The Effects of National Culture on Firms’ Environmental CSR and the Moderating Role of the Degree of Internationalization

(1)

Master Thesis

The Effects of National Culture on Firms’ Environmental CSR and

the Moderating Role of the Degree of Internationalization

Serena Riccomagno S3558452

s.riccomagno@student.rug.nl

Supervisor: Dr. O. Lindahl Co-assessor: Dr. R.W. de Vries

MSc International Business & Management

University of Groningen, Faculty of Economics and Business Date of Submission: 21st_{of January 2019}

(2)

Abstract

Cross-cultural CSR has increasingly been the focus of scholarly attention in International Business literature. Nonetheless, cross-country CSR literature often shows mixed and inconclusive results, only giving marginal importance to the environmental CSR pillar, which is almost exclusively studied in combination with social CSR. This thesis takes an alternative approach by investigating how cultural dimensions can affect the specific aspect of environmental CSR in firms, therefore focusing on how national culture can influence environmental action at the firm level of analysis. Furthermore, scholars often rely on the assumption that home-country culture is the sole reference culture guiding CSR action, neglecting the role that factors such as the degree of internationalization might have on

environmental CSR firms’ strategies. The main objective of this thesis is therefore to fill the

aforementioned research gaps by investigating the effect that Hofstede’s cultural dimension

Individualism, Power Distance, Masculinity and Uncertainty Avoidance have on firm’s environmental CSR of 100 companies originating from different home countries, along with the

moderating effect given by the degree of internationalization. Drawing from institutional theory, it is argued that different cultural dimensions will have different impacts on firms’ environmental CSR, and this impact will be negatively moderated by the degree of internationalization. The empirical analysis is carried out through OLS regression and find support for the positive impact of

Individualism on environmental CSR but, contrary to expectations, the hypothesized negative

moderating effect given by internationalization is not statistically significant. These findings allow for academic and practical implications in the field of environmental CSR.

Keywords: Corporate Social Responsibility (CSR), environmental CSR, cultural dimensions,

(3)

Acknowledgements

(4)

1.INTRODUCTION

The theme of Corporate Social Responsibility (CSR) has been the center of increasing scholarly interest in the last years (Halkos & Skouloudis, 2016). Nowadays, CSR is thoroughly institutionalized and globalized, with a growing number of umbrella organizations, CSR consultancy companies and CSR initiatives taking place (Carroll, 2008). Despite extensive effort for its conceptualization, there is no comprehensive and universally agreed upon definition for the term Corporate Social Responsibility (Hamidu et al., 2015). However, two main CSR pillars can be identified: social and environmental CSR pillars; nonetheless, this thesis will proceed to exclusively analyze the environmental one. This choice was dictated by the increasing importance that environmental concerns, such as global warming and gas emissions, are being given in the public debate arena. These issues call for immediate action, and firms are often the actors who are expected to intervene, by creating a set of more environmentally sustainable and green CSR strategies. More and more, firms are expected to conform to this “green imperative” (Menguc & Ozanne, 2003) that should be prioritized given the critical importance natural resources play in the survival of our species.

Scandinavian firms are “routinely cited as a global leader in corporate social responsibility (CSR) and sustainability” (Strand et al., 2015, p.1). Why is that so? Can cultural aspects, that characterize Nordic societies, be used as an explanatory factor for such a phenomenon? These are the questions that inspired the author in writing this thesis in the first place. The idea that national culture, defined as the “collective programming of the mind” (Hofstede, 2001, p.3), might play a role in what firms consider good CSR practices have increasingly been a topic of interest for scholars (Halkos & Skouloudis, 2016). Previous studies have successfully demonstrated how culture can be used as a determinant for CSR (Gallego-Álvarez & Ortas, 2017; Gallén & Peraita, 2018; Miska et al., 2018; Park, 2007). These studies have found a significant correlation between CSR performance and some cultural characteristics, giving support for the anecdotal intuition for which different cultures place different importance and meanings on CSR.

(8)

and mixed results. Particularly, the study of cultural determinants of environmental CSR is a niche field stemming from the already relatively new scholarly topic of cross-country CSR. Previous literature on the topic (Husted, 2005; Park, 2007) studied how cultural values are linked to environmental concerns at the societal level of analysis, leaving only a handful of studies to specifically addressing it at the firm level of analysis (Ioannou & Serafeim, 2012; Ringov & Zollo, 2007; Hartmann & Uhlenbruck, 2015). Even then, literature tends to treat CSR as a whole, without accurately differentiating between social and environmental CSR pillars. In fact, a targeted analysis exclusively focusing on the cultural antecedents of environmental CSR at the firm level of analysis is lacking, creating a research gap that this thesis will try to address. This thesis will attempt to merge and integrate different literature streams, trying to find out what kind of relationship, if any, exists between culture and the specific aspect of environmental CSR at the firm level of analysis. This thesis will also try to identify a possible moderating effect resulting from the degree of

internationalization of the firm. To the knowledge of the author, no existing literature takes into

consideration this aspect. In fact, past literature has relied on the unjustified assumption that firms’ home country culture is the exclusive culture determining the extent to which a firm decides to engage in environmental CSR. However, given the multiplicity of institutional settings to which highly internationalized firms are exposed, it would be reductive and simplistic to assume that a single culture affects environmental CSR, making the cultural determinants of highly internationalized MNEs a fascinating and complex topic of study.

It is therefore clear how the relationship between national culture and CSR, especially its

environmental pillar, is still an understudied topic, far from being clearly understood, especially at

the firm level of analysis. A layer of complexity is added by the interaction between possibly conflicting cultures to which highly internationalized firms are exposed to. This thesis will aim to investigate this very relationship through the following research question:

“Does home country national culture influence environmental CSR in firms? Is this relationship moderated by firms’ degree of internationalization?”

(9)

topic. In the first chapter, an overview of the existing literature on CSR, national culture and internationalization is discussed, along with an analysis of the interplay between these concepts. Subsequently, the hypotheses are developed along with the corresponding conceptual model. Secondly, the methodology and measurements are presented and hypotheses are tested using statistical regression analysis. Thereafter, the results are described and discussed. This thesis, other than a contribution to existing research, can be considered of managerial relevance. In fact, by showing that environmental CSR is not a homogenous concept, but rather a culturally driven one, managers can be helped in their practical decisions regarding strategies in cross-cultural environmental CSR.

2. THEORY

2.1 Literature Review

2.1.1 National Culture in International Business Literature

National Culture in the Context of Institutional Theory

In this thesis, national culture will be analyzed in the framework of institutional theory, which greatly emphasizes concepts such as legitimacy. Specifically, legitimacy, in the context of institutional

theory, can be defined as: “a generalized perception or assumption that the actions of an entity are

desirable, proper, or appropriate within some socially constructed system of norms, values beliefs and definitions” (Suchmann, 1995, p.574). This theory, therefore, places particular emphasis on how firms are shaped and are highly dependent on the institutions of the countries in which they operate, so much so a lack of legitimization within a national institutional setting might preclude a company’s own existence and survival (Suchmann, 1995). More in detail, literature distinguishes between formal and informal institutions (Helmke & Levitsky, 2004). While it is often implicitly expected for the firms to comply with formal institutions (such as laws and regulations), a full legitimization will only be achieved by complying with informal institutions (e.g.: national culture) (Miska et al., 2018). In fact, cultural and cognitive legitimization is defined by Scott (2001, p. 61) as “the ‘deepest’ level because it rests on preconscious, taken for granted understandings”.

(10)

In order to be studied in IB, culture first needs to be quantifiable and comparable, a series of frameworks attempting at such task have been developed and will be analyzed in the next section.

Cultural Frameworks in International Business Literature

Quantifying culture in order to make it comparable has always been a central topic of interest in IB literature and the most prominent cultural frameworks used are Hofstede (1980), Schwartz (1999) and the GLOBE study (House et al., 2004).

Schwartz (1999) and GLOBE (House et al., 2004) can be considered methodologically similar since they are theory driven. These frameworks, in fact, aim in the first place at defining salient, universally relevant problems (e.g.: the relationship between the individual and the group) that a society has to face (Schwartz, 2006). Subsequently, they proceed to hypothesize two polar opposite orientations and score each national culture based on this spectrum. Schwartz’s (1999) first hypothesizes 7 different cultural orientations that are then tested for validity with empirical data using a representative and recent sample (more than 70 000 surveys responses gathered from 1988 until 2000) (Schwartz, 2006). Nonetheless, literature is cautious in praising the superiority of Schwartz’s work compared to more well established and deeply ingrained cultural frameworks, such as Hofstede (Drogendijk & Slangen, 2006). In fact, out of all of the frameworks analyzed, Schwartz is the one that is most rarely used in the context of IB literature.

(11)

Hofstede’s Framework

This thesis chooses Hofstede’s framework and its cultural dimensions in order to quantify national culture. This choice is dictated by the fact that Hofstede’s framework is the most widely used, and still considered the reference work in IB studies (Beugelsdijk et al., 2016), making it easier to compare and contrast the results with existing literature. Furthermore, Hofstede’s framework resists time, and research has shown how its dimensions are still appropriate and valid to use (Beugelsdijk et al., 2015). The Dutch professor Geert Hofstede, in his book “Culture’s consequences: International differences in work-related values” first published in 1980, is among the first scholars who attempted to create a framework to make national culture comparable. Hofstede operationalized national culture through “national values” which he defines as “broad tendencies to prefer certain states of affairs over others” (Hofstede, 1991, p.35). The scholar based his framework on answers from 100 000 work-related surveys filled by IBM employees from 1967 to 1973, with a sample consisting of 66 countries (McSweeney, 2002). After factor analyzing the data, the Dutch scholar identified four central and independent “bi-polar dimensions of a national culture” (McSweeney, 2002, p.91) and assigned a national score on each of the identified dimensions. The initial four cultural dimensions identified by Hofstede were: Power distance, Uncertainty Avoidance, Individualism vs Collectivism and

Masculinity vs Femininity. Subsequently, two other dimensions were added, namely Short vs Long Term orientation and Indulgence vs Restraint (Hofstede, 2011). In this thesis, only four (Power Distance, Uncertainty Avoidance, Individualism vs Collectivism and Masculinity vs Femininity) of

the six dimensions will be analyzed. This decision can be justified by the fact that data availability for Short vs Long Term orientation and Indulgence vs Restraint is limited (Hofstede, 2011). Furthermore, existing literature on cross-cultural CSR tends to neglect the last two dimensions (Park, 2007; Husted, 2005; Vachon, 2010).

Power Distance

(12)

even older people becomes more formal and distant. These characteristics also tend to lead to higher corruption and higher acceptance for covering up consequent scandals (Hofstede, 2011).

Power Distance scores were originally calculated through the following survey questions:

“(a) the preference for one style of decision-making by one’s boss over other styles;(b) the perception

of the boss’s actual decision-making style; and (c) (for non-managerial employees only) the feeling that employees were afraid to disagree with their manager” (Hofstede, 2006, p.887)

High power distance countries include Mexico, China, India and France, while Finland, Austria and Israel exhibit low Power Distance scores (Hofstede, 1980).

Uncertainty Avoidance

Uncertainty Avoidance (UA) “deals with a society's tolerance for ambiguity” (Hofstede, 2011, p.10)

and measure how uncomfortable or comfortable people from a certain society will feel in a novel situation (Hofstede, 2011). Societies characterized by strong UA will be more “intolerant towards deviant persons and ideas” and “need for clarity and structure” is emphasized (Hofstede, 2011, p.10). These societies emphasize rules, that even if not followed, need to be present. This "emotional need for rules" (Hofstede, 2011, p.10) is correlated with a higher level of stress and neuroticism (Hofstede, 2011). Furthermore, the need for certainty leads people to stay in a job even if disliked leading to worsening well- being levels (Hofstede, 2011).

Hofstede calculated the Uncertainty Avoidance scores through three questionnaire items: “(d)

the feeling that company rules should not be broken, not even when the employee thinks it would be in the company’s best interests to do so; (e) the respondent’s intention to stay with the company for more than five years; and (f) the respondent’s feelings of stress at work” (Hofstede, 2006, p.888).

Examples of high UA countries include Japan, Greece and Portugal while countries such as Denmark, Sweden and Singapore are characterized by low levels of UA (Hofstede, 1980).

Individualism

Individualism (vs Collectivism) can be defined as “the degree to which people in a society are

(13)

"predetermined by in-group" (Hofstede, 2011, p.11) therefore personal opinions are not deemed important. Furthermore, in collectivistic societies “relationship prevails over task” (Hofstede, 2011, p.11) meaning that personal connections and ties are the real force guiding social interactions. This mechanism leads collectivistic societies to discourage voicing personal opinions in the name of harmony, paired up with a sense of belonging to the in-group (Hofstede, 2011).

Hofstede originally operationalized this dimension “based on IBM employees’ scores for the

importance of various job aspects for describing their ideal job” (Hofstede, 2006, p.888). Venezuela,

China and Thailand can be seen an example of highly collectivistic countries, while the US, the UK and the Netherlands score very high in the dimension of Individualism (Hofstede, 1980). It is interesting to notice how the dimension of Individualism strongly correlates with wealth, reason for which the Dutch scholar decides to control for national wealth each of the cultural scores (Hofstede, 2006).

Masculinity vs Femininity

The cultural dimension Masculinity (vs Femininity) is defined as a preference in society for achievement, heroism, assertiveness, and material rewards (Hofstede, 2011; Hofstede, 2018). In masculine societies, gender roles are more clearly defined. Men are expected to focus on material success while women on quality of life (Hofstede, 1998). On the other hand, in feminine societies, both genders are supposed to place the same importance on being modest and caring about life quality, exhibiting a lower level of competition (Hofstede,1998). Key values in masculine societies emphasize “material success and progress” at the expenses of quality of life or “solidarity” (Hofstede,1998, p.16). In the work environment, masculine societies expect managers to be “decisive and assertive” while in feminine ones they are expected to be more “intuitive” (Hofstede,1998, p.16). In feminine societies, conflicts should be dealt with "negotiation and compromise" (Hofstede,1998, p.17) while in masculines ones fighting is seen as the solution. Overall, Masculinity vs Femininity can be seen as a sort of “tough vs tender” approach to life (Hofstede, 1998, p.6).

The scores for this dimension were originally calculated through survey questions scores about matters such as: “work goals opposed the importance of the relationship with one’s direct

manager and with one’s colleagues, but also the possibility to live in a desirable area, receiving high earnings, recognition, advancement and challenge” (Hofstede, 2006, p.888). Countries such as

(14)

Hofstede’s Strengths and Limitations

Hofstede’s framework is often considered one of the most seminal works in IB literature. In fact, Hofstede’s book “Culture’s consequences: international differences in work-related values” (1980)

has more than 40 000 citations on Google Scholar, making it is the 15th_{most cited book in social}

sciences (Green, 2016). Despite its widespread use, it is important to mention that Hofstede’s work has been the subject of many critiques over the years. First and foremost, scholars have pointed out how Hofstede, just like many authors in social sciences, runs into the so-called ecological fallacy; the assumption for which the relationships observed at the group level of analysis will also be valid at the individual level (Grenness, 2012). In other words, it is assumed that single individuals will behave the same way as the general tendency observed at the country level of analysis. It is easily understandable how this could be a limitation of the framework. In fact, individuals’ variations are completely overlooked and a single, univocal, country tendency is identified. Through this unjustified assumption, Hofstede defends the very small sample he used for some countries (namely, Pakistan only had a total of 107 respondents) explaining how, even a small number of respondents, is, in his view, representative of the overall country tendency (Grenness, 2012). Secondly, given the specific sample used, critics have underlined how Hofstede might have identified a sort of IBM-specific cross-cultural variation, which may not be reflective of national culture. The sample was in fact exclusively made up by highly educated IBM employees that had to go through a very selective recruitment process by a US-based company. All of these factors might have distorted the measurement of national culture and critics have pointed out how “IBM employees might most likely diverged from the general population” (McSweeney, 2002, p. 101). Thirdly, it has been hypothesized that Hofstede’s dimensions might lack validity given the fact that they were measures more than 40 years ago (Beugelsdijk et al., 2015). Furthermore, it has been underlined how Hofstede overlooks intra-national diversity given by minorities and regional subcultures (McSweeney, 2002), which in countries that have more than 1 billion inhabitants, such as India and China, might be extremely pronounced (Dheer et al., 2015).

(15)

fact, despite the apparent differences, Beugelsdijk et al. (2016) manage to integrate the three frameworks by creating a “dendrogram of the cultural profile clusters integrating Hofstede, Schwartz and GLOBE” (Beugelsdijk et al., 2016, p. 39). This means that the three frameworks lead to roughly similar conclusions about the supra-national clustering of countries.

2.1.2 Corporate Social Responsibility

CSR in the Context of Institutional Theory

Institutional Theory (IT) has been generally neglected by past literature when analyzing CSR.

Scholars preferred other frameworks, such as stakeholder theory, when faced with the challenge of understanding related phenomena. Nowadays, IT is gaining relevance in the field of CSR-related studies, which especially in the context of IB literature, is more and more interested in understanding why CSR practices are so diverse among regions and countries (Brammer et al., 2012). IT supports the idea for which companies, in order to survive, are obliged to act in a way which will grant them institutional legitimacy. CSR is, oftentimes, one of the tools chosen by companies in order to obtain such institutional legitimization, also through systematic compliance with informal institutions, such as national culture (Honkanen, 2013). It will therefore be important for firms to be able to align and “tune” CSR activities to the culture that permeates the company’s identity. Hence, CSR is conceptualized, in this context, as something more than a simple “business-centered” practice aimed at improving firms’ profit (Brammer et al., 2012) but rather as a tool which is often used to obtain institutional legitimization.

As stated above, institutional theory lends itself particularly well to analyze CSR diversity in the global arena. This diversity not only stems from different legislations and regulations (the so-called formal institutions) (DiMaggio & Powell, 1983) but it can be attributed to “far more important,

informal institutions such as religious norms, customary practices or tribal traditions” (Brammer et

(16)

CSR- Evolution and Definition of the Concept

The concept of CSR as we know it today first developed in the 1960s, simultaneously with increased societal attention to themes such as workers’ rights and environmental protection (Carroll, 2015). CSR was initially implemented on an entirely voluntary basis and mostly consisted of philanthropic activities (Carroll, 2015). Subsequently, in the 1970s a series of “social regulations” and agencies were set up in order to formalize the duties of companies in a more systematic manner (Carroll, 2015, p.88). CSR evolved during the 1980s and 90s thanks to the increased internationalization undergone by many companies. The concept of CSR was exported globally and had to be adapted to be responsive to the host environments; globalization leading to internationalization was therefore one the forces that actively shaped CSR and its evolution especially during the 1990s (Carroll, 2015).

Even if it is not the main focus of this thesis, it is worth mentioning that despite various attempts at defining it, literature is faced with a lack of universal and comprehensive definition for the concept of CSR (Hamidu et al., 2015). It is important to underline how scholars advise for “differentiated” frameworks to be used (van Marrewijk, 2003). Each firm can, therefore, decide which definition to apply when it comes to CSR in order to mirror its strategy and intentions (van Marrewijk, 2003). Nonetheless, to make the concept easily measurable, it is necessary to choose one of the many definitions. For the sake of this thesis, CSR will be defined as: “a concept whereby companies integrate social and environmental concerns in their business operations and in their interaction with their stakeholders on a voluntary basis” (European Commission, 2001, p.6).

Environmental CSR

(17)

Uhlenbruck, 2015; Ringov & Zollo, 2007; Ioannou & Serafeim, 2012 ) or simply “environmental CSR”.

The overwhelming majority of current literature measures how cross-country settings affect “Corporate Social Responsibility” or “sustainability”. These are both umbrella terms that include

environmental and social CSR without specifically investigating each and single one of them

(Hartmann & Uhlenbruck, 2015). This thesis, on the other hand, aims at shading light on the particular pillar of environmental CSR. A couple of other studies (Ioannou & Serafeim, 2012; Ringov & Zollo, 2007) mention the distinction between cross-country environmental and social CSR. However, in line with previous literature, these studies do not specifically address the two pillars separately but simply acknowledge their existence, and proceed to study them in combination, without any distinction. Additionally, most of the definitions of environmental CSR are contingent on the measurement used to quantify it. For instance, Ioannou & Serafeim (2012) only give a general and comprehensive definition for CSR and no specific definition of environmental or social CSR is given. The scholars, only briefly mention the two concepts simply because Asset 4 the Thomson Reuters ESG database (their reference CSR score database) does so. This thesis will therefore follow the definition suggested by Hartman et al. (2015) for environmental CSR: “set of firm policies and activities intended to protect the natural environment as well as their outcomes” (Hartmann & Uhlenbruck, 2015, p. 729).

2.1.3 Degree of Internationalization

(18)

Despite the competitive advantage that internationalization can provide, it has also been studied how this phenomenon can be a source of disadvantage. In the context of institutional theory, scholars have observed how firms engaging in internationalization can face additional challenges, conceptualized thought the construct of Liability of Foreignness (LOF) which can be defined as “all additional costs a firm operating in a market overseas incurs that a local firm would not incur” (Zaheer, 1995, p.343). It is argued that one of the main causes of the LOF can be found in the cognitive, normative and regulatory institutional distance between the home and host country (Eden & Miller, 2004). In order to partially mitigate the LOF, a series of strategies can be engaged, among them, the so-called institutional isomorphism (Maggio & Powell, 1983). Institutional isomorphism can be therefore deemed a valid approach for multinational companies in order to obtain institutional

legitimization, and subsequently mitigate their LOF, in any given institutional environment (Kostova

& Zaheer, 1999). In other words, the firm will have to engage in a series of strategies aimed at adapting to the local institutional environment (including local culture and practices) in order to acquire local legitimacy and obtain approval to operate in the local foreign context. Institutional

theory therefore supports the idea that CSR can be used as a tool get to legitimacy, through a process

of institutional isomorphism (Aluchna & Roszkowska-Menkesa, 2017). For internationalized firms, this process can be arduous, given the multiplicity of potentially contradicting settings in which the firm is embedded (Westney, 1993). It is therefore clear how, internationalization is a sort of "double-edged sword”: on the one hand, it might improve firms’ performance giving a competitive advantage, but on the other hand it might become a burden, forcing firms to seek legitimization in multitude of unfamiliar environments that will increase overall costs, therefore becoming a liability.

2.2 Hypotheses Development

2.2.1 Cultural Dimensions and CSR

(19)

Nonetheless, literature on the topic is still fairly recent and reveals mixed findings about which cultural dimensions can be used as a predictor for CSR performance (Miska et al., 2018). For instance, in their review on the topic, Miska et al. (2018) find that, while some studies found evidence for a positive impact of the Hofstede’s cultural dimension of Power Distance on CSR performance, others found the exact opposite effect. Uncertainty Avoidance and Masculinity similarly seem to have mixed effects (Miska et al., 2018). It has been hypothesized that the inconclusiveness of the results might stem from the fact that there is no agreed upon conceptualization for what CSR performance actually is. In fact, literature names the concept in various ways such as "sustainability" (Miska et al., 2018) “CSR commitment” (Peng et al., 2014) or “ethical policies of firms” (Scholtens & Dams, 2008).

For the sake of this thesis, the firms’ home country culture will be assumed to be the reference culture for each firm. This choice is justified by the fact that national culture has been recognized as a pervasive determinant factor for firms’ strategic decisions: “actions of corporate employees may have more to do with national cultural roots than organizational agendas” (Gulev, 2009, p.275). Furthermore, institutional theory (DiMaggio & Powell, 1983) recognizes how informal institutions, such as the home country culture, permeate the firm in every aspect. It could be argued that this is not the case for MNEs, since, given their internationalization, are prone to lose ties with their home countries costumes and values. However, it is worth mentioning how MNEs “manage to promote a unitary understanding of the organization” (Raitis, 2015, p.3) which is often characterized by the home country culture, leading to a so-called country of origin effect in the way companies identify themselves (Noorderhaven et al., 2003).

Cultural Dimensions and Environmental CSR

(20)

to reduce carbon emissions, but is unrelated to corporate philanthropy, a typical social activity" (Hartmann & Uhlenbruck, 2015, p.730).

An alternative literature stream (Husted, 2005; Park, 2007) focuses on the relationship between cultures and environmentalism. Nonetheless, these scholars seem to prefer a country level of analysis, investigating how specific cultural dimensions are linked to a country increased preoccupation and care for environmental protection. Specifically, Husted (2005) finds support for the role of culture in determining societal environmental values. The scholar determines that societies characterized by “low levels of power distance, high level of Individualism and low level of

Masculinity have higher social and institutional capacity” (Husted, 2005, p. 363) which seems to link

to a better environmental consciousness. Similarly, Park (2007) find evidence supporting the fact that low Masculinity and low Power Distance are related to a better score in the “Environmental Sustainability Index” (ESI) at the country level. It is important to note how both of these studies decide to focus on the country level of analysis and investigate how culture is correlated with different levels of “Environmental Sustainability Index". ESI is defined as an index that “benchmarks the ability of nations to protect the environment over the next several decades” (Esty, 2005, p.1). It is clear how this index is meant to measure an overall country tendency, and is not suited for the firm level of analysis.

Summing up, psychology literature has highlighted how different environmental attitudes significantly vary across culture on the individual level (Schultz, 2002). At a societal level, IB literature has investigated how different level of environmental concerns in a country can be predicted by cultural values (Husted, 2005; Park, 2007). At the firm level of analysis, the focus shifts on both

social and environmental CSR, without distinguishing between the two (Ioannou & Serafeim, 2012;

Ringov & Zollo, 2007) and as underlined by researchers, "less is known about the role of country-level factors shaping corporate environmental performance” (Hartmann & Uhlenbruck, 2015, p.729). This thesis aims to contribute to current literature by filling the understudied aspect of how cultural dimensions are linked to the environmental CSR at a firm level of analysis, relying on the assumption that home country values are the ones shaping and actively molding firms’ strategies.

Masculinity and Environmental CSR

(21)

cultural-green or sustainable value” (Husted, 2005, p.363). Moreover, interestingly, Hofstede himself in his book “Masculinity and Femininity: The Taboo Dimension of National Cultures” (Hofstede, 1998) presents the result of a survey aimed at measuring the extent to which each society is concerned with the preservation of the environment and its tradeoff with economic growth. The scholar decided to investigate the topic through a questionnaire answered by a representative sample of people from 20 countries (see Appendix A, table 1). Hofstede found a negative correlation between higher levels of

Masculinity and the preference for the preservation of the environment over economic growth, hence

validating its definition for the dimension, assigning to more masculine societies a preference for material success (economic growth) over quality of life (environmental preservation).

These results have been validated at the societal country level of analysis and, drawing from

institutional theory, it is fair to assume that, it might be applicable at the firm level as well, given that,

as previously argued, a country of origin effect exists. Referring back to the definition given by Hofstede, it is not far-fetched to hypothesize that firms originating from more masculine societies will be more competitive and oriented towards material achievements, likely neglecting aspects such as environmental performance, which might come at expenses of financial results. These firms, pressured by increased levels of competition and assertiveness would potentially resort to a more disinhibited and mastery-oriented use of natural resources, mostly finalized at achieving a financial return and not at the protection of the environment per se. On the other hand, firms from feminine societies, given their greater emphasis on aspects such as quality of life, would potentially place greater importance on matters such as environmental performance given the impact the latter has on general well-being and living conditions.

Therefore, it can be argued:

Hypothesis 1: The higher the Masculinity of the firm’s home country culture, the lower the firm’s environmental CSR will be.

Power Distance and Environmental CSR

(22)

Moreover, as shown by Waldman (2006), high PD cultures encourage managers to show little concern for stakeholders and their expectations, such as environmental performance. It has also been demonstrated how societies characterized by high PD have lower “social and institutional capacity for sustainability” (Husted, 2005, p. 349) potentially leading to a decreased concern for the environment.

The high respect for authority present in high PD societies possibly hinders the possibility for debate leading to weaker receptivity for issues such as environmental problems (Katz et al., 2001) possibly in a company context as well. Furthermore, high power distance cultures are more prone to “frequent corruption” and scandals are more easily covered up (Hofstede, 2001, p.9). This might discourage environmental CSR, given the fact that if an environmental scandal was to happen, it could be easily be swept under the carpet. Moreover, in a high PD society, public opinion would not dare voice their discontent regarding environmental performance given the superior hierarchical position that a company has compared to individuals.

Therefore, it can be argued:

Hypothesis 2: The higher the Power Distance of the firm’s home country culture, the lower the firm’s environmental CSR will be.

Individualism and Environmental CSR

(23)

the widespread opinions might often be cut out and anti-conformism might be discouraged (Husted, 2005). Therefore, in collectivistic countries, it will be harder to find environmental groups, movements or dissenting voices that are willing to demand firms to go beyond regulations and make a tangible effort for a good environmental CSR (Husted, 2005; Vachon, 2010). Similarly, at the firm level of analysis, literature (Ringov & Zollo, 2007; Ioannou & Serafeim, 2012) has highlighted how firms originating from more individualistic societies tend to have a better overall (social and environmental) CSR performance.

As institutional theory suggests, firms are entities that are closely linked, connected and influenced by the societies and institutional environment they operate in. An increased level of

collectivism, leading to a homogenous and univocal opinion on a matter such as environmental

performance at the societal level, might therefore translate at the firm level as well. A lack of external stimuli calling for an improvement in environmental performance might hinder the development of an effective and relevant environmental CSR strategy from the firms’ part. Furthermore, loyalty to the opinions determined by the “in-group”, typical of collectivistic societies, would discourage employees from suggesting new and improved environmental CSR strategies that might lead to better CSR performances.

Therefore, it is hypothesized:

Hypothesis 3: The higher the Individualism of the firm’s home country culture, the higher the firm’s environmental CSR will be.

Uncertainty Avoidance and Environmental CSR

This dimension “deals with a society's tolerance for ambiguity” (Hofstede, 2011, p.10) and measure how uncomfortable or comfortable people from a certain society will feel in a novel situation (Hofstede, 2011). Literature has found extremely contradicting results for what concerns the relationship between uncertainty avoidance and environmental performance. Both positive, negative and non-significant relationships have been found (Miska, Szőcs, & Schiffin, 2018 ).

(24)

performance designed to minimize the risk of scandals or problem of environmental nature involving the firm.

Therefore, it can be argued that:

Hypothesis 4: The higher the Uncertainty Avoidance of the firm’s home country culture, the higher the firm’s environmental CSR will be.

2.2.2 The Moderating Effect of the Degree of Internationalization

As previously mentioned, MNEs “manage to promote a unitary understanding of the organization” (Raitis, 2015, p.3) through a series of founding values, practices and a common culture, a phenomenon known as country of origin effect (Noorderhaven & Harzing, 2003). Nonetheless, it is often underlined how important differences within subsidiaries of the same MNE do exist, originating from “adaptation to local circumstances" (Noorderhaven & Harzing, 2003, p.50).

It is possible to extend this logic to CSR, and specifically to environmental CSR. In fact, according to Campbell et al. (2012), CSR activities targeted at host countries are often engaged by companies in order to cope with the liability of foreignness caused by a foreign institutional and cultural context, making CSR an important tool to acquire legitimization in the host countries. Internationalized firms will have to therefore align with a variety of institutional settings (what scholar define as multiple embeddedness) which poses a significant challenge in terms of CSR, given the likely divergence from the host and home countries leading to potentially contradicting forces of

institutional isomorphism (Marano & Kostova, 2016). These divergent institutional contexts, might

(25)

will therefore negatively moderate the relationship between home country culture and environmental CSR activities.

Therefore, it can be hypothesized that:

Hypothesis 5: The negative relationship between home country level of Masculinity and environmental CSR is negatively moderated by the degree of internationalization of the firm.

Hypothesis 6: The negative relationship between home country level of Power Distance and environmental CSR is negatively moderated by the degree of internationalization of the firm.

Hypothesis 7: The positive relationship between home country level of Individualism and environmental CSR is negatively moderated by the degree of internationalization of the firm.

Hypothesis 8: The positive relationship between home country level of Uncertainty Avoidance and environmental CSR firm is negatively moderated by the degree of internationalization of the firm.

(26)

3. RESEARCH METHODOLOGY

In this section, the methodology behind this research will be explained. Firstly, the sources for the data will be introduced, subsequently the sampling method will be presented and finally, the measures for the variable will be illustrated.

3.1 Data Collection

This thesis chooses a quantitative approach in order to explore the relationship between cultural dimensions, environmental CSR and the moderating effect of internationalization. To investigate the hypothesized relationships, secondary data, both publicly available and provided by the University of Groningen are used. Values of dependent and control variables were lagged by one year in order to control for possible reverse causality (dependent variable is for the year 2018 while the control variables are for the year 2017). The data for the independent variable (Hofstede’s cultural dimensions of Individualism, Uncertainty Avoidance, Power distance, Masculinity vs Femininity) is retrieved to the publicly accessible website: www.hofstede-insights.com. Data regarding the dependent variable environmental CSR is obtained through the Thomson Reuters Asset 4 ESG database, only scores regarding the environmental pillar are taken into consideration for the purpose of this thesis. Data regarding the moderator degree of internationalization are retrieved from Orbis database. For what concerns the control variables of firm size, firm profitability and industry data are retrieved through the Orbis database while data regarding GDP/capita are retrieved from the World Bank website.

3.2 Sample

(27)

minimum of 90 observation should be included. Additional 10 observations are included, adding up to a sample of 100 firms.

3.3 Measures

3.3.1 Dependent Variable- Environmental CSR

Literature has reached no definitive conclusion regarding the best method to measure CSR. In order to measure and operationalize CSR, therefore a variety of indicators exist. At the firm level, the most common indicators are composite indexes created by CSR analyst organizations that use both publicly and privately disclosed information in order to create an evaluation for the CSR performance for each firm (Paredes-Gazquez et al., 2016). In fact, academic works seem to mostly refer to measurements provided by rating agencies such as MSCI ESG data (former KLD) or Thomson Reuters Asset4 ESG database when confronted with the challenge of operationalizing CSR. KLD STATS (statistical tool for analyzing trends in social and environmental performance) are the social, environmental and governance performance indicators of the biggest 3000 US firms (KLD & STATS, 2010) which have been generally deemed methodologically valid (Callan & Thomas, 2009). One major issue with KLD indicators is that they are only available for US firms (KLD & STATS, 2010), making it non-suitable for cross-country comparison. Therefore, for the sake of this thesis, an alternative measure has been found on the Thomson Reuters Asset 4 ESG ratings, available for 6000 companies globally, measuring the social, environmental and governance aspects of CSR (Thomson Reuters, 2018). It is clear how, for this thesis, using Thomson Reuters Asset 4 ESG ratings will be more convenient given the representative cross-cultural sample available and the specific measurement for the environmental pillar.

(28)

incorporates three main aspects of environmental CSR, namely: Resource use, Emissions, and Innovation in the field of environmental sustainability.

3.3.2 Independent Variables- Hofstede’s Cultural Dimensions

Consistent with prior literature (Husted, 2005; Vachon, 2010; Park, 2007; Gallego-Álvarez & Ortas, 2017) this thesis will use Hofstede’s cultural dimension to operationalize culture. As previously mentioned, Hofstede’s framework keeps being extensively used and cited in academia, despite being first introduced almost 40 years ago (Beugelsdijk et al., 2016). In this thesis, only four (Individualism,

Uncertainty Avoidance, Power distance, Masculinity vs Femininity) of the total six dimensions are

considered. As previously stated, this choice can be justified by the fact that the additional dimensions of Indulgence vs Restraint and Long-Term orientation have limited data availability. Each of the considered cultural dimension uses a single country as a level of analysis, the scores range from 1 to 100, 100 representing the possible maximum score.

3.3.3 Moderator- Firm’s Degree of Internationalization

Literature underlines how “there are strong differences of opinion as to what indicators are suitable for measuring corporate internationalization” (Dorrenbacher, 2000, p.120). Mainly, these indicators are either “structural” (e.g.: the proportion of foreign assets or subsidiaries) “performance” oriented (e.g.: foreign sales) or “attitudinal” (e.g.: international experience of top managers), oftentimes these indicators are combined in order to create composite and more comprehensive indexes (Dorrenbacher, 2000). Despite the fact that a composite “performance” and “structural” index integrating measures such as proportion of foreign sales and foreign affiliates would be a more accurate metric to assess the degree of internationalization, time constraints and difficulties in finding data, make the task of creating a composite internationalization index of this sort quite hard. Hence, in this thesis, the structural indicator of “proportion of foreign affiliates” (number of foreign-owned

subsidiaries/ total number of subsidiaries) (Dorrenbacher, 2000; Ramaswamy et al., 1996) will be

used to determine the degree of internationalization of a firm. Additionally, as suggested by Grüning et al. (2011) the number of foreign countries in which the company operates in is also taken into consideration. A single indicator integrating both standardized values of proportion of foreign

affiliates and number of foreign countries in which the company operates is therefore created to

(29)

3.3.4 Control Variables

Firm size: Firm size will be included in the control variables given that bigger firms are exposed to a

multitude of stakeholders and therefore will be subject to more expectations to what concerns CSR activities (Gallego-Álvarez & Ortas, 2017). It might be expected that environmental CSR will be affected by the size of the firm and bigger firms will engage in a more extensive environmental CSR. Literature has underlined how bigger firms are characterized by an increased visibility and therefore they will benefit from an increased reputational and legitimacy enhancing effects thanks to the CSR activities, including the environmental one (Udayasankar, 2008). This variable will be measured through the number of employees of the company, consistent with Ali et al. (2014).

Firm profitability: Financial performance and CSR activities are found to have a correlation in

previous literature (e.g: Campbell, 2007; Ioannou & Serafeim, 2012). It has to be expected that more profitable firms will have the economic resources to engage in more environmental CSR activities, therefore it is important to control for this aspect. Given the fact that different industries (which might be more or less capital-intensive) are analyzed, ROA (Return on Asset) might not be considered a suited indicator to analyze firm profitability. In fact, research has shown that, on average, companies operating in highly capital-intensive industries tend to show lower ROAs (Selling & Stickney, 1989). Therefore, ROE (Return on Equity) is deemed a more accurate measure for cross-industry profitability comparison.

Industry: Industry plays a major role in how firms decide to face environmental CSR, in fact,

companies operating in services will be less exposed to issues such as carbon emission and resource use compared to companies operating in manufacturing. Companies are first categorized based on the “European Classification of Economic Activities” (NACE) (see Appendix B; table 1). Subsequently, a dummy variable is created in order to control for the industry related effect (see Appendix B; table 2).

GDP/capita: National wealth has been found to have a significant relationship with the ESI index

(30)

3.4 Empirical Data Analysis

The mentioned hypotheses are tested performing a multiple linear OLS (Ordinary Least Squares) regression using the software STATA (version 15.0). In order to simplify the interpretation of the results, all of the independent variables and the moderator are mean centered.

Before conducting the regressions, the assumptions regarding the multiple regression analysis are tested. In order to check for normality of the variables, the Shapiro-Wilk test and a graphical representation of the Kernel Density Estimation are used. None of the variables seem to be normally distributed. However, normal distribution of the variables it is not one of the main pre-assumptions in order to conduct a multiple regression and therefore should not pose a problem in our analysis (Li et al., 2012). Normality of the residuals is, on the other hand, a key pre-assumption that has to be met. In order to meet this condition, the variable environmental CSR is square transformed:

environmental CSR squared= (raw environmental CSR)2_{, as suggested by the STATA tool}

“ladder” (STATA, 2018). After the transformation of the variable, the normality of the distribution of its residuals it is checked through the Shapiro-Wilk test, Skewness and Kurtosis test (Appendix C,

table 1) and with the graphical Kernel Density Estimation (Appendix C, table 2). Despite a slight

negative skew (-0.311) and an imperfect kurtosis value of 2.66 this distribution can be considered roughly normally distributed, also given the Shapiro Test result (p=0.215) allowing to accept the normality of the distribution hypothesis. The linearity assumption is checked through a series of scatter plots that confirm no major departure from this assumption. Possible outliers are checked using Cook’s distance (see Appendix C, table 3). No potential outliers are found given the fact that all of the cases show a value <1 (Lomax & Hahs-Vaughn, 2013). Heteroskedasticity of the residuals is checked for through White test, the value of 0.330 (see Appendix C, table 4) allows to confirm the presence of homoskedasticity and therefore no correction is needed. Multicollinearity is tested using the variance inflation factor (VIF), all of the factors present a VIF below the value of 10, and the tolerance level (1/VIF) is always bigger than 0.10 (Acock, 2010); therefore, multicollinearity can be excluded (see Appendix C, table 5). Finally, the Durbin-Watson test confirms the independence of errors, given the fact that its value is close to 2 (Appendix C, table 6).

4. RESULTS

(31)

4.1 Descriptive Statistics

The sample consists of 100 companies randomly chosen from the Forbes 2000 list. The MNEs originated from 30 different countries (see Appendix D) operating in 11 different industries which are coded 1 for services and 0 for manufacturing (see Appendix B, table 2). This coding is chosen in light of the obvious effects on environmental CSR caused by the nature of manufacturing activities, which are intrinsically more polluting than service activities. Table 1 presents the descriptive statistics; it can be noticed how the environmental CSR varies substantially among countries, from a minimum of 15.42 to a maximum of 99.25. Nonetheless, the mean is of 76.16, indicating that the distribution is negatively skewed, leaning towards the right (see Appendix E, table 1). Hofstede’s dimensions of Power Distance ranges from a minimum of 11 to a maximum of 93, with an average of roughly 50, nonetheless, a standard deviation of 20.86 indicates that the data are quite dispersed.

Individualism seems to follow a similar pattern (min:13, max:91, SD: 21.32), but its average it is

slightly higher (61.87). Masculinity, once again seems to follow a similar pattern for which the dimension ranges from 5 to 95 with an average of roughly 50, but a high standard deviation (21.77) indicating a high dispersion in the data. Finally, Uncertainty Avoidance has the smallest rage of all of the dimensions (min:23, max:95) with the highest average and standard deviation (mean:63.10, SD: 23.20). The moderator degree of internationalization is first analyzed in its two unstandardized components (number of countries in which the MNE operates and its percentage of foreign

subsidiaries): the number of foreign countries in which the sample MNEs operate varies greatly,

ranging from 1 to 100, nonetheless, an average of roughly 34 indicates an extreme positive skewness, leaning towards the left. The percentage of foreign subsidiaries is highly heterogenous, ranging from less than 1% to almost 100% (min: 0.024, max: 0.977, mean: 0.588) the standard deviation of 0.25. Subsequently, the resulting composite standardized internationalization degree is analyzed and ranges from -3.59 to 3.59 being centered on 0 and with a standard deviation of 1.81, all in all, this is the variable with the most symmetrical distributed variable. Finally, the control variable of GDP per

capita (PPP) assumes the minimum value for India (7 055 $) and its maximum value for Norway (61

(32)

Table 1: Descriptive Statistics

Variable Min Max Mean SD

Environmental CSR (raw) 15.42 99.25 76.16 18.27 Environmental CSR (squared) 237.87 9851.22 6132.21 2436.70 PD 11 93 50.09 20.86 IDV 13 91 61.87 21.32 MASC 5 95 50.49 21.77 UA 23 95 63.10 23.20 Number of Foreign Countries in which the MNE operates in 1 100 34.91 25.06 Percentage of foreign subsidiaries (%) .024 .977 .588 .251 Internationalization degree (standardized) -3.59 3.59 0 1.81 GDP per Capita 7 055 61 504 40 230.6 14 758.5

Firm size (employees) 17 300 642 300 329 800 116 649.4

Firm Profitability (ROE)

-88.157 494 16.99 51.19

(33)

4.2 Correlation

In the correlation matrix (Table 2) it can be observed how environmental CSR significantly correlates with the Hofstede’s dimension of Individualism (0.236**). Furthermore, environmental CSR significantly correlates with the moderator internationalization degree (0.320**) and the control variable GDP/capita (0.409***).

A strong significant correlation (-0.578**) between the dimensions of Power Distance and

Individualism can be observed; this is consistent with Hofstede’s own remarks regarding these

dimensions. The Dutch scholar is in fact aware of this correlation, but nonetheless decides to treat these dimensions as separate concepts (Hofstede, 2014). Furthermore, Hofstede also underlines how these dimensions are significantly correlated with wealth (Hofstede, 2014). Evidence supporting this claim is also found in this correlation matrix in which PD significantly and strongly correlates with GDP/capita (-0.674***) just like Individualism (0.732***). In spite of these multiple correlations, Hofstede defends this classification by stating that “if we compare rich with rich and poor with poor societies, the negative correlation between IDV and PDI almost disappears” (Hofstede, 2014, p.8). It is important to underline that even these two dimensions are correlated they do not cause multicollinearity as it is shown in the VIF table presented in Appendix C, table 5. The weaker correlation between GDP/capita and Masculinity (-0.198**) and Uncertainty Avoidance (-0.168*) seem to confirm past literature’s findings, for which these dimensions are the most stable over time, given the fact that they are only marginally correlated with wealth (Cox et al., 2011).

(34)

Table 2: Correlation Matrix

Note: N=100, p<0.01***, p<0.05**, p<0.10* Environmental

CSR

(35)

4.3 Regression Results

Table 4 shows the results of the regression analysis. Model 1 presents the effects of the control variables on the dependent variable environmental CSR (squared). Values of dependent and control variables were lagged by one year in order to control for possible reverse causality (dependent variable is for the year 2018 while control variables are for the year 2017). It can be observed how, only two of the control variables result statistically significant. In particular, GDP per capita (log) shows extreme positive significance (p=0.000), along with the industry variable (p=0.070). It appears that companies originating from wealthier countries and operating in services tend to exhibit higher levels of environmental CSR (squared). On the other hand, both firm profitability (p=0.321) and especially firm size (p=0.818) do not appear to be statistically significant. For this reason, the control variable firm size it is excluded from the analysis given the neglectable impact in the regression model.

Model 4 explores the effect of Masculinity on the dependent variable which was hypothesized to be negative. Contrary to expectations, no statistically significant results were found (p=0.778) and the mixed effects (both positive and negative) of Masculinity in model 7 and model 10 firmly reject hypothesis 1.

Model 3 aims at measuring the effect of the dimension Power Distance, which according to hypothesis 2 should negatively impact environmental CSR. Despite the negative coefficient (B=-11.61) this effect is not statistically significant (p=0.412). A weak negative statistically significant effect of Power Distance on environmental CSR can be witnessed in model 9 (B=-26.58; p<0.1) but nonetheless hypothesis 2 cannot be accepted.

Model 2 individually integrates the independent variable Individualism to the control variables. Individualism seems to be positively statistically significant in the model 2 (p=0.011; B=37.73) as well as in model 6,7 and 8, therefore supporting hypothesis 3, which can be accepted.

(36)

reason for which internationalization degree would be better fitted as an independent/control variable rather than a moderator, furthermore, model 7 also has the highest Adjusted R2 _(0.237).

Internationalization degree as a moderator (shown from model 8 to model 11) does not show

any statistically significant effect, nonetheless a negative interaction effect with Power Distance (p=0.182; B=-8.93) Masculinity (p= 0.465; B=-5.75) and Uncertainty Avoidance (p=0.150; B=-7.61) is present, even if non-significant. For this reason, hypotheses 5 6 7 and 8 cannot be accepted.

Table 3: Variations in R2 _{and Adjusted R}2

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7

R2 _.186 _.239 _.191 _.186 _.205 _.246 _.287 Change in R2 compared to the previous model +0.053 -0.048 -0.005 +0.019 +0.041 +0.041 Adjusted R2 _.152 _.207 _.157 _.152 _.171 _.182 _.237 Change in Adjusted R2_{compared to the} previous model +0.052 -0.053 -0.005 +0.019 +0.011 +0.055 .

Table 3 presents the variations in R2 _{and Adjusted R}2 _{of the models presented in a hierarchical}

fashion. It can be noticed how both R2 _{and Adjusted R}2 _{improve as the independent variable IDV is}

added in model 2. R2 _{and Adjusted R}2_{seem to decrease when PD (model 3) and MASC (model 4)}

are added only to slightly improve in model 5, testing for UA, which despite a non-significant effect shows a p=0.128, which can be considered borderline significant. Model 6, integrating all of the independent and control variables seem to have an improved R2 _{and Adjusted R}2_{, which is furtherly}

(37)

Table 4: Regression Analysis

Environmental CSR (squared)

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Model 8 Model 9 Model 10 Model 11 Independent Variables IDV 37.73** (14.41) 33.51** (12.09) 33.86** (16.12) 38.32** (13.51) PD -11.61 (14.36) -.798 (15.14) -8.34 (15.04) -26.58* (15.67) MASC 2.98 (10.55) -1.95 (10.90) 2.59 (10.79) 7.34 (11.21) UA 14.93 (9.72) 3.04 (12.09) -1.37 (11.93) 15.40 (9.54) Moderating Variables Internationalization Degree 341.08** (141.12) 321.25** (134.91) 382.38* (143.11) 313.64** (140.46) 278.63* (138.07) Interactions IDV x Internationalization 1.292 (6.29) PD x Internationalization -8.93 (6.64) MASC x Internationalization -5.75 (6.95) UA X Internationalization -7.61 (5.24) Control Variables Firm Size .0001 (.000)

Excl. Excl. Excl. Excl. Excl. Excl. Excl. Excl. Excl. Excl.

Firm Profitability 4.38 (4.46) 4.25 (4.31) 4.74 (4.47) 4.31 (4.46) 5.29 (4.44) (4.47) 4.51 5.73 (4.41) 5.54 (4.26) 6.46 (4.39) 5.53 (4.40) 6.75 (4.40) Industry 841.33* (499.80) 997.71** (484.76) 803.01* (576.86) 815.30* (498.05) 889.86* (492.72) 1005.66* (498.00) 1170.00** (496.12) 1217.98** (485.15) 937.26** (494.41) 968.29* (501.16) 1022.27* (491.42) GDP/Capita (log) 1811.34*** (434.21) 2730.18*** (547.74) 2123.44*** (576.86) 1836.34*** (440.55) 1820.68*** (428.95) 2648.32*** (726.28) 2530.51*** (714.43) 2226.83*** (601.96) 1670.28*** (583.40) 1393.46*** (478.18) 1284.76*** (470.12) Constant -13261.22** (4564.79) -22888.47*** (5767.30) -16472.21*** (6061.81) -13435.82*** (4629.85.74) -13327.79** (4510.63) (7565.94) -23202.6** -20751.07*** (7440.18) -17676.8*** (6346.24) -11645.28* (6136.08) -8916.19** (5011.32) -7747.98 (4928.68) R Square .186 .239 .191 .186 .205 .246 .287 .287 .253 .235 .260 Adjusted R Square .152 .207 .157 .152 .171 .182 .237 .232 .205 .185 .212 F-stat 5.45*** 7.48*** 5.61** 5.43** 6.13*** 6.13*** 6.14*** 4.56*** 5.26*** 4.76*** 5.46***

N=100; Standardized Errors in parenthesis; ***p£0.01 **p£0.05 *p£0.1 Standardized Beta coefficients:

(38)

Table 5: Robustness test

Environmental CSR reporting

N=97 Standardized Errors in parenthesis; ***p£0.01 **p£0.05 *p£0.1

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Model 8 Model 9 Model 10 Model 11 Independent Variables IDV -.001 (.003) -.003 (.004) -.003 (.004) -.0005 (.003) PD -.003 (.003) -.004 (.003) -.004 (.003) -.002 (.003) MASC .002 (.003) .002 (.002) .002 (.003) .001 (.003) UA -.004* (.002) -.005* (.002) -.005* (.002) -.004* (.033) Moderating Variables Internationalization Degree -.0009 (.034) -.023 (.034) -.002* (.001) -.021 (.034) -.250 (.333) Interactions IDV x Internationalization .0001 (.001) PD x Internationalization -.002 (.001) MASC x Internationalization -.0003 (.001) UA X Internationalization -.002 (.001) Control Variables

Firm Size 6.58e-07

(4.53e-07) Excl. Excl. Excl. Excl. Excl. Excl. Excl. Excl. Excl. Excl.

The Effects of National Culture on Firms’ Environmental CSR and the Moderating Role of the Degree of Internationalization