Staying Legitimate: How Cultural Distance
Influences the Coordination of CSR Practices
Wendy Streumer 10057269 29 June 2015
MSc Business Administration: Strategy University of Amsterdam
Final Version Master Thesis Supervisor: Daniel Waeger
2
Statement of originality
This document is written by Wendy Streumer, 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.
3
Abstract
CSR has become more important in an MNE’s decision-making. Nowadays, firms not only need to legitimate their activities to their shareholders, but also to the broader society. A challenge for MNEs is that they have to deal with a large variety of different stakeholder pressures when deciding on their CSR practices. Culture is an important factor influencing the expectations of stakeholders and cultural differences can create different perspectives on what is considered as legitimate. Public criticism towards MNEs can make these different cultural expectations more visible and will put MNEs under pressure to react to them. To stay legitimate, MNEs have to make sure that they do not engage in contradicting practices. This thesis argues that by using clear coordination mechanisms for their CSR practices, namely centralization and formalization, MNEs can make sure that no contradictory practices will occur, and that they will thus remain legitimate. This study examines the relation between cultural distance and the centralization and formalization of CSR practices. Furthermore, this study also looks into the moderating effect of criticism on the relation between cultural distance and the centralization and formalization of CSR. A sample of 221 multinationals has been used for which data from the year 2011 was combined from different secondary databases. The findings of this thesis show that cultural distance is significantly related to the centralization and formalization of CSR. The more culturally distant the countries are in which an MNE is active, the more formalized and centrally managed its CSR practices are. In general, no support has been found for a moderating role of criticism. However, a significant negative moderation effect has been found for the proportion of transnational criticism on the relation between cultural distance and the centralization of CSR practices, although this effect has to be taken with caution.
Key words: Centralization; Corporate Social Responsibility; Criticism; Cultural Distance;
4
Table of contents
1. Introduction ………. 5
2. Literature review ……….. 8
2.1. Corporate social responsibility ………... 8
2.2. Culture ……….... 9 2.3. Criticism ………. 10 2.4. Coordination mechanisms ……….. 12 2.5. Conclusion ……….. 13 3. Theoretical framework ……… 14 4. Research methods ………... 20
4.1. Sample and data collection ……… 20
4.2. Measures ……… 22
4.2.1. Dependent variables ……….... 22
4.2.2. Independent variable ……… 24
4.2.3. Moderating variables ……… 25
4.2.4. Control variables ……….. 26
4.3. Statistical analysis and results……….. 27
5. Discussion ……… 36
5.1. Academic relevance ……… 36
5.2. Managerial implications ………... 39
5.3. Limitations and suggestions for future research ………... 40
6. Conclusion ………... 42
Acknowledgment ………... 45
5
1. Introduction
Corporate social responsibility (CSR) is becoming more and more influential in the strategic decision-making of multinational enterprises (Muller, 2006). It is not enough anymore to legitimate business activities to shareholders. Firms also need to legitimate their activities to the broader society (Crane, Matten and Spence, 2013). Besides the influence of the state in today’s globalized world, also nonpolitical and state actors, for example non-governmental organizations (NGOs), have political power. Therefore, multinational enterprises (MNEs) should think of their CSR within the context of emerging governance institutions (Scherer and Palazzo, 2007). A difficulty that multinationals face here is that they have to deal with a large variety of stakeholder pressures when deciding on their CSR practices. What is perceived as legitimate behavior in one country, does not have to be legitimate in another (Sharfman, Shaft and Tihanyi, 2004). This depends especially on the cognitive and normative parts of the institutions in a country (Kostova and Zaheer, 1999). Culture is in the background of the multiple stakeholder expectations (Katz, Swanson and Nelson, 2001), and can therefore create different legitimacy perceptions. When MNEs face public criticism towards their CSR practices, the different cultural expectations of its stakeholders become more visible. Public criticism can threaten an MNE’s reputation and put an MNE under pressure to change its behavior.
One way to achieve legitimacy as an MNE is to put in place high standards to make sure that every stakeholder expectation is met (Sharfman, Shaft and Tihanyi, 2004). When not following shared, high standards, MNEs face the risk that their subsidiaries create lower standards than what is required from the other countries in which they are active (Muller, 2006). The case of Nike is an example of what happens when an MNE simply adheres to local expectations without maintaining a certain minimal standard amongst its activities. When the Japanese market matured, Nike moved its activities to cheaper suppliers from Taiwan and
6
Korea. However, when demand raised over the years, these suppliers further outsourced their practices to countries with less-developed labor markets. In 1999, Nike’s athletic wear was produced in over 500 factories in 45 countries. With this outsourcing, Nike’s control over and knowledge of the working conditions in the different factories declined (Bell DeTiene and Lewis, 2005). In June 1996, Nike became the center of attention after Bob Herbert, columnist at the New York Times, accused Nike of working with ‘slave’ labor in so called ‘sweatshops’. Soon after this column, Nike was the target of different groups, among others NGOs, activists, consumers and international corporations. Nike’s reputation was damaged, and it faced high pressures to react (Bell DeTiene and Lewis, 2005).
The case of Nike shows the importance for MNEs to not get involved in contradictory practices deriving from compliance with different expectations among the different countries in which MNEs are active. Involvement in these kinds of contradictory practices can be a serious threat to an organization’s existence. This thesis argues that MNEs can use clear coordination mechanisms for their CSR practices to make sure that no such contradictory practices will arise, and thus to make sure that they will maintain their legitimacy. To do this, managers need to have an understanding of how cultural aspects can influence their corporate responsibilities (Katz, Swanson and Nelson, 2001). However, little is known in the literature so far about the influence of culture on a firm’s coordination mechanisms. Therefore, this thesis looks into the relation between cultural distance and the use of coordination mechanisms by MNEs for their CSR practices. Furthermore, the case of Nike shows the importance of criticism in the media. Legitimacy only became a problem for Nike after being criticized in the New York Times. Therefore, this thesis will also look into the effect of criticism on the relation between cultural distance and the coordination of MNEs’ CSR practices.
7
the existing literature and leads to the research questions of this thesis. In the third section, a theoretical framework is proposed, leading to five hypotheses. The fourth section explains the methodology used and presents the results of this study. Finally, the findings are discussed and a conclusion is given.
8
2. Literature Review
In the following section a brief review of the existing literature on the topic is given. First, the concept of CSR, and the potential legitimacy problems that can arise here, are shortly explained. Second, attention is paid to the idea of cultural distance and its possible influence on an MNE’s legitimacy. Third, the importance of criticism directed at an MNE is discussed. Fourth, the focus will shift to coordination mechanisms that can be of help in addressing legitimacy problems. Finally, a short conclusion is given, leading to the research questions.
2.1. Corporate social responsibility
Corporate social responsibility is a concept that gained extensive attention recently and although the literature on this topic is large and expanding, there is not one clear definition (Crane, Matten and Spence, 2013). Generally, CSR can be explained as the clearly expressed and communicated procedures and practices of firms, which have to do with a firm’s responsibility for a broader societal good. Yet, it depends on the firm in what precise form this takes shape (Matten and Moon, 2008). CSR steps away from two broad ideas. First, the idea that a firm is only focused on making profits. Second, that social responsibilities are only borne by the government (Matten and Moon, 2008). At first sight, it might seem illogical for a firm to invest money in CSR. In contrast to other, more traditional, organizational functions, it does not directly contribute to the creation or sales of a product or service with the goal of making a profit, which is supposed to be reason of existence of firms. However, nowadays firms face increasing demands to legitimate their activities to the broader society, and not just to their shareholders (Crane, Matten and Spence, 2013). In today’s globalized society, not only the state has political power, but also non-political and non-state actors such as NGOs, intergovernmental organizations, and transnational corporations (Scherer and Palazzo, 2007). These actors are all in the position to influence an MNE’s legitimacy and should thus be taken
9
into account by firms. Therefore, CSR is becoming more and more important in the strategic decision-making of MNEs.
Due to being active in multiple countries, a major challenge that MNEs face when making decisions about their CSR practices is that they have to deal with a large amount and variety of different stakeholder pressures that might conflict. This makes the question for corporate responsibility very complex (Scherer and Palazzo, 2007). Countries differ in what they perceive as legitimate behavior (Sharfman, Shaft and Tihanyi, 2004). Christmann and Taylor (2001) show, for example, that there are variations in environmental regulations between countries due to differences in the local valuation of environmental quality. Thus, within an MNE, each subsidiary faces its own institutional environment, with its own legitimacy requirements (Kostova and Zaheer, 1999). 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, p. 574).
2.2. Culture
An MNE’s legitimacy might be influenced by different factors. Kostova and Zaheer (1999) argue that, among others, a higher complexity in the institutional environment of an MNE leads to more difficulty for firms to develop and keep their legitimacy. More specific, they argue that especially the cognitive and normative parts of the institutional environment, i.e. its culture, contribute to legitimacy challenges for MNEs (Kostova and Zaheer, 1999).
Hofstede (1980a, p. 24) defines culture as “the collective programming of the human
mind that distinguishes the members of one human group from those of another. Culture, in this sense, is a system of collectively held values”. As software enables computers to function,
culture can be seen as the mental software for humans. It shapes the way we think, feel and act (Hofstede, Hofstede and Minkov, 2010). This means that culture influences the different
10
stakeholders of a firm and the expectations that these stakeholders have (Katz, Swanson and Nelson, 2001), and thus whether a firm is perceived as legitimate or not. The greater the variety of countries in which an MNE is active, and the greater the cultural distance between the home and host countries of an MNE, the more difficulties an MNE will have with its legitimacy (Kostova and Zaheer, 1999). This suggest that cultural distance is an important factor to take into account for managers.
Cultural distance refers to the difference between the cultural systems of a firm’s home and host countries. Hofstede (1980b) proposes four dimensions of national cultures along which cultural distance can be measured, namely power distance, uncertainty avoidance, individualism/collectivism, and masculinity/femininity. Power distance refers to the extent that people at the lower levels of a society accept and expect the unequal distribution of power. Uncertainty avoidance deals with the degree to which people feel at ease with uncertainty and ambiguity. Individualism fits with a society where people are only expected to take care of themselves and their close family. Collectivism, in contrast, refers to a society where people can expect their relatives and other people around them to take care of them. It refers to a society where loyalty is an important concept. Masculinity describes a society focused on achievement, heroism, assertiveness and material rewards. Femininity, alternatively, refers to a society characterized by cooperation, modesty, caring for the weak and quality of life (Hofstede, 1980b).
When MNEs would simply adapt their CSR practices to the variety of demands they face in the different countries in which they are active, due to cultural differences between these countries, they risk engaging in contradicting practices (Surroca, Tribó and Zahra, 2013). External contestants, for example activists, could easily take advantage of these contradictions and threaten an MNE’s legitimacy.
11
Campbell (2007) argues that organizations are more willing to engage in CSR when there are independent organizations in a firm’s environment who monitor their actions and, if needed, try to change these actions. These independent organizations - for example NGOs, social movement organizations, institutional investors, and the media - are stakeholders of MNEs. According to King (2008a), collective action is needed for these stakeholders to have influence. Collective action is defined as the “coordinated behavior among two or more
people that, at least in some minimal way, satisfies individual goals and produces a jointly experienced outcome” (King, 2008a, p. 22). This collective action is needed for the creation
of a stakeholder identity. Without this, the different stakeholders would be disconnected individuals and managers would not perceive them as being consequential (King, 2008a). The monitoring of a firm’s actions by external actors through collective action is an important factor to take into account when thinking about CSR decisions. The media plays a particularly powerful role here, which will be explained below.
Dyck and Zingales (2002) argue that the media shape the public image of managers and directors and force them to behave socially. People gain much of their information through the media, also about firms. The media decide which parts of information are communicated to the public. By doing this, the media have an important role in creating and accumulating a firm’s reputation (Dyck and Zingales, 2002). The reputation of a company tells something about its quality and reliability, it expresses a certain status in the market, and sets the firm apart from its competitors (King, 2008b). Reputation is connected to a firm’s legitimacy (King, 2008b). By communicating their grievances through the media and rebuilding the public image of a company, external contestants may become a threat for a firm’s legitimacy. Especially when an organization’s reputation depends on its CSR, it is vulnerable to these attacks (King, 2008b). An example of threat by external contestants through the media can be found in the Nike case from the introduction of this thesis. Here
12
Nike’s reputation was put under pressure after being accused of using ‘slave’ labor in the New York Times. After this news item, other groups also started to attack Nike, putting even more pressure on the organization. Therefore, it is important to take criticism, or more specifically, criticism through the media, into account when thinking about CSR practices.
2.4. Coordination mechanisms
Thus, compliance to cultural differences between an MNE’s home and host countries might lead them to engaging in contradicting CSR practices, which might be problematic for their legitimacy. External contestants can take advantage of this through public criticism, and in this way threaten an MNE’s reputation. To make sure that they do not engage in contradictory CSR practices, that might lead to legitimacy problems, MNEs need to coordinate their activities. A mechanism of coordination can be defined as “any administrative tool for
achieving integration among different units within an organization”(Martinez and Jarillo,
1989, p. 490). Every company faces particular levels of specialization or differentiation among their parts that ask for some form of coordination. Large and complex companies, like MNEs, however, need more coordination among a variety of dimensions. The need for coordination is thus not specific to MNEs, but their international context gives it a different element (Martinez and Jarillo, 1989). Two constructs mostly used when studying the coordinating mechanisms within MNEs are centralization and formalization (Ghoshal and Nohria, 1989). Centralization is a coordination mechanism in which the decision-making process takes place at the higher levels of the organization, or, in other words, within the headquarters of a multinational organization. The opposite of centralization is decentralization, where the decisions are being made lower in the organization, thus within the subsidiaries (Martinez and Jarillo, 1989). Formalization, sometimes also known as standardization, is the degree to which, among others, policies and rules are written down and procedures have become routines (Martinez and Jarillo, 1989).
13
Formalization, in the form of MNE-wide standards for social, environmental and general ethical issues, can increase legitimacy by making sure that everyone acts according to the same principles so that no contradicting practices arise. Centralization can improve legitimacy by making it easier to align practices with each other, because decisions are made in one place. A risk with a decentralized CSR strategy is, namely, that subsidiaries create lower standards than what is expected by the organization’s home country (Muller, 2006), or by other host countries. This, in turn, might lead to legitimacy problems.
Research has already been done on the relation between cultural distance and entry mode choices, international diversification and performance (Kogut and Singh, 1988; Hennart and Larimo, 1998; Reus and Lamont, 2009; Slangen and Beugelsdijk, 2010; Tihanyi, Griffith and Russel, 2005). Yet, little is known about the relation within multinationals between cultural distance and coordination mechanisms, especially with regard to CSR.
2.5. Conclusion
Concluding, when making decisions about their CSR, MNEs face a large variety of different stakeholder pressures. Different cultures between the countries in which MNEs are active play an important role in setting these different stakeholder expectations and can lead to legitimacy problems for MNEs. It might be needed for MNEs to clearly coordinate their CSR practices across all their host countries to make sure that no contradictory practices arise. No research has been done, however, on the relation between cultural distance and the coordination of an MNE’s CSR practices. Therefore, the research question of this thesis is as follows: What is the influence of cultural distance on the centralization and formalization of CSR? Because criticism towards an MNE’s CSR practices can threaten an MNE’s reputation and put it under pressure, it is an important factor to take into account. Therefore, a second research question arises: How does criticism influence the relation between cultural distance and the centralization and formalization of CSR?
14
3. Theoretical Framework
In this section, five hypotheses are developed regarding the research questions stated before. First, the expected relation between cultural distance and the coordination of CSR practices is explained. After, the potential influence of criticism on this expected relation is discussed. Following Suchman’s (1995) definition of legitimacy, as given in the literature review, legitimacy is based on cultural characteristics. The more culturally distant the countries are in which an MNE operates, the more different expectations an MNE faces regarding its legitimacy. Katz, Swanson and Nelson (2001) support this idea by showing that there are varying perspectives on five social issues at the different extremes of Hofstede’s cultural dimensions. Some examples of these are the following: cultures with high uncertainty avoidance have more environmental regulation than cultures with low uncertainty avoidance; in masculine cultures economic growth has priority, while in feminine cultures environmental conservation is more important; and in cultures characterized by individualism consumers have more advocacy than in cultures characterized by collectivism, which means that consumers are more likely to make their opinions on products and services explicit and they are more likely to create strong public initiatives (Katz, Swanson and Nelson, 2001). As already discussed in the previous section, the variety of stakeholder pressures that this creates might ask for a clear coordination of MNEs’ CSR practices. Simply implementing your home country standards in every host country is not the way to go. Different countries have different ethical standards and implementing your home country standards in other countries could lead to resistance and disastrous mistakes in other countries (Donaldson, 1996). However, adjusting to every different culture a company faces is not the best approach as well, since this can lead to contradicting and unethical practices (Donaldson, 1996). There should be a mid-way between the two extreme approaches and a clear coordination of CSR practices might be needed here to prevent MNEs from engaging in contradictory activities,
15
and thus to prevent them from potential legitimacy problems. Two ways in which this coordination can take place are already explained in the literature review, namely centralization and formalization (Ghoshal and Nohria, 1989). First, in a centralized organization the probability that contradictory practices arise is expected to be smaller than in a decentralized organization. The reasoning behind this is that when all decisions are made in once place, it is easier to align them than when decisions are made in multiple places. Second, formalization is expected to lower the chance of MNE’s subunits engaging in CSR practices that contradict each other. This is because MNE-wide standards for social, environmental and general ethical issues make sure that everyone in the organization is on the same page. It creates an understanding throughout the whole organization of what it values and how things should be done (Vlaar, Van den Bosch and Volberda, 2006), and therefore ensures consistency among practices. Besides creating consistency, organizations might also maintain or create legitimacy by setting high standards, because this allows them to meet, or go beyond, every stakeholder expectation they face (Sharfman, Shaft and Tihanyi, 2004). However, caution should be exercised with this last argument, since it does not apply to every practice. As Donaldson (1996) already argued, countries have different ethical standards. What might be ethical in one country, could lead to resistance in the other.
So, MNEs that are active in countries that are culturally different from each other will face a larger variety of stakeholder expectations than MNEs that are active in countries that are culturally close to each other. More CSR coordination in the form of centralization and formalization is expected to be needed in this first case to make sure that an MNE’s CSR practices stay aligned. This leads to the following two hypotheses:
Hypothesis 1: Cultural distance is positively related to the centralization of CSR practices. Hypothesis 2: Cultural distance is positively related to the formalization of CSR practices.
16
As already explained in the literature review, monitoring, and the powerful role of media here in specific, are of importance for an MNE’s decisions on its CSR practices. Support for this can be found in the case of Nike, shortly described in the introduction of this thesis. Nike faced problems because the labor conditions in the factories in which its products were produced, were at variance with the idea of fair labor conditions in other countries in which Nike was engaged. However, this case showed that Nike only had a high incentive to react to this problem after it became an important media item (Bell DeTiene and Lewis, 2005). In this same line, Lee, Plambeck and Yatsko (2012) found that some firms see small NGOs only as a nuisance and that they did not pay attention to its suppliers’ environmental violations until they became known in the media. King (2008b) looks more specific into the phenomenon of boycotts and shows that boycotters are more likely to have an impact on firms when the boycott gets a lot of media attention. This underlines once again the importance of criticism in the media when considering the CSR practices of an MNE.
An important effect of public criticism is that it can make stakeholder expectations more salient. Mitchell, Agle and Wood (1997) describe stakeholders as being ‘latent’ when they do not have much power, legitimacy and urgency. The more power, legitimacy and urgency stakeholders acquire, the more salient they become (Mitchell, Agle and Wood, 1997). The media might be used as a tool to help the different stakeholder groups in creating more salience. The distinctive stakeholder expectations that MNEs, operating in culturally distant countries, are facing, are often latent. However, when MNEs receive criticism in the media, and thus are put under public pressure, these differences become visible. This salience is important, because only then will stakeholders be able to get management’s attention (Mitchell, Agle and Wood (1997).
In general, the more criticism an MNE receives in the media, the bigger the chance that contradictions in CSR activities will become visible. Therefore, firms have to align their
17
CSR practices through coordination mechanisms to make sure that their legitimacy won’t become threatened. This suggests that the relation between cultural distance and the coordination of CSR practices will become stronger when an MNE faces more criticism in the media, leading to the following hypothesis.
Hypothesis 3: The total amount of criticism faced by an MNE positively moderates the relation between cultural distance and the centralization/formalization of CSR practices, such that this relationship is stronger for higher levels of criticism.
After this more general argument, more specified arguments can be made based on where criticism is coming from. First, attention will be paid to criticism coming from host countries, i.e. from parties that are active outside of an MNE’s home country. These can be, for example, local governments or labor unions. After, the focus will shift to transnational criticism, i.e. criticism that crosses the national boundaries. This can be criticism by a host country on something that has happened somewhere else, or criticism exercised by parties that are active across multiple countries, for example NGOs and social movements organizations. MNEs might face criticism in their host countries because their actions are contradicting the local norms and values. Cultural differences might be so substantial, that they cannot be overcome by an overarching approach. Public criticism in host countries might put pressure on an MNE to behave differently in these countries than they do in their home country or in other host countries. An example of this would be that gift giving is seen as favorable in some countries, but as a form of unethical bribery in others (Steidlmeier, 1999). Therefore, when MNEs face criticism in their host countries, it becomes harder for them to coordinate their CSR practices well. Centralization might reduce legitimacy at the local level (Muller, 2006) and formalization might lead to resistance in some of the countries in which an MNE is engaged (Donaldson, 1996). To stay legitimate in the host countries in which they are
18
being criticized, MNEs might need to loosen their standards and decentralize their decision-making on CSR.
Because public criticism faced in host countries by MNEs makes it more difficult for them to coordinate their CSR practices, it is expected to work as a moderator between cultural distance and CSR coordination. The larger the proportion of criticism coming from host countries, the weaker the relation will be between cultural distance and the coordination of CSR practices. This leads to the fourth hypothesis.
Hypothesis 4: The proportion of criticism in host countries faced by an MNE negatively moderates the relation between cultural distance and the centralization/formalization of CSR practices, such that this relationship is weaker for higher levels of host country criticism.
Whereas criticism in host countries is expected to derive from MNE’s actions that are contradicting the local norms and values, transnational criticism is expected to derive from the fact that companies are engaged in contradicting practices within the MNE itself, or that practices are below the standard in multiple countries. There are independent organizations who monitor the CSR practices of MNEs. For contradictions to become visible among an MNE’s CSR practices, these external monitors have to be connected throughout the different countries in which an MNE is active. Transnational monitors can harm an MNE’s reputation by making contradicting practices, or practices that are overall below the standard, public. As already discussed in the literature review, negative media attention can be seen as a strong threat to an organization’s legitimacy. It allows people to communicate their grievances and therefore it can damage the public image of an organization (King, 2008b). By making sure that CSR practices are aligned, MNE’s can maintain or regain their legitimacy. This suggests that if external contestants take advantage of contradicting practices through the media, MNEs will have a stronger tendency to coordinate their practices well in order to save their reputation than when this is not the case. Therefore, transnational criticism towards an MNE’s
19
CSR practices might work as a moderator in the expected relation between cultural distance and the coordination of CSR practices.
Thus the stronger the proportion of transnational criticism in the media towards an MNE’s CSR practices, the stronger cultural distance is expected to lead to a clear coordination of these CSR practices. This leads to the last hypothesis.
Hypothesis 5: The proportion of transnational criticism faced by an MNE positively moderates the relation between cultural distance and the centralization/formalization of CSR practices, such that this relationship is stronger for higher levels of transnational criticism.
An overview of the expected relations can be found in figure 1.
20
4. Research methods
In the following section, first, a description of the sample and the data collection is given. After, the different variables used in the statistical analysis are explained. Finally, the statistical tests used are explained and the results of the study are presented.
4.1. Sample and data collection
The focus in this thesis is on multinational enterprises, i.e. companies that are active in multiple countries. The reason behind this is that organizations that are active in multiple countries are expected to face different cultures in their different host countries, and therefore face a bigger challenge to coordinate their CSR practices than organizations that are active in only one country. To test the hypotheses, and give answers to the research questions, cross-sectional data of the year 2011 from different secondary sources is combined. The data for the two dependent variables ‘centralization’ and ‘formalization’ is retrieved from the Asset4 database. The independent variable ‘cultural distance’ is computed by combining Hofstede’s cultural dimension scores with data from the Carbon Disclosure Project. To compute the different criticism variables, data is obtained from Covalence Ethicalquote. Finally, the control variables are computed by using the Worldscope database. The different databases are discussed in more detail during the explanation of the different variables in section 4.2. The year 2011 has been chosen because this is the most recent year for which data is available in all the different databases.
The original sample contained 541 firms, which were all traced by the Covalence Ethicalquote. However, after combining the different databases for the purpose of this study, the sample was reduced to 221 MNEs, for which complete data was available. To gain some general information about the companies in the sample, data from the Worldscope database is used. This database offers fundamental data on the leading public and private firms. Table 1
21
shows the division among industries in which the companies in the sample are active. These industries are classified by the General Industry Classification from Worldscope. As can be seen in table 1, the majority of the companies in the sample is active in the industrial field (76,9%), followed by the utility industry (9%) and the ‘bank/savings and loan’ industry (6,8%). Most companies in the sample originate from the United States of America (40,7%), followed by Japan (11.3%) and the United Kingdom (10,4%). The complete list of home countries can be found in table 2.
Table 1: General industry classification
Industry Frequency Percentage
Industrial 170 76,9
Utility 20 9,0
Transportation 3 1,4
Bank/savings and loan 15 6,8
Insurance 9 4,1
Other financial 4 1,8
Total 221 100,0
Table 2: Home countries
Country Frequency Percentage
Australia 8 3,6 Belgium 1 0,5 Brazil 1 0,5 Canada 6 2,7 Denmark 1 0,5 Finland 3 1,4 France 15 6,8 Germany 13 5,9 Italy 2 0,9 Japan 25 11,3 Luxembourg 1 0,5 Netherlands 7 3,2 Norway 3 1,4 South Africa 4 1,8 South Korea 3 1,4 Spain 7 3,2 Sweden 3 1,4 Switzerland 5 2,3 UK 23 10,4 USA 90 40,7 Total 221 100,0
22
An advantage of using secondary data is that it allows to test the hypotheses set in this thesis with a large sample, and therefore more reliable conclusions can be drawn. The use of cross-sectional secondary data also makes it possible to compare data from different countries in the relatively short time period of this thesis and with the limited resources (Saunders, Lewis and Thornhill, 2009). A disadvantage from using secondary data is that the original data is collected for different purposes (Saunders, Lewis and Thornhill, 2009). The Worldscope database contained so many missing values on the geographic segments in which MNEs are active that, for the purposes of this study, additional data had to be collected through the Carbon Disclosure Project. Besides, the various databases do not contain the same companies, and because companies have to be present in all databases for the purpose of this research, this limits the sample size. However, despite these disadvantages, the use of secondary data is still the best option available to address the research questions of this thesis.
4.2. Measures
In this part, the different variables will be discussed. First, the dependent variables will be explained. Then, the independent variable will be discussed, followed by the moderating variables and the control variables.
4.2.1. Dependent variables
This study has two dependent variables, namely the ‘centralization of CSR’ and the ‘formalization of CSR’. To gain data on these two variables, the ASSET4 database is used. This database provides information on environmental, social and governance (ESG) practices by companies. To compose the variable ‘centralization’, first, there has been a search for all indicators for centralization in the database. Three measures were found, namely the availability of a CSR sustainability committee, the availability of an employees’ health and safety team, and the availability of an environment management team. When companies have
23
such a team in place, they score a 1 on these measures. If they do not have such a team, they get a 0. To compute the overall dependent variable ‘centralization’, these three scores are first added to each other and then divided by three. Thus, the scores on centralization can range from 0 to 1. The higher the value, the more centrally is CSR managed in a company. Decentralized MNEs leave the decision-making on CSR issues to the subsidiaries in the different countries in which they are active. Centralized MNEs, on the other hand, create committees at the headquarters with the specific function of dealing with CSR practices, such as the three teams mentioned above.
For the development of the dependent variable ‘formalization’, data on companies’ policies is used. In the ASSET4 database, companies receive a score, ranging from 0 to 100, on the existence of policies in three domains, namely environmental, social and governance. In total, policies on 15 subject matters are scored. These policies are on the subjects of resource reduction, emission reduction, product innovation, employment quality, health and safety, training and development, diversity and opportunity, human rights, community, product responsibility, board structure, compensation, board functions, shareholder rights, and vision and strategy. To compute the dependent variable ‘formalization’ in this study, the different scores explained above are added for each company and then divided by 15. This leads to an average score on formalization for each MNE that can range from 0 to 100. The higher the value, the more formalized CSR practices are. Because the data on formalization is highly skewed (skewness = -1,804), it is transformed by using the following formula: Y = Log10 (K + 1 – X), where K is the highest value for formalization found in the data. This formula, however, reverses the values. Whereas, originally, a higher value would imply more formalization, now a higher value implies less formalization. Following Field (2009), the variable is reversed again afterwards by using Y = K - X , to make sure that no mistakes will
24
be made while interpreting the results. The new, transformed variable is significantly less skewed (skewness = -0.091).
4.2.2. Independent variable
The independent variable in this study is the ‘cultural distance’ an MNE faces between its home country and host countries. To find out in which countries the sampled companies are active and how important these countries are to the companies, data is obtained from the Carbon Disclosure Project (CDP). Every year, the CDP asks firms to fill in a detailed questionnaire about their greenhouse gas emissions. Of special interest for this study, MNEs are asked by the CDP to report both their total greenhouse gas emissions and to decompose their total emissions by country. To find out the importance of a country to an MNE, the amount of greenhouse gas emissions emitted in this country was divided by the total amount of greenhouse gas emission of the MNE.
Hofstede’s national scores are used to gain information on the cultural characteristics of the different countries in which the MNEs are active. There is some controversy regarding the measurement of culture by Hofstede’s four dimensions (Jones, 2007). Hofstede’s (1980b) dimensions are accused of being too simplified. Nations are not always the proper unit of analysis and a country’s population is wrongly assumed to be homogenous. Finally, the data is critiqued as being outdated (Jones, 2007). Despite these critiques, however, it is still one of the best options available and the method most widely used. Several studies have replicated Hofstede’s method and overall the accuracy of Hofstede’s dimensions is confirmed (Jones, 2007).
Following Kogut and Singh (1988), the independent variable ‘cultural distance’ is calculated based on the following composite index:
25
Using Hofstede’s index, deviations by an MNE’s home and each of its host countries are computed for each cultural dimension. These values are squared and divided by the variance of the cultural dimension. The values are then weighted by multiplying them with the importance of the county for the MNE. After, the weighted distances for every home country-host country pair of an MNE are added up, leading to a firm’s weighted cultural distance on one of the four dimensions. Finally, the scores on the four dimensions are added to each other and divided by four. The final score per MNE is its weighted cultural distance. This weighted measure of cultural distance is highly skewed (skewness = 1,566). Therefore, the independent variable ‘cultural distance’ used in the analysis is a transformation of the weighted cultural distance by the formula Y = Log10 (X + 1). The value for skewness of this transformed variable is 0,643.
4.2.3. Moderating variables
Three moderating variables are tested in this study. The data for these variables is obtained from Covalence EthicalQuote. This is an ethical reputation index of the world’s largest firms on ESG, CSR, ethics and sustainability. It integrates online news items based on 50 CSR-criteria inspired by the Global Reporting Initiative (EthicalQuote, 2015). The first moderating variable is ‘total criticism’. This variable consists of the number of times that an MNE was the target of CSR-oriented public criticism in the year 2011. Because this variable is highly skewed (skewness = 4.648), it is transformed by using the formula Y = Log10 (X + 1). The new value for skewness is 0.158. The second moderating variable is the ‘proportions of criticism in host countries’. Host country criticism is the number of times an MNE is target of CSR-oriented public criticism outside of its home country in the year 2011. To create this
26
variable, first the proportion of criticism in the home country had to be computed. This is done by dividing the amount of times an MNE is publically criticized on CSR issues in its home country in 2011 by the total amount of CSR-related criticism an MNE faced in 2011. This value is then detracted from 1, which leads to the proportion of criticism in host countries. The third moderating variable is the ‘proportion of transnational criticism’. Criticism is classified as transnational if the origin of the criticism is not located in the same country as the country in which the criticized event took place. This variable is computed by dividing the amount of times that an MNE was the target of transnational CSR-oriented public criticism in 2011 by the total amount of criticism an MNE faced in that year on CSR-related issues.
4.2.4. Control variables
When studying the influence of cultural distance on the centralization and formalization of CSR, it is important to control for other variables that might have an impact on an MNE’s CSR coordination. The first control variable is firm size. This variable is measured by taking the logarithm, with a base of 10, of the number of employees in an MNE. Firm size should be controlled for as large and visible MNEs have more problems with maintaining their legitimacy than small and less visible MNEs, because they are more vulnerable to attacks from external contestants (Kostova and Zaheer, 1999). The second control variable is firm profitability, which is measured by taking the square root of the return on assets (ROA). Firms with strong financial performances have more resources available to spend on CSR than firms with weak financial performances, which influences the CSR decisions a company makes (Adams and Hardwick, 1998; Campbell, 2007). Therefore the profitability of an MNE should be controlled for. Whereas higher firm profitability can increase the resources firms use on CSR, high levels of leverage put a constraint on the resources a firm has available for its CSR due to high debt contracting costs (Adams and Harwick, 1998). Therefore, leverage will also
27
be taken into account as a third control variable. An MNE’s leverage is computed by dividing its total liabilities by its total assets. The fourth and final control variable is industry. Depending on an MNE’s industry, more or less attention might be needed for certain social and/or environmental issues. Sweeney and Coughlan (2008), for example, found a clear industry effect in the reporting on CSR practices by different firms. The MNEs in the sample are therefore classified based on the general industry classification by Worldscope. This classification distinguishes between firms that are active in industrial, utility, transportation, banks/savings and loan, insurance, and other financial industries. To be able to use industry as a control variable in the regressions, dummy variables are created. The ‘industrial’ industry is the baseline here, because this is the industry in which the majority of the MNEs in the sample is active. All data on the control variables is obtained from the Worldscope database.
4.3. Statistical analysis and results
To conduct the statistical analysis, the Statistical software Package for Social Sciences (SPSS) is used. After cleaning the dataset, the normality of the data is tested by computing the values of skewness and kurtosis. Variables that scored very high or low on these measures were transformed, as already explained in section 4.2. Skewness was considered as acceptable when the values were between -0,8 and 0,8. Kurtosis was considered as acceptable when the values were between -3 and 3. Table 3 presents the descriptive statistics of the dependent, independent, moderating and control variables. The variables are tested on multicollinearity by computing the correlations between the independent and control variables. None of the predictor variables have a correlation above 0,5. Therefore, at first sight, no multicollinearity issues seem to be at place. To find out about the more subtle forms of multicollinearity, the variance inflation factor (VIF) is used. Following Field (2009), a VIF value higher than 10 or tolerance levels below 0,2 are considered as critical. All of the VIF values found are between the 1 and 2, and the tolerance levels are all above 0,6. This supports that multicollinearity is
Table 3: Descriptive statistics: Means, standard deviations and correlations Variable Mean SD 1 2 3 4 5 6 7 8 9 10 11 12 13 14 1. Centralization 0,884 0,201 1 2. Formalization 1,016 0,473 0,434** 1 3. Firm size 4,751 0,453 0,104 0,344** 1 4. Firm profitability 3,939 0,717 -0,034 -0,072 0,168* 1 5. Leverage 0,628 0,195 0,025 0,123 0,246** 0,394** 1 6. Industrial vs. utility 0,091 0,288 0,078 0,102 -0,029 -0,142* 0,101 1 7. Industrial vs. transportation 0,014 0,116 0,068 0,087 0,090 -0,115 0,040 -0,037 1 8. Industrial vs. banks/savings and loans 0,068 0,252 0,007 0,142* 0,164* -0,272** 0,424** -0,085 -0,032 1 9. Industrial vs. insurance 0,041 0,198 -0,109 -0,114 -0,054 -0,233** 0,285** -0,065 -0,024 -0,056 1
10. Industrial vs. other financial 0.018 0,134 -0,147* -0,137* -0,113* -0,117 0,181** -0,043 -0,016 -0,037 -0,028 1
11. Cultural distance 0,206 0,172 0,274** 0,194** 0,061 -0,044 -0,068 0,078 -0,043 -0,069 -0,086 -0,050 1
12. Total criticism 0,869 0,633 0,158* 0,425** 0,472** -0,033 0,073 0,146* -0,035 -0,012 -0,128 0,009 0,107 1
13. Proportion criticism host countries 0,682 0,350 0,109 -0.058 -0,091 -0,086 -0,037 0,077 0,028 -0,105 0,050 -0,054 0,167* -0,314** 1 14. Proportion transnational criticism 0,359 0,363 0,219** 0,214** -0,025 0,038 0,055 0,087 -0,064 -0,133* 0,048 -0,017 0,298** 0,409** 0,168* 1 * p < 0,05, ** p < 0,01
29 not a problem in this study.
The MNEs in the sample have on average 98.372 employees. Their ROAs have a mean of 7,030 and their average leverage is 0,628. Note that these numbers for firms size and profitability are different from the ones stated in table 3, because in the table the transformed variables are used. The average centralization measured within the sample is 0,884. Because theoretically this measures can range from 0 until 1, it suggests that there is a lot of centralization of CSR at the sampled MNEs. The average formalization measured is 71,380. Theoretically, this measure can range from 0 until 100, suggesting that in general there is also substantial formalization of CSR in place. The average amount of criticism for an MNE is 19,48. This means that firms gain on average 19,48 negative mentions with respect to environmental, social and governance issues in the year 2011. The amount of criticism in the sample, however, deviates strongly from this mean, with 0 and 321 as the extremes. Table 3 shows some significant correlations that should be acknowledged during this data analysis. First, there is a high correlation between the dependent variables ‘centralization’ and ‘formalization’. It is important to keep this in mind when testing hypothesis 1 and 2. The table also shows that there is a significant correlation between two of the moderating variables, namely the ‘total amount of criticism’ and the ‘proportion of transnational criticism’ and the two dependent variables. Moreover, there is a significant correlation between cultural distance and the proportion of criticism in host countries, and between cultural distance and the proportion of transnational criticism.
To test whether cultural distance is related to the centralization and formalization of CSR practices, first a multivariate regression is used. This type of regression is chosen, because it incorporates the correlation between the two dependent variables. Using Pillai’s trace, there is a significant effect of cultural distance on the centralization and formalization of CSR practices, V = 0,068, F(2, 202) = 7,324, p < 0,01. This finding seems to support
30
hypothesis 1 and 2. However, this test says nothing about any difference between the two different dependent variables. Therefore, two separate hierarchical regressions are also performed.
The first hierarchical regression, with centralization as a dependent variable, supports hypothesis 1. This means that cultural distance is positively related to the centralization of CSR practices. In the first step of this regression, the control variables are entered. When adding cultural distance to the model, the R2 changes by 0,056 (p < 0,01). Table 4 shows the results of this first regression. In the second model, the influence of cultural distance on centralization is statistically significant (B = 0,277, SE B = 0,077, p < 0,01), supporting
hypothesis 1.
The second hierarchical regression, with formalization as dependent variable, supports hypothesis 2. This implies that cultural distance is positively related to the formalization of CSR practices. Here, again, the control variables are entered first into the regression. By adding cultural distance to the model, the R2 changes by 0,026 (p < 0,05). The results of this second regression can be found in table 5. In the second model, the influence of cultural distance on formalization is statistically significant (B = 0,445, SE B = 0,173, p < 0,05), supporting hypothesis 2.
To test the moderating effects of hypothesis 3 to 5, the two hierarchical regressions used above are extended by including the different moderators and their interactions with the independent variable. Before creating the interaction terms, the independent and moderator variables are centered to reduce problems of multicollinearity. The complete hierarchical regression with centralization as dependent variable can be found in table 4. Table 5 shows the complete hierarchical regression with formalization as dependent variable. First, the moderator ‘total criticism’ is entered as a control variable (model 3). Total criticism does not significantly influence the centralization of CSR practices (B = 0,031, SE B
31
= 0,025, p = 0,209), but it does significantly influence the formalization of CSR practices (B = 0,267, SE B = 0,053, p = 0,000). This means that when a firm faces more criticism, it is more likely to coordinate its CSR practices through formalization. In model 4, the interaction of ‘cultural distance’ and ‘total criticism’ is added to the regression. The interaction of cultural distance and total criticism does not significantly relate to CSR centralization (B = -0,022, SE B = 0,014, p = 0,103), nor does it significantly influence CSR formalization (B = -0,004, SE B = 0,030, p = 0,885). Together, these two findings show no support for hypothesis 3. Model 5 introduces the moderator ‘proportion of criticism in host countries’ as control variable. The proportion of criticism in host countries does not significantly relate to the centralization of CSR practices (B = 0,051, SE B = 0,039, p = 0,194), nor to the formalization of CSR practices (B = -0,061, SE B = 0,089, p = 0,497). In model 6, the interaction of ‘cultural distance’ and ‘the proportion of criticism in host countries’ is included in the regression. This interaction term does not significantly influence the centralization of CSR practices (B = -0,008, SE B = 0,014, p = 0,595). The interaction of cultural distance and the proportion of criticism in host countries also does not significantly influence the formalization of CSR practices (B = -0,010, SE B = 0,032, p = 0,750). These two findings reject hypothesis 4.
Next, the proportion of transnational criticism is added as a control variable in model 7. The proportion of transnational criticism is significantly related to the centralization of CSR practices (B = 0,080, SE B = 0,032, p = 0,039). Thus the larger the proportion of transnational criticism an MNE faces, the more centrally managed its CSR practices are expected to be. The proportion of transnational criticism is also significantly related to the formalization of CSR practices (B = 0,253, SE B = 0,086, p = 0,004). This implies that MNEs are more likely to formalize their CSR practices when they face a larger proportion of transnational criticism. In model 8, the interaction of ‘cultural distance’ and ‘the proportion of
32
transnational criticism’ is added to the regression. The interaction of cultural distance and the proportion of transnational criticism has an significant influence on CSR centralization (B = -0,028, SE B = 0,013, p = 0,035). However, this interaction term does not significantly influence CSR formalization (B = -0,029, SE B = 0,030, p = 0,332). Although these findings do not support a moderating influence of the proportion of transnational criticism on the relation between cultural distance and the formalization of CSR practices, they do support that the relation between cultural distance and the centralization of CSR practices is moderated by the proportion of transnational criticism. Figure 2 shows that the higher the proportion of transnational criticism is, the weaker the relation between cultural distance and CSR centralization becomes. This is contrary to the expected influence of the proportion of transnational criticism stated in hypothesis 5. This finding, together with the fact that the interaction of cultural distance and the proportion of transnational criticism does not significantly influence the formalization of CSR practices, show no support for hypothesis 5. A cautious note should be placed here, since these significant relations found in model 7 and 8 disappear when all variables are entered into the regression (model 9).
33
Finally, model 9 includes all the variables used in this analysis. In this model, cultural distance remains to have a significant effect on the centralization of CSR practices (B = 0,219, SE B = 0,080, p = 0,007), and the formalization of CSR practices (B = 0,346, SE B = 0,175, p = 0,049). Besides, total criticism still has a significant influence on CSR formalization (B = 0,276, SE B = 0,066, p = 0,000).
34
Table 4: Results of hierarchical regression with centralization as dependent variable
Dependent variable: Centralization Model 1 Model 2 (H1) Model 3 Model 4 (H3) Model 5 Model 6 (H4) Model 7 Model 8 (H5) Model 9 Control variables Firm size 0,025 (0,032) 0,018 (0,031) -0,003 (0,035) 0,000 (0,035) 0,022 (0,031) 0,023 (0,031) 0,018 (0,031) 0,017 (0,030) 0,008 (0,036) Firm profitability -0,013 (0,022) -0,006 (0,021) -0,007 (0,021) -0,004 (0,021) -0,003 (0,021) -0,003 (0,021) -0,009 (0,021) -0,008 (0,021) -0,002 (0,021) Leverage 0,130 (0,090) 0,138 (0,088) 0,136 (0,088) 0,139 (0,087) 0,141 (0,088) 0,135 (0,088) 0,112 (0,088) 0,100 (0,087) 0,105 (0,088) Industrial vs. Utility 0,024 (0,049) 0,016 (0,047) 0,006 (0,048) 0,007 (0,048) 0,013 (0,047) 0,011 (0,048) 0,012 (0,047) 0,005 (0,047) -0,007 (0,048) Industrial vs. Transportation 0,075 (0,116) 0,101 (0,113) 0,114 (0,114) 0,128 (0,113) 0,097 (0,113) 0,092 (0,114) 0,116 (0,113) 0,122 (0,112) 0,125 (0,112) Industrial vs. Banks/savings and
loans
-0,066 (0,068) -0,048 (0,066) -0,041 (0,067) -0,032 (0,067) -0,039 (0,067) -0,038 (0,067) -0,027 (0,067) -0,016 (0,066) 0,001 (0,067) Industrial vs. Insurance -0,163* (0,076) -0,139 (0,074) -0,129 (0,074) -0,124 (0,074) -0,141 (0,074) -0,139 (0,074) -0,143 (0,073) -0,134 (0,073) -0,122 (0,073) Industrial vs. Other financial -0,264* (0,106) -0,245* (0,103) -0,254* (0,103) -0,261 (0,103) -0,236* (0,103) -0,229* (0,104) -0,238* (0,102) -0,215* (0,102) -0,216* (0,103)
Independent variables
Cultural distance 0,277** (0,077) 0,273** (0,077) 0,266 (0,077) 0,259** (0,078) 0,263** (0,078) 0,229** (0,080) 0,232** (0,079) 0,219** (0,080)
Total criticism 0,031 (0,025) 0,031 (0,025) 0,028 (0,030)
Cultural distance x Total criticism
-0,022 (0,014) -0,022 (0,017)
Proportion of criticism in host countries
0,051 (0,039) 0,047 (0,040) 0,056 (0,045)
Cultural distance x Proportion of criticism in host countries
-0,008 (0,014) -0,015 (0,015)
Proportion of transnational criticism
0,080* (0,032) 0,089* (0,038) 0,062 (0,045) Cultural Distance x Proportion
of transnational criticism -0,028* (0,013) -0,018 (0,015) R2 0,066 0,122 0,129 0,141 0,130 0,131 0,141 0,160 0,176 Adjusted R2 0,029 0,084 0,086 0,094 0,087 0,083 0,098 0,114 0,113 F-statistic 1,804 3,148 3,000 2,994 3,013 2,755 3,312 3.475 2,805 P-value 0,078 0,001 0,001 0,001 0,001 0,002 0,001 0,000 0,001
35
Table 5: Results of hierarchical regression with formalization as dependent variable
Dependent variable: Formalization Model 1 Model 2 (H2) Model 3 Model 4 (H3) Model 5 Model 6 (H4) Model 7 Model 8 (H5) Model 9 Control variables Firm size 0,310** (0,070) 0,300** (0,070) 0,119 (0,075) 0,119 (0,075) 0,295** (0,070) 0,296** (0,070) 0,299** (0,068) 0,298** (0,068) 0,124 (0,079) Firm profitability 0,015 (0,048) 0,026 (0,048) 0,020 (0,045) 0,021 (0,045) 0,022 (0,048) 0,022 (0,048) 0,017 (0,047) 0,018 (0,047) 0,024 (0,046) Leverage 0,241 (0,201) 0,254 (0,198) 0,235 (0,187) 0,236 (0,187) 0,251 (0,198) 0,243 (0,200) 0,173 (0,196) 0,160 (0,197) 0,206 (0,192) Industrial vs. Utility 0,165 (0,108) 0,152 (0,107) 0,069 (0,102) 0,069 (0,102) 0,155 (0,107) 0,151 (0,108) 0,140 (0,105) 0,133 (0,105) 0,055 (0,104) Industrial vs. Transportation 0,245 (0,259) 0,287 (0,256) 0,396 (0,242) 0,398 (0,244) 0,292 (0,256) 0,286 (0,258) 0,332 (0,252) 0,338 (0,252) 0,401 (0,245) Industrial vs. Banks/savings and
loans
0,184 (0,152) 0,212 (0,150) 0,277 (0,142) 0,279 (0,143) 0,200 (0,151) 0,201 (0,151) 0,280 (0,149) 0,291 (0,150) 0,318* (0,146)
Industrial vs. Insurance -0,269 (0,168) -0,229 (0,167) -0,147 (0,158) -0,146 (0,159) -0,226 (0,167) -0,223 (0,168) -0,243 (0,164) -0,233 (0,164) -0,147 (0,160) Industrial vs. Other financial -0,402 (0,235) -0,371 (0,232) -0,450* (0,120) -0,452* (0,221) -0,382 (0,233) -0,373 (0,235) -0,349 (0,228) -0,325 (0,230) -0,404 (0,225)
Independent variables
Cultural distance 0,445* (0,173) 0,408* (0,164) 0,407* (0,164) 0,466** (0,176) 0,471** (0,178) 0,292 (0,178) 0,295 (0,178) 0,346* (0,175)
Total criticism 0,267** (0,053) 0,267** (0,053) 0,267** (0,066)
Cultural distance x Total criticism
-0,004 (0,030) 0,007 (0,037)
Proportion of criticism in host countries
-0,061 (0,089) -0,066 (0,091) 0,071 (0,098)
Cultural distance x Proportion of criticism in host countries
-0,010 (0,032) 0,000 (0,034)
Proportion of transnational criticism
0,253** (0,086) 0,262** (0,087) 0,082 (0,097) Cultural Distance x Proportion
of transnational criticism -0,029 (0,030) -0,029 (0,034) R2 0,178 0,204 0,293 0,294 0,205 0,206 0,236 0,240 0,303 Adjusted R2 0,146 0,168 0,258 0,255 0,166 0,162 0,198 0,198 0,249 F-statistic 5,513 5,765 8,389 7,591 5,222 4,735 6,248 5,765 5,697 P-value 0,000 0,000 0,000 0,000 0,000 0,000 0,000 0,000 0,000
36
5. Discussion
This study examined the influence of cultural distance on the coordination of an MNE’s CSR practices. Due to the cultural differences between the countries in which they are engaged, MNEs face a large variety of different stakeholder expectations towards their CSR practices. These different expectations might lead to legitimacy problems for MNEs. Public criticism towards an MNE’s CSR practices can threaten an MNE’s reputation by making these different stakeholder expectations more visible. To make sure that they do not engage in contradictory practices, and thus that they stay legitimate, MNEs can use coordination mechanisms for their CSR practices. By centrally managing, or formalizing their CSR practices, MNEs can keep them aligned. Therefore, a positive relation between cultural distance and the centralization and formalization of CSR practices was expected. This expectation is supported by the findings of this study. The more culturally distant an MNE’s host countries are from its home country, the more centrally managed and the more formalized its CSR practices are. Moreover, this study tried to find out how criticism influences the relation described above. No clear answer, however, can be given to this second question. The findings show that criticism, in five out of six models, does not significantly moderate the relationship between cultural distance and the coordination of CSR practices. Thus, in general, criticism does not have a moderating role in this study, but it might have a different influence. This idea will be further discussed in section 5.1.
In the subsections below, the academic relevance of the findings are discussed first. Then, the managerial implications of this study are explained. Finally, the limitations of the study are given and recommendations for future research are presented.
37
Different studies have been performed on the effects of cultural distance for firms, for example its effects on entry modes, international diversification and performance (Kogut and Singh, 1988; Hennart and Larimo, 1998; Reus and Lamont, 2009; Slangen and Beugelsdijk, 2010; Tihanyi, Griffith and Russel, 2005). However, little was known so far about the relation between cultural distance and the coordinating mechanisms within MNEs. This thesis adds to the understanding of the influence of cultural distance on an MNE’s decision-making, by extending it to the coordination of CSR practices.
The first hypothesis, stating that cultural distance and the centralization of CSR practices are positively related, is supported by the findings of this study. The larger the cultural distance between the home and host countries of MNEs, the more CSR practices are centralized in special teams at the headquarters. This suggest that centrally managing their CSR practices helps MNEs to keep these practices aligned, and therefore to stay legitimate. The second hypothesis, stating that cultural distance and the formalization of CSR practices are positively related, is also supported. The larger the cultural distance, the more policies a firm has with regard to its CSR activities. This suggests that formalization also helps MNEs to stay away from contradictory practices.
This study finds no support for hypotheses 3, 4 and 5. The total amount of criticism an MNE faces does not positively moderate the relation between cultural distance and the coordination of CSR practices, nor does the proportion of transnational criticism. Also, the proportion of criticism in host countries does not negatively moderate the relation between cultural distance and CSR coordination. Therefore, at first sight, it seems as if criticism might not be as important in this case as stated in the literature. However, when taking a closer look, criticism might take a different role than expected in this thesis. First, although the results suggest that criticism is not an important moderator in this study, the findings in table 5 show that there is a positive main relation between the total amount of criticism and the
