National Culture and CSR: The Influence of National Culture on CSR Performance and the Moderating Effects of Headquarters’ Home Country Sustainable Development and CEO Foreignness

(1)

Master Thesis

National Culture and CSR: The Influence of National Culture on

CSR Performance and the Moderating Effects of Headquarters’

Home Country Sustainable Development and CEO Foreignness

Nina Onkenhout S2888335

c.m.t.onkenhout@student.rug.nl

Supervisor: Dr. O. Lindahl Co-assessor: Prof. Dr. H.J. Drogendijk MSc International Business & Management

University of Groningen, Faculty of Economics and Business Date of Submission: June 15th_{, 2020}

(2)

ABSTRACT

During the past decades the relationship between national culture and CSR performance has gained increased attention worldwide. However, the CSR performance of companies deviates between firms from different countries. Furthermore, to date, the outcomes of research on this relationship are contradicting and further research is needed. Therefore, the key objective of this thesis is to fill the research gap by examining the four cultural dimensions of Hofstede – power distance, uncertainty avoidance, individualism, and masculinity – and including two non-accounted moderating effects which could potentially explain why research still finds inconclusive results. Based on the institutional, stakeholder, and Upper Echelons’ theory, the country-level moderator, headquarters’ home country sustainable development, and the individual-level moderator, the CEO Foreignness are explained as moderators on this examined relationship. Using multiple regression analysis, national culture on CSR performance of the world’s largest 165 multinationals located in 23 countries and operating in 8 different industries are investigated, this leads to surprising results. It is found that power distance seems to be an important predictor of a firm’s CSR performance, however, the other cultural dimensions do not seem to have a meaningful impact in this thesis. Additionally, the interaction term of power distance shows a negative significant relation and uncertainty avoidance a positive significant relation on headquarters’ home country sustainable development. Furthermore, only the interaction-term of individualism of CEO foreignness is tested to be significant and create a stronger negative relationship between individualism and CSR performance. These findings thereby add to the body of knowledge on national culture, CSR performance, headquarters’ home country sustainable development and, the CEO foreignness and hence provide important theoretical and managerial implications.

(3)

2.8 Conceptual Model ... 20 3. RESEARCH METHODOLOGY ... 20 3.1 Data Collection ... 20 3.2 Data Sample ... 21 3.3 Measurement of Variables ... 21 3.3.1 Dependent Variable ... 21 3.3.2 Independent Variables ... 23 3.3.3 Moderating Variables ... 23 3.3.4 Control Variables ... 24 3.4 Data Analysis ... 26 3.5 Robustness Test ... 26 4. EMPIRICAL RESULTS ... 27

4.1 Preliminary Requirement Tests for the Moderated Multiple Regression Analysis ... 27

4.1.1 Outliers ... 27

4.1.2 Normality ... 27

4.1.3 Homoskedasticity ... 27

4.1.4 Linearity ... 27

4.1.5 Multicollinearity ... 28

4.1.6 Independence of residual values ... 28

(4)

4.3 Correlation ... 29

4.4 Multiple Regression Analysis ... 31

4.5 Robustness Tests ... 35 4.5.1 Bootstrapping ... 35 4.5.2 Country Representation ... 36 5. DISCUSSION ... 37 6. CONCLUSION ... 42 6.1 Theoretical Implications ... 42 6.2 Practical Implications ... 43

6.3 Limitations and Future Research ... 43

REFERENCES ... 46

APPENDICES ... i

Appendix A: Overview of Previous Results of National Culture on CSR ... i

Appendix B: Research Methodology Overview ... i

Appendix C: Descriptive of Countries, Industries and Cultural Scores ... ii

Appendix D: Outliers ... v

Appendix E: Assumption of Normality ... vi

Appendix F: Assumption of Heteroskedasticity ... viii

Appendix G: Assumption of Multicollinearity ... ix

Appendix H: Assumption of Independence of Residuals Values ... ix

(5)

List of Figures

Figure 1: Conceptual Model ... 20

Figure 2: Interaction term CEO foreignness with individualism ... 32

Figure 3: Interaction term CEO foreignness with masculinity ... 33

List of Tables Table 1: Descriptive Statistics ... 29

Table 2: Correlation Matrix ... 30

Table 3: Summary of Results - Main Regression Analysis ... 33

Table 4: Regression Analysis ... 34

Table 5: Summary of Results – All Analyses ... 37

List of Abbreviations

CEO Chief Executive Officer CSP Corporate Social Performance CSR Corporate Social Responsibility

e.g. exempli gratia

EPI Environmental Performance Index

HCSD Home Country Sustainable Development

IDV Individualism

IVR Indulgence

LTO Long-Term Orientation

MAS Masculinity

MNE Multinational Enterprise

OLS Ordinary Least Square

PDI Power Distance

ROA Return on Assets

ROE Return on Equity

ROW Rest of the World

SDG Sustainable Development Goals

UET Upper Echelons Theory

UK United Kingdom

UAI Uncertainty Avoidance

USA United States

(6)

1. INTRODUCTION

Environmental disasters that are caused by organizations’ unchallenged expansion and the increasing global focus on environmental and economic sustainability have resulted in growing public attention to corporate social responsibility (CSR) (Boulouta & Pitelis, 2014; Liu, 2019). Some researchers have defined CSR as “actions that appear to further some social good, beyond the interests of the firm and that which is required by law” (McWilliams & Siegel, 2001, p. 117). Therefore, CSR can be described as voluntary behaviour through which firms try to contribute to social wellbeing and environmental conservation that goes beyond economic and legal concerns (European Commission, 2001). Currently, firms need to not only pursue their profit objectives; they are also responsible for their impact on society and the environment (Boulouta & Pitelis, 2014). In this thesis, corporate environmental and social responsibility is referred to as CSR, and CSR performance is the extent to which companies implement their CSR activities (De Bakker, Groenewegen, & Den Hond, 2005). According to prior studies, CSR performance could lead to good reputations (Melo & Garrido‐Morgado, 2012); offers competitive advantage (Du, Bhattacharya, & Sen, 2011); increases the potential to motivate, attract, and retain employees (Greening & Turban, 2000); and enhances firm legitimacy (Zheng, Luo, & Maksimov, 2015). Such potential beneficial outcomes are valuable for firms and society; therefore, it is important to understand what drives CSR.

(7)

According to Ringov and Zollo (2007), national culture influences how people expect businesses to behave. Therefore, national culture seems to play an important role and could sometimes cause differences in an organization’s CSR performance since national culture could influence expectations regarding CSR in both positive and negative ways. In this thesis, the national culture of the firm is defined as the country where the firm’s headquarters is located. Since a multinational corporation’s headquarters plays a significant role in the formulation and implementation of the corporate strategy, it is assumed that national culture is taken into consideration when determining the corporate strategy (Petruzzella, Salvi, & Giakoumelou, 2017). Therefore, the national culture of the headquarters’ country has a particularly strong influence on the firm.

In cross-cultural research, most scholars have applied the Hofstede model of national culture, because it is the most widely accepted representation of national cultural dimensions that characterize societies (AlAnezi & Alansari, 2016). However, in the literature there are contradicting findings regarding the relationship between national culture and CSR performance. For example, power distance (PDI) and masculinity (MAS) have conflicting outcomes in different studies. According to Peng, Dashdeleg, and Chih (2012) and Petruzzella et al. (2017), both dimensions have a negative influence on firms’ CSR engagement and environmental performance; however, prior study results from Ho, Wang, and Vitell (2012) show a positive effect of PDI on corporate social performance (CSP), and Naeem and Khurram (2019) found a positive effect of MAS on CSR. Additionally, Halkos and Skouloudis (2017) have concluded that these factors and individualism (IDV) are insignificant. Furthermore, Ioannou and Serafeim (2012) found that IDV has a positive relationship with CSR performance. According to Peng et al. (2012), uncertainty avoidance (UAI) is also a predictor of CSR performance. However, according to Petruzzella et al. (2017), UAI and IDV are both insignificant, although their study suggests that UAI negatively influenced CSR. The fact that these studies all have different outcomes indicates that further research on the relationship between national culture and CSR performance is needed to close the literature gap.

(8)

economic and social development on the environment and human societies are unsustainable (Tan, Shuai, Jiao, & Shen, 2017). Therefore, country institutions are including principles and rules of sustainable development in their policies (Banerjee, Gupta, & McIver, 2019). National institutions set the “rules of the game” and therefore influence a firm’s decision making; however, institutions vary significantly across countries. Since companies are embedded in institutions, external and internal pressures on firms concerning CSR may differ around the world (Ioannou & Serafeim, 2012). For example, in countries with poorly developed institutions, standards for social and environmental indicators may be low and affect the company. Thus, it is interesting to examine the influence of headquarters’ HCSD since the headquarters has a pivotal role within the firm and represents the company’s legal domicile (Ciabuschi, Dellestrand, & Holm, 2012).

It is also interesting to consider CEO foreignness since the CEO is the leader of the organization and thus plays an important role in making decisions about CSR (Slater & Dixon-Fowler, 2009). In recent decades, increasing globalization and borderless global careers have resulted in multinational enterprises (MNEs) acquiring more foreign CEOs, because MNEs believe that this will benefit them (Hymowitz & White, 2004) since these CEOs have cosmopolitan views, multilingual skills, and global visions (Arp, Hutchings, & Smith, 2013). Today, it is necessary for firms to develop into multicultural MNEs due to competition. Foreign CEOs have different values, behaviours, and beliefs than their colleagues. Consequently, it is interesting for organizations to get a better understanding of foreign CEOs in terms of whether they affect the relationship between national culture and CSR performance. Accordingly, these two factors might explain the differences in the aforementioned conflicting findings.

Therefore, the purpose of this thesis is to fill the research gap by proposing the following research question:

(9)

Performance Index (EPI) was used. Finally, data from BoardEx was gathered to discover the CEO’s foreignness. To answer the research question, a multiple regression analysis was run. This thesis makes a theoretical contribution to existing research by studying the relationship between national culture and CSR performance to better understand why CSR performance is still different across firms. To explain why prior research has indicated conflicting findings in the direct relationship, two moderators – headquarters’ HCSD and CEO foreignness – were added which further expand the current results.

This thesis is structured as follows: first, the literature is reviewed, and based on the relevant concepts, the relations between the concepts are presented. The hypotheses are then listed, and the conceptual model is shown. Next, the research methodology is explained and described, including data collection and measurements of constructs. Statistical analyses are then analyzed in-depth and discussed, followed by a conclusion. Finally, the limitations, contributions, and recommendations for future research are examined.

2. LITERATURE REVIEW

This section reviews the literature on CSR, CSR performance, national culture, the headquarters’ HCSD, and CEO foreignness followed by the interrelationship between these constructs. Subsequently, hypotheses are formulated and the conceptual model is shown.

2.1 Corporate Social Responsibility

(10)

The concept of CSR is not new; in 1946, Fortune magazine asked business leaders how they were acting regarding their social responsibilities. In the 1950s, the evolution of the CSR construct began (Carroll, 1999). Since then, the field of CSR has grown significantly, and this has resulted in a great proliferation of approaches, terminologies, and theories (Garriga & Melé, 2004). According to Carroll (1999, p. 270), Howard Bowen is considered the “Father of Corporate Social Responsibility”. According to Bowen, CSR “refers to the obligation of businessmen to pursue those policies, to make those decisions, or to follow those lines of action which are desirable in terms of objectives and values of our society” (Bowen, 1953, p. 6). In the beginning, CSR was more referred to as the social responsibility of individuals engaging in businesses but later became “CSR” as we know today since corporation prominence in the business sector was still in its early stages (Carroll, 1999). In the 1950s through the 1970s, several scholars conceptualized different and new definitions focusing more on the role and perspective of a corporation (Kang, Lee, & Yoo, 2016). Carroll (1979, p. 500) has defined CSR as follows: “The social responsibility of business encompasses the economic, legal, ethical and discretionary expectations that society has of organizations at a given point in time”. Therefore, businesses should strive to “make a profit, obey the law, be ethical, and be a good corporate citizen” (Carroll, 1991, p. 283). These four categories can be depicted as a pyramid with economic responsibilities as the foundation, because without these responsibilities, the other goals cannot be achieved (Carroll, 1991). Carroll’s four categories have been applied by various theorists and remain a dominant paradigm of CSR (Schwartz & Carroll, 2003). In contrast, Milton Friedman was against the view that social matters should be a concern of firms; he argued that companies have one objective: “to make as much money as possible while conforming to the basic rules of society, both those embodied in the law and those embodied in ethical custom” (Friedman, 1970, p. 1). This only rejects Carroll’s philanthropic category. Due to increasing stakeholder pressures, business people are unlikely to exclude the philanthropic category in their programs (Carroll, 1991).

2.2 CSR Performance

(11)

it; if you can’t understand it, you can’t control it; if you can’t control it, you can’t improve it”. However, there is no single method to measure CSR performance (Wolfe & Aupperle, 1991). The broad concept of CSR and its dimensions make the creation of a performance framework a complex but necessary process (Carroll, 2000). Subjective measurement of performance remains a key point of the measurement process (Carroll, 2000). Several rating agencies measure CSR performance in different ways, and these scores are reliable sources of CSR performance (Cho, Lee, & Park, 2012; Cho, Lee, & Pfeiffer Jr, 2013). Since firms can define opportunities for improvement, identify their strengths and weaknesses, and adapt their strategies by measuring their CSR activities into performance (Giannarakis & Theotokas, 2011), thus measuring the CSR performance of firms is relevant. In this thesis, the definition of CSR performance is congruent with CSRHubs’ definition, which examines how a company performs in relation to its employees, environment, community, and governance (CSRHub, 2020a). Firm scores from this database are used to measure CSR performance.

2.3 National Culture

In the field of international business, a country’s culture is seen as a fundamental determinant underlying systematic differences in people’s behaviour (Hofstede, 2001; Steenkamp, 2001). Over the years, several scholars have tried to identify an appropriate definition of national culture. One of the most well-known sociologists of culture is Hofstede (1991, p. 6), who defines national culture as “the collective programming of the mind that distinguishes the member of one group or category of people from others”. These patterns of feeling, thinking, and acting are acquired in one’s childhood and remain relatively stable over time. Moreover, Schwartz (2006) emphasizes that values are the principle features of culture. People within a given culture share sets of values that sequentially translate into commonly shared identities, attitudes, and beliefs rooted in societal practices and norms (Halkos & Skouloudis, 2017). It provides the basis for interactions among group members (Beugelsdijk, Kostova, & Roth, 2017). Moreover, cultural values emphasize shared perceptions and behaviour about what is right and desirable in society. They justify and shape individual and organizational behaviour (Schwartz, 2006). Moreover, national culture, including its norms, values, beliefs, practices, and meanings, implies that one way of doing is preferable to another (Newman & Nollen, 1996).

(12)

was the first to establish a national cultural framework including six cultural dimensions and providing country scores on these dimensions (Beugelsdijk et al., 2017). Inspired by Hofstede, GLOBE developed nine independent dimensions of culture based on the relationship between leadership effectiveness and culture (Magnusson et al., 2008). Not only do these dimensions capture cultural values, but they also capture present practices in a society. Furthermore, both Schwartz and Trompenaars have indicated that culture has a shared set of core norms and values which guide the behaviour of their members (Magnusson et al., 2008). The difference lies in the values that they both indicate as capturing national differences (Magnusson et al., 2008). Schwartz (2006) has indicated seven cultural value orientations summarized into three cultural dimensions. Trompenaars’ work originates from Parsons’ five dimensions (universalism, individualism, neutral-emotional, specific-diffuse, and achievement-ascription), to which Trompenaars added “attitude toward the environment” and “attitude toward time” (Trompenaars, 1996).

Hofstede’s cultural framework still dominates cross-cultural research (Beugelsdijk, Maseland, & Van Hoorn, 2015; Tang & Koveos, 2008). According to previous studies, Hofstede’s cultural dimensions are still the most prominent variables to operationalize national culture (Kang et al., 2016). However, Hofstede’s framework also received criticism concerning the applicability of his framework (Jones, 2007; Magnusson et al., 2008), since, according to the modernization theory, countries will develop economically and this will result in a shift in cultural values (Inglehart & Baker, 2000), and the change of culture over time is not captured in the indices (Kirkman, Lowe, & Gibson, 2006). However, research done by Beugelsdijk et al. (2015) has determined that countries’ cultural scores have not changed much relative to the scores of other countries. Thus, Hofstede’s cultural dimensions can still be used for cross-cultural research.

2.3.1 Hofstede’s Cultural Dimensions

Hofstede (1980a) cultural framework is based on the assumption that human beings around the world are guided by different beliefs, attitudes, and ethical standards. He derived his framework from an employee survey about job satisfaction and attitude towards work done at IBM. According to Hofstede, the definitions of the first four cultural dimensions are as follows:

• Power distance refers to “the degree of inequality among people which the population of a country considers as normal” (Hofstede, 1994, p. 5)

(13)

• Individualism versus collectivism refers to “the degree to which people in a country prefer to act as individuals rather than as members of groups” (Hofstede, 1994, p. 6)

• Masculinity versus femininity (MAS) refers to “the division of emotional roles between women and men” (Hofstede, 2011, p. 5)

Later, two new dimensions were added. To include these dimensions, Hofstede used the World Values Survey (Hofstede, Hofstede, & Minkov, 2010).

• Long-term (LTO) versus short-term orientation (STO) refers to “the choice of focus for people’s efforts: the future or the present and past” (Hofstede, 2011, p. 5)

• Indulgence versus restraint (IVR) refers to “the gratification versus control of basic human desires related to enjoying life” (Hofstede, 2011, p. 5)

In this thesis, the last two added cultural dimensions are neglected for the following reasons: first, Hofstede (2011) has mentioned that the data for LTO and IVR is of limited availability since it does not have the same numerous country cultural scores as the initial four dimensions. Second, in line with prior studies that also neglected these two dimensions (e.g. Peng et al., 2012; Ringov & Zollo, 2007; Thanetsunthorn, 2015), it gives this thesis enough reasons to consider that these dimensions have less empirical importance than the initial four cultural dimensions. Therefore, the scores of power distance, uncertainty avoidance, individualism, and masculinity represent national culture in this thesis.

2.4 Headquarters’ Home Country’s Sustainable Development

(14)

principles of sustainable development in their national programs and policies (Moran, Wackernagel, Kitzes, Goldfinger, & Boutaud, 2008). However, even though the goals have a global approach, the extent to which the goals are achieved depends on the level of priority each country gives them. Consequently, in 2015, the Sustainable Development Goals (SDG) were designed, because the Development Goals are always set for 15 years (Sachs, 2012). Therefore, today these policies and national programs are included in countries’ institutions, which set “the rules of the game” to which firms must adapt to survive (Rodriguez & Perez, 2016).

2.5 CEO Foreignness

In 1984, Hambrick and Mason developed the upper echelons theory (UET). This theory refers to senior management persons such as CEOs. According to the main premise of this theory, each top manager has his or her own values, perspectives, and cognitive biases which influence managerial perceptions, and these likewise lead to strategic decisions (Hambrick, 2007).

In most firms, the CEO is the captain of the ship, making him or her the symbolic and substantive leader of the company. Moreover, the CEO also has the final responsibility and authority for setting and maintaining the company’s strategy, making decisions, distributing information, and allocating resources (Papadakis & Barwise, 2002; Roth, 1995; Thomas & Simerly, 1994). The CEO also sets the standards, ethics, and values of a firm (Godos‐Díez, Cabeza‐García, Fernández‐Gago, & Nieto‐Antolín, 2019). The CEO therefore influences organizational outcomes, because he or she views business situations through his or her personalized lens (Lee, Sun, & Moon, 2018). Frequently, top executives’ observable characteristics and demographics can be used as an indicator of their value-based and cognitive lens (Hambrick & Mason, 1984). One such characteristic is the CEO’s national culture, which is determined by his or her nationality. In a firm, the CEO may share the same or a different nationality than the headquarters location. In line with Thams (2013), a foreign CEO is an individual who was born and has spent (most of) his or her formative1_{years in a country other}

than that in which the firm is headquartered. Since the CEO is in a unique position to influence the firm’s strategic decisions, the CEO’s nationality is important to consider. Thus, CEO foreignness is examined as a moderator.

(15)

2.6 National Culture and CSR Performance

National culture has been recognized as a fundamental determinant of differences among not only individuals but also firms from other cultural backgrounds (Hofstede, 2001). Thus, national culture also affects the behaviour of people within the company and consequently has a recognizable impact on the strategic behaviour firms display (Kreiser, Marino, Dickson, & Weaver, 2010). Considering that CSR activities are seen as a type of strategic choice, they influence CSR implementations (Lee et al., 2018). The behaviour and values of the headquarters’ firm are likely to correspond with national cultural values (Drogendijk & Holm, 2010). In this thesis, it is assumed that the national culture of the country where the headquarters is located is a reasonable indicator of the national culture displayed within the firm since the headquarters plays a pivotal role in the company by formulating its strategy and being responsible for the success of the firm (Ciabuschi et al., 2012; Petruzzella et al., 2017).

2.6.1 Power distance and CSR performance

Hofstede (1980b, p. 45) has referred to power distance as “the extent to which a society accepts the fact that power in institutions and organizations is distributed unequally”. High PDI cultures could be described as countries where subordinates wait to be told what to do, corruption is common, scandals are covered up, and “power is a basic fact of society antedating good or evil: its legitimacy is irrelevant” (Hofstede, 2011, p. 6).

Prior research concerning the relationship between the cultural dimension of PDI and CSR has provided mixed results. Most studies have found a negative relationship between PDI and social and environmental performance (Ringov & Zollo, 2007); CSR engagement (Peng et al., 2012); employee-, community-, and environment-related CSR performance (Thanetsunthorn, 2015). Even though Ho et al. (2012) and Ioannou and Serafeim (2012) both hypothesized the relationship in a negative direction, the results show interesting findings since the relationship demonstrates an opposite effect, namely a significant positive relationship with CSP. An overview of the results of previous scholars on national culture and CSR is given in Table 6, Appendix A.

(16)

autocratic management style (Bochner & Hesketh, 1994) since this autocratic management style is rooted in the mental programming of its members in society (Krokosz-Krynke, 1998). In contrast, low-PDI societies value equality between their group members and stimulate democratic forms of participation (Rinne et al., 2012).

From this perspective, prior research has concluded that individuals from high-PDI cultures are more likely to view a questionable business practice as ethical compared to individuals from low-PDI cultures (Cohen, Pant, & Sharp, 1996). Similarly, high-PDI countries are more likely to behave unethically since superiors do not have to defend or justify their decisions to the larger organization and stakeholders (Khatri, 2009). This also leaves room for corruption, because subordinates remain loyal to their superiors. Scandals are covered up as long as these individuals are in power, which could result in unethical and negative environmental actions (Takyi-Asiedu, 1993). Thus, this supports the argument that high-PDI countries are less likely to participate in CSR activities that influence CSR performance. Based on this, the first hypothesis is as follows:

H1. The higher the power distance of the country in which the company is headquartered, the lower the CSR performance.

2.6.2 Uncertainty avoidance and CSR performance

Hofstede (1980b, p. 45) has referred to uncertainty avoidance as “the extent to which a society feels threatened by uncertain and ambiguous situations”. It is important to mention that UAI should not be confused with risk avoidance. Societies characterized by strong UAI could be described as countries where individuals need clarity and structure, people do not tolerate what is different because it is dangerous, and people believe in the achievement of expertise and absolute truths (Hofstede, 1980b; Hofstede, 2011). Moreover, UAI also concerns how cultures handle an unpredictable future (Rinne et al., 2012).

(17)

Individuals in high-UAI cultures prefer clear structures in their organizations, take fewer risks, and place high importance on keeping everything certain and accountable (Gray, 1988; Hofstede, 1983; Williams & Zinkin, 2008). In these countries, individuals are socialized with the belief that everyone should behave according to the accepted guidelines (Kim, Lee, & Kang, 2018). The behaviour of individuals in organizations in these countries must be perceived as loyal, desirable, and positive to remain part of the organizational “membership” (Blodgett, Lu, Rose, & Vitell, 2001). Thus, individuals in the organization follow the norms and guidelines, because that is assumed to be desirable to the organization’s members. Given the current unknown environmental and social challenges, organizations in high-UAI countries want to structure their operating framework to keep everything certain and accountable (Disli, Ng, & Askari, 2016). To do so, firms in high-UAI countries are more likely to implement a business strategy that could help them to develop a long-term sustainable relationship with their stakeholders (Flammer & Bansal, 2017; Kim et al., 2018). Since CSR is considered a long-term investment (Falck & Heblich, 2007; Yuan, Tian, Lu, & Yu, 2019), CSR could be a way to reduce these unknown environmental and social challenges. This supports the argument that the organizations in countries displaying more UAI engage more in CSR activities than low-UAI countries do. Therefore, the second hypothesis is as follows:

H2. The higher the uncertainty avoidance of the country in which the company is headquartered, the higher the CSR performance.

2.6.3 Individualism and CSR performance

Hofstede (2009, p. 2) refers to individualism versus collectivism as “the degree to which individuals are integrated into groups”. Individualistic countries could be described as “I” consciousness, others are classified as individuals, and tasks prevail over relationships. Collectivistic countries could be described as “we” consciousness, others are classified as in-group or out-in-group, and relationships prevail over tasks (Hofstede, 1980b; Hofstede, 2011).

(18)

immediate family and personal interests rather than group interests (Hofstede, 1980a). In contrast, individuals in collectivistic countries place greater importance on group welfare and their interests rather than themselves, and relationships with stakeholders are more valued in collectivistic countries than in individualistic countries (Waldman, De Luque, et al., 2006).

The rationale of CSR is to “further some social good” by voluntarily contributing to social wellbeing and community (McWilliams & Siegel, 2001, p. 117) since in collectivistic countries (i.e., low-IDV countries) individuals see themselves more as part of a broader community than individualistic countries do. This implies that the positive effect of CSR creates more impact among firms based in collectivistic countries since they are more likely to incorporate the society’s well-being into their strategies and decision-making (Eisingerich & Rubera, 2010) since individuals in these organizations recognize the importance of having responsibility for the community and its welfare (Waldman, De Luque, et al., 2006). Furthermore, Hampden-Turner and Trompenaars (2011) have concluded that in highly individualistic countries, organizations put shareholders ahead of other stakeholders. Organizations do this by focusing on the core value that their primary responsibility is the interest of their shareholders. Therefore, most CSR activities are perceived as the exceeding expenses of the organization, making CSR not a priority. Moreover, research done by Wang and Young (2014) indicates that business managers in collectivistic countries are more concerned about environmental ethics than those in individualistic countries are. In line with this study and the arguments given, it is expected that unless it is in their own personal recognized self-interest, individuals from individualistic countries are likely to be less concerned about the broader impact of their decisions on the community; they are more focused on their personal goals. This results in less interest in CSR activities and therefore negatively influences CSR performance. Following this argument, the third hypothesis is as follows:

H3. The higher the individualism of the country in which the company is headquartered, the lower the CSR performance.

2.6.4 Masculinity and CSR performance

(19)

Previous research done concerning the relationship between masculinity and CSR has also led to contradictory findings. Most of the prior studies found a negative relationship between MAS and CSR engagement (Peng et al., 2012), social and environmental performance (Ringov & Zollo, 2007), and employee-, community-, and environment-related CSR performance (Thanetsunthorn, 2015). On the other hand, Naeem and Khurram (2019) and Ho et al. (2012) have concluded that masculinity has a positive influence on CSR and CSP; however, Ho et al. (2012) hypothesized the relationship in the opposite direction.

According to Burritt and Orij (2010) masculine societies have a lower stakeholder orientation because they are less socially oriented. Similarly, Blodgett et al. (2001) concluded that individuals in masculine societies value their self-interest more compared to those of various stakeholders. Since the values of personal achievement and success are dominant in a masculine society and these individuals care more about these values, they focus less on benefit to their community and stakeholders. Therefore, it is assumed to be unlikely that firms from masculine cultures will engage in CSR. To be more specific, the strive to attain their own objectives will sometimes come at any means and by any costs what might be necessary, only to obtain what they want (Vitell & Festervand, 1987). In addition, individuals in masculine societies tend to focus more on economic responsibilities instead of non-economic responsibilities, such as environmental conservation (Husted, 2005). Therefore, this is assumed to lead to reduced responsiveness to societal and environmental problems, which also indicates that these societies engage less in CSR activities. In contrast, Burritt and Orij (2010) concluded that feminine societies have a high stakeholder orientation, because caring for others such as stakeholders (e.g., employees, the community, etc.) is a prevalent value in these societies. Because they care about the wellbeing of others, organizations from these countries tend to care more about the society and the environment and are thus expected to be more engaged in CSR activities than organisations from masculine societies. Firms in masculine countries are expected to be less likely to engage in practices that encourage the wellbeing of others, such as CSR, which negatively influences companies’ CSR performance. Thus, masculinity is assumed to negatively affect CSR performance. Based on this, the fourth hypothesis is as follows:

(20)

2.7 Moderating Effects on the Relationship Between National Culture and CSR Performance

As previously mentioned, the opposing findings of prior research on national culture and CSR performance indicate that national culture can have varying outcomes in relation to CSR performance. As a result, scholars have begun to explore potential contingencies that could influence the relationship, such as industry type, the degree of globalization (Hawn & Burbano, 2018), and top management commitment (Jian, Jaaffar, Ooi, & Amran, 2017). In this thesis, the headquarters’ HCSD and the CEO’s foreignnessare analyzed to determine whether these contingencies influence the relationship between national culture and CSR performance.

2.7.1 The Moderating Role of The Headquarters’ Home Country’s Sustainable Development

In the last decades, sustainable development has gained increasing urgency in the institutional environment in both rich and poor countries across the world, because of significant environmental problems that are caused by rapid economic and social development. In this thesis, the definition of sustainable development is “to maintain the level of human-wellbeing so that it might improve but at least never declines (or, not more than temporarily)” (Beckerman, 1994, p. 195). By meeting the needs of the present generation, without compromising the ability of generations in the future to meet their personal needs (Emas, 2015).

(21)

2010). As result of the varying institutions, there are differences in the degree of stakeholder pressures on organizations around the world to engage in CSR initiatives (Banerjee et al., 2019).

Brinkerhoff and Goldsmith (1992) have concluded that sustainable development depends on the quality of national institutions. As a result, laws, rules, and regulations deriving from institutions differ in companies with high compared to low levels of sustainable development. For instance, in poor institutional environments, countries’ sustainable development is lower, and this results in lower expectations of stakeholders because the country’s agenda to include stimuli towards sustainable development is lower compared to high-sustainable development countries (Batty, Davoudi, & Layard, 2012; Haleem, Farooq, Boer, & Gimenez, 2015; Salvia et al., 2019). In contrast, in high-quality institutional environments, the sustainable development of countries is higher, and this results in higher stakeholder expectations since sustainable development is higher on the country’s agenda (Dutt, 2009; Salvia et al., 2019). This means that, even though organizations are equally inclined to meet stakeholders’ expectations, the expectations of stakeholders differ per environment. In a certain environment, stakeholders’ expectations are higher, therefore firm needs to make more effort to remain legitimate, whereas in other environments the firm also remains legitimate with less effort since the legitimacy of companies refers to the ability of the firm to meet stakeholders’ expectations (Leyva-de la Hiz, Hurtado-Torres, & Bermúdez-Edo, 2019). Thus, the sustainable development in a country is assumed to differ based on the different national institutions and stakeholder pressures, and since a country’s sustainable development can be thought to affect how firms make strategic decisions, this also differently impacts the relationship between national culture and CSR performance.

Even though it has been proven that a country’s sustainable development does not fully influence the cultural values of firms employees, because cultural values are stable over time (Hofstede, 2011), in certain situational contexts such as a country’s sustainable development, it can strengthen or weaken the influence of cultural values on CSR performance (Meglino & Ravlin, 1998). Therefore, it is important to consider the headquarters’ HCSD.

(22)

unknown social and environmental challenges. In addition, this relationship could be strengthened by the headquarters’ HCSD if it is high. In a headquarters’ home country with high sustainable development, the institutions enforce stricter social and environmental laws and regulations; therefore, since sustainable development is a priority on the country’s agenda, it is mainstream for stakeholders to expect firms to engage in CSR initiatives. In contrast, in low-sustainable development headquarters’ home countries, institutions enforce less strict environmental and social laws and regulations; thus, stakeholder pressure to engage in CSR initiatives is less mainstream. Consequently, in a highly sustainably developed country, organizational members want to meet stakeholder expectations; therefore, the way these members can follow their cultural values is strengthened, because the positive cultural effect on CSR results in the fact that they can act out. However, in a low-sustainable development country, members of organizations feel less of an urge to meet stakeholders’ expectations, because the stakeholders’ pressures to engage in CSR is less mainstream and thus they can afford to hold back; therefore, the way these members can follow their cultural values is weakened. Overall, the positive relationship between uncertainty avoidance and CSR performance is strengthened by a high-sustainable development context. This results in the following hypothesis:

H5a. The higher the sustainable development of the country in which a company is headquartered, the stronger the positive relationship between uncertainty avoidance and CSR performance will be.

(23)

H5b. The higher the sustainable development of the country in which a company is headquartered, the weaker the negative relationship between power distance and CSR performance will be.

H5c. The higher the sustainable development of the country in which a company is headquartered, the weaker the negative relationship between masculinity and CSR performance will be.

H5d. The higher the sustainable development of the country in which a company is headquartered, the weaker the negative relationship between individualism and CSR performance will be.

2.7.2 The Moderating Role of The CEO Foreignness

Already mentioned in section 2.5, is according to the UET, the behavior of firms are a reflection of the experiences and choices that CEOs will make (Wang, Holmes Jr, Oh, & Zhu, 2016). Nowadays, more CEOs do not share the same nationality as the headquarters’ country (Thams, 2013). Companies who want to become more international, seek to hire a non-native CEO (Blonigen & Wooster, 2003). This results in CEOs having a different culture than the headquarter.

Considering that firms differ in their CSR performance, which is assumed to be influenced by the cultural values of employees in the headquarters, the impact of the nationality of the CEO on this relationship is also expected to differ. According to Arp (2013), a person is influenced in his/her formative years by the cultural values of the country situated, such that a country’s national culture is reflected in a person’s nationality. Therefore, it can be said that a CEO is influenced by and reflects his or her native country’s cultural values. Consequently, foreign CEOs do not share the same national culture as most employees in their firm, thus there will be a difference for firms between having a foreign or native CEO. This implies that any foreign CEOs may have a strong national cultural imprint of a nation that is not the headquarters’ home country. Hence, they are likely to have a word view or mindset that is less concerned about organizations’ domestic settings compared to native-born CEOs (Thams, 2013).

(24)

compared to their national counterparts. On the other hand, native CEOs have the same local vision as the headquarters’ home country, thus it is assumed that this CEO will more likely to comprehend the local cultural importance than the foreign-born CEO. Since the CEO is in a position to influence strategic decisions in the organization, and as CSR is seen as a strategic decision (Lee et al., 2018), it is therefore expected that a native CEO influence the firm in such a direction that the way organization members’ are able to follow their cultural values is strengthened. Since the native CEO is more likely to share the same local cultural visions with their organizational members within the headquarters’ office.

In contrast to a native CEO, a foreign CEO will lead the organization in such a direction that the company focus more on global visions than on local visions. This less inclined local cultural importance of foreign CEOs, no matter which countries the CEO is from, can influence the behavior of organizational members (Huang, Cheng, & Chou, 2005). In this regard, a foreign CEO could influence the firm, so that the way organization members’ are able to follow their cultural values is weakened for all the cultural dimensions. This weakening effect will result in the fact that the negative relationships of power distance, individualism and masculinity will be weakened, as the negative relationship will become more negative and the positive relationship of uncertainty avoidance will be weakened, as the positive relationship will become less positive. Thus, the CEO’s foreignness is expected to influence the relationship between national culture and CSR performance. More specifically, it is assumed that the impact any foreign CEO can make, will affect the way the firm can perform on their cultural values. The foreign influence results in a weaker effect, therefore the negative relationships and positive relationship are weakened. This results in the following hypotheses:

H6a. Having a foreign CEO will create a weaker negative relationship between power distance and CSR performance compared to a local CEO.

H6b. Having a foreign CEO will create a weaker negative relationship between individualism and CSR performance compared to a local CEO.

H6c. Having a foreign CEO will create a weaker negative relationship between masculinity and CSR performance compared to a local CEO.

(25)

2.8 Conceptual Model

The conceptual model in Figure 1 is derived from the literature review and hypotheses.

3. RESEARCH METHODOLOGY

In this section, the data collection method and sample are explained. Moreover, the measurement of constructs is discussed. Finally, a statistical analysis is explained.

3.1 Data Collection

To test the presented hypotheses, a quantitative approach was used. Secondary data collected from several databases made available to students from the University of Groningen was used in addition to the publicly available information. Data for the dependent variable CSR performance was retrieved from CSRHub’s website. The data for the independent variable of national culture was collected from the Hofstede Insights website. Moreover, the moderator, the headquarters’ HCSD, was determined by the 2016 EPI. Finally, the CEO’s foreignness was collected from the BoardEx database. Data for the control variables firm size, firm profitability, and firm industry were collected from the Fortune Global 500, CompStat, or Orbis (Table 7, Appendix B).

(26)

3.2 Data Sample

The sample of this thesis focuses on the 2017 Fortune Global 500. This sample was chosen since according to Carroll (2010), CSR applies to all firms regardless of their size. However, the focus is likely to be more on large firms, because they have more clout and are more often in the public eye. This dataset covers the world’s largest public MNEs based on their revenue. The firms are from different countries and industries around the world, which makes the outcomes more generalizable to large firms. Moreover, since these companies are publicly listed, a large amount of information is available.

For such a sample size, Hansen, Andersen, and Parner (2014) have proposed using 10 observations per independent variable. Therefore, the ideal number of observations should at least be more than 90. However, to minimize the risk of ending up with less than 90 observations due to missing data, more firms were added. Thus, the sample size is 165 firms from 23 countries (Table 8, Appendix C) and eight major categorized industry groups (Table 9-10, Appendix C). These 165 observations are randomly sampled firms that were not excluded2_{from the initial sample. Only 66 observations were already in the sample since they}

have a foreign CEO. The random sampling was done in Excel using the formula (=RAND). This formula provides an evenly distributed random real number between 0 and 1 (Microsoft, 2020). Subsequently, these values were sorted by ‘random’, resulting in a random dataset where the first 99 observations were chosen, on top of the fixed 66 observations.

3.3 Measurement of Variables 3.3.1 Dependent Variable

The dependent variable is the firm’s CSR performance. Over the years, many methodological tools have been used to measure CSR performance. The KLD database, GRI index, Fortune reputation index, and CSRHub index are broadly recognized and used CSR indexes and databases. However, most of these measurement tools have some disadvantages. The KLD database is one of the most widely accepted and used measurements of CSR performance (McGuire, Dow, & Ibrahim, 2012) and includes the social records of more than 3,000 US companies including seven indicators (Manner, 2010). One major weakness is that the information is mainly from American organizations. Since this thesis compares countries, this database is not suitable (Costantini & Mazzanti, 2012). Furthermore, the GRI index

(27)

recommends guidelines for organizations’ corporate sustainability reporting (Charitoudi, Giannarakis, & Lazarides, 2011). Nevertheless, criticism of this index indicates that companies could manipulate their performances to appear more transparent by overemphasizing their real performance (Moneva, Archel, & Correa, 2006). Fortune reputation index is based on a survey of approximately 5,000 CEOs, senior executives, and financial analysts. It is also this questionnaire survey element that is a major disadvantage. According to Wood (2010, p. 74), the Fortune rating is “itself a fatal blow to using the measure in any objective sense. It seems the same as asking the foxes how well they take care of things down at the henhouse”. Besides the disadvantages mentioned above, these measurement tools also have specific advantages. However, Wolfe and Aupperle (1991) have argued that there is no single best way to measure CSR performance; therefore, this thesis uses the CSRHub index. This index provides advantages to address some of the disadvantages mentioned above.

First, the CSRHub index covers more than 17,268 organizations from 143 countries in 134 industries (CSRHub, 2020a). Second, it covers information from more than 618 data sources, including ESG, ASSET4, and KLD, and this database also covers information from NGOs. The CSRHub propriety system aims to remove inconsistency and bias by mapping onto a central schema. Thus, regarding a large number of data sources and observations as well as the way the data is adjusted and processed, this index is a representative measurement tool and contains the best accessible information that fits into the main goal of this thesis (Thanetsunthorn, 2015). CSRHub’s schema contains four key categories:

• Community (e.g., community development and human rights, etc.)

• Employees (e.g., compensation and benefits, diversity and labour rights, etc.) • Environment (energy and climate change, resource management, etc.) • Governance (board, transparency, reporting, etc.)

The CSRHub score is based on a rating scale from 0 (extremely negative) to 100 (extremely positive In this thesis, no category has been given greater priority than the others.

Time lags

(28)

year; therefore, 2019 CSR performance scores of firms are used. Furthermore, according to this logic, the moderator headquarters’ HCSD is assumed to influence the organization at the time when the management makes CSR decisions. Thus, the 2016 EPI is used. The 2017 ranking reported in 2018 cannot be used since, for example, in February firms do not know what their sustainable development will be in December 2017; however, they know the sustainable development from 2016. For the control variables, the same logic holds for firm profitability, meaning that data from December 2016 was used. For firm size, the 2017 data from Fortune Global 500 was used, because managers know at the moment of making decisions what the number of employees is. If this data was not available, the latest accessible data was included. Finally, the cultural values and the control variable – the firm industry – do not change over time.

3.3.2 Independent Variables

To measure the independent variable – the national culture of the headquarters – the country of the firm’s headquarters first had to be identified. This data was retrieved from BoardEx and double-checked with the information available on the Fortune Global 500 website. Afterwards, the four cultural dimensions were collected from Hofstede Insights to match the national culture score with the headquarters’ location to create four variables that represent the cultural dimensions of Hofstede. The cultural scores range from 1 to 100, and the higher the score is, the more the cultural dimension is presented in that country.

3.3.3 Moderating Variables

3.3.3.1 Headquarters’ home country’s sustainable development

(29)

ecosystem (e.g., climate, energy, and forests) (Environmental Performance Index, 2016). Moreover, air quality is an example of a threat to public health. In 2016, the Institute for Health Metrics and Evaluation estimated “that diseases related to airborne pollutants contributed to two-thirds of all life-years lost to environmentally related deaths and disabilities” (Environmental Performance Index, 2016, p. 2). This index reflects the priorities that countries have concerning these human and natural environmental issues. The EPI scale ranges from 0 (lowest sustainable development) to 100 (highest sustainable development).

3.3.3.2 CEO foreignness

For CEO foreignness, each CEO’s nationality was collected from director profile reports on the BoardEx website. The data concerning the nationality of CEOs was only available from April 2017 for the UK, Europe, and North America and from October 2016 for the rest of the world (ROW). Since these two datasets have a difference of six months, the researcher manually examined each firm’s website and annual reports to determine whether the same CEO as in 2016 was still in place for the ROW. A categorical variable was created: 0 for a local CEO and 1 for a foreign CEO. To obtain a representative sample to test a potential moderating effect, 66 firms have a foreign CEO and 99 firms have a local CEO. Even though a 50-50 distribution creates the most generalizability (Dodd & Epstein, 2012), it should also be considered that the larger the sample size, the more statistical power the dataset has (Suresh & Chandrashekara, 2012). Therefore, 60% foreign CEOs and 40% non-foreign CEOs was chosen as the distribution, because the number of foreign CEOs in the Fortune Global 500 will not change.

3.3.4 Control Variables

Three control variables were included to ensure that some factors previously found to affect CSR performance did not influence the proposed hypotheses.

Firm size

(30)

to engage in CSR activities. In this thesis, firm size is measured in numbers of employees gathered from the 2017 Fortune Global 500 (e.g. Baumann-Pauly, Wickert, Spence, & Scherer, 2013; Ioana-Maria & Adriana, 2012). The number of employees is either the yearly average number or the number published by the company at the end of a fiscal year (Fortune, 2017).

Firm profitability

The economic profitability of a firm could be considered an indicator of the organization’s financial performance (Godos‐Díez et al., 2019). In the past, both return on assets (ROA) and return on equity (ROE) have been commonly used financial performance measures (Griffin & Mahon, 1997; Waddock & Graves, 1997). In this study, the ROE of companies is used because it is expected that firms make CSR investments based on how profitable the firm is, and this ratio is frequently used by shareholders to measure a firm’s success (Hidayat, 2017). ROE is calculated as a ratio by dividing net income by the amount of shareholder equity. According to Campbell (2007), firms with greater profitability are more likely to implement their CSR activities and strategies more aggressively since the organization has greater slack resources available. Furthermore, Muttakin, Khan, and Subramaniam (2015) found that more profitable firms disclose more environmental information since they have larger resources available to engage in CSR and are more motivated to disclose CSR, because they want to favourably distinguish themselves. Regarding the time-lag, this control variable was taken from the end of the 2016 fiscal year since companies incorporate prior firm profitability while making strategic decisions, including CSR decisions (McGuire, Sundgren, & Schneeweis, 1988).

Firm industry

(31)

et al. (2016) have categorized sensitive industries by the following SIC codes: 800–899 (forestry), 1000–1099 (metal mining), 1200–1399 (coal mining and oil and gas extraction), 2600–2699 (paper and allied products), 2800–3099 (chemicals and allied products, etc.), 3300– 3399 (primary metal industries), and 4900–4999 (electric, gas, and sanitary services) (Tables 8 & 9, Appendix B).

3.4 Data Analysis

To test the proposed hypotheses, IBM SPSS Statistics version 24 was used. Since this thesis examines multiple independent variables to explain the continuous dependent variable CSR performance, a multiple linear regression analysis based on the ordinary least square (OLS) method was appropriate to use (Slinker & Glantz, 2008), because this method is usually used in national culture and CSR studies (Giannarakis, 2014; Ringov & Zollo, 2007; Thanetsunthorn, 2015). Moreover, considering the moderators in this study, a moderated multiple regression analysis was used (Dawson & Richter, 2006). Therefore, the independent variables and the continuous moderator are mean-centred for this analysis, as well as the computing of interaction terms to test the moderating effect. This thesis uses a hierarchical regression. Before running the multiple regression analysis, the preliminary requirements for the analysis were examined to guarantee an adequate interpretation of the results since problems in the dataset could be corrected before the main analysis (Fidell & Tabachnick, 2003).

3.5 Robustness Test

(32)

2000, p. 881). In the past years, this method has been represented as a promising new data method, especially since traditional statistical methods have been violated. Furthermore, a second robustness test was performed to test if the overrepresentation of the USA, the UK, and Japan, which account for 48.2% of the companies in this sample, biases the results.

4. EMPIRICAL RESULTS

In this section, the outcomes of the preliminary tests are outlined, followed by the empirical results of the moderated multiple regression analysis.

4.1 Preliminary Requirement Tests for the Moderated Multiple Regression Analysis 4.1.1 Outliers

To test for potential outliers, Cook’s distance was used. The precondition is that the created corresponding variable is > 1 (Dhakal, 2017). This test is illustrated in a scatter plot. Two potential outliers were identified and excluded from the dataset (Figure 4, Appendix D).

4.1.2 Normality

To test for normality, a P-P plot and histogram were made (Figure E.1-E.6, Appendix E). The data in the P-P plot is close to the normality line, assuming that the data is normally distributed. Moreover, the histogram confirms a relatively normal distribution. Concerning the skewness and kurtosis, the histogram is a bit negatively skewed, and distributions are a bit peaked.

4.1.3 Homoskedasticity

For this assumption, a scatter plot was made. Models 8 and 10 include the interaction terms (KEAN University, 2004). Figure F.1-3 in Appendix F shows that the distributions of the scatter plots are equal. To confirm this, the Breusch-Pagan test was conducted, because this is the most common statistical test of homoskedasticity (Table 11, Appendix F)(Shukur, 2002). These results also indicate homoskedasticity, because no value is significant (p ≤ 0.05).

4.1.4 Linearity

(33)

residuals and the standardized predicted value is also a method for detecting linearity. Additionally, this residual plot shows no sign of non-linearity (Figure F.1-3, Appendix F).

4.1.5 Multicollinearity

For this assumption, the variance inflation factor (VIF) was tested for all the main variables since it tests the degree of multicollinearity of the independent variable with other independent variables (O’brien, 2007). Since all the VIF values are below the threshold of 10 and above the tolerance level of 0.1, there is no multicollinearity (Table 12, Appendix G) (O’brien, 2007).

4.1.6 Independence of residual values

This assumption is tested with the Durbin-Watson tests and checks for autocorrelations. Since these results are close to a value of 2, they are between the accepted range of 1.5 and 2.5 (StatisticsSolutions, 2020). Therefore, this assumption is met (Table 13, Appendix H).

4.2 Descriptive Statistics

(34)

4.3 Correlation

The correlation matrix is presented in Table 2. In this table, power distance, individualism, headquarters’ HCSD, CEO foreignness, and the control variables firm size and firm industry correlate significantly with CSR performance. This indicates that an increase in one variable results in a similar increase in the other variable. Only for power distance due to the negative correlation is an increase accompanied by a decrease in CSR performance. A strong correlation can be observed between power distance and individualism. However, this is consistent with Hofstede’s own remark that “lower power distance are more often individualist and countries with higher power distance are more often collectivist” (Hofstede, 2014 min. 10:00). Since this is the effect of wealth, taking that effect out means that it almost disappears (Husted, 2005). Because Hofstede treats PDI and IDV as separate dimensions, this thesis does so as well. Moreover, headquarters’ HCSD correlates strongly with individualism and power distance. Some scholars have argued that a correlation with a value > 0.5 could be problematic (Cohen et al., 1996), while others have argued that a correlation higher than 0.9 is problematic (Tabachnick, Fidell, & Ullman, 2007). Even though these three correlations might be interpreted as high, the multicollinearity test confirms that these factors do not cause any violation (Table 11, Appendix F). Furthermore, headquarters’ HCSD correlates significantly with CSR performance and power distance and individualism, potentially affecting these variables. Moreover, the same dimensions (PDI and IDV) plus uncertainty avoidance also correlate significantly with the second moderator: CEO foreignness. Thus, this variable could also indicate an influencing effect on these dimensions.

(35)

(36)

4.4 Multiple Regression Analysis

After testing the six assumptions of the preliminary analysis, which were all met, and explaining the descriptive statistics and correlations, the main analysis was conducted, the results of which are presented in Table 3. As this table illustrates, 10 different models were used for this regression analysis, whereby a hierarchical regression was used by adding each predictor one by one in steps. This was done because hierarchical regression is an appropriate tool to use for analysis when the study has correlated variables, as shown in Table 2 (Lewis, 2007).

In the first model, the effects of the control variables on the dependent variable were tested, and firm size (p = .009) and firm industry (p = .005) both had significant explanatory power regarding CSR performance. These three control variables have an R² of 10.7%. This indicates how much variance in CSR performance is explained by the independent variables – in this case, the three control variables (Miles, 2014).

To test Hypothesis 1, the independent variable power distance was added to the control variables separately. The results in Model 2 indicate that it has a negative effect (B = -0.138). This means that one unit change in this independent variable results in a negative change of -0.138 in CSR performance. Moreover, Model 6 (B = -0.140) also shows a significant level of p ≤ 0.01. Therefore, both models support Hypothesis 1 and can be accepted.

In Model 3, the independent variable uncertainty avoidance was tested separately from the control variables to test Hypothesis 2. The results show no significance (B = 0.028; p = 0.240); however, the results in the general regression on Model 6 display a positive significant effect (B = 0.058; p = 0.013). While the dimension is significant in the general regression, it is not significant on its own; therefore, the effect is unstable and Hypothesis 2 is not supported.

Model 4 assesses the impact of individualism on CSR performance. Contrary to the hypothesized effect, individualism seems to have a positive significant effect on the dependent variable (B = 0.071; p = 0.000). However, Model 6 did not confirm the same result; it produced a positive insignificant outcome (B = 0.014; p = 0.644). Even though this variable has strong statistical significance, it lost its effect in Model 6 when the other predictors were added. Thus, this variable is also unstable, resulting in no support for Hypothesis 3.

(37)

Moreover, the R² in Model 6 compared to Model 1 increased to 24.4%. Therefore, the four independent variables add significant influence to the explanation in the variation.

Next, Model 8 showed that the moderating variable headquarters’ HCSD has a positive significant effect on CSR performance (B = 0.339, p = 0.002), and this was also confirmed in Model 10 when the interaction terms were added (B = 0.993, p = 0.000). The negative relationships between this moderator and power distance and individualism are clear; however, only power distance is significant (B = -0.023; p = 0.004). Further, the interaction with masculinity (B = 0.013, p = 0.084) is positive and very small but significant. Nonetheless, the relationship was hypothesized to be negative. Moreover, the interaction term uncertainty avoidance (B = 0.009, p = 0.034) also influences CSR performance. Thus, Hypotheses 5c and d are rejected and a and b supported. Since this moderator is continuous, no graphs are made to plot the interaction terms. Because according to Grace-Martin (2020) the graphs would be a mess, therefore it would be difficult to see any pattern.

(38)

Figure 3: Interaction term CEO foreignness with masculinity

(39)

(40)

4.5 Robustness Tests 4.5.1 Bootstrapping

Bootstrapping was performed in SPSS, resulting in an automatic robustness test. The preference for bootstrapping has been explained (Section 3.5). A “different” sample was created since simulating the initial dataset results in some observations appearing once, some twice, etc., and some not at all (Pardoe, 2000). This thesis has a bootstrap sample of 2,000, because according to Gignac (2019, min 1:20), this is necessary for multiple regression analysis. The bootstrap has a confidence interval of 95% and bias-corrected accelerated was applied, leading compared to percentile interval to a more general bootstrap confidence interval (Stine, 1989). The typical “standard error” is studied in this method; if it becomes smaller, this indicates that the mean in the sample is a better reflection of the actual population mean (Lee, In, & Lee, 2015).

National Culture and CSR: The Influence of National Culture on CSR Performance and the Moderating Effects of Headquarters’ Home Country Sustainable Development and CEO Foreignness