is run over
environmental aspect of CSR. Financial ratios are winsorized at the 1% and 99% level. * denotes significance at the 10 percent level; ** denotes significance at the 5 percent level; *** denotes significance at the 1 percent level.
Appendix 4. Marginal Effects of Aggregate CSR and CEO Power on CEO Turnover with Additional Controls
Dependent variable: CEO turnover
(1) (2) (3) (4) CSRt-1 -0.001 -0.001 -0.001 -0.001 (0.003) (0.002) (0.003) (0.029) CEO Power -0.071*** -0.071*** (0.018) (0.018) CSRt-1 x CEO Power -0.003* -0.002* (0.003) (0.001)
Industry-adjusted CSR No No Yes Yes
Controls Yes Yes Yes Yes
Year and Industry Fixed Yes Yes Yes Yes
Wald Test 84.88 126.46 84.88 126.46
Observations 1,253 1,253 1,253 1,253
Note: This table reports the results of logit regression models where the dependent variable is CEO turnover during the period of 2007 – 2017. Each number represent the marginal estimate of the corresponding variable. Clustered robust standard errors are shown in parentheses. Financial ratios are winsorized at the 1% and 99% level. Models (1) and (2) is run over firm-specific CSR. Models (3) and (4) is run over industry-adjusted CSR. * denotes significance at the 10 percent level; ** denotes significance at the 5 percent level; *** denotes significance at the 1 percent level.
28
Appendix 5. Marginal Effect of High/Low CSR and CEO Power on CEO Turnover
Dependent variable: CEO turnover
____ High CSR Low CSR (1) (2) (3) (4) CSRt-1 -0.004 -0.003 0.015 0.014 (0.006) (0.006) (0.014) (0.016) CEO Power -0.073** -0.081** (0.031) (0.034) CSRt-1 x CEO Power -0.026* -0.003 (0.022) (0.039)
Controls Yes Yes Yes Yes
Year and Industry Fixed Yes Yes Yes Yes
Wald Test 59.74 92.49 83.63 105.18
Observations 753 753 589 589
Note: This table reports the results of logit regression models where the dependent variable is CEO turnover during the period of 2007 – 2017. Each number represent the marginal estimate of the corresponding variable. Clustered robust standard errors are shown in parentheses. Financial ratios are winsorized at the 1% and 99% level. Models (1) and (2) is run over a subsample of observations in which the net score value of CSR is above 0. Models (3) and (4) is run over a subsample of observations in which the net score value of CSR is below 0. * denotes significance at the 10 percent level; ** denotes significance at the 5 percent level; *** denotes significance at the 1 percent level.
Appendix 6. Firm Governance Characteristics in High and Low CSR Firms Variables High CSR (1) Low CSR (2) Difference (3) T-test (4) CEO Ownership 3.46 [1.22] 2.34 [0.504] 1.42 (0.19) 7.42*** CEO Duality 0.57 [1.00] 0.55 [1.00] 0.02 (0.02) 1.31 CEO Tenure 13.74 [11.00] 11.06 [9.00] 2.67 (0.29) 9.08*** CEO Compensation 7.82 [7.89] 8.37 [8.32] -0.55 (0.04) -14.93*** Independent Directors on Board (%) 0.63
[0.78] 0.69 [0.8] -0.06 (0.01) -5.85*** Common Stock Owned by all Directors
(%) [0.02] 0.04 [0.02] 0.02
-0.02 (0.01)
-4.01***
Common Stock Owned by Institutional Shareholders (%) 0.16 [0.07] 0.16 [0.07] -0.001 (0.03) -0.001 No. of observations 1,395 1,194 2,589 2,589
Note: This table reports the summary statistics and sample paired t-test of firm governance
characteristics during the period of 2007 – 2017. Each number represents the marginal estimate of the corresponding variable. Median of the corresponding variable are shown in brackets. Clustered robust standard errors are shown in parentheses. Models (1) and (2) is run over a subsample of observations in which the net score value of CSR is above 0. Models (3) and (4) is run over a subsample of observations in which the net score value of CSR is below 0. * denotes significance at the 10 percent level; ** denotes significance at the 5 percent level; *** denotes significance at the 1 percent level.
