Master Thesis
CEO Narcissism and CSR dimensions: The Moderating Role
of Resource Dependency
Groningen, January 18th, 2021
Garini Retno Utami
S4155602
University of Groningen Faculty of Economics and Business
g@student.rug.nl
Supervisor: dr. P.J. Steinberg Co-assessor: dr. C. Carroll
ABSTRACT
Social and environmental issues have been emerging in recent years that drive companies to perform Corporate Social Responsibility (CSR). Drawing upon Upper Echelon Theory (UET) and Resource Dependency Theory (RDT), this study examines the effect of CEO narcissism on two dimensions of CSR – internal and external CSR – with the role of moderating effect of slack and market competition. Using panel data of S&P 100 companies from 2008 to 2018, slack was found to strengthen the inverse relationship between CEO narcissism and internal CSR. Although some hypotheses were not supported, this study contributes to the upper echelon and resource dependency literatures to better understand about the relationship between personality traits, role of resources, and CSR dimensions.
Keywords:
Table of Contents
1. INTRODUCTION ... 4
2. THEORETICAL BACKGROUND AND HYPOTHESES BUILDING ... 8
2.1 Disentangling Corporate Social Responsibility (CSR) ... 8
2.2 CEO Narcissism and CSR Oriented Activities ... 10
2.2.1 CEO Narcissism and Internal CSR Activities ... 12
2.2.2 CEO Narcissism and External CSR Activities ... 13
2.3 The Moderating Role of Resource Dependence ... 13
2.4 The Moderating Role of Firm’s Slack ... 14
2.5 The Moderating Role of Market Competition ... 16
3. METHODOLOGY ... 17
3.1 Sample and Data Collection ... 17
3.2 Measurements ... 18
3.2.1 Dependent Variables: Internal and External CSR Actions ... 18
3.2.2 Independent Variables: CEO Narcissism ... 19
3.2.3 Moderator ... 20
3.2.4 Control Variables ... 21
3.3 Analytical Method ... 23
4. RESULTS ... 23
4.1 Descriptive Statistics and Correlations ... 23
4.2 Regression Results and Hypothesis Testing ... 27
4.3 Robustness Analysis ... 28
5. DISCUSSION AND CONCLUSION ... 29
5.1 Theoretical Implications ... 31
5.2 Practical Implications ... 33
5.3 Limitations and Future Research ... 33
References ... 35
4
1. INTRODUCTION
The increasing phenomenon of environmental damage and ethical issues among individuals and organizations (e.g., Inappropriate treatment of employees) (Harrison & Freeman, 1999) nowadays have driven the emergence of voluntary corporate responsibility actions to respond towards these issues (Waddock S., 2008). They aim to make changes among organizations to remedy their business’ impacts on “people, the planet, and societies” (Waddock, S., 2008, p.87). It is then suggested that companies should not only generating profits but also commit on focusing environmental and social matters (Kenton, 2020). Consequently, companies perform Corporate Social Responsibility (CSR) actions as a medium to show endowment in the well-being of communities and environment that they affect due to the business activities. In return, companies generally obtain positive economic return; it is assumed because CSR enhances positive brand image of being responsible to the community (Freeman, 1984). This argument was empirically supported in which CSR and firm performance have a positively significant relationship (Hawn & Ioannou, 2016; Margolis et al., 2009).
In conducting CSR activities, companies disentangled their focus on the external and internal oriented CSR activities (Hawn & Ioannou, 2016). The external activities are related to companies’ effort in widening their “branding, disclosure, and partnership”, while the internal scope refers to “training or forming board committees” (Hawn & Ioannou, 2016, p.1). However, the proportion of these CSR dimensions were not equally balanced (Visser, 2011). Some firms decided to engage more in external actions than internal, and vice versa. This imparity may be detrimental because emphasizing too much on external actions may create a green-washing image for the company, while stressing more on the internal actions may find the company be unable to capture the CSR value due to low “transparency and credibility on the firm’s internal actions” (Hawn & Ioannou, 2016, p.1). Instead of highlighting certain stakeholders, it should align the needs of anyone who matter to the company (Al-Shammari et al., 2019). Therefore, examining two CSR dimensions is important to establish more reasonable CSR strategies (Al-Shammari et al., 2019).
5
(Aguilera et al., 2007; Petrenko et al., 2015; Al-Shammari et al., 2019). This is surprising because scholars supported the argument raised by upper echelon theory (UET) in which personality traits effect organization’s decision making (Al-Shammari et al., 2019). Other studies have indeed analyzed the effect of company’s members on CSR (e.g., Petrenko et al., 2015; Chin et al., 2013), but only few studies focused on the CEO’s psychological traits. Due to the lack of evidence on the relationship between the CEO’s personality and external and internal CSR activities, the role of the CEOs and their personality trait were considered to be underexplored.
To examine personality trait, I adapt the UET (Hambrick and Mason, 1984), in which psychological character of top managers and CEOs influence company’s actions. This is because of their power, previous experiences, and disposition (Chin et al., 2013) that create credibility. One of the prominent psychological character among literatures is narcissism (Al-Shammari et al., 2019). Empirical evidences found that a narcissistic CEO leads to more successful firms (Maccoby, 2000; Wales et al., 2013) due to their interest in self-promoting and high confidence that leads to dramatic action for the firms which brings positive breakthroughs for the firms. On the other hand, a narcissistic CEO rewards the employees subjectively rather than accomplishment which result in “higher turnover rates and lower employee enthusiasm” (Resick et al., 2009; Lin et al., 2018).
As CEO narcissism plays a big role in the business’ success, the findings on the relationship between CEO narcissism and CSR should be more explored (Ahn et al., 2020). The decision to perform CSR is derived from the leaders’ personal interests for “attention and image reinforcement” (Petrenko et al., 2015, p.2). Since CSR is perceived to be socially favorable by various stakeholders (Freeman, 1984; Masulis and Reza, 2015; Tang, Mack, and Chen, 2017), it drives the CEOs to engage in CSR activities. As a result of the desire of attention, a narcissistic CEO may have a great influence on company to decide whether or not to emphasize on internal or external oriented CSR activities. Hence, the link of narcissism and CSR is vital to be further explored to add more perspectives of a specific driver of CSR initiatives. Therefore, this paper will examine the influence of CEO narcissism on internal and external CSR.
6
to identify any resources available to execute the strategies; like CSR. These resources are generally obtained from the existing stakeholders, which makes it important to maintain relationships with stakeholders for resource accessibility (Matuleviciene & Stravinskiene, 2015). Based on the resource dependency theory, the relationship between stakeholders and companies are reinforced by the needs of providing resources and distribute profits to stakeholders, respectively. This put the stakeholders to become more superior in terms of having power on controlling the resources availability. In return, such superiority put the company in a “weaker position in regard to stakeholders” (Matuleviciene & Stravinskiene, 2015, p. 77; Neville, Bell, & Menguc, 2005) which creates high dependency on stakeholders for resources. Greater dependency indicates that CEOs of company rely on the stakeholders for resources in which they will attempt more on continuously giving good impression to the stakeholders by performing CSR.
Resources in companies are divided into internal resources and external resources (Tang et al., 2014; Nemati et al., 2010). The internal resource indicates what the company already has, while the external resources refer to the market situation that can influence the success of the company itself. Internal resources include firm’s slack and external resources refer to market competition (Tang et al., 2014). Since some companies may perceive CSR as a non-core business activity (Rekom et al., 2013), companies may need to identify whether they are able to perform CSR given limited resources. If a company has slack as a resource, it indicates that the company is more financially flexible (Cheng et al., 2014). Thus, slack is important to be analyzed whether it drives company to conduct CSR activities. Besides that, in a competitive market company may or may not be able to perform CSR as a tool to achieve competitive advantage because CSR may lead to positive or negative outcome. Positive outcome is related to better reputation, while negative result refers to greenwashing; hence, company needs to consider the outcome carefully. As a result, market competition caught an increasing attention to literatures whether it influences company’s consideration on CSR engagement (Tang et al., 2014).
7
organization’s performance such as participating in CSR activities (Thompson, 1967) and market competition was viewed as the most powerful force that determines the viability of a company (Meng et al., 2016).
These resources may explain better the relationship between the CEO’s personality traits and CSR activities. For example, other kind of personality trait named hubris was previously examined (Tang et al., 2014). The findings suggest that companies with more slack leading a hubristic CEO is less likely performing CSR activities because company does not require to satisfy stakeholders’ interests; whereas firms with less slack strengthen the relationship between CEO hubris and CSR as the company depend stakeholders for resources and support (Tang et al. 2014). External resources such as market competition also influenced the CEO hubris and CSR relationship. When the market competition is high, firms tend to rely more on stakeholders for resources because they need to address greater complexity (Hambrick and Finkelstein, 1987). In this situation, hubristic CEOs would lower their perception that they can deal with problems and lower enthusiasm in engaging in socially irresponsible activities (Tang et al., 2014).
Accordingly, the research question (RQ) in this paper is the following: “How CEO narcissism influence the internal and external CSR oriented activities, and how does resource dependency moderate this relationship?”. I expect that narcissistic CEOs will have positive effect on external CSR activities because they seek attention and acknowledgment of their good deeds from massive audiences. However, narcissistic CEOs may be less likely to conduct internal CSR activities because of smaller scope of admiration.
The research question will be further analyzed by using a quantitative method by using a dataset of S&P 100 companies from 2008 to 2018 by using multiple resources for data collection. The result of this study will contribute to the UET which defines relationship between executives’ characteristics and organizational decisions (Hambrick, 2007). The findings also contribute to resource dependence theory that sheds light on how two distinct resource types act as contingent factor in this relationship between CEO narcissism and CSR dimensions. Academicians and practitioners can use the empirical evidence on how internal and external resource dependency as moderating variables can strengthen and/or weaken the link of CEO narcissism and CSR activities.
8
of this study. The last chapter comprises the discussion, conclusion of this study and implications, as well as limitations and future research suggestions.
2. THEORETICAL BACKGROUND AND HYPOTHESES BUILDING
The following literature reviews will describe the relevant theories as the basis for the arguments raised in this study. First, the concept of Corporate Social Responsibility and its components will be discussed. Next, the Upper Echelon Theory (UET) on the personality traits and how CEO narcissism affect the company’s decision on internal and external CSR practices will be elaborated, along with my hypothesis developments. It follows with the discussion on the underlying theories on Resource Dependence Theory (RDT) and hypothesis on the role of moderating variables.
2.1 Disentangling Corporate Social Responsibility (CSR)
Companies generally formulate strategies in order to survive in the market competition (Al-Shammari et al., 2019) by integrating both market and nonmarket strategy (Baron 1995). Literatures argued that market strategies focus on company’s relationship with market related organizations such as customers, competitors, and suppliers (Cadogan et al., 2002; van Raaij and Stoelhorst, 2008). Conversely, nonmarket strategy refers to a range of activities that involve social engagement (Delmas and Montes-Sancho, 2010). In regards to market strategy, nonmarket strategy is thought to be an important complementary (Doh et al., 2012; Henisz and Zelner, 2012; Kingsley et al., 2012; Sawant, 2012). One of vital nonmarket strategy in firms that has growing acknowledgment in the literatures is Corporate Social Responsibility (CSR) (Chin et al., 2013).
9
to create productive relationship with stakeholders (Waddock and Smith, 2000; Bhattacharya et al., 2009).
Numerous scholars have extensively studied the aggregate CSR actions and solely categorize the CSR actions when necessary, typically into environmental, social, and governance matters (Hawn & Ioannou, 2016). The neo-institutional theory suggested that companies should perform two types of actions in order to gain legitimacy (e.g., King et al., 2005), which are internal and external oriented actions (Hawn & Ioannou, 2016). Disentangling the CSR is driven by two factors (Al-Shammari et al., 2019). First, aggregate CSR may offer less accurate description of company’s CSR activities engagement (Aguinis and Glavas, 2012; Aguinis and Glavas 2019). For instance, firm ABC scored high on the category of community relations while scored low on employee relations ratings, and overall, “it will still have the same aggregate score as another firm that may have moderate emphases on both of these dimensions” (Al-Shammari et al., 2019, p.109). Second, each social dimension contains distinguished characteristics and deserves to be investigated independently (Al-Shammari et al., 2019). Therefore, disaggregating CSR activities lead to better understanding in the tradeoffs of social performance and resource allocation for CSR activities (Wang et al., 2016). Furthermore, researchers also define both types of actions. The internal actions reflect “conformity with legitimized structure” (Hawn and Ioannou, 2016, p. 4) which support the company to obtain legitimacy (e.g., Fligstein, 1985), while the external actions aim at certain audiences who grant legitimacy such as through public endorsement (Hawn & Ioannou, 2016). Since legitimacy is critical for firm’s profitability and survival, it is notable to distinguish the CSR actions into internal and external actions (Hawn & Ioannou, 2016).
10
In contrast, external actions aim at obtaining organizational support by external parties and seek public endorsement (eg. McDonnell and King, 2013; Sine et al., 2007). This indicate that external actions value more on “highly visible initiatives” (Hawn & Ioannou, 2016, p.4) from external audiences. In CSR context, external actions involve statements that accentuate their actions undertaken by the firm; typically, through reports or sustainability reports that highlight firm’s environmental initiatives in order to signal external audiences (Hawn & Ioannou, 2016). The most salient external CSR oriented activities that firms nowadays participate in is protecting the nature through reducing the pollution percentage and sustainability initiatives in regards for healthier and safer earth (El Akremi et al., 2018; Turker, 2009). Moreover, being committed to consumers also denotes external CSR in which firms consistently provide high quality products (El Akremi et al., 2018; Farooq et al., 2017; Jia et al., 2019).
2.2 CEO Narcissism and CSR Oriented Activities
A firm’s CEO has been considered to be a dominant member and have significant impacts on the firm’s strategic formulation (Hambrick and Mason, 1984); this is because of CEOs’ highest position that they have in the firm which enable them to have absolute authority in giving order (Rijsenbilt, 2011). Upper echelons theory also agrees that the key members in decision-making process are the top management teams (TMT) and CEOs (Hambrick and Mason, 1984; Miller and Toulouse, 1986). Moreover, the personality of CEOs was also emphasized to be another fundamental factor to the firm’s success (Finkelstein and Hambrick, 1996). The personal characteristics of CEO shape to what extent they are able to influence the organizational strategy and future organization’s performance (Rijsenbilt, 2011).
11
& Rhodewalt, 2001); especially the CEOs personality have significant influence on organizational decisions (Peng and Fang, 2010; Wang et al., 2016).
CEOs are likewise treated as special individuals due to their highest rank within the company which deliver uplifting self-confidence (Raskin et al., 1991). They prefer to prioritize their own self-interests above company’s interests (Jensen and Meckling, 1976) and more responsive to social praise than to objective indicator (Chatterjee and Hambrick, 2011). Based on general population’s narcissistic tendencies graph, narcissism tends be found in CEOs (Emmons, 1984). This finding is in accordance with other evidences that narcissism was identified as a core of CEO’s personality characteristic (e.g., Judge et al., 2006; Chatterjee and Hambrick, 2007; Resick et al., 2009). Narcissistic CEOs often seek admiration from external audiences as a form of validation regarding their capabilities (Chatterjee and Hambrick, 2007) by performing challenging and bold actions that audiences can notice easily (Rijsenbilt, 2011) to hold their narcissistic supply (Kernberg, 1985).
One of the bold actions that can help CEOs to maintain their status quo is CSR; which was deemed to be a tremendous practice where narcissistic CEOs can become the center of attention and social admiration along with another favorable outcome such as more positive self-image and self-enhancement (Kim, 2018). CSR engagement also frequently leads to praise or criticism for the CEO rather than to the company itself; leading to more willingness for the CEO to engage in CSR (Petrenko et al., 2015). Three arguments were found on the CSR initiatives as practices that give narcissistic CEOs a narcissistic supply (Petrenko et al., 2015). Firstly, CSR is a value-loaded program to encourage social benefits and provide an individual with attention to him/herself through participating in socially acceptable actions (Bogart et al., 2004; Petrenko et al., 2015). Secondly, CSR includes a range of “value sensitive audiences in adulation, media attention, and praise which are external sources of narcissistic supply” (Petrenko et al., 2015, p. 265; Wallace and Baumeister, 2002). Third, CSR provides ways to alter an individual’s status quo and acts as medium for narcissistic CEOs to gain responsive and attentive audiences (Petrenko et al., 2015).
12
activities (El Akremi et al., 2018; Jones et al., 2018). Accordingly, it would be valuable to take into account the effects of CEO narcissism on each CSR dimensions (Al-Shammari et al., 2019).
2.2.1 CEO Narcissism and Internal CSR Activities
Internal stakeholders (e.g., employees) are considered to be targeted by the company’s CSR activities (Al-Shammari et al., 2019) and vital assets for the firm (Mory et al., 2015). Mory et al. (2015) suggest that internal CSR activities are generally related to behavior that largely focus on employee-oriented; for example, fostering employment stability, a positive working environment, and diversity.
Taking into account the internal stakeholders into the company’s behavior is important; however, this idea might be difficult to achieve if the company is led by a narcissistic CEO because they have the tendency to overlook others’ issues (Dubrin, 2012; Kernberg, 1985) which consider themselves to be more valuable than others (Chen et al., 2009). They also underestimate the importance of others who have different social status level (Morf and Rhodewalt, 2001, Rhodewalt and Morf, 1998). As narcissists always attempt to find more considerable admiration from greater audiences (Rijsenbilt & Commandeur, 2013; Rosenthal & Pittinsky, 2006), they are less likely to gather knowledge from others, especially those that they consider to be below them in rank (Chen et al., 2009).
Chen et al. (2009) identified a CEO that has low self-esteem is called reactive narcissist; who pretend to have high self-regard but in fact they are lacking of it. To enhance their self-esteem, they tend to adhere with those who have equal or higher social status rank, yet employees are not considered to be in this rank (Chen et al., 2009). Consequently, they give less attention to the employees’ welfare while utilizing the existing resources for increasing themselves with “external visibility and image” (Al-Shammari et al., 2019, p.110). Because of these arguments, narcissistic CEOs are expected to be less likely engage in internal CSR actions. Hence, I hypothesize the following:
Hypothesis 1a (H1a): Narcissistic CEO will have negative effect on the internal CSR oriented
13
2.2.2 CEO Narcissism and External CSR Activities
CSR practices are also aimed at external stakeholders such as the consumers, suppliers, community, and the environment; who are considered to affect or be affected by the company’s behavior (Story et al., 2016). This is called external CSR activities which is more outward-activities oriented. According to the social identity theory, external CSR helps the company by increasing its prestige (Farooq et al., 2017) and prosocial image (Brammer et al., 2006) because external actions tend to be more visible and allow greater public endorsement for the company (Hawn and Ioannou, 2016). Being acknowledged by massive audiences allow higher legitimacy for the company (Baum and Oliver, 1991) which is in line with narcissistic’ desire.
Narcissists limit their interactions to stick in a group of high-status members (Brockner, 1988). These members are perceived to be authentically in top level social status which influences the narcissistic CEO to show “outward confidence and seek to maintain it” (Al-Shammari et al., 2019, p.110). Narcissistic CEOs engage in practices of showing off to impress others (Al-Shammari et al., 2019). In return, they expect to keep the narcissistic supply continuously given to them through higher attention and praise (Kernberg, 1985).
Gaining admiration by doing CSR may uplift the firm’s status. Nevertheless, public will be most likely to acknowledge and appreciate the CEO rather than the company (Petrenko et al., 2015). This drives narcissistic CEOs to engage in outward behavior even more (Chatterjee and Hambrick, 2007) as they demand “drama and an attentive audience” (Al-Shammari et al., 2019, p.110). Therefore, engaging in bold actions is a visible way that can be easily noticed by larger segment of people and appropriate for narcissistic CEOs as they are considered as exhibitionists (Chatterjee and Hambrick, 2007; Raskin and Terry, 1988). Hence, narcissistic CEOs are expected to be more likely engage in external CSR actions. I hypothesize the following:
Hypothesis 1b (H1b): Narcissistic CEO will have positive effect on the external CSR oriented
activities.
2.3 The Moderating Role of Resource Dependence
14
narcissistic CEO, such boundary circumstance can help determining to what extent the company is able to conduct the strategies, especially CSR actions. These boundaries can be interpreted as company’s dependence on stakeholders (Tang et al., 2014). Due to its dependency, company needs to continuously satisfy stakeholders’ expectations for resources and supports availability by engaging CSR which will benefit the shareholders (Margolis et al., 2007). However, the presence of narcissistic CEOs may switch the situations for company differently (Al-Shammari et al., 2019). Therefore, it is important to determine “any contingent factor that may curb a CEO’s inflated ego” (Tang et al., 2014, p.6).
According to resource dependence theory (RDT), the contingent factors also include resources which are often linked to power that can define the company’s behavior (Pfeffer and Salancik, 2003). This is because when firms need resources, they will experience dependence with stakeholders (Denktas-Sakar & Karatas-Cetin, 2012). RDT also explains the degree of boards’ dependence on resources (Pfeffer, 1972b); in which the boards refer to top management team (TMT) and CEOs. The term ‘dependence’ is interpreted as a condition where an individual relies on other’s actions to meet certain outcomes (Pfeffer, 1982). As previously explained, depending on stakeholders mean increasing the requirement of company to satisfy shareholders’ interests (Tang et al., 2014). In the context of CSR, literatures previously found that hubristic CEOs tend to have “lower dependence on stakeholders for resources and support” (Tang et al., 2014, p.6) due to their high confidence on their own capabilities and become less likely conduct CSR
The level of dependency on stakeholders is still vague. Thus, classifying resources into internal (e.g., slack) and external (e.g., market competition) can better understand firm’s dependency on stakeholders for resources (Wang and Qian, 2011). Accordingly, the boundary circumstances in the relationship between CEO narcissism and CSR dimensions can be explained through the firm’s resource dependence on stakeholders (Tang et al., 2014).
2.4 The Moderating Role of Firm’s Slack
15
required to maintain the company” (Cyert and March, 1963, p.36); (2) slack acts as a resource for disputes resolution which can be used to solve any issues when the slack is considered to be sufficient (Tan and Peng, 2003); (3) slack is utilized as a buffer that separate “the technical core of the organization from environmental turbulence” (Tan and Peng, 2003, 1250); (4) slack as a facilitator of strategic enabling the company to experiment in novel strategies (Thompson, 1967). Organizational slack, derived from financial instruments (Greve, 2003), is excess resources that firm prepares to protect itself from unexpected external shocks and environmental changes (Bourgeois, 1981; Bromiley, 1991; Singh, 1986).
Companies that have greater slack as resources tend to have more flexibility in performing CSR (Cheng et al., 2014). The availability of slack was considered to be the main determinant in firm’s decision of CSR engagement (Clarkson et al., 2011; Reverte, 2009) because slack determines whether a company is able commit in CSR activities (Lin et al., 2019) as it indicates that the company is financially independent (Orlitzky et al., 2003). Others, however, argued if a company has more organizational slack, CEOs are less likely to conduct CSR because they can prosper without shareholders’ resources and support (Tang et al., 2014) hence there is no need to satisfy shareholders’ interests. On contrary, if the company is lack of resources, CEO may in favor conduct CSR in order to gain trust of stakeholders and then obtain the required resources and support in return (Tang et al., 2014). s
16
availability will boost narcissistic CEOs decision on external CSR. Therefore, I hypothesize the following:
Hypothesis 2a (H2a): The negative relationship between CEO narcissism and internal CSR
oriented activities will be strengthened when higher slack is available.
Hypothesis 2b (H2b): The positive relationship between CEO narcissism and external CSR
oriented activities will be strengthened when higher slack is available.
2.5 The Moderating Role of Market Competition
Market competition is another determinant that can influence company’s strategies; as it produces uncertainties and risks for the company (Mia and Clarke, 1999). Other argued as competition becomes more intense, companies perform better (The Australian Financial Review, 1995). This is because companies are racing to perform as best as they can, given substantial availability of substitutes exist from competitors (Lensen, 2011) drives them to win over others. Several factors were considered to determine the competition intensity, such as the concentration of market supply and the size market demand (Raith, 2003; Karuna, 2007; Lensen, 2011). The intensity of competition within an industry is also based on the aggregate strength of various actions in the industry itself (Porter, 1979).
17
In regards to CEO narcissism and CSR dimensions, market competition is expected to moderate this relationship. When market competition is high, a narcissistic CEO may be motivated to maintain its reputation by having continuous solid relationship with shareholders in order to secure his or her access to valuable resources. This is achieved by performing CSR as it was considered to be an effective way for company to gain competitive advantage and to show off the company’s performance to shareholders (Kemper et al., 2013). Since the praise or criticism from CSR engagement is mostly aimed at the CEO (Petrenko et al., 2015), in a competitive market narcissistic CEOs will be more likely to engage in CSR to boost their reputation in the market. This is in line with the conception of external CSR as greater audiences are involved. However, in a competitive market, narcissistic CEOs are less inclined to value the employees’ interests because they will use this circumstance as an opportunity to increase narcissistic CEOs’ brand image externally. According to these arguments, I hypothesize the following:
Hypothesis 3a (H3a): The negative relationship between CEO narcissism and internal CSR
oriented activities will be strengthened in a more competitive market.
Hypothesis 3b (H3b): The positive relationship between CEO narcissism and external CSR
oriented activities will be strengthened in a more competitive market.1
3. METHODOLOGY
This chapter elaborates the methodology choices of testing hypotheses in this research. Firstly, the sample and data collection are described. Secondly, the description of each variables and their measurement procedures are discussed. Lastly, I describe the analytical method used in this research.
3.1 Sample and Data Collection
Initially, the research sample consisted of 139 companies of United States (US) based firms listed in Standard and Poor’s (S&P) 100 companies for 11 years period from 2008 to 2018. The S&P 100 index is a subset of S&P 500 index that list1 a set stock markets index. The reason of using
18
this sample is because the S&P 100 is the largest public companies in US which makes the availability of the data across databases is accessible. These companies are also deemed to be stable and have diverse groups of industry in which the result of this research can be relevant to a wide array of companies’ matters. During data collection, some required company’s information was missing due to a merger or acquisition by other firms, or even went bankrupt. In order to avoid bias results, I deleted 17 companies and reached a final sample of 122 companies.
I carried out a quantitative analysis using secondary data from multiple reliable databases and sources. First, I started by collecting data on the independent variable: CEO narcissism. I obtained the data from my supervisor’s overarching project. This data is proprietary for my supervisor and is classified. I am only allowed to use this data for the purpose of this thesis. I used ORBIS to collect company identifier, such as company ISIN number and two-digit Standard Industry Code (SIC). I also collected some of control variables from ORBIS: firm age, CEO age, and CEO tenure. Second, I used Eikon Thomson Reuters database to collect data on the dependent variable (Internal and External CSR actions) and one of my control variables: firm size. This database has been proven by CSR literatures (Cheng et al., 2014; Ioannou and Serafeim, 2012) that it concentrates on providing relevant and auditable CSR information (Hawn and Ioannou, 2016). For firm size, I used Thomson Reuters to collect number of employees that became the final measure of firm size. Third, I used Wharton Research Data Services (WRDS) to collect data of my moderating variables, slack and market competition, as well as the rest of control variables. In WRDS, I used COMPUSTAT to collect data on cash holding for slack measurement and cash and accounts receivable for robustness check. I also collected data on sales of every company in the industry based on 2-SIC to measure the market competition. The rest of control variables, such as CEO gender, ROA, and R&D Intensity data are available on COMPUSTAT.
Subsequently, after all the data has been collected, I combined each excel files into one Data Editor in Stata to get the final dataset for regression. The final dataset consists of 122 companies with 1,366 observations.
3.2 Measurements
3.2.1 Dependent Variables: Internal and External CSR Actions
19
to indicate which actions refer to internal and external CSR category. The internal CSR actions consists of “corporate policy questions as well as several quantitative indicators of CSR implementation” (Hawn and Ioannou, 2016, p.9). The external CSR actions includes “claims that firms make and disclosures they issue” (Hawn and Ioannou, 2016, p.9). These claims and disclosures point out communication activity which indicate company’s engagement for public attention (Hawn and and Ioannou, 2016). The list of both types of CSR action is divided into 21 internal CSR actions and 24 external CSR actions. However, Thomson Reuters has an update which alter the access of ASSET4 to Eikon from Datastream; resulting some missing lists of CSR actions. Consequently, ASSET4 only provides 21 internal CSR actions and 20 external CSR actions as the final composition of both CSR actions for period of 2008 to 2018 as shown in Appendix B.
In order to measure both CSR actions, all datapoints firstly should be presented in binary variables which are 0 or 1. To measure internal CSR actions, the first four datapoints that consists of scores had to be converted to binary variables as they were originally presented in percentages. First, this was conducted by calculating the industry average of the score per year based on two-digit SIC codes. Then, I categorized the score that below the industry average as “0”, while those that are above the industry average are categorized as “1”. When the first four datapoint scores have been transformed to binary variables, the final measurement of internal CSR actions is dividing the sum of all internal CSR datapoints by 21 and multiplied by 100 to get a percentage score. To measure external CSR actions, all datapoints are summed up and then divided by 20 and multiplied by 100 to get a percentage score.
3.2.2 Independent Variables: CEO Narcissism
20
this thesis. Following Chatterjee and Hambrick (2007), CEO narcissism will be measured by computing the use of first-person singular pronouns and first-person plural pronouns (“I” vs “We”). Raskin and Shaw (1988) suggested that the use of person singular pronouns and first-person plural pronouns depict an individual’s self-centeredness which is an indicator of narcissism. This method was considered to be related with the Narcissistic Personality Inventory (NPI) (Raskin and Hall, 1979) that measure a narcissistic individual level based on scores (Rhodewalt and Morf, 1995).
The CEO narcissism data was firstly obtained from digital company’s conference earning calls transcripts, specifically in the question-and-answer (Q&A) session that are available on Thomson Reuters Eikon. The decision to use Q&A session is because CEOs can talk freely in which their statements are not priorly arranged through transcript. Then, the Q&A session provide CEOs statement in using first-person singular pronouns and first-person plural pronouns such as “I” and “We”. LIWC software was used to analyze how many times these pronouns mentioned by CEOs. To measure the CEO narcissism score, the use of first-person singular pronouns (“I”) is divided by the use of first-person plural pronouns (“We”) (Raskin and Shaw, 1988). Since this research is annually based, thus the mean of CEO narcissism data of the four quarters will be used.
3.2.3 Moderator
Two moderators are used in this research, namely slack and market competition. As companies often face uncertainty within and outside of the firm, slack is used to address these issues by allowing the company to adapt to changes (Kim and Bettis, 2013). Following Tang et al. (2014) and Kim and Bettis (2013), slack was measured by company’s cash holding which consist of cash and short-term investment; as slack was considered to be proportionate to cash holding (Kim and Bettis, 2013). I obtained the data from COMPUSTAT by using on the data range from 2008 to 2018. Then, I selected the company that has active status on the market. COMPUSTAT already provided a variable of Cash and Short-term Investment (CHE) which simplified the data collection. Therefore, slack was measured by adding company’s cash and short-term investment from 2008 to 2018.
21
and Wiseman, 1999; Harris, 1998). In a natural logarithm, the four-firm concentration is calculated by “dividing the combined sales of four largest firms in the industry by total sales of that industry” (Tang et al., 2014, p.10) based on 2-SIC codes. I decided to use two-digit SIC codes to present the company’s industry group as it captures bigger market scope of one industry. I obtained the sales data of all industry in the dataset from 2008 to 2018 in COMPUSTAT. Then, I calculated the four largest sales in each year in each industry and summed up the total sales of that industry in each year. Lastly, I used the four-firm concentration ratio formula to arrive at the final score of market competition (Tang et al., 2014).
3.2.4 Control Variables
I controlled firm-level and CEO-related variables to be observed whether they have impact on the dependent variable. The decision to use these control variables are following prior literatures (e.g., Al-Shammari et al., 2019; Hawn and Ioannou et al., 2016).
CEO Gender. To control CEO Gender, I used the company’s CEO profile to identify the gender and it is available in COMPUSTAT. Literatures argued that gender differences between male and female have different effects on CSR ratings and performance (McGuinness et al., 2016). When women are in the board position, the CSR activity tends to be more apparent (Kahreh et al., 2014). Hence, male and female CEO may have different emphasis level on internal and external CSR actions.
22
CEO Tenure. I controlled CEO tenure by collecting data on the period length of CEO is in position from ORBIS. In selecting main company’s strategies, tenure has an important effect (Simsek, 2007) that it presents CEO’s ability as a leader to the market which leads to career concern (Chen et al., 2019; Gibbons and Murphy, 1992; Holmstorm, 1982). The newly appointed CEOs will have stronger willingness to show their concern on CSR as they tend to address the career concern problem by participating in CSR activities (Hong et al., 2016). Therefore, as CEOs have been in position for longer period, their CSR engagement interest decreases.
Financial Performance. The financial performance of company will be measured by Return-on-Assets (ROA). ROA was computed by dividing net income by total assets and this dataset is available on COMPUSTAT. ROA is another important variable to be controlled since company with better financial performance (relatively higher ROA) is more likely to conduct CSR activities because they have sufficient financial resources (Tang et al., 2014).
Firm Size. Firm size was controlled by measuring the number of employees. Size indicates company’s visibility in the eye of shareholders; as larger firms are more visible in the market, they are more responsive to social matters (Udayasankar, 2008). Hence, bigger firms boost their CSR participation as their scale of activities are impartial (Cowen et al., 1987; Udayasankar, 2008).
Firm Age. Prior literatures suggest that firm age influence company’s social activities participation (e.g., Adam and Hardwick, 1998; Waddock and Graves, 1997). As firms getting older, they participate more in CSR because of stable cash flows and profitability (Jiraporn and Withisuphakorn, 2015); sufficient financial allow them to engage in CSR. Firm age is controlled by calculating the difference between the entire period (2008 to 2018) and company’s year of establishment.
R&D Intensity. Following Hawn and Ioannou (2016), R&D intensity is an intangible variable that is used as control variable in this thesis. R&D intensity is measured by dividing R&D expense by total sales.
23
3.3 Analytical Method
This thesis analyzes a panel data that requires to test two types of analysis, namely Fixed Effect (FE) and Random Effect (RE). FE model allows omitted variables to be correlated with the variables in the model to control for the omitted variable bias. This model “control for the effects of time-invariant variables with time-invariant effects” (Williams, 2018, p.2). It also measures changes within the company. Schurer and Yong (2012) suggest that FE model is often perceived to be the “the golden standard”. On the other hand, in RE model the omitted variables are not allowed to be correlated with explanatory variables in the model. It measures changes across the firms.
This study tests the changes of CEO within company; because my data shows that companies experienced changes of CEO position from 2008 to 2018. In order to ensure which model is more suitable, I ran a Hausman test in Stata suggested by Greene (2008). The null hypothesis was that random effects is preferred while the alternative hypothesis prefers fixed effects. After running the Hausman test, the significance value is 0.000 (p-value = 0.000) which means the null hypothesis is rejected. Therefore, I used fixed effects model to run the regression analysis.
4. RESULTS
This section explains the result that begins with describing the descriptive statistics and correlations among variables. Subsequently, it elaborates the regression analysis and result of hypothesis testing. This chapter is concluded by testing the robustness analysis.
4.1 Descriptive Statistics and Correlations
The descriptive statistics and correlations matrix are presented in Table 1 and Table 2 respectively. Table 1 presents that the sample of 122 companies consists of 1,366 observations and all variables observed the equal number of observations. From the descriptive statistics, I noticed that on average the CEOs are male (μ = 0.0674) with 5.9 year of term of office. The companies used in the sample generally have been in the market for more than 40 years with some relatively positive financial performance (μ of ROA = 13.33).
24
insignificant with External CSR (r=-0.0134). Slack is positively related to Internal CSR (r=0.114, p<0.01) and External CSR (r=0.143, p<0.01). On the contrary, Market Competition is significantly and negatively related to Internal CSR (r=-0.0463, p<0.1), while only negatively related and not significantly related to External CSR (r=-0.00238).
25
Table 1. Descriptive Statistics
ln = Natural Logarithm
Table 2. Correlations Matrix
Variable (1) CEO Narcissism (2) Internal CSR (3) External CSR (4) Slack (5) Market Competition (ln) (6) CEO Gender (7) CEO Age (8) CEO Tenure (9) ROA (10) Firm Age (11) Firm Size (12) R&D Intensity CEO Narcissism 1 Internal CSR 0.0794*** 1 External CSR -0.0134 0.730*** 1 Slack 0.0192 0.114*** 0.143*** 1 Market Competition (ln) 0.00190 -0.0463* -0.00238 -0.138*** 1 CEO Gender -0.0439 -0.0287 0.0409 -0.0710*** 0.0458* 1 CEO Age 0.154*** 0.212*** 0.164*** 0.0374 0.0589** 0.00944 1 CEO Tenure 0.139*** 0.0680** -0.0189 0.0134 0.0183 -0.0700** 0.306*** 1 ROA 0.0305 0.178*** 0.198*** -0.260*** 0.213*** -0.0299 0.146*** 0.0268 1 Firm Age 0.0649** 0.311*** 0.291*** -0.0532* -0.00659 0.0391 0.233*** 0.00494 0.0358 1 Firm Size -0.100*** 0.144*** 0.172*** 0.0913*** 0.241*** -0.00469 0.0658** -0.0234 0.0714*** 0.0561** 1 R&D Intensity -0.0357 -0.0497* -0.0547** -0.00856 -0.0295 -0.00736 0.0135 -0.0264 0.00865 0.0384 -0.0135 1 * p < 0.1, ** p < 0.05, *** p < 0.01, ln = natural logarithm
Variable N Mean Std. Dev. Min. Max.
26
Table 3. Regression Results
* p < 0.1, ** p < 0.05, *** p < 0.01
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
VARIABLES Internal CSR External CSR Internal CSR External CSR Internal CSR External CSR Internal CSR External CSR Internal CSR External CSR
CEO Narcissism 0.00567 0.000393 0.00584 0.000511 0.00558 0.000302 0.00574 0.000415
(0.00688) (0.00660) (0.00686) (0.00660) (0.00689) (0.00659) (0.00686) (0.00658)
CEO Narcissism x Slack -0.00774** -0.00538 -0.00808** -0.00573
(0.00362) (0.00428) (0.00365) (0.00437)
CEO Narcissism x Market Competition (ln) -0.00325 (0.00911) -0.00345 (0.00767) -0.00392 (0.00906) -0.00392 (0.00761) Slack 0.0103 0.0221 0.00963 0.0220 0.0131 0.0244* 0.00978 0.0222 0.0135 0.0248* (0.00961) (0.0137) (0.00954) (0.0137) (0.00896) (0.0136) (0.00968) (0.0139) (0.00897) (0.0138) Market Competiton (ln) 0.0101 0.0470 0.00969 0.0469 0.00999 0.0472 0.0102 0.0474 0.0106 0.0477 (0.0287) (0.0320) (0.0283) (0.0321) (0.0282) (0.0322) (0.0287) (0.0322) (0.0286) (0.0324) CEO Gender 0.117** 0.0385 0.119** 0.0387 0.120** 0.0396 0.119** 0.0388 0.120** 0.0398 (0.0579) (0.0334) (0.0578) (0.0336) (0.0577) (0.0335) (0.0579) (0.0338) (0.0578) (0.0337) CEO Age 0.00705*** 0.00466*** 0.00691*** 0.00465*** 0.00683*** 0.00460*** 0.00690*** 0.00465*** 0.00682*** 0.00459*** (0.00216) (0.00151) (0.00214) (0.00153) (0.00214) (0.00153) (0.00213) (0.00151) (0.00212) (0.00151) CEO Tenure -0.00518*** -0.00447*** -0.00514*** -0.00447*** -0.00505** -0.00440*** -0.00510*** -0.00443*** -0.00500** -0.00436*** (0.00193) (0.00141) (0.00194) (0.00140) (0.00194) (0.00140) (0.00193) (0.00139) (0.00192) (0.00139) ROA 0.000700 0.00102* 0.000671 0.00102* 0.000652 0.00101 0.000655 0.00100 0.000632 0.000986 (0.000969) (0.000612) (0.000959) (0.000613) (0.000955) (0.000612) (0.000959) (0.000615) (0.000954) (0.000614) Firms Size 0.000245* 0.000300* 0.000245 0.000300* 0.000244* 0.000299* 0.000245* 0.000300* 0.000244* 0.000299** (0.000148) (0.000153) (0.000149) (0.000153) (0.000146) (0.000152) (0.000147) (0.000152) (0.000144) (0.000150) Firms Age 0.0117*** 0.0124*** 0.0117*** 0.0124*** 0.0118*** 0.0124*** 0.0117*** 0.0124*** 0.0118*** 0.0124*** (0.00205) (0.00177) (0.00205) (0.00177) (0.00207) (0.00178) (0.00205) (0.00176) (0.00206) (0.00178) R&D Intensity -3.09e-05*** -1.68e-05*** -3.04e-05*** -1.68e-05*** -3.01e-05*** -1.66e-05*** -3.01e-05*** -1.65e-05*** -2.98e-05*** -1.62e-05***
(1.42e-06) (1.15e-06) (1.47e-06) (1.22e-06) (1.50e-06) (1.26e-06) (1.72e-06) (1.44e-06) (1.75e-06) (1.47e-06)
27
4.2 Regression Results and Hypothesis Testing
Based on the result of Hausman test, I performed a fixed effect model to test the hypotheses. Table 3 presents the regression results of each model with standardized variables. Model 1 and Model 2 only presents the effects of all control variables on dependent variables. Model 3 and Model 4 add the independent variable (CEO Narcissism) to the model. Model 5 and Model 6 include the interaction term between CEO Narcissism and Slack, then the interaction term of CEO Narcissism and Market Competition is presented in Model 7 and Model 8. The last two models exhibit the result of all variables.
Model 3 is used to test Hypothesis 1a (H1a). With the independent variable – CEO Narcissism- is added to the model, the R2 is 0.338 which does not show much differences with Model 1 and Model 2 that have R2 of 0.337 and 0.294 respectively. It means that adding CEO narcissism to the model does not explain more of the proportion of the variance for Internal CSR. I expected that the coefficient will be negative, but the result shows a positive coefficient and not significant (β = 0.00567). Therefore, H1a is rejected. Model 4 explains the Hypothesis 1b (H1b) testing. Compared to the previous model, Model 4 examines the relationship with the dependent variable – External CSR. Similarly, the R2 is 0.294 which does not explain more of variability. As expected, the coefficient is positive but not statistically significant (β = 0.000393). Although the coefficient shows a direction of the hypothesis, H1b is rejected.
Model 5 and Model 6 add moderator – Slack – in which the result for CEO narcissism on Internal CSR and External CSR do not change, with both positive coefficients and insignificant (β = 0.00584 and β = 0.000511). The R2 for Model 5 is 0.340 and for Model 6 is 0.295 which do not show significant difference of variability compared to previous models. These models add the interaction effect between CEO Narcissism and Slack on Internal CSR and External CSR to test the result of Hypothesis 2a (H2a) and Hypothesis 2b (H2b). Model 5 shows the negative coefficient and significant (β = -0.00774, p<0.05). Therefore, H2a is supported. Additionally, Model 6 shows negative coefficient and insignificant (β = -0.00538), hence H2b is rejected.
28
Hypothesis 3b (H3b) respectively. The coefficients in both models are negative – in Model 7 (β = -0.00339) and Model 8 (β = -0.00294) – and both are insignificant; hence H3a and H3b are rejected.
4.3 Robustness Analysis
In order to strengthen the model analysis, I performed robustness analysis to examine whether the results would change under different assumption. The robustness is checked by using Random Effect model and Slack by using different measurement. The result of the first robustness analysis is shown in Appendix D and several different results are found in random-effects model.
As random-effect model measures the data across firms, I included industry dummy as a control variable as a baseline to have more accurate robustness analysis. First, the coefficients of the effect of CEO narcissism on External CSR become negative (Model 14, Model 16, Model 18, and Model 20) and on Internal CSR is similarly positive (Model 13, Model 15, Model 17, Model 19), but both relationships remain insignificant. Hence, H1a and H1b are rejected. Second, the interaction effect of Slack with CEO narcissism shows similar result on Internal and External CSR. In Model 15, this interaction effect on Internal CSR presents negative coefficient and statistically significant (β = -0.00699, p<0.05). Similar to Model 5, Model 15 also show supporting evidence on H2a. Model 16 exhibits the interaction effect of Slack on External CSR. The coefficients remain negative and insignificant (β = -0.00418), thus H2b is also rejected. Third, slight changes occur on the coefficient of interaction term between CEO narcissism and Market Competition. The relationship of this interaction term with Internal CSR in Model 17 (β = -0.00142) and in Model 19 (β = -0.00199) are negative and not significant. Similarly, with External CSR as shown in Model 18 (β = -0.00219) and Model 20 (β = -0.00255) are negative and not significant. H3a and H3b are therefore also rejected.
29
of the models remain the same after different assumptions are applied. The difference in model 5 and model 21 is lies on the significance level, but the coefficients in both models are the same.
5. DISCUSSION AND CONCLUSION
The final chapter starts with the general discussion and summary about the research, including the conclusion, based on the empirical evidences and support the findings based on theoretical and methodological reasonings. The next section discusses the theoretical and practical implication. Lastly, this chapter concludes with the limitations found in this study and future research suggestions.
According to the upper echelon theory, CEOs and their personality traits needs to be more emphasized on examining the CEOs’ participation on the CSR dimensions: internal and external CSR (Al-Shammari et al., 2019; Petrenko et al., 2015). One of the most prominent personality characteristics that scholars have suggested is narcissism (e.g., Chatterjee and Hambrick, 2007). Moreover, I also combined a resource dependence theory into this study to examine the effect of slack and market competition as moderators (Tang et al., 2014) in the relationship between CEO narcissism and CSR dimensions. By adapting these theories, this study aims to answer the research question: “How CEO narcissism influence the internal and external CSR oriented activities, and how does resource dependency moderate this relationship?”. In order to answer this research question and test the hypotheses, I performed an empirical test by using panel data with a total observation of 1,366 for a period of 11 years from 2008 to 2018 with fixed-effects model regressions.
30
giving positive and great admiration to the narcissistic CEOs, external CSR engagement may become a liability and threat for the CEO’s career concern.
The methodological reasoning behind H1a and H1b result is due to different level of narcissism may affect the emphasis on CSR dimension importance. As this current sample has changes of CEOs across the year (2008 to 2018), there might be potential differences of narcissism level; categorizing the level of narcissism may explain the emphasis on internal CSR differently. Previous dissertation by Al-Shammari (2017) found that when categorizing the CEO narcissism, high narcissistic CEOs are found to engage more in internal CSR than lower narcissistic CEO. Moreover, the datapoints used in this study is imbalance between internal and external CSR. Al-Shammari et al. (2019, p.114) suggest that accurate mechanism to measure both types of CSR are needed by “equally emphasized rather than one at the expense of the other”. Due to the time constraint and unavailability of data provided from Thomson Reuters ASSET4, the datapoints might have measured the CSR activities less accurately.
In addition, this study supports H2a, but not for H2b. In regards to H2a, this finding shows that the availability of slack in organization is used by the narcissistic CEO’s as a resource to boost their image and maintain reputation (Al-Shammari et al., 2019). Also, the organization’s slack is not used for improving the well-being of the company’s working condition or employee-related, and narcissistic CEOs do not value this issue to be important. This shows that the internal CSR will less likely to be conducted by narcissistic CEOs when higher slack is available. Moreover, one potential explanation for the rejection of H2b is as conducting external CSR involves greater audiences and usually involves CSR report disclosure that include financial resource (e.g., slack), there might be higher risk if the CSR activities do not correspond with the existing slack. Narcissistic CEOs need to identify whether the slack is intended for external CSR engagement purposes. If narcissistic CEOs still use slack that is not for CSR engagement, there might be a probability the narcissistic receive high criticism and bad reputation, which is contradict with what narcissistic CEO desire. Another reason for the different result between H2a and H2b is the measurement of slack, which is cash holding. It might not be intended for CSR-related purposes which is perceived to be less accurate in determining the moderating effect of slack.
31
resulting low quality of CSR and may create a greenwashing image to shareholders. As a result, narcissistic CEOs may perceive market competition as another potential reputational threat due to the existing risk that CSR engagement planning is not well prepared due to external pressure. Especially in competitive market, shareholders pay high attention to companies. If narcissistic CEOs decide to conduct low quality of CSR because they are too focus on enhancing their own reputation, shareholders can easily notice this. To avoid this risk, narcissistic CEOs might decide to be free-riders in the industry (Samuelson, 1954) and less likely to conduct internal and external CSR activities.
From methodological reasoning, my data used four-firm concentration ratio which measures the ratio of competition within the industry based on 2-SIC code. However, Kontesa (2020) suggest that the measurement of market competition is classified into high and low market competition. Due to time constraint, this study does not categorize the market competition which might have provide more accurate findings. Therefore, market competition classification might examine two different competition situations more precisely and moderate the relationship between narcissistic CEO and CSR activities dimensions.
5.1 Theoretical Implications
32
findings show contrast results which add another new aspect to the literatures. This inconsistency finding can be explained by different sample and year settings. Also, while prior research use companies that have constant CEO in the entire period (Al-Shammari et al. 2019) my data involves changes in CEOs position. Therefore, the result of this study demonstrates add knowledge about personality trait on CSR dimensions.
Secondly, the current study sheds light on resource dependency theory which suggest that resources as boundary limitation (Tang et al., 2014) can influence company’s decision making (Pfefer and Salancik, 2003). This study also highlights resources as a tool to demonstrate one’s power which allow him or her to satisfy personal interests (Fiol et al., 2001; Kim et al., 2005; Salancik and Pfeffer, 1977; Vredenburg and Brender, 1988; Weber, 1947; Whitmeyer, 2001) and attracts narcissists to be leader (Campbell and Campbell, 2009), where leader corresponds to CEO. This study examines narcissism as another personality trait and its effect with slack to better understand each of CSR dimension. Moreover, this study present contrast result as what expected. Since vast literatures measure slack as moderator on CSR as a one-dimensional construct, this study contributes to the role of slack in two CSR dimensions. This study, therefore enriches the theory that slack as internal resources have different effect in the relationship between CEO narcissism and each CSR dimensions.
Furthermore, market competition as an external resource (Tang et al., 2014) was also found to have insignificant effect on the relationship between CEO narcissism and CSR. In a competitive market, companies might compete to utilize CSR as a differentiation mechanism to achieve competitive advantage (Fernandez-Kranz and Santalo, 2010). On the other hand, the result of this study suggest that market competition show insignificant result as a moderator between CEO narcissism and CSR dimensions. Market competition was argued to lead to improper behavior such as greenwashing (Shleifer, 2004) which can create negative market reactions (Du, 2015). Yook and Lee (2020) noted that narcissistic CEOs tend to avoid criticism on pay lip service including greenwashing. This study is in line with this theory. Therefore, the finding of this study adds to the extant literatures about market competition in relation to upper echelon theory and CSR literatures.
33
measurement might not deliver slack as a significant moderating variable for the CEO narcissism and CSR dimensions relationship.
5.2 Practical Implications
It is important for company to acknowledge whether their CEO has a narcissistic character considering it influences the outcomes of decision-making. This finding shows that narcissistic CEOs do not lead to internal and external CSR engagement, and slack and market competition have different contribution to these relationships. This shows that leaders with narcissistic characters may find other factors to influence their actions in emphasizing each CSR dimension. Thus, practitioners can take into account of other variables that can affect narcissistic CEOs actions to ascertain the company has favorable outcomes.
Additionally, practitioners can apply other attributes correlated with resources dependency theory to enrich our knowledge about how narcissistic CEO’s dependence on theory can influence their action on CSR dimensions. As Tang et al. (2014) differentiate resources into internal and external resources, practitioners can consider these resources in determining CEO’s personality to examine whether it has consistent result with my finding or not. The outcome of this can be used for practitioners to better understand on what drives company to act be socially responsible.
5.3 Limitations and Future Research
34
Another limitation in this study is the data and measurement of internal and external CSR activities. Originally, Hawn and Ioannou (2016) classified 120 datapoints from Thomson Reuters ASSET4. However, due to some updates on Thomson Reuters, several variables measuring internal and external CSR activities are not available which make the measurement of internal and external CSR in this study is less thoroughly examined. Future research may use other resources to obtain full datapoints of CSR. Moreover, the concept that classify CSR dimension across countries are different (Al-Shammari et al., 2019). The datapoints that represent CSR dimension classification might not be equal in one single country such as USA (the country scope in this study). Hence, future research needs to ensure that the conception of what constitutes CSR is relevant in certain country to increase the reliability of the study result.
In addition, I used cash holding as the main measurement of slack. However, this is found not to be the sole measure of slack as literatures suggest various financial equation (Lin et al., 2019). Moreover, the cash holding of companies in my data might not be used for CSR-related purposes. Sayekti (2011) suggest that not all slack is intended for CSR activities. Therefore, it is suggested for future research to firstly acknowledge the company’s financial category to identify which type of slack that can be considered to be used for CSR activities. Classifying the company’s financial category can be conducted by disentangling the component of slack into: available, recoverable, and potential (e.g., Bouregois, 1981). Future research is also encouraged to use other financial equation such as free cash flow and total revenue (Lin et al., 2019) or the total of cash and cash equivalents divided by firm size (Zhang et al., 2018).
35
References
Adams, M., & Hardwick, P. (1998). Determinants of the leasing decision in United Kingdom listed companies. Applied Financial Economics, 8(5), 487-494.
Agle, B., Mitchell, R., & Sonnenfeld, J. (1999). Who matters to CEOs? An investigation of stakeholder attributes and salience, corporate performance, and CEO values. Academy of Management Journal, 507-525.
Aguilera, R., Rupp, D., Williams, C., & Ganapathi , J. (2007). Putting the S back in corporate social responsibility: a multilevel theory of social change in organizations. Academy of Management Review, 32, 836-863.
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. Aguinis, H., & Glavas, A. (2019). On corporate social responsibility, sensemaking, and the
search for meaningfulness through work. Journal of Management, 45(3), 1057-1086. Ahn, J. S., Assaf, A. G., Josiassen, A., Baker, M. A., Lee, S., Kock, F., & Tsionas, M. G. (2020).
Narcissistic CEOs and corporate social responsibility: Does the role of an outside board of directos matter? International Journal of Hospitality Management, 1-8.
Al-Shammari, M. A. (2017, March). CEO NARCISSISM, PRIOR EXPERIENCE, BOARD PRIOR EXPERIENCE, AND CORPORATE SOCIAL RESPONSIBILITY (CSR): AN INTEGRATED FRAMEWORK USING UPPER ECHELONS THEORY,
STAKEHOLDERS’ THEORY, AND A DECOMPOSED MEASURE OF CSR. Texas: University of Texas at Arlington.
Al-Shammari, M., Rasheed, A., & Al-Shammari, H. (2019). CEO narcissism and corporate social responsibility: Does CEO narcissism affect CSR focus? Journal of Business Research, 106-117.
Arora, P., & Dharwadkar, R. (2011). Corporate Governance and Corporate Social Responsibility (CSR): The Moderating Roles of Attainment Discrepancy and Organizational Slack. Corporate Governance: An International Review, 19(2), 136-152.
Baron, D. P. (1995). Integrated Strategy: Market and Nonmarket Components. Californian Management Review, 37(2).
Becker, S. (2019, March 5). What is resource dependence and what does it have to do with transitions? Retrieved from Sustainable transitions blog:
36
Bhattacharya, C., Korschun, D., & Sen, S. (2009). Strengthening stakeholder-company relationships through mutually beneficial corporate social responsibility initiatives. Journal of Business Ethics, 85, 257-272. doi:10.1007/s10551-008-9730-3
Bingham, J. B., Dyer, W. G., Smith, I., & Adams, G. L. (2011). A stakeholder identity
orientation approach to corporate social performance in family firms. Journal of Business Ethics, 99(4), 565-585.
Bogart, L., Benotsch, E., & Pavlovic, J. (2004). Feeling superior but threatened: the relation of narcissism to social comparison. Basic and Applied Social Psychology, 26, 35-44. Bourgeois, L. (1981). On the measurement of organizational slack. Academy of Management
Review, 6(1), 29-39.
Brammer, S., Millington, A., & Pavelin, S. (2006). Is philanthropy strategic? An analysis of the management of charitable giving in large UK companies. Business Ethics: A European Review, 15(3), 234-245.
Brockner, J. (1988). Self-esteem at work: Research, theory, and practice. Lexington: Lexington Books.
Bromiley, P. (1991). Testing a casual model of corporate risk taking and performance. Academy of Management Journal, 34, 37-59.
Cadogan, J. W., Sundqvist, S., Salminen, R. T., & Puumalainen, K. (2002). Market-oriented behavior: comparing service with product exporters. European Journal of Marketing, 36(9), 1076-1102.
Campbell, W. (1999). Narcissism and romantic attraction. Journal of Personality and Social Psychology, 77, 1254-1270.
Campbell, W., & Campbell, S. (2009). On the self-regulatory dynamics created by the peculiar benefits and costs of narcissism: A contextual reinforcement model and examination of leadership. Self and Identity, 8(2), 214-232.
Carroll, A. B. (1979). A three-dimensional conceptual model of corporate performance. Academy Management Review, 4, 497-505.
Chatterjee, A., & Hambrick, D. C. (2007). It's all about me: Narcissistic chief executive officers and their effects on company strategy and performance. Administrative Science
Quarterly, 52, 351-386.
Chatterjee, A., & Hambrick, D. C. (2011). Executive personality, capability cues, and risk taking: How narcissistic CEOs react to their successes and stumbles. Administrative Science Quarterly, 56, 202-237.
