National Culture and CSR: The Influence of National Culture on CSR and the Moderating Effect of State of Development

National Culture and CSR: The Influence of National

Culture on CSR and the Moderating Effect of State of

Development

MSc International Business & Management

University of Groningen, Faculty of Economics and Business Date of Submission: January 20th_{, 2019}

Abstract

The relationship between national culture and CSR has received increased attention over the past few decades. However, to date, the results of this relationship are inconclusive as they show positive, negative, and non-significant results. Moreover, the results are outdated as only the first four dimensions of the Hofstede framework have been researched thoroughly, while the two newest dimensions have received very little attention with regards to their relationship to CSR. The key objective of this study is to contribute to the debate by testing the influence of power distance, uncertainty avoidance, individualism, masculinity, long-term orientation, and indulgence on CSR. This is done by analyzing 215 companies from 23 different countries, which are on the Forbes 2000 list. Furthermore, the moderating role of state of development is assessed in this context, by using the distinction made in the MSCI ACWI Index. By conducting a moderated multiple regression analysis, mixed results were found on the influence of national culture on CSR, which could be ascribed to a difference in sample. Moreover, the influence of state of development was not significant in the main regression, however, it gave interesting opportunities for future research. The findings of this study thereby add to the body of knowledge on national culture, CSR, and state of development, and provide several theoretical and managerial implications.

Acknowledgments

Table of Contents

2.1 Corporate Social Performance ... 9

2.2 National Culture ... 11 2.2.1 Power Distance ... 13 2.2.2 Uncertainty Avoidance ... 13 2.2.3 Individualism ... 15 2.2.4 Masculinity ... 15 2.2.5 Long-term Orientation ... 16 2.2.6 Indulgence ... 17 2.3 State of Development ... 19 2.4 Conceptual Model ... 21 3. Research Methodology ... 21 3.1 Data Collection ... 21 3.2 Sample... 22 3.3 Measurement of Variables ... 23 3.3.1 Dependent Variable ... 23 3.3.2 Independent Variable ... 24 3.3.3 Moderating Variable ... 25 3.3.4 Control Variables ... 26 3.4 Data Analysis ... 26 3.4.1 Robustness Test ... 27 4. Results ... 27

4.1 Test of Basic Assumptions ... 27

4.2 Descriptive Statistics ... 28

4.3 Correlations ... 29

4.4 Regression Results ... 31

5. Discussion ... 36

6. Conclusion ... 40

6.1 Theoretical Implications ... 40

6.2 Practical Implications... 41

6.3 Limitations and Future Research ... 42

References ... 44

Appendices ... 49

Appendix A: Fragment of companies in India over the years 2013 - 2018 ... 49

Appendix B: Industry Types ... 49

Appendix C: P-Plot to test for Normality ... 50

Appendix D: A scatterplot to test for Homoscedasticity ... 51

Appendix E: The VIF values to check for Multicollinearity ... 52

Appendix F: Correlation Matrix 2016 ... 53

Appendix G: Regression Analysis DV + EM 2016 ... 54

Appendix H: Regression Analysis DV 2017 ... 55

Appendix I: Regression Analysis EM 2017... 56

List of figures

Figure 1: Conceptual Model Figure 2: MSCI ACWI Index

Figure 3: Hofstede scores for Brazil, China, the Netherlands, and Sweden.

List of tables

Table 1: Descriptive Statistics. Table 2: Correlation Matrix 2017

Table 3: Regression Analysis DV + EM 2017 Table 4: Overview Regression Analyses

List of abbreviations

ACWI All Country World Index CSP Corporate Social Performance CSR Corporate Social Responsibility DS Descriptive Statistics

DV Developed Market

EM Emerging Market

ESG Environmental, Social, and Governance

GLOBE Global Leadership and Organizational Behavior Effectiveness

HQ Head Quarters

MNE Multi National Enterprise

MSCI Morgan Stanley Capital International SOD State of Development

1. Introduction

Over the last few decades, there has been a growing body of research emphasizing the importance of Corporate Social Responsibility (CSR hereafter) (Carroll, 1999; Matten & Moon, 2008; McWilliams & Siegel, 2001; Wood, 1991). Many have researched whether CSR has an influence on for instance consumer behavior (Becker-Olsen, Cudmore & Hill, 2006), competitive advantage (Porter & Kramer, 2006), or financial performance (Ruf, Muralidhar, Brown, Janney & Paul, 2001). However, a topic that has been somewhat missing from this body of research is national culture. National culture in this paper is defined as the influence of the environment on an individual’s behavior, which distinguishes people in that environment from other environments. Only a few researchers have looked at the relationship between national culture and CSR, and their results are contradicting.

Ringov & Zollo (2007) used Hofstede’s (2011) dimensions, and looked at countries with a high level of power distance, individualism, masculinity, and uncertainty avoidance. What they found was that power distance and masculinity have a negative effect, while individualism and uncertainty avoidance have no significant effect. Peng, Dashdeleg & Chih (2012) looked at the same Hofstede (2011) dimensions and indeed found that power distance and masculinity have negative influences, however, they also found that individualism and uncertainty avoidance have positive influences on CSR commitment. Ho, Wang & Vitell (2012) also made use of these dimensions and found that Hofstede’s dimensions are significantly associated with CSR, but they also noted that European companies outperform companies in other regions and countries with regards to CSR. As can be seen from these results, there is still some contradiction with regards to the influence of national culture on CSR.

Moreover, this paper will research another aspect, namely whether the state of development of a country influences the relationship between national culture and CSR. As Peter, Muller & Kusyk (2011) note, there are certain ways of doing business that work in developed countries that may not work in emerging markets. Reasons for this might for instance be differences in political risk (Henisz & Zelner, 2010) or infrastructure (Khanna, Palepu & Sinha, 2005). Moreover, as Jamali & Karam (2018) state, there is a “parallel universe” in literature in developed markets vs. emerging markets (p. 50). These differences might also explain why the aforementioned results were contradicting. The reason why, for instance, individualism and uncertainty avoidance have no significant effect in some papers, while in others they have a positive effect.

Therefore, will this paper not only look at the relationship between national culture and CSR, but also at how the state of development of a country (either emerging or developed) influences it. The research question is as follows:

Does national culture influence MNE’s CSR and is this relationship influenced by the MNE’s home country’s state of development?

This question will be answered firstly, by looking at national culture with the six Hofstede dimensions (power distance, individualism, masculinity, uncertainty avoidance, long-term orientation, and indulgence). Secondly, as CSR is an abstract concept, it will be measured using CSP, because this results in measurable scores per MNE. CSP is defined as the actions and decisions taken by companies to not only be economic and legal, but to also voluntarily be socially responsible and responsive in order to create higher living standards for its stakeholders. CSP will then be tested on economic, social, and environmental indicators (ESG scores), which will be accessed through the Thomson Reuters ESG database. And finally, the countries will be compared by the distinction made in the Morgan Stanley Capital International (MSCI) ACWI Index, to find out whether state of development influences the relationship between national culture and CSP. This data will then be used to run a regression analysis, in order to find an answer to the research question.

dimensions. Furthermore, by looking at the influence of state of development on this relationship, knowledge will be gathered on the different states of economic development, and how this might influence the results. Moreover, this information will also be useful for companies. With more information about the influence of national culture and the state of development, MNEs will have better insight into how this would affect their CSR policies, their decision-making process, and the signals they send to stakeholders.

The remainder of this paper is structured in the following way. First, the literature review on the concepts national culture, CSP, state of development, and the connections between them will be presented. From this section the hypotheses and conceptual model will follow. Then, the methodology section will explain how these hypotheses will be gathered and tested using a regression analysis, which will be followed by the data section that interprets the results. After this the discussion, conclusion, limitations, and suggestions for future research will round up the paper.

2. Literature Review

This section will explain the core concepts of the research question before outlining the hypothesized relationships between them. First, CSP and its indicators will be defined, after which the concept of national culture and its six dimensions will be explained, and finally, the moderator, state of development, and its influence on this relationship will be clarified. Throughout this section, the hypotheses will be formed and at the end, a conceptual model will be presented to show the expected relationships.

2.1 Corporate Social Performance

interchangeably and defined similarly in literature (Orlitzky, Schmidt & Rynes, 2003). In the remainder of this paper the term CSP will be used, because CSP results in measurable scores per MNE and thus provides the information needed to answer the research question. Firstly, CSP will be defined by looking at different definitions of researchers who have looked into the concept previously, before coming to a definition that will be used throughout the rest of this paper. After this, national culture and its six dimensions will be defined.

One of the first researchers to look at the concept of CSP is Keith Davis (1960). He defines it as “decisions and actions taken for reasons at least partially beyond the firm’s direct economic and technical interests” (Davis, 1960, p. 70). Another researcher who looked into the concept is Carroll (1979), who states that CSP is an inclusive and global concept that embraces “corporate social responsibility, responsiveness, and the entire spectrum of socially beneficial activities of business” (Carroll, 1991, p. 40). Moreover, he states that it includes not only economic and legal responsibilities, but also ethical and discretionary responsibilities (Carroll, 1979). Wartick & Cochran (1985) defined CSP as “the underlying interaction among the principles of social responsibility, the process of social responsiveness, and the policies developed to address social issues” (p. 758). They also showed how several competing perspectives could be incorporated into a CSP framework (economic responsibility, public responsibility, and social responsiveness) (Watrick & Cochran, 1985). Wood (1991) extends this last definition, and defines CSP as “a business organization's configuration of principles of social responsibility, processes of social responsiveness, and policies, programs, and observable outcomes as they relate to the firm's societal relationships” (Wood, 1991, p. 693).

(Bagnoli & Watts. 2017; Gamerschlag, Möller & Verbeeten. 2011; Jain, Keneley & Thomson, 2015; Rodríguez & LeMaster, 2007). Based on these papers, the definition for CSP that will be used in the rest of this paper is the following: CSP is the actions and decisions taken by companies to not only be economic and legal, but to also voluntarily be socially responsible and responsive in order to create higher living standards for its stakeholders.

As CSP is quite a broad concept, researchers often make subcategories in order to measure it. For instance, Ho et al. (2012) used four categories; social issues related to the environment, strategic governance, labor relations, and stakeholder management. Ringov & Zollo (2007) measured social/environmental performance, and Muller & Kolk (2009) measured it in three dimensions; environment, labor, and community. In this research CSP will be measured by using ESG scores, which consist of three subcategories; Environmental, Social, and Governance. This is done because it is believed that these subcategories give insight into the broad concept of CSP. More information on how ESG will be measured and used will follow in the methodology section. Hereafter national culture and its six dimensions will be defined, before elaborating on the moderator state of development.

2.2 National Culture

National culture over the years has been defined by many different researchers in many different ways but, as Venaik & Brewer (2008) state, there isn’t a general consensus on one definition. Therefore, this section will start by looking into several definitions of national culture to come to a comprehensive one that will be used throughout the rest of this paper. After this, the separate dimensions of national culture will be described.

share certain cultural traits that make them recognizable as belonging to that society (Hofstede, 2001). Schwartz (2008) conceptualizes national culture by stating that it “influences the distribution of individual beliefs, actions, goals, and styles of thinking through the press and expectations to which people are exposed.” (p. 5). He states that it is something outside the individual, instead of in the minds and actions of individual people, and that it refers to what people are exposed to by living in a specific social system (Schwartz, 2008).

Even though these definitions slightly differ, they have a common ground in that they all mention the environment/social system, and the influences on an individual’s behavior, which distinguish it from other environments/social systems. Therefore, the definition of national culture used in this paper will be the following: national culture is the influence of the environment on an individual’s behavior, which distinguishes people in that environment from other environments.

2.2.1 Power Distance

Hofstede (2011) defines the first dimension, power distance, as “the extent to which the less powerful members of organizations and institutions (like the family) accept and expect that power is distributed unequally” (p. 9). What should be noted about this definition is that it states that inequality is endorsed by both followers and leaders; he formulates it from below, not from above (Hofstede, 2011). Examples of this dimension are that in countries with low power distance, parents treat children as equals, education is student-centered, subordinates expect to be consulted, corruption is rare, and income distribution is quite even. In countries with high power distance, contrarily, parents teach children obedience, education is teacher-centered, subordinates expect to be told what to do, corruption is frequent, and income distribution is very uneven (Hofstede, 2011, p. 9).

There have been a few researchers who have looked into the relationship between national culture and CSP. For instance, Ho et al. (2012) looked at the original four Hofstede dimensions, and found that a higher CSP is significantly associated with cultures characterizes by higher power distance. Ringov & Zollo (2007), on the other hand, found that power distance has a negative effect on CSP. Peng et al. (2012) agree with the findings of the latter, and conclude that power distance has a negative effect. Thanetsunthorn (2015) also agrees with these findings, as her research also found a significant negative effect of power distance on employee, community, and environment related CSP.

As CSP is focused on creating higher living standards for a company’s stakeholders by being social and ethical, it is expected that having a high degree of power distance has a negative influence on it. This is because in cultures with a high-power distance, corruption is frequent and income distribution is uneven. The less powerful members of organizations also accept and expect that power is distributed unequally. Therefore, it is believed that a high power distance does not create higher living standards for all its stakeholders, and thus the following hypothesis is formed:

H1: A higher degree of power distance will have a negative influence on CSP.

2.2.2 Uncertainty Avoidance

situations” (p. 10). It is about how a society tolerates ambiguity. Unstructured situations as described in this definition are situations that are new, unknown, surprising, or different from usual. Cultures that try to minimize these situations have, for instance, strict laws and rules, disapproval of opinions that are different, and behavioral codes (Hofstede, 2011). Cultures with high uncertainty avoidance have an intolerance of deviant people and ideas, teachers should have all the answers, an emotional need for rules, and a need for clarity and structure. On the other hand, cultures with low uncertainty avoidance have a tolerance of deviant people and ideas, teachers are allowed to say that they don’t know the answer, there is a dislike of rules, and they are comfortable with ambiguity and chaos (Hofstede, 2011, p. 10).

Uncertainty avoidance is one of the dimensions previously examined by researchers, however, their findings have been slightly contradicting. Ringov & Zollo (2007) did not find any significant results between uncertainty avoidance and CSP. Peng et al. (2012), on the other hand, found positive influences between uncertainty avoidance and CSR commitment. Thanetsunthorn‘s (2015) findings agree with the latter, as she found that cultures with high uncertainty avoidance have “a tendency to exhibit high performance on community and environment” (p. 51). Ho et al. (2012) also agree with this, as they found that cultures with higher uncertainty avoidance have a higher CSP.

The aim of CSP is to create higher living standards for the company’s stakeholders, while also being economic and legal, keeping the organization profitable. As there are so many people dependent on the survival of the organization, it is expected that a high degree of uncertainty avoidance will lead to higher CSP. This is, because in cultures with high uncertainty avoidance, they have strict laws, rules, and behavioral codes. These codes provide guidelines for the organization, and people might be expected to behave socially responsible. They thus not only lead the organization to profitability, but they will also give insight into how to act socially responsible, and what is expected from stakeholders. Therefore, the following hypothesis is formed:

2.2.3 Individualism

The third dimension has two opposites; individualism and collectivism. This dimension is defined as “the degree to which people in a society are integrated into groups” (Hofstede, 2011, p. 11). In individualistic cultures, the ties between individuals are loose and everyone only looks after him/herself and their immediate family. In collectivistic cultures people are integrated from birth on into strong in-groups, which will always protect them in exchange for loyalty (Hofstede, 2011). Examples of characteristics of an individual culture are that a personal opinion is expected, others are classified as individuals, there is a right of privacy, and tasks triumph relationships. Characteristics of a collectivistic culture on the other hand are for example that opinions and votes are predetermined by the in-group, others are classified as in-group or out-group, they stress belonging, and relationships triumph tasks (Hofstede, 2011).

The relationship between individualism and CSP has so far resulted in contradicting findings. Ringov & Zollo (2007) found that individualism has no significant effect, while Peng et al. (2012) found that individualism has a positive effect on CSR commitment. Thanetsunthorn (2015) on the contrary, found that there was less concern for issues related to employees, communities, and environments in cultures with high degrees of individualism, thereby indicating a negative relationship. And lastly, Ho et al. (2012) found that cultures characterized by collectivism have a higher CSP, so there is a negative relationship between individualism and CSP.

In line with the findings of Thanetsunthorn (2015) and Ho et al. (2012), it is expected in this paper that a high degree of individualism has a negative effect on CSP. This is because in cultures with high individualism, people only look out for themselves and their immediate family. They thus have less regard for the community and the environment. Based on this reasoning the following hypothesis is formed:

H3: A higher degree of individualism will have a negative influence on CSP.

2.2.4 Masculinity

modest and caring, while masculine cultures are more competitive and assertive. Some example characteristics of feminine cultures are sympathy for the weak, there is a balance between work and family, both mothers and fathers deal with facts and emotions, and many women are elected for political positions. Examples of characteristics of masculine cultures are admiration for the strong, work prevails over family, fathers deal with facts, mothers deal with feelings, and only few women are elected for political positions (Hofstede, 2011, p. 12).

As mentioned before, there have been a few papers that looked into the relationship between national culture and CSP. One of these papers is by Esteban, Villardón & Sanchez (2017), who found that companies located in feminine cultures “are characterized by the pursuit of the common good and place greater emphasis on social and environmental issues” (p. 18). Thus, what they found was that companies in feminist cultures have better CSR practices. Ringov and Zollo (2007) and Peng et al. (2012) looked into the same dimensions, and found that masculinity has a negative influence on CSR commitment. In contrast, Ho et al. (2012) found that in more masculine cultures there is higher CSP. And finally, Thanetsunthorn (2015) looked at the impact of national culture on CSP across geographic regions, and found that there are lower levels of environment related CSP and community related CSP in cultures with a high degree of masculinity.

In this paper, it is expected that a high degree of masculinity has a negative influence on CSP. This is because, while masculine cultures are more competitive and assertive, feminine cultures are more modest and caring. They also have more sympathy for the weak. Therefore, it is expected that more feminine cultures care more about improving the living standards for the company’s stakeholders and thus have a higher CSP. Based on this reasoning the following hypothesis is formed:

H4: A higher degree of masculinity will have a negative influence on CSP.

2.2.5 Long-term Orientation

and fulfilling social obligations” (Hofstede, 2001, p. 359). These differences can be seen in that long-term oriented cultures say that good and evil depends on circumstances, students attribute success to effort and failure to lack of effort, and traditions are adaptable to changing circumstances. Short-term oriented cultures, on the other hand, have universal guidelines about what is good and bad, students attribute success and failure to luck, and traditions are sacred (Hofstede, 2011).

As this is one of the newest dimensions of the Hofstede (2011) framework, it has not been researched as extensively yet. Therefore, a few papers that have researched the overall area of long-term orientation and CSP are examined, in order to form expectations about the relationship. Bae & Kim (2013), for instance, state that people with a long-term orientation are inclined to support the importance of legal and ethical responsibility “because they put a great emphasis on following and respecting social norms and expect for-profit companies to do the same” (p. 74). They thus state that people with a high long-term orientation are more likely to prefer companies with high CSP. Wang & Bansal (2012) looked at it from a company perspective. They found that new companies, who have a long-term orientation, are able to counteract their liability of foreignness and generate positive economic returns when acting socially responsible. Thus, their long-term orientation is of positive influence on their socially responsible activities (Wang & Bansal, 2012).

While these researchers have looked into the area of long-term orientation and CSP, no clear answer to the research question is given yet. Therefore, it is important to test this relationship. In this paper it is believed that long-term orientation has a positive influence on CSP. This is because long-term oriented cultures foster virtues aimed at future rewards, due to perseverance and thrift. By working towards higher CSP now, perseverance and thrift, they will be rewarded in the end by positive economic returns. By having a long-term orientation they will be rewarded in the future, as having a long-term orientation would then encourage CSP. Based on this reasoning, the following hypothesis is formed:

H5: A higher long-term orientation will have a positive influence on CSP.

2.2.6 Indulgence

desires” (Beugelsdijk, Maseland & van Hoorn, 2015). On the one hand, indulgence is defined as “a society that allows relatively free gratification of basic and natural human desires related to enjoying life and having fun” (Hofstede, 2011, p. 15). On the other hand, restraint is defined as “a society that controls gratification of needs and regulates it by means of strict social norms” (Hofstede, 2011, p. 15). The differences in these cultural dimensions can be seen in the fact that in cultures with high indulgence, there is a higher percentage of people saying that they are very happy, there is higher importance of leisure, there are more people who participate in sports, and freedom of speech is seen as important. Contrarily, in cultures with high restraint, there are fewer very happy people, there is less importance of leisure, less people who participate in sports, and freedom of speech is not a primary concern (Hofstede, 2011). Thus, in cultures with high indulgence leisure is important and there are few norms and suppressions with regards to gratification of human desires, while in cultures with high restraint there are strong norms regulating and suppressing the gratification of human desires.

As this sixth dimension is the newest in the Hofstede (2011) framework, there are only a few papers that have looked into it, especially with regards to CSP. One of the few papers is by Sun, Yoo, Park & Hayati (2019). They tested the relationship between CSP and Corporate Financial Performance, and found that there was a weaker effect in companies located in indulgent countries. They also recommend managers to allocate more resources to CSP initiatives if the firm is operating in a culturally restrained country (Sun et al., 2019). Halkos & Skouloudis (2017) also looked into the relationship between CSR and national culture, and found that the dimension indulgence has a positive effect on their CSR index. Thus, countries with a high level of indulgence seem to foster CSR.

As these findings are contradicting and there is very little research available, it is important to test this relationship. In this paper that will be done by assuming that a high level of indulgence has a positive effect on CSP. This is because by engaging in CSP MNEs aim to create higher living standards for its stakeholders. By enhancing the living standards for its stakeholders, it could be stated that it enhances their basic and natural human desires related to enjoying life and having fun. Having a high level of indulgence is then connected to a higher CSP, as it would make their stakeholders happier. Based on this rationalization the following hypothesis is formed:

2.3 State of Development

As stated before, it is important to not only verify the first four dimensions of Hofstede’s framework, as there are some contradicting findings, to test the newest two dimensions, as they have not been researched as extensively yet, but also to test the influence that the state of development of a country has on these relationships. State of development in this research indicates whether a country is either a developed or emerging market. This distinction is important because firstly, EMs have a growing influence on the global economy (Jamali & Karam, 2018; Peters et al., 2011). Secondly, even though CSP is nowadays seen as an important topic for firms to engage in, most research still focuses on developed countries (Muller & Kolk, 2009). Finally, this distinction is important because different types of governance systems need different approaches. As Peters et al. (2011) state, there is no “one size fits all approach” with regards to CSP systems. Muller & Kolk (2009) agree with this, as they state that EMs have different priorities, less access to technology, and weaker implementation systems, and they might need different CSP policies than Western countries. Therefore, it is important to include the concept of state of development when looking at the relationship between national culture and CSP. First, results from previous research will be examined, before forming a hypothesis about the expected relationship. After this section the conceptual model will be presented.

factors such as stages of economic development, culture, and institutions explain a large part of the variation across countries. They also found that firms in developed countries have higher CSP ratings than those from emerging countries (Cai et al., 2016). Muller & Kolk (2009) confirm this, with their research of companies in the Mexican auto industry. They state that while traditional research suggests that CSP in EMs is underdeveloped, new research looks at context specificity and suggests that it is different due to the specific and political background (Muller & Kolk, 2009). These findings thus confirm that economic development, the state of development, has an influence on national culture and CSP.

Even though these papers have not looked into the precise question that will be researched here, the current paper believes in the same rough direction of the relationship. What is thus expected, based on these previous papers, is that the state of development of a country indeed influences the relationship between national culture and CSP. This is, because while national culture has a direct influence on CSP, the state of development of a country influences this relationship. In other words, while an MNE in an EM can have approximately the same score on a certain national culture dimension as a country in a DV, it might still have a very different CSP. This might be due to many differences such as the aforementioned political backgrounds, institutions, and different governance systems, which in this paper together are gathered under the term state of development. Moreover, it is expected that a high state of development has a positive influence on CSP. As DVs have a more progressed state of development, it is expected that they have more resources available to direct towards increasing their CSP. Based on this reasoning, the following hypotheses are formed:

H7a: A higher state of development will have a positive influence on the negative relationship between power distance and CSP.

H7b: A higher state of development will have a positive influence on the positive relationship between uncertainty avoidance and CSP.

H7c: A higher state of development will have a positive influence on the negative relationship between individualism and CSP.

H7e: A higher state of development will have a positive influence on the positive relationship between long-term orientation and CSP.

H7f: A higher state of development will have a positive influence on the positive relationship between indulgence and CSP.

2.4 Conceptual Model

From the literature review and the hypotheses, the conceptual model in figure 1 can be derived.

Figure 1: Conceptual model

3. Research Methodology

This section will explain the research methodology that will be used to test the hypotheses. Firstly, the data collection will be explained, after which the sample and measurement of variables will follow, and lastly, the data analysis will be described.

3.1 Data Collection

variable, CSP, will be retrieved through the Environmental, Social, and Government (ESG) pillars of Thomson Reuters DataStream. This will be done on a company level, as the companies receive an overall ESG score indicating their level of CSP. The data for the independent variable, national culture, will be gathered through the public website of Hofstede, which covers data on 76 countries. Countries receive a score between 0 and 120 on each dimension, 120 being the highest and 0 being the lowest. This is appropriate, as it allows to quantify the dimensions of national culture. Lastly, the moderator, state of development, will be determined by making use of the ACWI Index by Morgan Stanley Capital International (MSCI). This will be further elaborated in the measurement of variables section.

3.2 Sample

The scope of this study is on both a company and country level. While the scores for CSP will be on a company level, they will be used in clusters per country. This is done, in order to be able to compare between countries with regards to national culture. Moreover, the distinction between countries will be made by the aforementioned MSCI. They distinguish 23 countries as developed, and 26 as emerging (see figure 2).

Furthermore, the sample for this study will then be MNE’s that have their HQ in these countries. As the aim is to look at the influence of national culture on MNE’s CSP, and the MNE’s home country state of development, companies with their HQ in these countries will be used, as the employees of these MNE’s are influenced by the culture and state of development of that country. The companies used are from the Forbes 2000 list to ensure that the companies are publicly listed and, based on their size, are more likely to have CSP data available.

Comparisons will then be made between emerging and developed countries, with regards to the relationship between national culture and CSP. As there are ten variables, and a measurement of at least ten samples per variables is used to make sure it is representable (Harrell, 2001; as mentioned in Austin & Steyerberg, 2015), some extra firms per state of development are added in order to ensure that there is still enough data after missing data is removed. In total, 215 companies from 23 different countries are used, of which 109 companies are from DVs and 106 are from EMs. Within this sample, countries from many different regions are used within both states of development in order to make sure the possible differences are per state of development and not per region. Furthermore, the data for the years 2016 and 2017 will be used in order to make sure that the most recent data is used. This is especially important for EMs, as they are still rapidly changing and for some countries, only recently companies have begun to engage in CSP. An example of this can be seen in Appendix A, which shows a fragment of companies in India where a clear increase over the last few years can be seen. The data for 2018 was considered for this paper; however, as there were fewer companies with data available for that year, the choice for 2016 and 2017 was made in order to ensure there would be less impact of missing data.

3.3 Measurement of Variables 3.3.1 Dependent Variable

these methods (Jo & Na, 2012; Muller & Kolk, 2009). However, the problem with this index is that it only provides information about American companies. Another example is the Global Reporting Initiative (GRI). The issue with this index is that it is based on reports from the companies themselves, making them subjective. Therefore, it is believed that the ESG index by Thomson Reuters is the best fit for this research, as it provides information on not only American companies, but on companies all over the world, thereby making it possible to compare countries from emerging and developed markets, and because it includes many different indices in order to make their measures transparent and objective. Moreover, researchers who have explored the concept of CSP before made use of several subcategories in order to measure it, but all of these were slightly different (Ho et al., 2012; Muller & Kolk, 2009; Ringov & Zollo, 2007). The ESG scores that will be used in this paper are continuous and three-fold, and cover the categories environmental, social, and governance. The ESG scores are then calculated on over 400 company-level ESG measures, which are grouped into ten sections and fall into the three main subcategories. The environmental category consists of the resource use, emissions, and innovation sections. Social consists of workforce, human rights, community, and product responsibility. And lastly, governance, consists of management, shareholders, and CSR strategy. These three categories are then merged together, to ultimately give one ESG score per company, per year, on a scale of 0 to 100. This gives insight into how that company performs on environmental, social, and governance areas, its CSR policy, for that particular year.

3.3.2 Independent Variable

culture. These examples show how the continuous scores of Hofstede’s dimensions are given and the differences that exist between countries.

Figure 3: Hofstede scores for Brazil, China, the Netherlands, and Sweden.

3.3.3 Moderating Variable

The moderating variable, state of development, will be measured by using the ACWI MSCI Index, which divides countries into developed and emerging markets (Morgan Stanley Capital International, 2019). As mentioned before, they divide 49 countries into DV and EM, of which 23 will be used in this paper. While there are several institutions that distinguish states of development, in this research the one from MSCI will be used because firstly, they review their index quarterly so the results are up to date, and secondly, because their method is designed to “take into account variations reflecting conditions across regions, marketcap sized, sectors, style segments and combinations” (Morgan Stanley Capital International, 2019). They thus aim to represent the different markets broadly and fair (Hau, Massa & Peress, 2009). Moreover, the MSCI Index is regarded as a leading provider of indexes (Hau et al., 2009), and therefore it is believed to provide the correct data needed in the current research to be able to compare between DVs and EMs.

3.3.4 Control Variables

To make sure no other factors are influencing the proposed hypotheses, several control variables will be used. Firstly, the size of the MNE will be controlled for. This is done because some researchers claim that the size of an MNE influences their CSR policy (Leopoutre & Heene, 2006; Udayasankar, 2008). This is, according to Brammer & Millington (2006), because larger firms have higher public attention and thus a greater need to act responsibly. In order to control for this, firm size will be measured by the number of employees, as Camacho (1991) defines organization size as “the number of individuals participating in the activity of the firm” (p. 137), thereby making it a continuous variable.

Secondly, industry type will be controlled for. Several researchers have examined the relationship between type of industry and CSP (Casado-Díaz, Nicolau-Gonzálbez, Ruiz-Moreno, & Sellers-Rubio, 2014; Giannarakis, 2014; Wanderley, Lucian, Farache & de Sousa Filho, 2008). This is of importance, because some industries received earlier or higher pressures to confirm to socially responsible behavior, and therefore this factor will be controlled for. It is a categorical variable, and eleven different categories will be used in total. This will make sure that the industries have specific numbers, and it can be analyzed whether industry type has a significant influence.

Thirdly, MNE profitability will be controlled for. Giannarakis (2014) is an example of a researcher who found a significant relationship between the companies’ profitability and CSR disclosure, stating that a profitable company has the freedom and flexibility to show its CSR practices to stakeholders. As there are several other researchers who found a relationship between MNE profitability and CSP (Campbell, 2007; Mulyadi & Anwar, 2012), this factor will be controlled for by looking at MNE’s return on equity (ROE), and by measuring it as a continuous variable.

3.4 Data Analysis

variables and the continuous dependent variable need to be mean-centered for the analysis.

3.4.1 Robustness Test

In order to make sure the results are reliable, robustness tests will be performed. While there are many ways to test this, a model variation test will be used. Examples of this are a change in the functional form, a change in operationalization, or a change in the sample (Neumayer & Plümper, 2017). In this case, a change in the sample will be made to check whether the relationship is consistent. The first robustness test will make use of data from 2016, in order to make sure the results are stable over time, and to ensure there is no role of Time-Lags. By comparing the results from 2017 with the results of 2016, it is made sure that there are no other factors influencing the results of the changing variables. The second and third robustness tests will check the data for DVs and EMs separately, to verify the significance or insignificance of state of development.

4. Results

To make sure the results are the most recent possible, data from the latest years will be used. However, as mentioned before, the year 2018 has more missing data than the years beforehand, and therefore will the year 2017 be used for the main regression, and the year 2016 in order to test for robustness.

4.1 Test of Basic Assumptions

Before performing the analysis and checking if there is a relationship between the proposed hypothesizes, four assumptions will have to be tested first. This is done in order to make sure that the relationships represent the actual population by looking at the efficiency and correctness.

linearity. This is done in order to make sure that the independent variables have a straight-line relationship with the dependent variable. However, when the previous two assumptions are met, it can be assumed that the linearity is met as well (Statistics Solutions, 2019). Fourth and finally, there has to be tested for multicollinearity. This predicts how the variables are correlated. In this research, this is checked by looking at the variance inflation factor (VIF) values. As can be seen in Appendix E, all factors score below the maximum of 10, and thus is the assumption met. As all four assumptions have been tested and meet the requirements, the descriptive statistics, correlation matrix, and the regression analysis will now be presented.

4.2 Descriptive Statistics

The descriptive statistics of the years 2016 and 2017 are presented in table 1. Out of the 215 sampled companies 207 MNEs had ESG scores available in 2017, and in 2016 this number was 209. The sampled MNEs operate in eleven different industries, ranging from Technology, to Health Care, to Real Estate, with Finance having the highest frequency (see Appendix B).

The data for the 2017 sample shows that the CSP scores have a quite wide spread, with a range from 13.58 till 95.83. This demonstrates that the ESG scores differ quite substantially among MNEs. However, with the average score being 70.7361 it shows skewness towards the upper-value range. The dimensions for national culture each have a score on a scale of 120, which remain stable over the gathered years.

Table 1: Descriptive statistics

Note: N 2017 = 207; N 2016 = 209; Variables are presented in original form and not mean-centered

2017 2016

Variable Min Max Mean SD Min Max Mean SD

Uncertainty avoidance has the highest standard deviation (24.466) and also the highest mean (62.99), explaining the big difference in minimum and maximal score (17;91). However, masculinity has the biggest difference in minimum and maximum (5;95), but its mean (53.58) and standard deviation (20.304) are lower. Indulgence has the lowest average (49.82), with a range from 17 till 97. The sample’s MNE size in 2017 has an average of 108344 (rounded as it’s a number of people), with a wide range (Min = 1419, Max = 1290000), and its MNE profitability has an average of 14.6857 with also a wide range (Min = -29.42, Max = 29.15).

The 2016 sample is used for robustness purposes, and it shows that the ESG has increased slightly (from 69.5532 in 2016 to the aforementioned 70.7361 in 2017). This is also the case for the other changing variables. The average for MNE size was 106450 in 2016, and MNE profitability was 10.9683. As the scores for national culture, state of development, and industry type only change very slowly over time or not at all, they remain the same.

4.3 Correlations

In order to check the linear relationships between the variables a correlation analysis was performed. In table 2 the results for 2017 are presented and in Appendix F the results for 2016 are shown. In both years power distance, individualism, state of development, and the control variable industry type correlate significantly with ESG score. While some of these are quite high, there are no relationships above R ± 0.7, and therefore it can be assumed that there are no strong relationships between the variables. Another method to test this is by looking at the VIF scores, which test for multicollinearity. As presented in Appendix E, all factors score below the maximum of 10 and thus can it be stated that there are no strong relationships between the variables.

Table 2: Correlation Matrix 2017 Note: N=215; p<0.05*, p<0.01** (2-tailed) Correlations Power distance Uncertainty

avoidance Individualism Masculinity

Long-term

orientation Indulgence

State of

development ESG score Company size Industry type

Interestingly, while the correlation between individualism and ESG score was the only one with a positive effect, the correlation between this dimension and state of development is a negative one, indicating a negative moderation effect on this relationship. The scores in 2016 indicate the same effect, thereby verifying the possible effects.

4.4 Regression Results

In table 3, the results of the regression analysis for all data in 2017 are presented (DV + EM 2017). Firstly, the effects of the control variables on the dependent variable, CSR performance, are tested in model 1. Industry type is the only variable that has a significant explanatory power on CSP (B=-.188; p=.012; significance level p≤.05), indicating that there are industry types that influence an MNE’s CSP. In models 2 up until 7 the six culture dimensions are each tested in a separate manner to the control variables, in order test hypothesizes 1 through 6.

In model 2 the independent variable power distance is added in a separate manner to the control variables, to test hypothesis 1. The results are negatively significant (B=-.414; p=.000l; significance level p≤.01). This is in line with the findings of the correlation matrix in table 2, and is again confirmed in model 8 where the general regression model is run (B=-.276; p=.001). Therefore, hypothesis 1 can be accepted.

In model 3, the independent variable uncertainty avoidance is tested separately to the control variables, to test hypothesis 2. While the results here are not significant (B=.056; p=.231), the results in model 8, the general regression model, are significant (B=.130; p=.002). From this, it can be gathered that while the dimension on its own is not significant, it is significant when the general regression is run. However, as model 3 does not result in significant results, the effect is unstable, hypothesis 2 is rejected.

Table 3: Regression Analysis DV + EM 2017

Note: N=215; p≤0.1*, p≤0.05 **, p≤0.01***

Corporate Social Performance

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Model 8 Model 9 Model 10

Independent Variables Power distance -.416 (.000)*** -.276 (.001)*** -.250 (.006)*** -.239 (.202) Uncertainty avoidance .056 (.231) .130 (.002)*** .130 (.002)*** .185 (.021)** Individualism .258 (.000)*** .227 (.000)*** .206 (.002)*** .223 (.097)* Masculinity -.018 (.747) .015 (.763) .014 (.778) .013 (.852) Long-term orientation .058 (.232) .147 (.018)** .145 (.021)** -.056 (.562) Indulgence .060 (.265) -.008 (.908) .002 (.971) -.012 (.941) Moderating Variable State of development -2.090 (.478) -18.553 (.513) Interactions

Power distance X SOD -.057 (.795)

Uncertainty avoidance X SOD .009 (.933)

Individualism X SOD -.185 (.354)

Masculinity X SOD .252 (.117)

Long-term orientation X SOD .257 (.091)*

Indulgence X SOD -.040 (.841)

Controls

In model 5 the independent variable masculinity is tested separately to the control variables, in order to test hypothesis 4. The results of this test are not significant (B=-.018; p=.747) and are confirmed in the general regression in model 8 (B=.015; p=.763). Therefore, hypothesis 4 is rejected.

In model 6 the independent variable long-term orientation is tested separately to the control variables, in order to test hypothesis 5. The results are not significant in model 6 (B=.058; p=.232), but when tested in the general regression model they are significant (B=.147; p=.018). This indicates that while the dimension on its own is not significant, when tested in the general regression it is significant. However, as model 6 is not significant, the results are unstable, and hypothesis 5 is rejected.

In model 7 the independent variable indulgence is tested separately to the control variables, in order to test hypothesis 6. The results show that the effect from this dimension on CSR performance is very small and not significant (B=0.60; p=.265). This is confirmed in the general regression model with even more insignificant numbers (B-.008; p=.908). Based on these results, hypothesis 6 is rejected.

In model 9 the moderator state of development was tested, and the results show that the effect is negative, but not significant (B=-2.090; p=.478). Model 10 confirms this when the interaction terms are added and most of the interactions show no significance. Only the long-term orientation dimension shows a significant result when interacting with state of development (B=.257; p=.091). As the other five dimensions show no significance hypotheses 7a, b, c, d, and f are rejected, and only hypothesis 7e is accepted.

4.5 Robustness Test

CSP but the hypothesis expected a negative result, hypothesis 3 is rejected. With regards to the 2nd _{and 5}th _{hypothesis, uncertainty avoidance and long-term orientation,}

when the dimensions are run in the general regression model they show significant results (B=.111; p=.009; B=.171; =.007). However, as models 3 and 6 in the separate regressions show insignificant results (B=.027; p=.565; B=.071; p.144), it can be stated that their results are unstable and hypotheses 2 and 5 are also rejected. In model 5, in line with the 2017 regression, a negative relationship is found between masculinity and CSP. However, as this effect is very small and not significant, hypothesis 4 is rejected. And lastly, for the sixth dimension, indulgence, neither in the general regression nor in the separate model are the results significant and thus is hypothesis 6 rejected ((6)B=.071; p=144; (8)B=.006; p=.923). A difference with the 2017 regression is that in the 2016 analysis even less significance was found for the moderator state of development. In the 2017 regression uncertainty avoidance and individualism remained slightly significant, and long-term orientation showed significance when interacting with state of development. In the 2016 regression however, only uncertainty avoidance remains slightly significant, the other dimensions show no significance, and no dimensions show significance when interacting with state of development. While hypothesis 7e was accepted in the DV + EM 2017 regression, it shows no significance here and it can therefore be stated that the result is unstable. Based on these findings hypotheses 7a-f will be rejected.

When only testing the developed markets data in Appendix H (DV 2017) more significance is visible, and also more positive significant results are visible in comparison to the results from only the emerging data (EM 2017). The regression analysis shows significant results for four out of the six dimensions, and would confirm 3 of the 6 hypotheses. The results of model 2 in Appendix H confirm the negative relationship between power distance and CSP (B=-.338; p.000) and thus hypothesis 1 is accepted, which is in line with the results from table 3 (DV + EM 2017). Model 3 also shows significant results, between the independent variable uncertainty avoidance and CSP (B=.103; p=.061). As hypothesis 2 expected a positive relationship between these two, this is confirmed and thus hypothesis 2 can be accepted. Model 4 shows significant results for the separate test to the control variables between individualism and CSP (B=.321; p=.000). However, as hypothesis 3 expected a negative effect and the results show a positive relationship, hypothesis 3 is rejected. This is in line with the results in table 3 (DV + EM 2017). In models 5 and 6 insignificant results were found for the independent variables masculinity and long-term orientation when tested with the control variables ((5)B=.029; p=.569; (6)B=-.103; p=.115). And lastly, model 7 tests the independent variable indulgence separately to the control variables. In contrast with the findings in table 3, a significant positive effect was found (B=.247; p=.000). As hypothesis 6 expected this relationship, the hypothesis is accepted.

CSP (B=.173; p=.005). This significance was not found in the previous regressions. However, this result is in line with hypothesis 5 and therefore this hypothesis will be accepted in the case of emerging markets. Lastly, model 7 shows a negative significant result between indulgence and CSP (B=-.139; p=.062). While a significant result is in line with the findings of Appendix H, the results found for emerging markets is negative. As hypothesis 6 expected a positive relationship, this hypothesis will be rejected with regards to the result in Appendix I.

Although the results from the regression analysis in 2017 are confirmed in the first robustness check with the data from 2016, there are some differences with the second and third robustness check. The stable factor in all regressions is the results found for the independent variable power distance and the dependent variable CSP, which has a negative effect. The differences of significance found in the robustness checks will be discussed further in the discussion section. An overview of the findings is presented in table 4.

Table 4: Overview Regression Analyses

Note: DV + EM 2017/2016 N=215; DV 2017 N=109; EM 2017 N=106; p≤0.1*, p≤0.05 **, p≤0.01***

5. Discussion

The main objective of this study is to contribute to the body of knowledge on the influence of national culture on CSP. By examining whether state of development has a moderating role on this relationship, this study expands the knowledge on both EMs and DVs and how the tested relationship might have different outcomes in the

Overview DV + EM 2017 DV + EM 2016 DV 2017 EM 2017 Power distance -.416 (.000)*** -.421 (.000)*** -.338 (.000)*** -.257 (.048)** Uncertainty avoidance .056 (.231) .027 (.565) .103 (.061)* .122 (.076)* Individualism .258 (.000)*** .267 (.000)*** .321 (.000)*** -.123 (.243) Masculinity -.018 (.747) -.030 (.601) .029 (.569) -.156 (.179) Long-term orientation .058 (.232) .071 (.144) -.103 (.115) .173 (.005)*** Indulgence .060 (.265) .051 (.338) .247 (.000)*** -.139 (.062)** State of development -2.090 (.478) -3.230 (.275)

different states. Overall, the results of this study are similar to prior research that found mixed results regarding the effect of national culture on CSP. While power distance has a clearly negative effect on CSP and individualism has a positive effect, the effects for the other dimensions show contradicting results in the different regressions. Moreover, the effect of the moderator, state of development, shows a negative but insignificant relation when tested with the general regressions. However, when testing the data sets separately, the results did show some interesting differences.

As explained in the introduction and literature review, contradictions were found in the results of previous research papers. While there was also some contradiction in previous research with regards to the relation between power distance and CSP, this was only marginally. The findings in the current study are in line with the majority of the results (Peng et al., 2012; Ringov & Zollo, 2007; Thanetsunthorn, 2015), as power distance has a significant negative relationship with CSP. This result holds true regardless of the year in which it was tested, or the data sample that was used. It thus indicates that when a national culture has a higher degree of power distance, this has a negative influence on CSP. This finding could confirm the reasoning that in cultures with a high power distance income is distributed unequally and less powerful members of the organization accept and expect this, which negatively influences CSP as there is less regard for the creation of higher living standards for all its stakeholders.

Consistent with previous research, the findings for the relationship between uncertainty avoidance and CSP are contradicting. In line with the findings of Ringov & Zollo (2017), the results for 2017 in the separate regressions are not significant. However, in the general regression and in the separate states of development regressions the results are positively significant, which is in line with several other papers (Peng et al., 2012; Ringov & Zollo, 2007; Thanetsunthorn, 2015). This indicates that a higher degree of uncertainty avoidance results in higher CSP, which would confirm the reasoning that because many stakeholders are dependent on the survival of the MNE, a more certain operationalization will result in higher CSP, resulting in better living standards for more stakeholders.

Thanetsunthorn (2015) and Ho et al. (2012), the results found in the current paper show that most regression analyses have a significant positive relationship. While the regression of only EMs found no significant results, in line with the findings of Ringov & Zollo (2007), all three other regressions found a positive significant relation, conforming to the findings of Peng et al. (2012). These findings thus indicate that a higher degree of individualism has a positive influence on CSP. Peng et al. (2012) reason this by stating that in individualistic cultures, people take responsibility for their own actions. This could imply that because they are held responsible, they are more likely to engage in CSP, thus improving their ESG score. Ioannou & Serafeim (2012) argue in the same rough direction, as they state that in individualistic cultures managers are more likely to “make their mark” by choosing explicit decisions and actions in the CSP domain, and respond to the perceived expectations of a wide range of stakeholders (p. 19). It thus appears that a high degree of individualism has a positive influence on CSP because people are held responsible for their own actions and want their decisions and actions to be perceived well by the stakeholders.

Contrary to expectations and previous findings, masculinity was not found to have a significant effect on CSP. While the results in table 3 were negative, in line with several previous papers (Esteban et al., 2017; Peng et al., 2012; Ringov & Zollo, 2007; Thanetsunthorn, 2015), they were not significant. This indicates that the level of masculinity or femininity could not explain the level of CSP of an MNE. This difference could perhaps be explained by the difference in sample between this and former studies. For instance, Peng et al. (2012) used multinationals that were comparable and, among others, excluded firms based on insufficient liquidity. The current study aimed at comparing countries in both EMs and DVs, and therefore consists of a very diverse dataset. This difference in samples might explain the contradicting findings and presents opportunities for further research.

tested separately to the control variables. It also showed positively significant results in the regression with only EMs (2017), and not when tested with only DVs (2017). The positive findings are in line with the argumentation for hypothesis 5 and papers that previously looked into the same rough direction of this dimension (Bae & Kim, 2013; Wang & Bandal, 2012). Moreover, this difference in findings indicates that, while a long-term orientation doesn’t show an effect in DVs, a long-term orientation does have a positive relationship with CSP in EMs. This could be ascribed to, as Su et al. (2014) state, the signals that firms in EMs send to stakeholders when they participate in CSR practices. By engaging in CSR practices, they make investments for their stakeholders (a long-term orientation), which leads to a higher CSP. However, while this is the case for EMs, DVs show no significant results for this dimension. Therefore, forming definite conclusions about the possible relationship will require more extensive research.

Secondly, with regards to the sixth and final dimension of the Hofstede framework, the same conclusion as for the fifth dimension can be stated. Where the results above were either positively significant or not significant, the results for indulgence are even more diverse. While the results for the regressions of DV + EM 2017 and DV + EM 2016 are in line and indulgence does not have a significant effect on CSP, the results for the separate tests of 2017 (DV 2017, EM 2017) as robustness tests, show positive and negative influences. Whereas the data in DV 2017 shows that indulgence has a positive influence on CSP, the data sample in EM 2017 shows the opposite effect, a significantly negative influence. This difference would indicate that state of development does make a difference when it comes to indulgence and restraint; while in DMs indulgence has a positive effect on CSP, in EMs it has a negative effect. This difference could be ascribed to, as Muller & Kolk (2009) state, the different priorities EMs have. While DVs show a positive relationship between indulgence and CSP, a high level of for instance leisure leads to higher CSP, EMs show a negative effect here. As they have less access to technology and weaker implementation systems (Muller & Kolk, 2009), the opposite, a high level of restraint through for instance strict social norms, leads to higher CSP. Thus, a difference in development leads to different CSP outcomes.

certain differences can be observed. Firstly, the differences in descriptive statistics are notable. While DVs, on average, have a power distance of 47.21, EMs have an average of 71.84. The opposite can be observed for individualism, where DVs have a 67.80 on average, while EMs score a 36.76. Secondly, while these differences are quite large, these are two of the only dimensions where the influence on CSP was parallel in most regressions. When testing the data sets for DVs and EMs separately, the most notable differences can be observed for the two newest dimensions of the Hofstede framework. Whereas long-term orientation has an insignificant influence on CSP in DVs, it has a positive effect in EMs. Moreover, when testing indulgence separately, it has a positive influence on CSP in DVs, while it has a negative effect in EMs. These differences might indicate that, while the original regression did not show significant results, previous researchers (Cai et al., 2006; Muller & Kolk, 2009) and the argumentation in the current paper found interesting opportunities for future research and both opportunities and implications for MNEs.

And lastly, the control variables show some interesting results. While MNE size and MNE profitability do not show any significant results, industry type does show significant results. This indicates that, in line with previous findings (Casado-Díaz et al., 2014; Giannarakis, 2014;) industry type does indeed influence CSP. Thus, certain types of industries are more likely to behave more responsibly.

6. Conclusion

This master thesis contributed to the body of knowledge on corporate social responsibility and state of development by answering the research question: Does

national culture influence MNE’s CSR and is this relationship influenced by the MNE’s home country’s state of development? By answering this question, several

theoretical and practical implications have been made, which will be presented in this section. After this, the limitations and recommendations for future research are given.

6.1 Theoretical Implications