The effect of degree of internationalization on criticism : an empirical test of how civil society actors hold MNCs to account in a globalizing era

The effect of degree of internationalization on criticism: an empirical test of

how civil society actors hold MNCs to account in a globalizing era.

Master Thesis

Date: 29th of June 2015

Student: Myrthe Pijl, student number: 10086862

MSc in Business Administration – Strategy Track

University of Amsterdam, Faculty of Economics and Business Supervisor: Dr. Daniel Waeger

1

Statement of Originality

This document is written by Student Myrthe Pijl 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.

2

Abstract

As a result of globalization businesses have increasingly operated across borders and the role of multinational corporations (MNCs) has become more important. Simultaneously, as a consequence of globalization the regulatory power of governments has decreased. In order to compensate for the lack of governmental power, civil society actors are monitoring and exerting pressure on MNCs to hold them account. As a response to these pressures MNCs adjust their business behaviours and have started to engage in CSR activities. In order to hold MNCs that operate around the world accountable for their social and environmental behaviour, civil society actors need to be able to mobilize around the globe as well and in particular in the several host countries in which MNCs are engaged, as well as on a transnational level. This study tests empirically whether MNCs which are more internationally engaged indeed do receive more criticism in their host countries and whether they do receive more transnational criticism (i.e. criticism that appears in a media outlet of one country, but that criticizes the company for something it does in another country). Results show that MNC´s with a higher degree of internationalization do receive more transnational criticism but do not receive more host country criticism.

3

Table of contents

2.4 Challenges to international activism ... 12

2.5 Summary of literature review and research questions ... 14

3. Theoretical framework ... 16

3.1 Host country criticism ... 16

3.2 Transnational criticism ... 17 3.2 Environmental issues... 18 3.3 Cultural distance ... 19 3.4 Freedom of press ... 20 4. Methodology ... 22 4.1 Data collection... 22 4.2 Criticism variables... 22 4.3 Degree of internationalization ... 23 4.4 Cultural distance ... 23 4.5 Freedom of press ... 24 4.6 Control variables ... 25 5. Data analysis ... 27 5.1 Descriptive analysis... 27 5.2 Normality analysis... 28 5.3 Correlation analysis ... 29 5.4 Regression analysis ... 31 6. Discussion ... 35 6.1 Major findings ... 35 6.2 Implications ... 38

6.3 Limitations and future research ... 39

7. Conclusion ... 40

4

1. Introduction

During the past decades businesses have increasingly focused on operating in a socially, environmentally and ethically responsible manner (McWilliams, Siegel, and Wright, 2006; Scherer and Palazzo, 2007; Scherer and Palazzo, 2011). Firms engage in a wide variety of practices that contribute to issues such as health, education, protection of human rights, and counteract climate change (Matten and Crane, 2005). This engagement of firms is referred to as Corporate Social Responsibility (CSR) which has developed as a prominent term in the management business literature (Doh and Guay, 2006; Matten and Crane, 2005; McWilliams et al., 2006). In recent years there is an increasing expectation of firms to engage in CSR activities (Scherer and Palazzo 2007; Scherer and Palazzo, 2011; Scherer, Palazzo and Baumann, 2006). This demand is especially met by multinational corporations (MNCs) (Scherer and Palazzo, 2011). If firms do not meet the expectations of society they will be faced with social and governmental pressures (Baron, 2001). MNCs engage in CSR in order to respond and anticipate to the threat of these pressures (Baron, 2001; King and Pearce, 2010; Proffitt and Spicer, 2006; Scherer and Palazzo 2007; Scherer and Palazzo 2011).

Civil society actors use a wide range of tactics to exert pressure – such as letter writing campaigns, picketing, shareholder resolutions and boycotts – in order to influence the business practices of firms in a way that is consistent with their goals (Spar and La Mure, 2003). Firms react differently to the efforts activists and stakeholders make to express their criticism and influence firm behaviour. They can either pre-empt, capitulate or resists criticism considering the costs and benefits of compliance (Spar and La Mure, 2003). Furthermore, some firms are subject to more criticism than others. Much research has been done on this topic, for example Smith (1996) has found that firm size and level of institutional ownership positively affects the probability of being targeted by criticism. Also, the industry in which a firm operates appears to be an important indicator of criticism.

5 Epstein and Schnietz (2002) found that firms operating in contested industries such as mining, logging, oil, toy and apparel suffer from more criticism than firms operating in non-contested industries. Others have found that firms subjected to primary activism – activism initiated by a firm’s shareholders – are more severely affected compared to firms subjected to secondary activism – activism initiated by non-shareholders (Bogdan Vasi and King, 2012).

While there is much research which discusses how firms are affected by criticism and why some firms face more criticism than others, there is not much known on where it does occur. Knowing where criticism towards MNCs originates is critical for the theoretical perspective of ‘political CSR’. Theorists of political CSR have argued that MNCs do not only engage in CSR activities because of their value-creating contribution, but also because they assume to have a social and political responsibility to do so (Scherer and Palazzo, 2011). This is due to the process of globalization. Over the last decades, firms have increasingly operated across borders and the role of globally operating firms has become more important (Pisani, 2009; Scherer and Palazzo, 2011). Simultaneously, globalization has decreased the regulatory power of democratically elected governments (Nye, 2001; Scherer and Palazzo, 2011). Firms have become more mobile and operate on a global level. They are able to choose between different legal systems and have the capacity to shut down their operations in countries where regulations are deemed as too strict and shift these operations to countries with more favourable legal environments (Scherer & Palazzo, 2011; Surroca, Tribó, and Zahra, 2012). Therefore, the ability of governments to enforce regulations and to control for negative externalities of business activity has declined (King and Pearce, 2010; Korbin, 2009). In reaction, public actors such as national governments and international governmental institutions together with private actors such as MNCs and NGOs have launched governance initiatives and developed rules and criteria in an attempt to compensate for the lack of governmental power (Scherer and Palazzo, 2011). MNCs have sometimes adopted a

state-6 like role and are not only subjected to regulations but also co-develop regulations themselves (Scherer and Palazzo, 2007). Political CSR can be described as businesses entering the political sphere to react to social and environmental challenges and contribute to global regulation and the provision of public goods beyond compliance of legal regulations and stakeholder pressures (Scherer and Palazzo, 2007; Scherer and Palazzo, 2011). Key to these initiatives is that by the development of rules and criteria, external and critical voices are taken into account and that success is monitored independently through third party control (Scherer and Palazzo, 2007). Therefore, political CSR crucially depends on civil society actors being able to monitor where MNCs operate, i.e. in their different host countries and transnationally. Political CSR theorists have argued that in the absence of effective governmental regulations of MNCs, civil society actors have started to target MNCs in their host countries and transnationally and encourage them to engage in self-regulatory activities (King and Pearce, 2010; Scherer and Palazzo 2007; Scherer and Palazzo, 2011). Anecdotal evidence supports this claim. A well-known example is the case of Nike which is severely criticized for the working conditions and wage levels at its Indonesian factories (Spar and La Mure, 2003, Zadek, 2004). Nike was both criticized by major U.S. media networks as by Indonesian civil society actors (Spar and La Mure, 2003). In response, Nike started to adopt self-regulatory mechanisms such as raising the minimum wage, expanding monitoring systems and managing supply chain partners’ compliance with labour standards (Spar and La Mure, 2003). As activists are actively targeting firms such as Nike in their host countries and transnationally, MNCs which are more internationally engaged should logically expect more criticism outside their home environment than MNCs that are less internationalized. Also, they can expect information to travel from one country to another, certainly given the ongoing progress of telecommunication (Nachum and Zaheer, 2005). With respect to criticism, this means that firms can be criticized in one country for something they have done

7 in another country (i.e. Nike is set under pressure in the U.S. for bad working conditions at their suppliers in Indonesia only if the information that there are bad working conditions in Indonesia 'travels' to the American public) (Scherer and Palazzo, 2011; Spar, 1998; Zadek, 2004). MNCs which are more internationally active should expect more of this ‘transnational criticism’ than MNCs which are less internationalized.

This theoretical perspective presupposes that there is a vigilant and active civil society in all different countries around the world, which monitors the practices of globalized firms and therewith supplements state regulation. However, despite anecdotal evidence it should also be noted that civil society might encounter some challenges when mobilizing in an international context. Issues such as cultural differences, language differences, geographical distance and different institutional environments must be overcome in order to establish international collective action (King, 2008; McCarthy & Zald, 1977; Smith, Chatfield, and Pagnucco, 1997). Moreover, interests between actors on international, national and local levels might contradict which could constrain international mobilization (Kreamer, Whiteman, and Banerjee, 2013). Therefore, this study will test empirically whether indeed companies with a higher degree of internationalization receive more criticism outside their home country environment. And whether these companies receive more transnational criticism (i.e. criticism that appears in a media outlet of one country, but that criticizes the company for something it does in another country) than countries with a lower degree of internationalization.

Answering these questions will give more insight in the origins of criticism and will provide more knowledge on the drivers of CSR. Knowing whether MNCs receive criticism on an international and transnational level is important to indicate if MNCs are controlled by civil society. With regard to firms these insights can offer a basis for their strategic choices, as CSR can be a source of innovation, opportunity and competitive advantage (Porter and

8 Kramer, 2006). Finally, the findings of this study can contribute to the effectiveness of civil society actors exerting pressure in an attempt to holding MNCs to account (Spar and La Mure, 2003).

This study is structured as follows; first the literature with regard to these questions will be discussed. Second, a theoretical framework and hypotheses will be drawn. Subsequently, these questions will be answered through a data-analysis using a unique secondary dataset containing criticism data on 235 MNCs. Finally, the empirical results and implications of this study will be discussed.

2. Literature review

Beyond moral arguments, many CSR activities are undertaken for strategic reasons (Scherer and Palazzo, 2007). Reducing negative social and environmental impact may for example improve the overall reputation of a company (Ambec and Lanoie, 2008). A better reputation will increase consumer loyalty and will make a firm more attractive to potential employees (Aguinis and Glavas, 2012). Moreover, a better environmental performance will generally lead to cost reductions, as firms that reduce pollution decrease expenditures on raw materials, energy and services (Ambec and Lanoie, 2008; Porter and van de Linde, 1995). Furthermore, engagement in CSR can improve financial performance because it increases attractiveness to investors (Aguinis and Glavas, 2012).

Another strategic reason for firms to engage in CSR is related to stakeholder relationships. Better environmental performance increases a firm’s reputation and goodwill with external stakeholders (Aguinis and Glavas, 2012). This can foster relationships between firms and their stakeholders such as governments, activists, local communities and the media (Ambec and Lanoie, 2008). When firms perform well, they avoid potential costly fines and

9 lawsuits and they may avoid or anticipate to the risk of future government regulations (Ambec and Lanoie, 2008). Moreover, firms engage in CSR as a response to stakeholder pressure. In recent years there is a growing awareness of unfair firm behaviour. Social expectations have changed and firms are increasingly expected to contribute to social and environmental issues beyond compliance of current regulations (Scherer and Palazzo, 2007). Due to criticisms exerted by the public, firms have started to respond to issues which they formerly did not perceive as their responsibility (Porter and Kramer, 2006; Scherer and Palazzo, 2007). Over the past decades MNCs are increasingly criticized for the social and environmental harm that occurs along their supply chains and other negative externalities to which they are connected (Scherer and Palazzo, 2011; Zadek, 2004). Additionally, activist organisations have become more aggressive and effective in pressuring corporations to change their business practices in line with the expectations of the public (Porter and Kramer, 2006). These organisations often use the media to criticize firms, as generating negative media attention can be very disruptive to the reputation of a firm (King and Pearce, 2010). By improving social and environmental performance, MNCs may reduce the pressures exerted by social movements and other civil society actors (Baron, 2001; Delmas and Toffel, 2004; Scherer and Palazzo, 2007; Scherer and Palazzo, 2011).

2.2 Globalization

According to political CSR theorists globalization is at the origin of CSR (Scherer and Palazzo 2007; Scherer and Palazzo, 2011). Due to globalization there is an intensification of cross-border interaction and a growing interdependence of economic and social actors on many different topics, such as markets, environment, military and politics (Nye, 2001; Scherer and Palazzo, 2011). Therefore, national borders and mandates are becoming more ambiguous and the division between the public and the private sphere is less clear cut (Korbin, 2009; Scherer and Palazzo, 2007; Scherer and Palazzo, 2011). Because more

10 economic and social interactions are taking place beyond the reach of national jurisdictions, national governments are losing their regulatory power and global production takes increasingly place at locations that lack democratic control (Scherer and Palazzo, 2011). Simultaneously, the process of globalization has increased the power of MNCs (Pisani, 2009; Scherer and Palazzo, 2011). However, it has become more difficult to hold MNCs as a whole accountable for their operations and to control for negative externalities of their business practices (King and Pearce, 2010; Korbin, 2009; O’Rourke, 2003). In order to compensate for this regulatory gap, firms in cooperation with governments and civil society have started to fill these gaps and provide global public goods (Scherer and Palazzo, 2011). Businesses have started to fulfil functions which were formerly regarded as governmental responsibilities such as implementing and protecting citizenship rights (Scherer and Palazzo, 2011). MNCs have adopted political roles and contribute to global democratic regulation and control of economic activity (Scherer and Palazzo, 2011). CSR is concerned with firms fostering these public goods beyond government requirements (Doh and Guay, 2006).

2.3 Exerting pressure

MNCs have the objective to maximize profits, therefore they will not respond to social expectations in the absence of stakeholder pressure (Campbell, 2007). MNCs do thus not engage in CSR on a voluntary basis and stakeholder pressure is key in an attempt to compensate for the weakened state and therewith the declined democratic control of MNCs. Civil society actors exert pressure and advocate changes on certain issues by mobilizing resources, providing information and through activism (Spar and La Mure, 2003). Governments were previously their obvious target, but in the past decades the focus has moved from government to business, particularly to MNCs (King and Pearce, 2010; Scherer and Palazzo, 2011; Spar and La Mure, 2003). This is due to the fact that the regulatory power of governments has decreased and MNCs have gained importance in the global economy

11 (Scherer and Palazzo, 2011; Spar and La Mure, 2003). Civil society actors such as NGOs use the market as a platform for social change as MNCs have a considerable amount of economic influence which can be translated into political power (Spar and La Mure, 2003). Additionally, businesses are often more responsive to pressures than governments (King and Pearce, 2010). Targeting businesses can therefore be more effective than targeting governments directly (King and Pearce, 2010; Spar and La Mure, 2003).

In order to be most effective activists select a target that has the highest potential to address the issue on which they focus and which is highly sensitive to external pressure (Spar and La Mure, 2003). Often this is a firm with a high profile brand which is well recognized and convenient to the public rather than a government (Kostova and Zaheer, 1999; Spar and La Mure, 2003; Zadek, 2004). By focussing on such wealthy and generally known firms, an tangible target is attached to a broader cause (Spar and La Mure, 2003). Civil society actors target firms to portray the message that there is a broader problem (i.e. child labour, climate change, deforestation) which the targeted firm can address by adapting its business behaviour (Spar and La Mure, 2003). In response, firms will avoid this threat by bringing their business practices in line with the demands of the public (Spar and La Mure, 2003). Essential is that activists are able to harm the firm in question significantly, that firms respond by changing their firm behaviour, and that this change contributes to the goals of the activists (Spar and La Mure, 2003).

In effort to foster stakeholder pressure civil society actors have developed “market-oriented, nongovernmental standards and monitoring systems as a supplement to state regulation” (O’Rourke, 2003: 5). Examples are naming-and-shaming campaigns which can damage the image of a firm, or the creation of certification systems which can help to define appropriate corporate behaviour and signify which firms do not meet the standards (King and Pearce, 2010). The general objective of these civil society actors is to create negative public

12 attention which will harm the reputation of a MNC (Spar and La Mure, 2003). An important way to do is, is by drawing media attention (King and Pearce, 2010). Media attention creates public interest and support for a cause and frames the way in which the public regards an issue (King and Pearce, 2010). Also, media attention can influence third parties and bring them into a conflict or cause them to disassociate themselves from a MNC. This can be harmful when negative media attention for example leads governments to impose new regulations or provokes clients or suppliers to leave the company (King and Pearce, 2010). Media attention can be evoked by organizing events which are possibly newsworthy, by framing the perspectives of media representatives and by manipulating media coverage (McCarthy and Zald, 1977). The ability of social movements to induce the media is crucial to exert influence (McCarthy and Zald, 1977). MNCs that receive much negative attention in the media may face a loss of reputation and legitimacy which can be very disruptive (King and Pearce, 2010). Therefore, MNCs will limit this threat by bringing their business behaviour in line with the demands of the critics (King and Pearce, 2010; Spar and La Mure, 2003). In this way civil society is able to control for the negative externalities of the economic activities of MNCs.

2.4 Challenges to international activism

Given the fact that MNCs are increasingly operating across borders, pressure must be exerted on an international and transnational level in order to be effective. However, literature suggests that it is harder to exert pressure on these levels than on a national level. King (2008) argues that in order to influence corporations collective action of stakeholders is necessary as it binds individuals together. One of the main factors facilitating collective action are ‘mobilizing structures’ which can be distinguished into two types. ‘Formal organisations’ which form an organised basis for collective action and coordinate strategies among stakeholders and ‘informal networks’ which encourage and support individuals to

13 engage in collective action and facilitate convergence of interests among stakeholders. Activists are more likely to criticise companies when these mobilizing structures are present (King, 2008). However, on international and transnational level these structures are often missing. In developing countries, civil society often has a lack of know-how and technical resources (Keck & Sikkink, 1999). Moreover, in these countries civil society actors usually face oppression and do not have the political resources to influence governments (Kreamer et al., 2013). This indicates that civil societies particularly in the host countries of MNCs, often lack formal organisation for mobilization. This might also be the reason why some companies place irresponsible activities in countries with less stringent regulations. Some MNCs react to public pressures in their home countries by shifting their socially irresponsible activities to their subsidiaries located in countries where civil society lacks mobilizing structures and where their business activities are less susceptible to public scrutiny (Surroca et al., 2013). In this way the call for CSR in a company’s home country can result in corporate social

irresponsibility in a company’s host country (Surroca et al., 2013). If the transfer of socially

irresponsible activities goes unnoticed, this gives MNCs the appearance of compliance which can retain or even increase their legitimacy (Surroca et al., 2013).

Additionally, on a transnational and international level it is more difficult to establish informal networks. Because civil society actors on international levels do normally not meet in person, they cannot be bound through solidary selective incentives (McCarthy & Zald, 1977). Therefore, more time and resources are required to organise geographically distant individuals (McCarthy & Zald, 1977). Furthermore, in order to achieve activism on a transnational level, issues such as nationalism, language differences and cultural differences must be overcome to create strong informal networks (Keck & Sikkink, 1999). Because there are great ideological, historical, cultural and language differences throughout the world,

14 establishing a successful and widely embraced frame for collective action it is a more extensive process on a transnational level than on a national level (Smith et al,. 1997).

Another factor inhibiting international activism is related to group identity. According to Rowley and Moldoveanu (2003) action of stakeholder groups is driven both by the expected outcome of an action as well as by the expression of group identity which an action confers. Groups that have mobilized in the past have stronger group identities as relationships and norms have been established in past experience. Therefore, these groups are more likely to mobilize in the future (Rowley and Moldoveanu, 2003). Because in the host countries of MNCs there is no long history of public scrutiny over a MNC, civil society actors have not much past experience to build on and are therefore less likely to criticize these MNCs.

Furthermore, once transnational or international collaboration is established, relationships between different local, national and international actors have to be managed well. Often these relationships are characterized by conflict (Kreamer et al., 2013). Actors on different levels have different understandings on local issues and international actors can misinterpret the interests of domestic stakeholders. Additionally, the agenda of international, national or local stakeholders may conflict. Therefore, even though different actors on different levels adhere to a common cause, there can be tensions which constrain effective international collaboration (Kreamer et al., 2013).

2.5 Summary of literature review and research questions

One of the reasons MNCs have started to adopt CSR activities is to foster stakeholder relationships and to react to stakeholder pressures (Aguinis and Glavas, 2012; Scherer and Palazzo; 2007). Civil society actors have increased pressures towards MNCs as result of globalization. Due to globalization the regulatory power of national governments has declined and businesses operate increasingly without being democratically controlled (Scherer and Palazzo, 2011). Therefore, civil society actors exert criticism on MNCs to hold

15 them to account. Often this is done through the media as negative media attention influences consumers’ and other stakeholders’ perceptions of firms which could be harmful to their reputations (Aguinis and Glavas, 2012; Newell, 2008). In the theoretical perspective of political CSR it is argued that MNCs that operate on a global scale will also be held accountable by an internationally connected civil society (Scherer and Palazzo, 2007; Scherer and Palazzo, 2011). This argument seems intuitively compelling as companies that are more internationalized would expect more criticism that is not originating from their home country than companies which are less internationalized. Sharing economy companies such as Airbnb and Uber have for instance received much criticism when they entered the European market (Cusumano, 2015; Geradin; 2015). Soon after these companies were founded in the U.S. in 2008 and 2009, they received criticism from U.S. opponents and faced regulatory pressures from U.S. city governments, courts and unions (Cusumano, 2015). Also, their arrival in Europe has triggered European civil society actors to exert pressure and European media have argued that these firms engaged in unfair competition and claimed these companies did not comply with national regulations (Cusumano, 2015; Geradin; 2015). However, they only received criticism from European civil society actors once they had entered the European market. Thus, firms such as Airbnb and Uber are only criticized on international levels once they are more internationally active. However, as the discussed literature has shown, civil society may encounter important challenges when mobilizing internationally. Therefore, this study will examine where criticism directed at MNCs does occur and will empirically test whether there is indeed a connection between degree of internationalization and transnational and host country criticism. Hence, the research questions this study aims to answer are:

Do companies with a higher level of internationalization receive proportionally more criticism from host countries, rather than from their home country environment?

16

Do companies with a higher level of internationalization receive proportionally more transnational criticism (i.e. criticism that appears in a media outlet of one country, but that criticizes the company for something it does in another country)?

3. Theoretical framework

Over the last decades, firms have increasingly operated across borders. This is due to decreasing costs of communication and the transfer of capital, goods and people (Scherer and Palazzo, 2011). Also, the role of globally operating firms has become more important (Pisani, 2009; Scherer and Palazzo, 2011). Simultaneously, the regulatory power of nation states has declined. MNCs operate globally through a network of subsidiaries which are locally embedded (Korbin, 2009). Hence, it is more difficult for national governments to control MNCs because they cannot operate outside their own jurisdiction (Korbin, 2009). Therefore, firms should be democratically controlled by civil society in order to operate legitimately (Scherer and Palazzo, 2011). Civil society actors do this by criticizing MNCs and exerting pressure (Spar and La Mure, 2003). In return MNCs adjust their operations in order to avoid or respond to this criticism (Spar and La Mure, 2003). MNCs which are more internationally engaged and thus have a higher degree of internationalization, can logically be expected to receive more criticism outside their home environment than MNCs which have a lower degree of internationalization. Hence, it seems intuitively logical to expect that if a company such as Uber only has operations in the U.S. and Canada it is less likely that Uber comes under public scrutiny in European media than when Uber has expanded its operations and is actively operating in Europe as well.

Hypothesis 1: Companies with a higher degree of internationalization are exposed to more public criticism from their host countries than companies with a lower degree of internationalization.

17 3.2 Transnational criticism

Because of the declining costs of operating globally, firms shift their businesses to locations which are most profitable (Newell, 2001). Often, this is at the expense of the environment and communities of their host countries (Newell, 2001). As the regulatory power of governments has eroded and because of the lack of effective regulation on an international level, civil society actors are monitoring MNCs in an attempt to hold them accountable for their social and environmental responsibilities (King and Pearce, 2010; Newell, 2001). This is often done by civil society actors in the home countries of MNCs as they feel they have a moral duty to ensure that MNCs maintain the same standards worldwide (Muchlinski, 1999 as cited in Newell, 2001). Home county actors evaluate the overseas practices of MNCs based on standards applied in the home country of the MNC (Martin, 2002; Werther and Chandler, 2005). Some firms apply local social and environmental standards to their practices in host countries which often is not appreciated by the public in the home country. Notorious example is the case of Nike which ran its footwear plants in Indonesia according to local standards. In response Nike received extensive criticism from U.S. civil society which expected Nike to apply U.S. standards to its operations overseas, and regarded their behaviour as exploitative even though it was conform to local practices (Martin, 2002; Werther and Chandler, 2005; Zadek, 2004). Due to the declining costs of the exchange of information, home country actors are nowadays more easily able to track the corporate behaviours of MNCs and wrongdoings get worldwide coverage in the media (Nachum and Zaheer, 2005; Werther and Chandler, 2005). Therefore, MNCs which are more internationalized can expect more transnational criticism (i.e. criticism that appears in a media outlet of one country, but that criticizes the company for something it does in another country) than MNCs which are less internationally engaged.

18

Hypothesis 2: Companies with a higher degree of internationalization are exposed to more transnational public criticism than companies with lower degree of internationalization.

3.2 Environmental issues

When civil society actors are more strongly transnationally connected it is more likely that they collaborate effectively at the transnational level. When there is a common identity among these actors, and norms and practices are established, there is an increased likelihood that these actors will mobilize in an effective way (Kreamer et al., 2013; Rowley and Moldoveanu, 2003). An issue for which civil society actors are more transnationally connected is the environment. Environmental movements have gained a prominent place in world politics (Van der Heijden, 2006). There is a long history of civil society actors attempting to hold governments and corporations accountable for their environmental footprints (Newell, 2008). They often use confronting symbolic actions to reach an audience that is far away, which can foster the growth of their networks (Keck and Sikkink, 1999). Also, some environmental groups use more liberal strategies intended to reform business practices and regulations in order to achieve change (Newell, 2008). Traditional strategies of environmental movements aimed at governments (Newell, 2008). Tactics such as lobbying governments, influencing negotiating delegations and attracting media attention to expose non-compliance with commitments and targets are frequently used (Newell, 2008). Environmental groups encourage governments to sign up to international treaties which oblige them to take measures to tackle climate change such as the Kyoto Protocol (Newell, 2008). Recently, there is an increasing tendency to holding the private sector accountable for their environmental responsibilities as well (Newell, 2008). Because MNCs are able to operate within different jurisdictions and move between them they are often not democratically controlled and therefore environmental civil society actors have increased scrutiny to hold them to account (Newell, 2008). Environmental mobilization is characterized

19 by its transnational nature (Newell, 2008). Environmental activists do not only target national governments in an attempt to save the planet, they use transnational economic, social and cultural networks to achieve their goals (Wapner, 1995). They develop coalitions between multiple actors such as governments, cities, states, communities and civil societies to work together to establish change (Newell, 2008). Because environmental activists are more transnationally connected, it is more likely that they will exert international criticism and induce negative media attention. Therefore, MNCs can expect more transnational criticism related to environmental issues.

Hypothesis 3: The relationship between degree of internationalization and transnational public criticism is stronger for companies with a higher amount of criticism related to environmental issues.

3.3 Cultural distance

Acceptance of firms by the environment in which they operate is critical to their survival (Hannan and Freeman, 1984). Each environment has different economic institutions, political regulations and cultural norms, local adaptation to these issues is key to gain legitimacy (Husted and Allen, 2006). When firms start to do business abroad they encounter costs related to the need to coordinate across these geographical differences (Zaheer, 1995). As domestic firms do not face such costs, firms operating abroad face a ‘liability of foreignness’ (Kostova and Zaheer, 1999). Because MNCs operate in multiple countries, it might be challenging to gain legitimacy in all the different environments they are active in and to overcome liability of foreignness (Kostova and Zaheer, 1999). Each subsidiary is embedded in its own institutional environment, which might has a different perspective on legitimacy than the other environments a MNC is facing (Kostova and Zaheer, 1999). Therefore, it can be difficult to satisfy all the demands of varying environments because these demands can be mutually exclusive. This can be a source of criticism, as some business practices may be

20 acceptable in one country but unacceptable in a different country (Kostova and Zaheer, 1999). Also, organizational practices usually reflect norms of the institutional environment in which a firm has been developed. Therefore, it will be easier for firms to adjust in environments which resemble the institutional environment of their home countries (Kostova and Zaheer, 1999). In case MNCs operate in several countries which are culturally distant, it will thus be more difficult to adjust and to keep all stakeholders content, which leads to more pressure from civil societies.

Hypothesis 4: The relationship between degree of internationalization and host country criticism is stronger for companies which are active in culturally distant contexts.

On the other hand, cultural distance could lead to less criticism towards MNCs. For firms active in countries with varying cultures it appears to be more difficult for civil society actors to connect across borders. When there are large cultural differences, issues such as language, ideology and history can make the establishment of transnational collaboration more difficult (McCarthy & Zald, 1977; Smith et al., 1997). These differences can make the development of a common group identity more difficult and can also lead to different understandings which could create conflicts between actors on different levels (Kreamer et al., 2013; Rowley and Moldoveanu, 2003). Therefore, MNCs which face a variety of distant cultures in their operations can expect less criticism.

Hypothesis 5: The relationship between degree of internationalization and transnational criticism is weaker for companies which are active in culturally distant contexts.

3.4 Freedom of press

Civil society actors often seek media attention to make the irresponsible behaviours of MNCs known by the public. By organizing press conferences and newsworthy events they attempt to

21 influence consumers’ and other stakeholders’ perceptions which could harm the reputation of a firm (Aguinis and Glavas, 2012; Newell, 2008). Negative media attention has the greatest impact when it reaches a large number of stakeholders (Dyck, Volchkova, and Zingales, 2008). When there is more press freedom in a certain country, it is easier for civil society to exert criticism and to reach many people. Freedom of press is indicated by the freedom media has to publish or broadcast and by the freedom and access citizens have to media content (McQuail, 2000). Press freedom entails the absence of governmental and non-governmental restrains on the media and the assurance of the dissemination of different opinions and ideas to large audiences (Weaver, 1977). High levels of press freedom increases citizen participation as citizens can use free media to address problems in society (Becker, Tudor and Nusser, 2007). A free press is a crucial mechanism needed for external control (Brunetti & Weder, 2003). Possible wrongdoings of MNCs can be revealed when there is a free and active press. Also, when the press is free and competitive and journalists can investigate any type of irresponsible firm behaviour they are incentivized to discover such wrongdoings and publicize them (Brunetti & Weder, 2003). MNCs which are exposed to high levels of press freedom will therefore receive more criticism.

Hypothesis 6: the relationship between degree of internationalization and host country criticism is stronger for companies that are exposed to high levels of free press across all the countries it is active in.

When MNCs are exposed to more freedom of press it is more likely that information about irresponsible activities travels from one country to another. Therefore, companies which are exposed to high levels of press freedom will receive more transnational criticism.

Hypothesis 7: the relationship between degree of internationalization and transnational criticism is stronger for companies that are exposed to high levels of free press across all the countries it is active in.

22

4. Methodology

4.1 Data collection

For the purpose of this study a unique dataset is developed by combining four secondary data sources: Covalence EthicalQuote, the Carbon Disclosure Project (CDP), Thomson Reuters Worldscope and Thomson Reuters Datastream. The EthicalQuote reputation index tracks online news on topics relevant to CSR based on 50 CSR-criteria developed by the Global Reporting Initiative (EthicalQuote, 2015). Each news item is coded in terms of whether it constitutes criticism or praise directed towards a 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. CDP reports on the emission of greenhouse gasses of major corporations by sending out a detailed questionnaire to these companies. Particularly relevant for the present study, the CDP asks MNCs both to report their total greenhouse gas emissions and to break down their total emissions by country. Financial and company specific information is gathered from Thomson Reuters Worldscope and Thomson Reuters Datastream.

The original sample consisted of the 541 companies tracked by Covalence Ethicalquote for 2011. The year 2011 was chosen because it constitutes the most recent year full data from Covalence Ethicalquote was available. After combining the different data sources, the original sample of 541 companies was reduced to 235 companies for which full data is available.

4.2 Criticism variables

The criticism variables are constructed based on the total number of times a MNC was negatively mentioned in CovalenceEthicalquote's criticism database with respect to CSR topics. Host country criticism is operationalized bythe number of times the focal company was the target of CSR-relevant public criticism outside its home country during 2011.

23 Transnational criticism is operationalized by the number of times the focal company was the target of transnational CSR-relevant public criticism during 2011. Criticism is classified as transnational when the source of the criticism was not located in the country in which the criticized incident took place. Criticism on environmental issues is operationalized by the number of times the focal company was the target of public criticism concerning environmental issues during 2011.

4.3 Degree of internationalization

A MNC’s degree of internationalization is calculated based on the Carbon Disclosure Project database. This variable is calculated by dividing the total greenhouse gas emissions the focal company emits in its home country by the total greenhouse gas emissions the focal company emits in order to obtain the percentage of the greenhouse gas emissions the company emits in its home country. Subsequently, this percentage is subtracted from 1. This value indicates the percentage of greenhouse emissions the focal company emits outside its home country. The higher this percentage the more internationalized a firm is. Common used measures to calculate degree of internationalization are foreign assets over total assets or foreign sales over total sales (Sullivan, 1994). In this particular study these measures have not been used because for many companies in the Worldscope database that data was not available.

4.4 Cultural distance

To calculate cultural distance data from Hofstede’s cultural dimensions is combined with data from the Carbon Disclosure Project. To determine which countries a company is active in, data from CDP is 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. 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).

24 𝐶𝐷𝑗 = ∑{(𝐼𝑖𝑗 − 𝐼𝑖𝑢)²/𝑉𝑖}/4

4

𝑖 =1

Kogut and Singh (1988) use Hofstede’s cultural indices, which attribute a score on four dimensions of national culture (i.e., power distance, uncertainty avoidance, masculinity/femininity and individualism). For this study the deviation on these dimensions from the home country of a MNC compared to its host countries determines cultural distance. 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. In the next step, 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 on a specific dimension. Once the absolute cultural distance scores for each one of the four cultural dimensions had been calculated, these scores were added up and divided by four. This resulted in the variable absolute cultural distance.

4.5 Freedom of press

To determine the level of press freedom a MNC is exposed to throughout all the countries it is active in, data from Reporters Without Borders’ Press Freedom Index and data from the Carbon Disclosure Project are combined. To determine in which countries a company is active and how important these countries are to the company, CDP data is used. Companies reported in the CDP questionnaire their total greenhouse gas emissions and their emissions per country they are active in. The importance of the host countries is calculated by dividing the total amount of greenhouse gas emissions the company emitted in a country by the total amount of greenhouse gas emissions of that company. This information is then combined with Reporters Without Borders’ Press Freedom scores for 2011/2012. For each country a company is active in, the country’s Freedom of Press score was multiplied by that country’s importance for the company. This resulted in a weighted Freedom of Press score for each

25 country the company is active in. Once this weighted Freedom of press score had been computed for each one of the countries the company is active in, these scores were added up. This resulted in the variable ‘exposure to a free press’. A low score on the scale corresponds to greater exposure to freedom of press, a higher score to a lower exposure to of press freedom. Negative scores are possible.

4.6 Control variables

Profitability

Firms which control significant amounts of economic power are a logic target for civil society actors (Spar and La Mure, 2003). When economic power can be translated into political power, corporations can influence a country’s laws and regulations. Therefore, it will be interesting for civil society actors to target corporations with great economic power to indirectly exert pressure on governments and regulations (Spar and La Mure, 2003). In this respect, the more economic power a corporation has the more likely civil society will target this firm. To control for this effect, profitability will be used as a control variable. Hence, the more profitable a corporation is the more economic influence it has. The return on assets of the focal company is taken as indicator for profitability.

Firm size

When civil society actors decide to target a corporation, it is more likely that large corporations with high-profile brands are targeted (Zadek, 2004). The these corporations do usually not preform poorer than their competitors on social or environmental issues but these corporations are familiar to the public and by exerting pressure on these companies civil society can reach a wider public (Zadek, 2004). Therefore, MNCs which are familiar to the public are more likely to receive criticism. To control for this effect firm size is taken as indicator, firms of greater size are more visible to the public and therefore more prone to be criticized. The total assets of the focal firm is used as indicator of firm size.

26

Industry

Some industries are more likely to be criticized than others (Epstein and Schnietz, 2002; Maloni and Brown, 2006). MNCs which operate in industries that are more susceptible to CSR related criticism or MNCs active in industries which are more visible to the public are more likely to receive criticism. In order to take industry effects into account, this study will control for industry classifications via the item “General Industry Classification” from Thomson Reuters Worldscope. The MNCs are classified in six industry categories; industrial, utility, transportation, banking, insurance, and other financial company.

Regional differences

Regional differences could influence the extent to which companies are criticized in the media with regard to CSR issues. Economically advanced countries, have created structures and incentives that support CSR which are deeply embedded in their civil foundations (Martin, 2002; Soares, Lucian, Farache, and Milton de Sousa Filho, 2008). In Europe most countries have developed laws which enforce companies to report on CSR performance, have research centres and dedicate government budgets to support CSR practices (Soares et al., 2008). However, in developing countries CSR is often of less interest. In these countries civil society is less developed, governments do not promote CSR as strongly and media are criticizing corporations more moderately (Martin, 2002; Soares et al., 2008). These differences could indicate that MNCs operating in developing countries are less severely criticized by civil society. Next to level of economic development other country specific factors such as the extent of public regulations and the structure of the economy could influence the extent to which companies receive criticism on CSR issues (Gjølberg, 2009). Therefore this study controls for regional differences by classifying the MNCs by the world

27 region, in which they are headquartered. The regions distinguished are; Asia, Africa, Europe, North & Central America, South America and Australia & Oceania.

Total criticism

Because some companies are more susceptible to criticism than others (Epstein and Schnietz, 2002), they could receive more transnational criticism or criticism in their host countries because they receive generally more criticism than other companies. To control for this effect, total criticism will be used as control variable. Total criticism is calculated based on the total number of times a MNC was negatively mentioned in CovalenceEthicalquote's criticism database with respect to CSR topics in 2011.

5. Data analysis

Categorical variables

Table 1 presents descriptive statistics for the categorical variables used. From a total of 235 MNCs, the majority is headquartered in North & Central America (42.1 percent) and Europe (38.3 percent). 75.3 percent of the MNCs is classified within the Industrial industry, the remaining industries are less represented.

TABEL 1

Descriptive statistics categorical variables

Variable Level Frequency % Valid % Cumulative % World region Asia 30 12.8 12.8 13.2

Africa 4 1.7 1.7 14.9

Europe 90 38.3 38.3 53.2

North & Central America 99 42.1 42.1 95.3 South America 2 .9 .9 96.2 Australia & Oceania 9 3.8 3.8 100 Industry Industrial 177 75.3 75.3 75.3 classification Utility 22 9.4 9.4 85.0 Transportation 3 1.3 1.3 86.3

Banking 17 7.2 7.2 93.6

Insurance 9 3.8 3.8 97.4

28

Continuous variables

Table 2 presents the descriptive statistics for the continuous variables. The maximum number a MNC was negatively mentioned in the media of a host country is 246. The maximum number a MNC was negatively mentioned in the media outlet of one country about its activities in another country is 228. With respect to criticism on environmental issues, the maximum number a MNC received environmental criticism is 136. Based on Reporters Without Borders’ Press Freedom Index a MNCs exposure to free press across all the countries it is active in is calculated. A low score on the scale corresponds to greater freedom of press, a higher score to a lower level of press freedom (Reporters Without Borders, 2015). The lowest score a MNC in the sample received is -8.75, the highest score is 80,14.

TABEL 2

Descriptive statistics continuous variables

Variable N Min Max Mean Standard Deviation

Skewness Kurtosis Host country criticism 235 0 246 11.83 26.1 5.32 37.69 Transnational criticism 235 0 228 8.89 22.32 6.49 53.79 Degree of internationalization 235 0.00 1.00 0.45 0.32 .25 -1.24 Cultural distance absolute 235 0.00 265.7 29.63 40.88 -.25 -.73 Exposure to freedom of press 235 -8.75 80.14 14.98 13.42 .44 .93 Environmental criticism 235 0 136 3.67 13.04 1.49 2.3 Firm size 234 2437482 36451079000 976188561.3 3816387644 .90 .30 Profitability 226 -8.12 30 6.9 5.78 .10 1.44 Total criticism 235 0 321 18.86 37.71 .23 -.84

5.2 Normality analysis

A normality test is conducted to assess whether the variables are normally distributed. All the variables except for degree of internationalization have skewness and kurtosis coefficients above the threshold of .5 and 1 which indicate that these variables depart from normality (Pallant, 2013). To overcome this, the variables cultural distance absolute, criticism on environmental issues, firm size and total criticism are normalized with a logarithm. The variables exposure to freedom of press and profitability are normalized with a square root. The dependent count variables host country criticism and transnational criticism also violate the assumption of normality. These variables are not transformed because by choosing a

29 Negative Binomial Regression instead of a standard regression, the distribution of these variables is taken into account.

5.3 Correlation analysis

For the purpose of the correlation analysis the categorical variables world region and industry classification are transformed into dummy variables. A Pearson product-moment regression is performed to assess the relationship between the different variables. The results are shown in table 3. There is a weak positive correlation between the independent variable degree of internationalization and the dependent variables host country criticism (r = .22, n = 235 p < .001) and transnational criticism (r = .28, n = 235 p < .001) with high degrees of internationalization associated with more host and transnational criticism. The other correlations between the dependent, independent, moderators and control variables can be found in table 3.

30 TABEL 3

Pearson correlations

Variable 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

1. Host country criticism - 2. Transnational criticism .94** - 3. Degree of internationalization .22** .28** - 4. Cultural distance absolute .16* .17** .38** - 5. Exposure to freedom of press .04 .06 .40** .38** - 6. Criticism on environmental issues .85** .85** .85** .10 .02 -

7. Total criticism .89** .81** .81** .17** .06 .67** - 8. Firm size .14* .05 -.09 .09 -.07 .07 .13* - 9. Profitability .44** .39** .11 .15* .00 .32** .49** .5** - 10. Industrial -02 .05 .01 .17** .10 .07 -.02 -.34** -.21** - 11. Utility .12 .05 .03 -.10 .03 .02 .12 .04 .10 -.56** - 12. Transportation -.04 -.04 -.03 -.13* -.04 -.03 -.03 -.07 .02 -.20** -.04 - 13. Banking -.05 -.06 -.03 -.03 -.06 -.06 -.03 .38** .21** -.49** -.09 -.03 - 14. Insurance -.07 -.06 .07 -.06 -.10 -.05 -.08 .11 .08 -.35** -.06 -.02 -.06 -

15. Other financial company .02 -.02 -.07 -.06 -.09 -.05 .02 .16* -.08 -.28** -.05 -.02 -.05 -.03 -

16. Asia -.05 -.11 -.24** .02 -.09 -.04 -.10 .63** -.09 .13* -.08 -.04 -.06 -.08 .02 -

17. Africa -.05 -.05 -.03 -.12 .08 -.02 -.05 .02 -.13* .00 .07 -.02 -.04 -.03 -.02 -.05 -

18. Europe .21** .23** .51** .22** .04 .10 .13* -.07 .24** -.20** .12** .07 .08 .07 -.07 -.30** .10 -

19. North & Central America .12 -.11 -.29** -.11 .03 -.05 -.02 -.38** -.18** .17** -.13 -.02 -.11 -.04 .03 -.33** -.11 -.67** -

20. South America .01 .00 -.06 -.01 .13 .01 -.01 .07 .11 .05 -.03 -.01 -.03 -.02 -.02 -.04 -.01 -.07 -.08 -

21. Australia & Oceania -.07 -.06 -.07 -.23** -.15* -.02 -.07 .01 .03 -.14* -.06 -.02 .20** .08 .12 -.08 -.03 -.16* -.17** -.02 ** p < .001 (2-tailed)

31 5.4 Regression analysis

To analyse relationships with count variables as dependent variable either Poisson Regression or Negative Binomial Regression is most appropriate. For both dependent variables (host country criticism and transnational criticism) the variance is not equal to the mean which indicates that both variables are overdispersed. Therefore Negative Binomial Regression is the best fit to test the hypotheses. This study has two different dependent variables (host country criticism and transnational criticism), therefore two Negative Binomial Regressions are performed. The first regression with host country criticism as dependent variable will be used to test hypothesis 1, 4 and 6. The second regression with transnational criticism as dependent variable will be used to test hypothesis 2, 3, 5 and 7.

Host country criticism

The first Negative Binomial Regression is used to test hypothesis 1, 4 and 6, after controlling for firm size, profitability, total criticism, world region and industry classification. In the scatterplot of the standardized residuals some scores are above the threshold of 3,3 which indicate presence of outliers (Pallant, 2013). Therefore observations with a score above 150 are not included in the regression analysis. The results of the first Negative Binomial Regression are presented in table 4. To test the hypotheses, variables are entered in steps in predetermined order. In the first model the control variables are entered, in the second model the independent variable degree of internationalization is included. Model 2 shows that there is a positive relationship between degree of internationalization and host country criticism, however this relation is not statistically significant which rejects hypothesis 1. To test hypothesis 4, the interaction term between cultural distance absolute and degree of internationalization is included in model 3. This interaction effect is not statistically significant which leads to rejection of hypothesis 4. In model 4 the interaction term between

32 exposure to freedom of press and degree of internationalization is added. This effect is positive but not statistically significant. Therefore hypothesis 6 is also rejected.

TABEL 4

Results regression analysis host country criticism

Dependent variable: Host country criticism Model 1

Model 2 Hypothesis 1 Model 3 Hypothesis 4 Model 4 Hypothesis 6 Constant -2.571(2.115) -2.540(2.114) -2.694(2.129) -2.352(2.125) Independent variables Degree of internationalization .329(.319) .365(.335) .340(.372) Cultural distance absolute -.085(.168)

Exposure to press freedom -.038(.085) Cultural distance absolute x Degree of internationalization .060(.091)

Exposure to press freedom x Degree of internationalization .091(.098)

Control variables

Firm size -.004 (.215) -.013(.216) .003(.218) -.025(.218) Profitability .066 (.138) .072(.138) .078(.139) .061(.141) Total criticism 2.982***(.198) 2.963***(.199) 2.982***(.201) 2.999***(.207)

Control variables: industry

Industrial -.052(.591) -.120(.591) -.133(.591) -.103(.601) Utility .159(608) .160(.605) .118(.608) .141(.614) Transportation -.141(1.00) -.154(.996) -.182(.997) -.097(1.01) Banking -.248(.641) -.273(.639) -296(.639) -.215(.647) Insurance .290(.670) .243(.699) .232(.699) .320(.708) Other financial company

Control variables: world region

Asia 1.328(.749) 1.353(.749) 1.430(.748) 1.390(.748) Africa .144(.929) .155(.923) .138(.923) .225(.923) Europe 1.082(.584) .945(.596) 1.013(.598) .953(594) North & Central America .557(.580) .522(.579) .592(.580) .586(.582) South America 1.500(.945) 1.531(.946) 1.583(.943) 1.777(.991) Australia & Oceania

Omnibus test

Ratio Chi-Square 471.148 472.209 472.912 473.084 P-value .000 .000 .000 .000 *** p < .000 ** p < .01 * p< .05

33

Transnational criticism

The second Negative Binomial Regression is used to test hypothesis 2, 3, 5 and 7, after controlling for firm size, total criticism, profitability, world region and industry classification. Because there is no variation in the data for the effect of the world regions Africa and Australia & Oceania, these control variables are not included in the analysis. In the scatterplot of the standardized residuals some scores are above the threshold of 3,3 which indicate presence of outliers (Pallant, 2013). Therefore observations with a score above of 150 are not included in the regression. The results of the second Negative Binomial Regression are presented in table 5. Again the variables are entered in steps in predetermined order. In model 1, the control variables are included and the in model 2 the independent variable degree of internationalization is entered. Hypothesis 2 is supported as high levels of internationalization are associated with more transnational criticism (β = .793, p <.05). To test hypothesis 3 the interaction term between criticism on environmental issues and degree of internationalization is added in model 3. This interaction is not statistically significant meaning there is no support for hypothesis 3. To test hypothesis 5, the interaction of cultural distance absolute and degree of internationalization is included in model 4. This interaction is not statistically significant which rejects hypothesis 5. The interaction effect of degree of internationalization and exposure to freedom of press on transnational country criticism is not statistically significant which rejects hypothesis 7.

34 TABLE 5

Results regression analysis transnational criticism

Dependent variable: Transnational criticism Model 1

Model 2 Hypothesis 2 Model 3 Hypothesis 3 Model4 Hypothesis 5 Model 5 Hypothesis 7 Constant -4.256(2.205) -4.543*(2.208) -4.591*(2.245) -4.573*(2.223) -3.774(2.260) Independent variables Degree of internationalization .793*(.325) .784*(.329) .847*(.341) .970*(.391) Criticism on environmental issues .295(.288)

Cultural distance absolute -.040(.173)

Exposure to press freedom -.098(.084) Criticism on environmental issues x Degree of internationalization .008(.101)

Cultural distance absolute x Degree of internationalization _-.058(.097)

Exposure to press freedom x Degree of internationalization .124(.100)

Control variables

Firm size -.036(.214) -.018(.214) -.017(.220) -.011(.216) -.072(.219) Profitability .190(.157) .216(.159) .251(.162) .222(.167) .191(.161) Total criticism 2.874***(.205) 2.807***(.206) 2.664***(.241) 2.796***(.208) 2.861***(.216)

Control variable: industry

Industrial .549(.668) .380(.659) .178(.680) .381(.659) .421(.668) Utility .584(.688) .556(.677) .359(.698) .542(.676) .575(.687) Transportation -.445(1.152) -.433(1.128) -.534(1.127) -.476(1.131) -.385(1.149) Banking -.087(.728) -.249(.722) -.377(.719) -.245(.721) -.089(.737) Insurance .862(.781) .710(.773) .577(.778) .684(.778) .846(.787) Other financial company

Control variable: world region

Asia 1.782*(.739) 1.826*(.742) 2.015*(.796) 1.819*(.748) 1.898*(.740) Europe 1.977**(.581) 1.702**(.591) 1.845**(.620) 1.712**(.597) 1.599**(.589) North & Central America 1.318*(.571) 1.252*(.573) 1.356*(.594) 1.250*(.577) 1.293*(.569) South America 2.068*(.933) 2.156*(.937) 2.201*(.946) 2.156*(.939) 2.592**(.986)

Omnibus test

Ratio Chi-Square 421.815 427.768 428.972 428.188 429.845 P-value .000 .000 .000 .000 .000 *** p < .000 ** p < .01 * p< .05

35

6. Discussion

The aim of this study is to get more insight into where criticism directed at MNCs does occur. The underlying literature is reviewed and hypotheses were formulated based on this literature to address the question whether civil society actors are exerting more pressure on firms which are more globally active. This relationship is empirically tested with a unique dataset containing criticism data on 235 MNCs. This section will discuss the findings of these tests. Subsequently, the implications and limitations will be examined followed by suggestions for future research.

6.1 Major findings

Degree of internationalization

Typically MNCs do not voluntary engage in CSR activities but do so as a response to stakeholder pressures (Campbell, 2007). Since MNCs are increasingly operating across borders and within different jurisdictions, civil society actors must exert criticism on international and transnational levels in order to hold MNCs to account. This study shows that MNCs do not receive more host country criticism when they are more internationally active but they do receive more transnational criticism. This could be explained by differences between civil societies throughout the world. Most host countries in which MNCs operate are developing countries, in these countries civil societies are not as vivid as in most developed countries. Civil society actors in these countries often lack important mobilizing structures such as know-how and material resources (Keck and Sikkink, 1999). Therefore, in countries with weak civil societies MNCs face less pressure for conformity than in countries with well-developed civil societies (Surocca et al., 2012). Moreover, in developing countries CSR is of less interest than in economically advanced countries (Martin, 2002; Soares et al., 2008). In countries which are economically underdeveloped issues such as the environment and gender equality are less glaring than for instance labour standards or human rights