! !
! !
!
Explaining the Relationship between Cultural Distance and
Stakeholder Criticism
! ! ! ! ! ! ! ! ! ! ! !!Student: Nikkie Koop 10004473
University of Amsterdam, Faculty of Economics and Business Supervisor: Dhr. Dr. D.A. Waeger
University of Amsterdam, Amsterdam Business School Word count: 11620
Submitted: 28/06/15 Final deadline: 29/06/15
! "! Statement of originality
This document is written by student Nikkie Koop who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
! #!
Abstract
This research was conducted to examine the relationship between the accumulated distance of all home-host country pairs of a multi-national enterprise (MNE) and the amount of stakeholder criticism (with respect to CSR-related issues) the MNE is exposed to. In addition, the research took the possible moderating effects on this relationship of home country cultural variance and intra-firm cultural diversity in account. A negative binominal regression was conducted to analyse the above mechanisms. The findings suggest that there is a significant relationship between the cultural distance an MNE is exposed to and the amount stakeholder criticism the MNE receives. Contrary to the expectations, no moderating effect of cultural variance neither at the country level nor at the intra-firm level is found in this study.
! $!
Table of content
Abstract 3
I Introduction 5
II Literature review
1. Institutional theory and legitimacy
2. Corporate social responsibility legitimacy
3. Institutional distance, cultural distance and MNEs
7 7 8 9
III Theoretical framework 13
IV Methodology
1. Data collection and sample 2. Independent variable 2.1 Cultural distance 3. Dependent variable 3.1 Stakeholder criticism 4. Moderating variables 4.1 Cultural variance
4.2 Firm-level cultural diversity 5. Control variables 18 18 19 19 21 21 22 22 23 24 V Results 1. Descriptive statistics 2. Correlation analysis
3. Negative binominal regression 3.1 Model 1: control variables
3.2 Total criticism and absolute cultural distance 3.3 Total criticism and weighted cultural distance
25 25 27 30 30 31 32 IV Discussion 1. Findings
2. Limitations and further research
36 36 39 VII Conclusion 41 References 43 Appendix I 47 Appendix II 48
! %!
I Introduction
Last decades there has been a growing academic interest in the multinational enterprise (MNE) in its complex environment. Copious research is based on institutional theory (Aguilera-Caracuel et al., 2014; Hillman and Wan, 2005; Kostova et al., 2008; Kostova and Zaheer, 1999; Meyer and Rowan, 1977; Rosenzweig and Singh, 1991; Scott, 1995; Xu and Shenkar, 2002; Rodriquez et al., 2006) since it provides a broad theoretical base for explaining the effect of the environment on the MNEs subsidiaries. According to institutional theory, organizations must conform to the rules and belief systems prevailing in the environment to achieve and maintain environmental legitimacy (Kostova and Zaheer, 1999a). Current research focuses mostly on the influence of the institutional distance on the strategic behaviour of the firm (Kostova and Zaheer, 1999a; Xu and Shenkar, 2002). They argue that institutional distance triggers and enhances the conflicting demands for local responsiveness and global integration. Thus, institutional distance is an important construct for explaining, for example, the foreign investment decisions of a MNE. Other studies acknowledge the importance of normative and cognitive factors in the environment of the organization for explaining its behaviour (Arslan and Larimo, 2010; Kostova et al., 2008; Kostova and Zaheer, 1999b). They argue that a form of cultural distance is the main influence in the strategic behaviour of the firm. Overall, the majority of the studies focus on how either cultural or institutional distance affect the strategic behaviour of the firm.
However, so far none of the studies continued the work of Kostova and Zaheer (1999) on explaining how cultural distance influences the legitimacy of MNEs. Cultural distance has been used in relationship with performance or strategic behaviour of the firm, but there are no studies on the effect of cultural distance on legitimacy of the MNE. According to the institutional theory, environmental legitimacy is crucial for an organization to survive
! &! (Kostova and Zaheer, 1999b), thus it is a crucial topic to study. Also, studies have shown that cultural distance can have a significant effect on the behaviour of the firm (Brouthers and Brouthers, 2001; Kogut and Singh, 1988; Shenkar, 2001), thus it is important to explore the possible effect of cultural distance on the legitimacy of the firm. Next to contributing to the existing literature on MNEs, this thesis is also useful in practice. It can help managers by informing them about the possible legitimacy problems they might face when investing in countries with high cultural distance.
This research has the overarching goal to examine the relationship between the accumulated home-host cultural distance of the MNE and the total amount of stakeholder criticism (with respect to CSR related topics) it is exposed to. Based on past research, this study hypothesizes that an increase in cultural distance leads to a greater amount of stakeholder criticism that the MNE is exposed to. Since cultural variance in the MNE’s home country could increase its resources and capabilities in maintaining legitimacy, its moderating effect on the relationship between cultural distance and the amount of criticism will also be tested. And finally, the moderating effect of intra-firm diversity will be tested, to see whether this could have a weakening effect on the relationship between cultural distance and the amount of stakeholder criticism
The study will first provide a critical overview of the existing literature on the topic of MNEs in their complex environment and on legitimacy. Following this literature review, a theoretical framework is presented on which the hypotheses will be based. Finally, the results of the negative binominal regression will be presented and findings will be discussed. Finally, an overview of the limitations of this research will follow with advice for future research.
! '!
II Literature review
1. Institutional theory and legitimacy
A critical issue in the international business research is the organizational legitimacy of multinational enterprises (MNEs). A large amount of research on MNEs is based on institutional theory, since it provides a clear theoretical foundation for understanding how organizational legitimacy is affected. Institutional theorists argue that, in order to survive, organizations must conform to the rules and belief systems prevailing in the environment to achieve and maintain environmental legitimacy (Kostova et al., 2008; Kostova and Zaheer, 1999; Xu and Shenkar, 2002). This core idea of institutional theory has a great overlap with legitimacy theory, which states that in order to maintain its legitimacy, a firm has to conform with or attempt to alter social perceptions, expectations or values in their environment (Brown and Deegan, 1998; Gary O’Donovan, 2002; Pittroff, 2014). Legitimacy 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’ (Suchman, 1995: 574). When an organization lacks legitimacy, it can become the subject of attacks from stakeholders (King, 2007; Mitchell et al., 1997). Stakeholders are part of the environment of the firm and are identified through the actual or potential harms and benefits that they experience as a result of the firm’s actions or inactions (Donaldson and Preston, 1995). Thus, a firm has to fulfil the expectations of the stakeholders and maintain its environmental legitimacy; otherwise it risks becoming the subject of stakeholder criticism. This relationship between the MNE and its environment can be seen as a social contract, which means that the survival of the MNE depends on the boundaries and norms of the society. When the MNE meets the requirements of the society, it has a license to operate and continue its practices in the future (Brown and Deegan, 1998; Pittroff, 2014). When the MNE does not conform to the environment’s requirements, it breaches the social contract and a
! (! legitimacy gap emerges that is followed by stakeholder criticism (Brown and Deegan, 1998). This stakeholder criticism can take the form of extra-institutional tactics, such as protest or boycott and can have a negative effect on the financial performance of the firm (King, 2007, 2008).
MNEs engage in distant institutional environment and different cultural contexts, which means they will be exposed to multiple contradictory stakeholder demands. The assumption that companies should conform to the stakeholder expectations to maintain their legitimacy becomes very problematic because of the expansion of corporate activities into different countries and cultures. Palazzo and Scherer (2006) show that on the global level the idea of conformity to norms and rules of social communities is difficult to comprehend. They state that on the global level, there are no broadly accepted normative standards, neither in legal, nor in moral terms (Scherer & Palazzo, 2006). This growing complexity of the globalized world leads to a multiplicity of contradictory legal and moral requirements for MNEs which makes it even harder to establish legitimacy (Scherer & Palazzo, 2006).
2. Corporate social responsibility and legitimacy
As stated earlier, a firm is perceived as legitimate when the actions of the firm are desirable, proper, or appropriate within some socially constructed system of norms, values, beliefs, and definitions’ (Suchman, 1995: 574). Corporate social responsibility (CSR) of MNEs is an important issue concerning the legitimacy of the firm. CSR research focuses on what role corporations should play and in what CSR activities they should engage in order to create legitimacy (Palazzo and Scherer, 2006). Overall, institutional theory sees CSR as the result of a process whereby stakeholders (e.g. NGOs) put pressure on firms to adopt given social practices and apply legal, social and economic penalties to non-adopters (Misani, 2010). This means that legitimacy is provided by the institutional environment of the organization.
! )! According to Palazzo and Scherer (2006), current research is too much based on a positivist’s research paradigm to explain the role of business in society. This positivist view on CSR assumes that it is possible to develop a universal view on the creation and sustainability of legitimacy of a company. However, this view shows to be problematic in the case of MNE’s legitimacy, ‘[…] because in the end, they cannot develop a universal reference point from which to assess the ethical acceptability of norms or actions’ (Scherer & Palazzo, 2007: 1099). Scherer and Palazzo (Scherer and Palazzo, 2007) again show that because of the global character of MNEs it is difficult to understand and comply to the different stakeholder demands in different cultural contexts.
3. Institutional distance, cultural distance and MNEs
Because MNEs operate in multiple external environments and thus experience multiple institutional pressures, achieving and maintaining legitimacy is one of the most critical issues for MNEs. MNEs can face conflicting demands for external legitimacy in the host country and for internal consistency within the MNE system (Xu and Shenkar, 2002).
Authors in this research field provide different reasons for the possible legitimacy problems of MNEs, based on this theoretical background. Kostova and Zaheer (1999) argue that high complexity in the institutional environment, the organization and the process of legitimation challenges MNEs in establishing and maintaining legitimacy. In their article, Kostova and Zaheer (1999) made the first important steps in developing a theory of MNE legitimacy. They draw from institutional theory suggesting a set of institutional domains based on the three pillars of institutional environments suggested by Scott (1995): the regulatory domain, the normative domain and the cognitive domain. These domains form the complex environment of the MNEs and enhance the difficulty of maintaining legitimacy. The
! *+! regulatory domain is composed of regulatory institutions, meaning the rules and laws that exist in the society. Firms have to comply with these rules and laws in order to be legitimate, although in the long run firms often try and succeed in influencing the regulatory domain through interest intermediation (Kostova and Zaheer, 1999a). The normative domain consists of the existing wider societal values. Legitimacy exists when the organizational values overlap with the societal values and cultural support for the organization is established in the society (Kostova and Zaheer, 1999a). To be legitimate in the cognitive pillar, firms have to conform to what is ‘taken for granted’ in the society. What is taken for granted is decided by the society’s established cognitive structures (Kostova and Zaheer, 1999a). Kostova and Zaheer (1999) propose that, compared to the regulatory domain, the cognitive and normative domains of the institutional environment present a greater challenge to the MNE in establishing their legitimacy. This is because of the tacitness of the cognitive and normative domain and the difficulty to influence them. These domains together can be seen as a construct of the national culture the MNE is operating in.
Xu and Shenkar (2002) draw on this by proposing a framework that explains foreign direct investment by the MNE by focusing on institutional distance. They try to explain MNE behaviour using this construct by arguing that a large institutional distance triggers and enhances the conflicting demands for local responsiveness and global integration (Xu and Shenkar, 2002). When the difference between the host country and the home country is significantly high in all three pillars proposed by Kostova and Zaheer (1999), this might lead to diminishing legitimacy. Xu and Shenkar (2002) claim that because of the important effect of institutional distance on legitimacy of the MNE, this construct should be taken into account when explaining MNE behaviour.
Kostova and Zaheer (1999) thus implicitly discuss the concept ‘culture’, but so far the exact link between legitimacy and cultural distance has not been empirically researched yet.
! **! However, there is a growing literature on cultural distance and the effect on MNEs actions and performance (Kogut and Singh, 1988; Tihanyi et al., 2005). Kogut and Singh (1988) investigate how characteristics of national cultures influence the selection of entry modes of MNEs. They argue that because differences in national cultures have been shown to result in different organizational practices and employee expectations, it can be expected that the more culturally distant two countries are, the more distant their organizational characteristics are on average (Kogut and Singh, 1988). Thus, they stress the importance of cultural distance on strategic behaviour of the firm. Following Hofstede’s (1980) classification of culture, Kogut and Singh (1988) created a cultural distance index which measures the cultural distance between two countries. However, it solely focuses on the MNE’s strategic behaviour and does not explain the effect of cultural distance on legitimacy of the MNE. Most of the literature in cultural distance focuses on explaining foreign market investment location by MNEs, predicting the choice of mode of entry and effect on the performance of MNE (Shenkar, 2001). Kostova and Zaheer (1999) took the first steps in proposing a relationship between the normative and cognitive environment and the legitimacy of an organization.
We have seen that the concept of institutional distance is often used to explain MNE behaviour. Other researchers that focused on this are Arslan and Larimo (2010), who distinguish normative institutional distance from regulative distance. Regulative institutional distance is the difference in the legal institutions regulations in a MNEs home country and host country, whereas normative institutional distance refers to the difference in informal rules and expectations. These differences, defined by normative and regulative institutional distance, explain strategic behaviour of the MNE, according to Arslan and Larimo (2010). Most literature written on this topic is about the influence of these sorts of distances on the MNE behaviour. Aguilera-Caracuel et al. (2014) for example argue that institutional distance among countries influences environmental strategies of MNEs. Very few authors, however,
! *"! follow the footsteps of Kostova and Zaheer (1999) in trying to explain the effect on legitimacy of the MNE, while this is, as stated earlier, particularly interesting.
In this thesis I will try to fill this gap by answering the research question: what is the relation between cultural distance and legitimacy of the MNE? Based on the above discussed literature I hypothesize that the more a company is active in distant cultural context, the more it will be exposed to different demands. My expectation is that the larger the cultural distance, the more the firm’s legitimacy will diminish and the more it will be criticized for its overall activities. According to institutional theory, when an organization is no longer seen as legitimate, it will not survive. Thus, it is important to understand the link between cultural distance and the possible effect it may have on the legitimacy of the MNE. Also, the outcome of this thesis can be helpful in practice by informing MNEs over the possible risks of foreign investments with a large cultural distance.
! *#!
III Theoretical framework
In this part of the thesis three main hypotheses will be developed proposing an answer to the research question: what is the effect of cultural distance on the legitimacy of the MNE? Institutional theory can provide a clear theoretical framework in explaining this effect and formulating the first hypothesis. According to institutional theory, organizations must conform to the rules and belief systems prevailing in the environment to achieve and maintain environmental legitimacy (Kostova et al., 2008; Kostova and Zaheer, 1999; Xu and Shenkar, 2002). Building on this assumption, it can be stated that for a MNE that operates in multiple institutional environments (e.g., that of the home and that of the host countries), it is more difficult to maintain its legitimacy. The complexity of the different institutional environments increases the difficulty of understanding and conforming to the multiple demands. Balancing different, possible conflicting, demands, becomes a challenge for the MNE’s legitimacy. When this results in a legitimacy gap, it may lead to more criticism. For example, in China, giving gifts to your potential business partners is seen as part of doing business. In Western markets, however, Chinese managers must be cautious with this because it might be seen as bribery.
This is in line with the rationale of CSR research that focuses on what role MNEs should play in order to create legitimacy (Palazzo and Scherer, 2006). Palazzo and Scherer (2007) show that it is not possible to develop a universal view on the creation and sustainability of legitimacy of a company. Thus, it is difficult for MNEs to create legitimacy; there is no universal reference system from which to assess ethical acceptability of norms or actions (Scherer & Palazzo, 2007). When focusing on cultural distance, I hypothesize that when there is a high difference culture between the home and the host country of the MNE, the organization will find it more difficult to conform to the different demands and will
! *$! sooner become a target of stakeholder criticism as a result of its lack of legitimacy.
Hypothesis 1: The average cultural distance of the MNE is positively related to the total amount of criticism in the media (with respect to CSR-related issues) that the MNE is exposed to.
Figure 1 – Hypothesis 1
This positive relationship can become weaker when the cultural variance within the home country has higher values. Kostova and Zaheer (1999) argue that the greater the number and variety of countries in which an MNE operates, the less of a challenge it faces in establishing its legitimacy because of the extensive organizational experience in scanning different institutional environments, identifying important actors and making sense of its legitimacy requirements. In short, MNEs can, based on their past experience, better establish and maintain legitimacy. This means that when MNEs face great cultural diversity in their home country, they already have experience in facing different stakeholder demands because of the domestic cultural complexity. This cultural complexity can be understood as cultural variance in the home country of the MNE and, in this study, is measured by data on cultural diversity, ethnic fractionalization and the percentage of immigrants of the national population. When the cultural variance is high in the home country of the MNE, multiple ethnic groups
,-./-01.!234/1567!89! :;<! =>8-5/!89! 4/1?7@8.A70!603/3634>! 35!/@7!>7A31!B,CDE 07.1/7AF!
! *%! exist which each have their own fundamental cultural values (Barth, 1998). MNE have to face these different fundamental cultural values and the appurtenant stakeholder demands. This means that a heterogeneous, domestic culture in the home country can influence the relationship between the cultural distance of the MNE and the amount of stakeholder criticism because the MNE already has experience in identifying demands for establishing legitimacy. This is in line with the dynamic resource-based view which argues that based on experience, resources and capabilities may evolve over time and create a sustainable competitive advantage (Helfat and Peteraf, 2003). In this research, one can expect MNEs to develop the specific resources and capabilities, based on their experience, to more easily recognize different demands in the environment and conform to them in a successful manner. Furthermore, in a highly heterogeneous, domestic culture, the chances are higher that there are cultures present in the home country that have overlap with cultural characteristics in host countries. This makes it easier for the MNE to identify and conform to the stakeholder demands in the different environments.
Hypothesis 2: The positive relationship between the average cultural distance of the MNE and total amount of criticism in the media (with respect to CSR-related issues) that the MNE is exposed to, is moderated by the cultural variance in the home country of the MNE, so that this relationship is weaker for higher values of cultural variance in the home country.
! *&! Figure 2 – Hypothesis 2
Also important to take into account is the diversity level within the MNE itself. Proponents of intra-firm diversity argue that different opinions provided by culturally diverse groups enhance the quality of decision-making (Richard, 2000). Heterogeneity in decision-making and problem solving produces better decisions because of the wider range of perspectives and a more thorough critical analysis of issues (Richard, 2000). The resource-based view states that cultural diversity in the firm can lead to an enduring competitive advantage. This view argues that different types of employees will provide different beneficial resources to the firm. In other words, the more diverse the cultural background of the employees is, the more access to talent the MNE has, which in turn leads to better decision-making (Carter et al., 2010). For example, firms can increase their number of racio-ethnic minorities in the organization to better match the demographic characteristics of their environment to reflect the market they are attempting to reach (Richard, 2000). Firms hire employees who better understand different customer and stakeholder preferences and requirements. By increasing the intra-firm cultural diversity, one could expect that firms create a better understanding of the demands of the stakeholders in different environments. In short, the expectation is that firms with high intra-firm cultural diversity are more capable of recognizing and meeting the demands of
,-./-01.!G1031567!35! /@7!@8>7!68-5/0H! ,-./-01.!234/1567!89! :;<! =>8-5/!89! 4/1?7@8.A70!603/3634>! 35!/@7!>7A31!B,CDE 07.1/7AF!
! *'! stakeholders in different environments. This means that these firms are less likely to become a subject of stakeholder criticism. Following resource dependence theory, Pfeffer and Salancik (1978) also argue that one of the primary benefits of cultural diversity within the firm can create legitimacy for the firm in the external environment. This is especially the case in countries with an increasing growth in ethnic minority groups and thus with high cultural variance (Carter et al, 2010).
Hypothesis 3: The positive relationship between the average cultural distance of the MNE and total amount of criticism in the media (with respect to CSR-related issues) that the MNE is exposed to, is moderated by the cultural diversity on firm-level, so that this relationship is weaker for higher values of cultural diversity on firm-level.
Figure 3 – Hypothesis 3 ,-./-01.!234/1567!89! :;<! ,-./-01.!23G7043/H!85! I30>EJ7G7.! =>8-5/!89! 4/1?7@8.A70!603/3634>! 35!/@7!>7A31!B,CDE 07.1/7AF! !
! *(!
IV Methodology
In this chapter the research design of this study will be discussed, including its sample, the dependent and independent variable, the moderating variable and the different control variables.
1. Data collection and sample
For this thesis, a unique dataset is developed through the combination of three data sources: the Carbon Disclosure Project Data (CDP), the EthicalQuote Covalence Criticism data and Thomson Reuters Worldscope. Covalence EthicalQuote tracks online news on topics relevant to corporate social responsibility (CSR) based on 50 CSR-criteria developed by the Global Reporting Initiative (EthicalQuote, 2015). The analysts at Covalence EthicalQuote code each news item in terms of whether it constitutes criticism or praise directed towards the company, the source of the news item (the company itself or non-company sources, such as the media), which country the news item has been originally published in and which country is concerned with the incident described in the news article. The CDP sends out each year a very detailed questionnaire to companies regarding their greenhouse gas emissions. Particularly relevant for the present study, the CDP asks multinational corporations (MNCs) both to report their total greenhouse gas emissions and to break down these total emissions by country. Finally, the Thompson Reuters Worldscope database reports the financial and other relevant company-specific information used for the present study.
The original sample consisted of the 541 companies tracked by Covalence Ethicalquote for 2011. After combining the different data sources, the sample of 541 companies was reduced to 240 companies, for which full data was available..
! *)! The goal of this thesis is to explore the possible effect of cultural distance on the legitimacy of the firm. The selected MNEs are globally active in different countries, which is necessary for measuring cultural distance. The year 2011 is the latest year for which data from Covalence is available, so this is the most relevant existing data on the MNEs. To explore the relationship between stakeholder criticism and cultural distance, this study will perform a quantitative database analysis. A multivariate negative binominal regression will be conducted, using SPSS, to study if there is a correlation between the two variables and whether or not this correlation is moderated by a third variable.
2. Independent variable
2.1 Cultural distance
The independent variable ‘cultural distance’ is calculated using Hofstede’s (1983) theory on national cultures and data from the Carbon Disclosure Project. To assess which countries a company is active in and how important these countries are for the company, data from the Carbon Disclosure Project (CDP) was used. When filling in the CDP questionnaire, the companies report their total greenhouse gas emissions and they break these emissions down by the countries they are active in. In order to determine the importance of each one of the countries for the focal company, the amount of greenhouse gas emissions the company emitted in a country was divided by the total amount of greenhouse gas emissions of the company. This information was then combined with Hofstede’s cultural dimension scores to calculate companies’ cultural distance according to the formula developed by Kogut & Singh (1988). According to Hofstede (1983), culture can be measured by looking at four dimensions; power distance, uncertainty avoidance, individualism and masculinity. The power distance in a country refers to the way of dealing with power relationships between
! "+! superiors and subordinates. The more important the hierarchy, the higher the country scores on power distance (Hofstede, 1983). High scores on uncertainty avoidance mean a high anxiety level in a country and avoidance of behaviour that could increase this anxiety (Hofstede, 1983). Individualism refers to the relative importance in a country of freedom and independency. Finally, masculinity indicates the relative importance of recognition, advancement and earnings in a job. This stands in contrast with the importance of cooperation, desirable living area and/or employment security in a country (Hofstede, 1983). The scores of each dimension combined form the culture score of a country.
In this study Kogut and Singh’s (1988) formula (see below) is used to calculate the absolute cultural distance between the home country and the host country and the weighted cultural distance between the home country and the host country of MNEs. Both are used in this thesis and discussed below.
First, to calculate absolute cultural distance, the cultural dimension score (power distance, uncertainty avoidance, individualism and masculinity) for each MNE of the host country is subtracted from the cultural dimension score of the home country (Kogut and Singh, 1988). This score was then squared and divided by the variance of each dimension. After that, the squared differences divided by the variance were added up for each home country-host country pair. This resulted in the company’s absolute cultural distance scores on a specific dimension. Once the absolute cultural distance scores for each one of the four cultural dimensions had been calculated, the scores were added up and divided by four.
A similar procedure was used for the computation of Weighted Cultural Distance. Again, for each cultural dimension, the difference between a company’s home country and each one of its host countries was calculated. This value was then squared and divided by the
! "*! variance of the cultural dimension. These squared differences divided by the variance were weighted by multiplying the importance the host country has for the company. The importance of the host country was determined by dividing the amount of greenhouse gas emissions of the company. Then, the weighted distance for each home country-host country pair was added up. This resulted in the company’s weighted cultural distance on one specific dimension. Finally, these scores were added up and divided by four.
3. Dependent variable
3.1 Stakeholder criticism
The dependent variable in this research is defined as stakeholder criticism. The variable TotalCriticism measures the total number of times the company was negatively mentioned in the CovalenceThicalquote’s criticism database, with respect to environmental, social and corporate governance topics. This study is trying to understand the link between cultural distance and the amount of stakeholder criticism. As stated earlier, the complexity of the different institutional environments (cultural distance) enhances the difficulty of understanding and conforming to the multiple demands. Balancing different, possibly conflicting, demands, becomes a challenge for the MNE’s legitimacy. When this results in a legitimacy gap, it may lead to more criticism. This variable, TotalCriticism, indicates whether or not the legitimacy of the MNE diminishes when the level of cultural distance (the independent variable) increases.
! ""!
4. Moderating variables
4.1 Cultural variance
Hypothesis two suggests that the positive relationship between the average cultural distance of the MNE and total amount of criticism in the media (with respect to CSR-related issues) that the MNE is exposed to, is moderated by the cultural variance in the home country of the MNE. The relationship between the two variables is expected to be weaker for higher values of cultural variance. To measure home country cultural variance, three variables are used: ethnic fractionalization, cultural diversity and immigrants as percentage of national population.
Ethnic fractionalization is the most common variable used to measure ethnic diversity. Fearon (2003) defines ethnic fractionalization as the probability that two individuals, selected at random, from a country will be from different ethnic groups. He defines an ethnic group according to 7 conditions. First, membership in the group is reckoned primarily by descent by both members and non-members. Second, members are conscious of group membership and view it as normatively and psychologically important to them. Third, members share some distinguishing cultural features, such as common language, religion and customs. Fourth, these cultural features are held to be valuable by a large majority of members of the group. Five, the group has a homeland, or had one in the past. Six, the group has a shared and collectively represented history as a group. And finally, the group is potentially ‘stand alone’ in a conceptual sense (Fearon, 2003: 201). To calculate the score on ethnic fractionalization (i.e. the probability that to random individuals from a country will be from different ethnic groups) the following formula is used:
! ! ! ! ! !!! !!!
! "#! So, companies can score from 0 to 1 for its home country.
Fearson (2003) also constructed a measure of cultural diversity that modifies fractionalization to take in account the cultural distances between the groups. Cultural diversity is calculated the same way as ethnic fractionalization only this time you measure the probability that two groups’ languages come from completely different language families (e.g. Indo-European and Altaic).
! ! ! ! ! !!! !!!
If !! ! ! this means that the two groups’s languages come from completely different families, when ! ! ! the two groups speak exactly the same language. Again, the home country of the MNE can score from 0 to 1. To get the cultural diversity measure analogous to ethnic fractionalization, the expected cultural resemblance is subtracted from 1.
The variable immigrants as percentage of national population is based on the International Migration Report 2013. Since there is no data available for the year 2011, the numbers are used for the year 2013.
4.2 Firm-level cultural diversity
Hypothesis three states that the positive relationship between the average cultural distance of the MNE and total amount of criticism in the media (with respect to CSR-related issues) that the MNE is exposed to, is moderated by the cultural diversity on firm-level, so that this relationship is weaker for higher values of cultural diversity on firm-level. In this study the variable workforce/diversity and opportunity is used to measure cultural diversity on firm-level. This variable measures a company’s commitment and effectiveness towards maintaining diversity, regarding gender, age, ethnicity, religion and sexual orientation, in its workforce (Environmental, Social and Governance Performance., 2015).
! "$!
5. Control variables
This study controls for industry, firm size, and profitability. The MNEs differ in their industrial structures and this can influence the outcome of the regression analysis. It is possible that in some industries stakeholders are more active in criticizing MNEs on CSR related topics. For example, MNEs active in the industrial industry might get criticized more on environmental topics than MNEs in the banking industry. In the Worldscope dataset, each company is assigned a General Industry Classification, which defines whether a company is an industrial [1], utility[2], transportation[3], banking[4], insurance[5] or other financial company[6]. Also, it is important to control for the size of the firm because of the possibility that the larger the MNE, the more attention it gets from stakeholders. This can thus lead to an increase in stakeholder criticism. In this study the variable Total Assets indicates the size of the MNE. High profits can also lead to more attention from stakeholders and an increase in stakeholder criticism. On the other hand, a MNE with high profits has more financial possibilities to invest in CSR activities. This could then lead to a decrease in stakeholder criticism. Profitability of the MNE is indicated by Return on Assets in the dataset.
! "%!
V Results
This section reports on the results of this thesis. The first step will be to give the descriptive statistics for the continuous and categorical variables of this research. Then, a bivariate analysis is run to give an overview of the correlations between the different variables. To test the three hypotheses, formulated in the theoretical framework, a negative binominal regression is carried out. The results are presented in the last part of this chapter.
1. Descriptive statistics
In this thesis are 9 variables are used, from which 1 is categorical, namely the control variable general industry classification. The other 8 variables are continuous, the dependent variable total criticism, the independent variables workforce diversity and opportunity, ethnic fractionalization, cultural diversity, immigrants as percentage of total population and the control variables total assets and return on assets. A description of the categorical variable can be found in Table 1 and a description of the continuous variables is outlined in Table 2.
Table 1 shows that there is data for 239 companies for the variable general industry classification (N = 239, Missing = 1). The companies that were classified as operating in an industrial industry are estimated on 182 (76.2%). 22 (9.2%) Companies are identified as operating in the utility industry. 3 (1.3%) respondents are categorized as transportation firms. 17 (7.1%) of the 239 companies are operating in the insurance industry, leaving the remained 6 (2.5%) companies operating in other financial industries.
! "&! Table 1 – Descriptive Statistics for Categorical Variables
In Table 2 – Descriptive Statistics for Continuous Variables an overview of the used continuous variables can be found with information on summary statistics, such as the mean, the range and the standard deviation. The dependent variable total criticism has values ranging from 0 to 321, with a mean of 18.65 and a standard deviation of 37.343. The independent variable cultural distance has a minimum value of 0 and a maximum value of 265.69, a mean of 29.52 and a standard deviation of 40.72. The variable workforce diversity and opportunity has a range from 17.98 to 95.33, a mean of 83.94 and a standard deviation of 16.80. Ethnic fractionalization lowest value is 0 and has a highest value of 0.88, a mean of 0.32 and a standard deviation of 0.22. Cultural diversity has values ranging form 0 to 0.53, a mean of 0.2 and a standard deviation of 0.14. The variable immigrants as percentage of the total population has a minimum value of 0.30 and a maximum value of 43.3, a mean of 12.81 and a standard deviation of 6.37. The control variable total assets ranges from 2437482 to 36451079000, has a mean of 789133198.1 and a standard deviation of 3486464280. The control variable return on assets has a minimum value of -8.12, a maximum value of 30, a mean of 6.8661 and a standard deviation of 5.7081. Also important to note is that the variable workforce diversity and opportunity has 6 missing values. Ethnic fractionalization, cultural diversity and immigrants as percentage of total population all have 8 missing values. Total assets has 4 missing values and return on assets has 9 missing values.
Variable Level Frequency % Valid% Cumulative
% General Industry Classification 1- Industrial 182 75.8 76.2 76.2 2 - Utility 22 9.2 9.2 85.4 3- Transportation 3 1.3 1.3 86.6 4 - Banking 17 7.1 7.1 93.7 5 - Insurance 9 3.8 3.8 97.5 6 - Other financial company 6 2.5 2.5 100.0
! "'! Table 2 – Descriptive Statistics for Continuous Variables
2. Correlation analysis
Table 3 shows the results from the bivariate correlation analysis for all the combinations of variables. The correlations that are significant at 0.01 level (2-tailed) are marked with two asterix(**) and the correlations that are significant at the 0.05 level (2-tailed) are marked with one asterix (*). Before performing the correlation analysis, multiple variables have been transformed to avoid skewness. First, the independent variable absolute cultural distance was positively skewed (skewness score = 2.668, Std. Error = 0.157) and had to be transformed using Log10. All of the values of weighted cultural distance were squared (Sqrt) (skewness score = 1.576, Std. Error = 0.157). The variable total assets (skewness score = 8.092, Std. Error = 0.158) was transformed using Log10. Return on assets (skewness score = 0.957, Std. Error = 0.160) was transformed by squaring (Sqrt) all values. The variable workforce diversity and opportunity was negatively skewed (skewness score = -1.995, Std. Error = 0.159) and transformed using reflective Log10. The control variable general industry classification was transformed into a dummy variable. The dependent variable total criticism (skewness score = 4.761, Std. Error = 0.157) was not transformed because the negative binominal regression
Variables N Missing Min Max Mean Std Deviation
Cultural Distance 240 0 0.00 265.69 29.52 40.72 Total Criticism 240 0 0.00 321 18.65 37.343 Workforce Diversity and Opportunity 234 6 17.98 95.33 83.94 16.80 Ethnic Fractionalization 232 8 0.00 0.88 0.32 0.22 Cultural Diversity 232 8 0.00 0.53 0.20 0.14 Immigrants as percentage of total population 232 8 0.30 43.30 12.81 6.37 Total Assets 236 4 2437482 36451079000 789133198.10 3486464280 Return on assets 231 9 -8.12 30.00 6.8661 5.70810
! "(! takes into account the non-normal distribution of the dependent variable. In the following part the most important correlations will be discussed. First, the significant correlations with the dependent variable will be discussed. Then, the most important correlations between the independent variables will be outlined, followed by a report on the correlations between the control variables and independent variables.
The dependent variable total criticism appears to be positively correlated with the independent variable cultural distance (Pearson’s r = 0.17, N = 240, Sig. = 0.009). This is the only significant correlation between an independent and the dependent variable and thus, for now, seems to support hypothesis 1. The second variable that correlates with total criticism is the control variable total assets (Pearson’s r = 0.14, N = 236, Sig. = 0.031).
There are some important significant correlations between the independent variables. The strongest positive correlation is between cultural diversity and ethnic fractionalization (Pearson’s r = 0.921, N = 232, Sig. = 0.000. There is a weak negative correlation between cultural diversity and cultural distance (Pearson’s r = -0.13, N = 232, Sig. 0.044). In contrast, workforce diversity and opportunity has a weak positive correlation with cultural distance (Pearson’s r = 0.15, N = 234, Sig. 0.022).
Finally, total assets correlates with workforce diversity and opportunity (Pearson’s r = 0.204, N = 230, Sig. = 0.002), ethnic fractionalization (Pearson’s r = -0.435, N= 229, Sig. = 0.000), cultural diversity (Pearson’s r = -0.412, N = 229, Sig. = 0.000) and immigrants as percentage of total population (Pearson’s r = -0.375, N= 229, Sig. = 0.000). Return on assets correlates with the same variables; workforce diversity and opportunity (Pearson’s r = -0.155, N = 226, Sig. = 0.020), ethnic fractionalization (Pearson’s r = 0.357, N= 224, Sig. = 0.000), cultural diversity (Pearson’s r = 0.297, N = 224, Sig. = 0.000) and immigrants as percentage of total population (Pearson’s r = 0.150, N = 224, Sig. = 0.024). In addition, return on assets is negatively correlated with total assets (Pearson’s r = -0.479, N = 229, Sig. = 0.000).
! "#!
Table 3 – Bivariate Analysis
Variables Mean SD 1 2 3 4 5 6 7 8 9 10 11 12 13 14 1. Total Criticism 18.65 37.34 - 2. Cultural Distance 1.13 0.61 0.17** - 3. Workforce Diversity and Opportunity 83.94 16.80 0.12 0.15* - 4. Ethnic Fractionalization 0.32 0.22 0.03 -0.08 -0.12 - 5. Cultural Diversity 0.20 0.14 0.03 -0.13* -0.095 0.92** - 6. Immigrants 12.81 6.37 0.08 -0.10 -0.09 0.43** 0.557** - 7. Total Assets 7.88 0.82 0.14* 0.11 0.204** -0.44** 0.297** -0.375** - 8. Return on Assets 3.94 0.71 -0.03 0.04 -0.155* 0.357** 0.297** 0.15* -0.479** - 9. Industrial Industry 0.76 0.429 -0.026 0.163* -0.078 0.156* 0.092 -0.076 -0.359** 0.448** - 10. Utility Industry 0 0 - - - - 11. Transportation Industry 0.01 0.111 -0.026 -0.128* 0.03 -0.105 -0.076 -0.006 -0.065 -0.106 -0.199** - - 12. Banking Industry 0.07 0.257 -0.03 -0.031 0.155* -0.071 -0.038 0.096 0.369** -0.283** -0.489** - -0.031 - 13. Insurance Industry 0.04 0.190 -0.02 -0.054 -0.086 -0.094 -0.093 0.045 0.126 -0.22** -0.350** - -0.022 -0.054 - 14. Other Financial Industry 0.03 0.156 0.020 -0.054 -0.075 0.013 0.022 0.108 0.173** -0.126 -0.284** - -0.018 -0.044 -0.032 -
3. Negative binominal regression
Because the dependent variable total criticism is a count variable and is overdispersed, a negative binominal regression was performed to test the three hypothesis as formulated in the theoretical framework. For this study, two regressions are run; one using absolute cultural distance as independent variable and one using weighted cultural distance as independent variable. In this part an overview of the results of the tested hypotheses will be presented in three parts. First, the regression model will be discussed that only includes the control variables and the dependent variable total criticism. In the second part, an overview will be given of the results using absolute cultural distance as main independent variable. In the next step, weighted cultural distance was used as the main independent variable, taking in account the importance of each one of the host countries for the focal company.
3.1 Model 1: control variables
Model 1 only consists of the control variables and thus both tables show the same values (Table 4 – 5), because no independent variable (eg. absolute cultural distance or weighted cultural distance) is included yet. Both tables indicate that the control variable firm size (total assets) has a significant effect on total criticism (Sig. = 0.000)(Table 4-5). The B-value of 0.868 indicates that for each one-unit increase on firm size, the expected stakeholder criticism increases with 0.868 (Table 4-5). The dummy variables banking industry (Sig. = 0.000) and insurance industry (Sig. = 0.000) also have a significant effect on the dependent variable total criticism. When operating in a banking industry, the expected stakeholder criticism decreases with 1.792 (B = -1.792). The same negative effect goes for the insurance industry, that has a negative effect on the expected criticism with -2.542 (B = -2.542)(Table 4-5). Firm profit (Sig. = 0.311) does not seem to have a significant effect on the dependent variable.
! "#! 3.2 Total criticism and absolute cultural distance
Model 2 (Table 4) shows the regression model testing the effect of the control variables and the independent variable absolute cultural distance. The first hypothesis to be tested was: the accumulated cultural distance of all home-host country pairs of the MNE is positively related to the total amount of criticism in the media (with respect to CSR-related issues) that the MNE is exposed to. The results show that the independent variable absolute cultural distance is significantly correlated with total criticism (Sig. = 0.000) (Table 4) In short, this means that the first hypothesis is supported and that there is a significant positive correlation between cultural distance of the MNE and the total amount of criticism in the media after controlling for firm size, profit and industry.
The second hypothesis included the effect of a moderator: the positive relationship between the accumulated cultural distance of all home-host country pairs of the MNE and total amount of criticism in the media that the MNE is exposed to, is moderated by the cultural variance in the home country of the MNE, so that this relationship is weaker for higher values of cultural variance. To test this hypothesis three variables are used to measure cultural variance in the home country of the MNE; ethnic fractionalization, cultural diversity and amount of immigrants as percentage of the total population. After mean centering both the independent variable absolute cultural distance and the moderator variables, interaction terms were created to test the moderating effect of each of the variables. Model 3 (Table 4) shows that there is a no significant moderating effect (Sig. = 0.081) of ethnic fractionalization on the relationship between cultural distance and stakeholder criticism, after controlling for firm size, profit and industry. (Table 4). Model 4 (Table 4) shows that cultural diversity also has not got a significant moderating effect on the correlation between cultural distance and stakeholder criticism (Sig. = 0.083), after controlling for firm size, profit and industry. Model 5 (Table 4) indicates a significant moderating effect of the percentage of immigrants on the
! "$! relationship between cultural distance and stakeholder criticism (Sig. = 0.025, B = 0.074). The graph in Appendix I shows that a relatively large population of immigrants in the home country of the MNE can have a positive effect on the relationship between cultural distance and the total amount of stakeholder criticism.
The last hypothesis that was tested is: the positive relationship between the accumulated cultural distance of all home-host country pairs of the MNE and the total amount of criticism in the media (with respect to CSR-related issues) that the MNE is exposed to, is moderated by the cultural diversity on firm-level, so that this relationship is weaker for higher values of cultural diversity on firm-level. To run the negative binomial regression, first, the variable workforce diversity and opportunity had to be transformed using reflective Log10 (Skewness = -1.995, Std. Error = 0.159). Model 6 (Table 4) indicates that there is no significant effect of cultural diversity on firm-level on the relationship between absolute cultural distance and stakeholder criticism on the MNE (Sig. = 0.877).
3.3 Total criticism and weighted cultural distance
In this part the results of the same negative binominal regression will be discussed, but instead of using absolute cultural distance, as in the previous section, weighted cultural distance is used as the main independent variable. This can make a significant difference, because the variable weighted cultural distance takes in account the differences in importance of the host countries for the focal company.
Model 2 (Table 5) shows that there is a significant effect of the variable weighted cultural distance on stakeholder criticism (Sig. = 0.009). This means that there is a significant positive correlation between weighted cultural distance of the MNE and the total amount of criticism in the media after controlling for firm size, profit and industry. The variables firm size (Sig. = 0.000, B = 0.851), banking industry (Sig. = 0.000, B = -1.659) and insurance
! ""! industry (Sig. = 0.000, B = -2.381) are the only control variables with a strong significant effect. The coefficients of the control variables are all about the same as in Model 2 – Table 4, although there is a small increase in the B-value of firm size (total assets).
The second hypothesis focused on the effect of the moderator cultural variance. To test this hypothesis three variables are used to measure cultural variance in the home country of the MNE; ethnic fractionalization, cultural diversity and amount of immigrants as percentage of the total population. In order to test the moderating effect of each of the variables, new interaction terms were created using weighted cultural distance instead of absolute cultural distance as independent variable. Model 3, 4 and 5 (Table 5) show that the only significant moderating effect comes from the variable population of immigrants in the home country (Model 5) (Sig. = 0.042, B = 0.100). The graph in Appendix II shows that a relatively large population of immigrants in the home country of the MNE can have a positive effect on the relationship between cultural distance and the total amount of stakeholder criticism. The variables ethnic fractionalization (Model 3) (Sig. = 0.567) and cultural diversity (Model 4) (Sig. = 0.719) show to have no significant effect on the relationship between weighted cultural distance and stakeholder criticism.
The third hypothesis stated that the positive relationship between the cultural distance of the MNE and total amount of criticism in the media (with respect to CSR-related issues) that the MNE is exposed to, is moderated by the cultural diversity on firm-level, so that this relationship is weaker for higher values of cultural diversity on firm-level. Again, a new interaction term was created using weighted cultural distance instead of absolute cultural distance as independent variable. Model 6 (Table 5) indicates that there is no significant moderating effect of cultural diversity on firm-level on the relationship between weighted cultural distance and stakeholder criticism (Sig. = 0.458).
*** p<0.000 ** p<0.01 * p<0.05 Values in parentheses are standard errors
Table 4 – Absolute Cultural Distance as independent variable
Model 1 Model 2 or hypothesis 1 Model 3 or hypothesis 2a Model 4 or hypothesis 2b Model 5 or hypothesis 2c Model 6 or hypothesis 3 Constant -4.120** (1.408) -3.424* (1.142) -3.368* (1.431) -3.764** (1.427) -3.550* (1.401) -2.748 (1.493) Independent Variables
Cultural Distance Absolute 0.501***(0.128) 0.405** (0.142) 0.462** (0.137) 0.430** (0.136) 0.454** (0.139)
Ethnic Fragmentation 0.844* (0.380)
Cultural Diversity 2.174** (0.635)
Immigrants 0.093*** (0.020)
Workforce Diversity and Opportunity -0.270* (0.135)
Moderators
Ethnic Fragmentation x Cultural Distance Absolute
1.150 (0.660)
Cultural Diversity x Cultural Distance Absolute
1.770 (1.021)
Population of Immigrants x Cultural Distance Absolute
0.074* (0.033)
Workforce Opportunity and Diversity x Cultural Distance Absolute
-0.035 (0.225)
Control Variables
Firm Size (Total Assets) 0.868*** (0.152)
0.747*** (0.154) 0.807*** (0.154) 0.856*** (0.153) 0.885*** (0.149) 0.702*** (0.157) Firm Profit (Return on Assets) 0.131 (0.130) 0.055 (0.130) 0.016 (0.132) 0.021 (0.131) -0.101 (0.133) 0.108 (0.132)
Industry 1 -Industrial -0.308 (0.248) -0.373 (-0.252) -0.162 (0.254) -0.183 (0.254) -0.142 (0.253) -0.353 (0.259) 2- Utility - - - - 3- Transportation -0.418 (0.659) -0.225 (0.659) -0.100 (0.664) -0.073 (0.661) -0.094 (0.663) -0.277 (0.670) 4- Banking -1.792*** (0.386) -1.799*** (0.383) -1.747*** (0.391) -1.803*** (0.391) -1.960*** (0.391) -1.767*** (0.382) 5- Insurance -2.542*** (0.445) -2.337*** (0.449) -2.245*** (0.468) -2.160*** (0.471) -2.327*** (0.456) -2.224*** (0.453) 6- Other Financial -0.878 (0.566)) -0.879 (0.486) -1.243* (0.560) -1.390* (0.564) -1.735** (0.627) -0.930 (0.573) Model Fit
Likelihood Ratio Chi-Square 62.970 82.372 84.240 89.838 106.629 81.664
! "#! Model 1 Model 2 or hypothesis 1 Model 3 or hypothesis 2a Model 4 or hypothesis 2b Model 5 or hypothesis 2c Model 6 or hypothesis 3 Constant -4.120** (1.408) -4.357** (1.407) -4.229** (1.397) -4.688** (1.400) -4.288** (1.356) -4.102** (1.446) Independent Variables
Cultural Distance Weighted 0.440** (0.169) 0.498** (0.181) 0.500** (0.177) 0.454** (0.175) 0.363* (0.174)
Ethnic Fragmentation 0.915* (0.389)
Cultural Diversity 1.996** (0.644)
Immigrants 0.088*** (0.019)
Workforce Diversity and Opportunity -0.297* (0.137)
Moderators
Ethnic Fragmentation x Cultural Distance Weighted
0.515 (0.900)
Cultural Diversity x Cultural Distance Weighted
0.555 (1.541)
Population of Immigrants x Cultural Distance Weighted
0.100* (0.049)
Workforce Opportunity and Diversity x Cultural Distance Weighted
-0.216 (0.291)
Control Variables
Firm Size (Total Assets) 0.868*** (0.152) 0.851*** (0.152) 0.877*** (0.150) 0.926*** (0.150) 0.927*** (0.144) 0.825*** (0.154) Firm Profit (Return on Assets) 0.131 (0.130) 0.130 (0.129) 0.076 (0.131) 0.096 (0.131) -0.012 (0.132) 0.175 (0.130)
Industry 1 -Industrial -0.308 (0.248) -0.275 (0.251) -0.081 (0.259) -0.103 (0.257) -0.060 (0.256) -0.206 (0.260) 2- Utility - - - - 3- Transportation -0.418 (0.659) -0.422 (0.659) -0.196 (0.660) -0.278 (0.660) -0.197 (0.659) -0.415 (0.661) 4- Banking -1.792*** (0.386) -1.659*** (0.389) -1.548*** (0.402) -1.601*** (0.401) -1.715*** (0.401) -1.614*** (0.400) 5- Insurance -2.542*** (0.445) -2.381*** (0.449) -2.088*** (0.459) -2.049*** (0.460) -2.262*** (0.456) -2.264*** (0.452) 6- Other Financial -0.878 (0.566) -0.721 (0.491) -1.365* (0.559) -1.530** (0.567) -1.831** (0.602) -0.551 (0.571) Model Fit
Likelihood Ratio Chi-Square 62.970 74.247 77.556 81.529 99.919 75.126
Sig. 0.000 0.000 0.000 0.000 0.000 0.000
*** p<0.000 ** p<0.01 * p<0.05 Values in parentheses are standard errors
VI Discussion
This research was set out to test how cultural distance of the MNE is related to the total amount of criticism in the media (with respect to CSR-related issues) that the MNE is exposed to. Three hypotheses were formulated to explore this relationship and the possible moderating effects of other variables. In the following section, the findings of this study are discussed per hypothesis. Finally, an overview will be given of the limitations of this research and the advice for future research.
1. Findings
Hypothesis 1 expects that the accumulated cultural distance of all home-host country pairs of a MNE is positively related to the total amount of criticism in the media (with respect to CSR-related issues) that the MNE is exposed to. The results show that the independent variable absolute cultural distance (Sig. = 0.000) as well as the independent variable weighted cultural distance (Sig. = 0.009) is significantly correlated with total criticism. This is in line with the expectations based on the literature as discussed in the theoretical framework and literature review of this research. Institutional theory states that organizations must conform to the rules and belief systems prevailing in the environment to achieve and maintain environmental legitimacy (Kostova et al., 2008; Kostova and Zaheer, 1999; Xu and Shenkar, 2002). MNEs that operate in multiple institutional environments (i.e. that of the home and that of the host countries) can expect to find more difficulty in maintaining its legitimacy. The complexity of the different institutional environments enhances the difficulty of understanding and conforming to the multiple demands. If the MNE fails to conform to the demands in the environment, this might result in a legitimacy gap and lead to more criticism. The results of the analysis show that increase in complexity of different institutional environments of the
! "#! MNE (eg. weighted/absolute cultural distance) leads to a relative higher positive amount of stakeholder criticism.
Hypothesis 2 stated that the positive relationship between the cultural distance of the MNE and total amount of criticism in the media (with respect to CSR-related issues) that the MNE is exposed to, is moderated by the cultural variance in the home country of the MNE, so that this relationship is weaker for higher values of cultural variance. Based on the literature review of this research one can expect that MNEs can, based on their past experience, better establish and maintain legitimacy. When MNEs face great cultural diversity in their home country, they already have experience in facing different stakeholder demands because of the domestic cultural complexity (Kostova and Zaheer, 1999). The results of this research do not support this view. The cultural variance in the home country of the MNE was operationalized through three variables; cultural diversity, ethnic fractionalization and the percentage of immigrants of the national population. The percentage of immigrants of the national population was the only moderating variable with a significant effect on the relationship between cultural distance (i.e. absolute and weighted) and total criticism. Furthermore, the results of the regression analysis show that the moderator effect positively influences the relationship. This means that the when the percentage of immigrants increases, the positive effect of cultural distance on the amount of criticism also increases. These findings contradict the expectation that cultural variance would weaken the relationship between cultural distance and amount of criticism. One possible explanation for this positive moderating effect could be that instead of creating experience in creating legitimacy, a high population of immigrants increases the environmental complexity. Resource-based theory states that the greater the number of factors in the business environment a manager perceives he or she must deal with, and the greater the differences among those factors, the more complex the business environment is (Aragón-Correa and Sharma, 2003). Managers experiencing this complexity
! "$!
have difficulty determining the key factors that are important for maintaining the MNE’s legitimacy (Aragón-Correa and Sharma, 2003). In a domestic culture with highly diverse stakeholder demands, because of a large immigrant population, the likelihood decreases that a firm will use its capabilities and resources to develop a effective CSR strategy (Arag ón-Correa and Sharma, 2003).
Also, ethnic fractionalization, cultural diversity and the population of immigrants all have a significant positive main effect on the dependent variable stakeholder criticism (see table 4 and 5). This means that an increase in cultural variance in the home country of the MNE leads to higher values of stakeholder criticism. Institutional theory provides one possible explanation for this positive effect on the amount of stakeholder criticism. According to institutional theory, organizations must conform to the rules and belief systems prevailing in the environment to achieve and maintain environmental legitimacy (Kostova et al., 2008; Kostova and Zaheer, 1999; Xu and Shenkar, 2002). When there are a lot of different cultures in a country, MNEs face plural norms and values as an important part of these cultures. MNEs might find it difficult conforming to all of these, sometimes contradicting, norms and values and become a subject of stakeholder criticism. In short, high cultural variance in the domestic culture can lead to more stakeholder criticism. This is one way of explaining this relationship, but it would be interesting to study this effect of cultural variance on stakeholder criticism more in the future.
Hypothesis 3 focused on the effect of intra-firm cultural diversity and its effect on the relationship between cultural distance and the total amount of stakeholder criticism. The results show that intra-firm cultural diversity does not have a significant moderating effect on the relationship between cultural distance (i.e. weighted and absolute) and the amount of stakeholder criticism on the MNE. This goes against the expectations based on the literature review of this study. The expectations were based on the resource-based view that argues that
! "%! different types of employees will provide different beneficial resources to the firm. In other words, the more diverse the cultural background of the employees is, to more access to talent, which in turn leads to better decision-making (Carter et al., 2010). A firm with high cultural diversity would be better in recognizing and conforming to the different needs and demands of the complex environment. The results of this study go against this view and show no significant effect of intra-firm diversity on the relationship between cultural distance and the total amount of stakeholder criticism on the MNE. However, in both regressions there is a significant negative main effect of Workforce Diversity and Opportunity on the dependent variable stakeholder criticism. So, even though there is no significant moderating effect on the relationship between cultural distance and the amount of stakeholder criticism, intra-firm diversity does affects the amount of stakeholder criticism directly. This means that, in line with the literature, it can be expected that firms with high intra-firm diversity better understand different customer and stakeholder preferences and requirements. This results in less stakeholder criticism than firms with low scores on intra-firm diversity.
2. Limitations and future research
It is important to note that the current research has several methodological limitations, which deserve attention. First, although it provided a sufficient ground for research, the sample was relatively small. The targeted number of respondents was 541 MNEs around the world, whereas in the end there was only data available for 240 MNEs. It is possible that this has reduced the significance of the outcomes of this research. If the data was collected among more MNEs around the world, it could improve the generalizability of the findings.
The second limitation is that this study focused solely on the year 2011, because this was the latest year with relevant data. Although this is quite recent data, it could improve this
