The effect of institutional voids on the CSR reporting intensity : the difference between multinational enterprises from emerging and developed countries

The effect of institutional voids on the CSR

reporting intensity

The difference between multinational enterprises from emerging and developed

countries

Name: Wendy Venema Student number: 10222901

Date of submission of final version: January 27, 2017

MSc. In Business Administration – International Management Track University of Amsterdam

First supervisor: Dr. L. Divito

2 Statement of originality

This document is written by Wendy Venema 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 Table of Content

Abstract ... 4

1 Introduction ... 5

2 Literature ... 8

2.1 Corporate Social Responsibility ... 8

2.2 CSR disclosure ... 9

2.3 Developed and emerging country institutions ... 12

2.4 Institutions and legitimacy ... 14

2.5 Institutional view on CSR ... 15

2.6 Institutional voids and CSR reporting ... 17

2.7 Internationalization ... 20

3 Method ... 23

3.1 Research design ... 23

3.2 Data frame ... 24

3.3 Variables and measures ... 26

3.3.1 Dependent variable ... 26

3.3.2 Independent variables ... 27

3.3.3 Moderator ... 28

3.3.4 Control variables ... 28

4 Results ... 30

4.1 Analytical strategy, descriptive statistics, and correlation ... 30

4.2 Hypothesis testing ... 33

5 Discussion ... 39

5.1 Institutional voids and CSR reporting in emerging and developed countries ... 39

5.2 Managerial implications ... 42

5.3 Limitations and future research ... 43

6 Conclusion ... 46

7 References ... 48

4 Abstract

This study examines the relationship between home country institutional voids, internationalization and the CSR reporting intensity of MNEs from emerging and developed countries. Research into the CSR reporting of emerging market multinational enterprises (EM-MNEs) is still underdeveloped. As emerging countries catch up, CSR becomes increasingly important. However, the institutional voids in the home country of EM-MNEs might influence CSR reporting. This study proposes that institutional voids foster legitimacy practices like CSR reporting and that internationalization moderates this relationship. Furthermore, it is hypothesized that MNEs from emerging countries report more intensive on CSR than MNEs from developed countries. To evaluate these propositions, a quantitative database research study was performed for the 183 largest MNEs from emerging and developed countries ranked by foreign assets by the United Nations Conference on Trade and Development. A hierarchical regression analysis with moderator shows a negative relationship between institutional voids and the CSR reporting intensity for low internationalized MNEs. In addition, the results show that MNEs from developed countries report more intensive on CSR than MNEs from emerging countries. The fact that no evidence for the proposed relationship was found might be because CSR reporting intensity is based upon other factors than institutional quality, for instance stakeholder pressures. The results of this study shed light on the influence of internationalization on CSR reporting and provide interesting gaps for future research and several implications and contributions for theory and practice.

Keywords: Corporate Social Responsibility (CSR), institutional voids, emerging market multinational enterprises, emerging market, CSR intensity, CSR reporting,

5 1 Introduction

In recent years, multinational enterprises (MNEs) have become increasingly aware of the importance of corporate social responsibility (CSR) in their businesses. More frequently, the governments, media, and activists hold MNEs responsible for the social and environmental consequences of their activities (Porter & Kramer, 2006). Global environmental consequences, especially, are germane to MNEs because of the international concerns around this issue (Pinkse & Kolk, 2012). While MNEs from developed countries have always been the drivers behind CSR practices, an increasing number of emerging market multinational enterprises (EM-MNEs) are catching up (Li, Fetscherin, Alon, Lattemann, & Yeh, 2010; UNCTAD, 2011). Engaging in CSR practices is more challenging for EM-MNEs, because their home countries often fall short in providing the institutions necessary for developing CSR practices (Amaladoss & Manohar, 2013). Recently, research has put more emphasize on the CSR practices of EM-MNEs (Amaladoss & Manohar, 2013; Belal, Momim, 2009; Marano, Tashman, & Kostova, 2016) However, since this research is still in its infancy, there is need for more understanding on the link between the institutional environment and CSR reporting.

Recent results of a study from Marano et al. (2016) show a positive relationship between institutional voids in EM-MNEs’ home countries and their CSR reporting. According to Marano et al. (2016) institutional voids foster CSR practices, as CSR can help EM-MNEs escape from their home country liabilities (Marano et al., 2016). However, because their data only includes EM-MNEs, whether this relationship also applies when MNEs from developed countries are included is unknown. There are no comparative studies that examine the influence of the institutional voids on CSR practices of both developed countries and emerging countries. This study aims to fill this gap by examining the following question:

6 What is the influence of institutional voids on the CSR reporting intensity of MNEs from emerging and developed countries?

The question is answered using quantitative database research. This research method enables the inclusion of multiple MNEs from different countries, making the results of the study generalizable. The database consists of data about the largest firms from emerging and developed countries ranked by the United Nations Conference on Trade and Development. A regression analysis between CSR reporting of these firms and their home countries institutional voids is preformed to examine the relationship. Furthermore, internationalization is added as a moderator variable. The results of this study show that MNEs from developed countries report more intensive on CSR than EM-MNEs. Furthermore, institutional voids have a negative influence on CSR reporting intensity when MNEs have a low level of internationalization.

The primary contribution of this study is to extend the CSR literature beyond its traditional focus on developed countries and thereby provide clarification about the CSR reporting of emerging countries. The study includes multiple emerging countries, making the findings more generalizable than that of previous studies, which often focused on a specific emerging country or continent (Amaladoss & Manohar, 2013). Second, the study adds to the CSR literature by comparing the CSR reporting between MNEs from emerging and developed countries (Kolstad & Wiig, 2010). In this way, the study extends the study of Marano et al. (2016), which only focused on EM-MNEs. The results of this study may be able to confirm whether the positive relationship between institutional voids and CSR reporting, found in the study of Marano et al. (2016), also applies when MNEs from developed countries are included. The information in this study can be used in practice for investors and stakeholders to increase awareness of differences in CSR reporting intensity between EM-MNEs and MNEs from developed countries. Lastly, the study can help managers create awareness about

7 national differences in CSR reporting, which is specifically relevant for MNEs operating in distinct national environments. According to Fukukawa and Teramoto (2008), managers often see CSR through a global lens and are not aware of local pressure on CSR. Viewing CSR in a broader context of culture and business practices makes it possible to understand differences in CSR practices (Fukukawa & Teramoto, 2008).

This paper proceeds in four parts. In the Chapter 3, the literature review gives an overview of the relevant existing literature about MNEs’ CSR practices from developed and emerging countries in relation to institutional voids. Then the study’s research method is explained, together with the data framework and measurement. Chapter 5 discusses the results of the study. In the final chapter, the results and implications of the study are discussed, and directions for future research are given.

8 2 Literature

2.1 Corporate Social Responsibility

Corporate Social Responsibility (CSR) is a concept whereby firms voluntarily integrate social and environmental concerns into their business operations and their interactions with their stakeholders (Aguinis & Glavis, 2012). This definition emphasizes the triple bottom line of CSR, which consists of economic, social, and environmental concerns. MNEs that integrate the triple bottom line into their business activities positively contribute to a social and environmental cause and thus do not only aim for simple profit maximization or compliance to the law (Crane, Matten, & Spence, 2013).

There are different views on the incorporation of CSR into MNE business activities. In the economic view, CSR should not be a firm’s objective; the only social responsibility of a firm is to increase its profits (Friedman, 1970). Including social and environmental concerns in an MNE’s business is equal to theft, since MNEs only have a responsibility to their employees and shareholders (Friedman, 1970). However, authors who emphasize the stakeholders’ theory argue that CSR is a priority for firms (Epstein & Roy, 2001; Porter & Kramer, 2006). The stakeholder theory explains the nature of the relationship between firms and people with a “stake” in the operations and outcomes of business activities. In this theory, a manager’s principal function is to handle the expectations, needs, and demands of stakeholders (Arenas, Lozano, & Albareda, 2009). Making the stakeholder’s needs core to the firm’s strategy can create competitive advantages since firms need healthy societies and societies need healthy corporations (Porter & Kramer, 2006). Hence, CSR can create shared value for both firms and stakeholders. Finally, a third stream of research argues that CSR disclosure is simply the legalization of the power of large corporations (Banerjee, Mukherjee,

9 & Maulik, 2008). Banerjee et al. (2008) state that CSR is solely used by firms to enhance their profit and not to serve society.

Despite the contradictory views, the expansion of voluntary CSR disclosures by MNEs is clear evidence that managers perceive a direct value from engaging in CSR practices. Results on the relationship between a firm’s CSR activity and its financial performance are still ambiguous, although most studies indicate a positive relationship (Ambec & Lanoie, 2008; Barnett & Salomon, 2012; Hillman & Keim, 2001; Orlitzky, Schmidt, & Rynes, 2003; Porter & van der Linde, 1996). From a stakeholder perspective, CSR is beneficial since it improves the relationship with primary stakeholders, such as employees, customers, local communities, and NGOs (Hillman & Keim, 2001). Furthermore, empirical evidence shows that implementing CSR can decrease costs and increase revenue. (Ambec & Lanoie, 2008; Porter & van der Linde, 1995). Firms can, for instance, decrease their costs by preventing pollution and reducing waste and increase their revenue by differentiating into high-value markets (Ambec & Lanoaie, 2008;). This evidence is in line with the resource-based view, which argues that CSR is an important resource that can create competitive advantage through preemption of markets and differentiation (Hart, 1995).

2.2 CSR disclosure

Considering the potential positive impact of CSR activities on financial performance, investors and shareholders demand that firms provide more information about CSR activities and long-term strategies. Moreover, the popularity of the stakeholder approach made firms increasingly aware of the importance of transparency towards stakeholders (Boesso & Kumar, 2007). An increasing amount of CSR disclosure is achieved by formally reporting CSR efforts (KPMG, 2013). CSR reporting refers to the public disclosure of a firm’s corporate governance and its CSR activities (Maignan & Ralston, 2002). Larger firms in developed countries started CSR reporting in the 1970s by disclosing their CSR activities in annual reports (Fifka, 2008).

10 Nowadays, firms from all over the world disclose their CSR activities in both annual reports and separate CSR reports. These CSR reports present additional information to the host country and global stakeholders for a more rational evaluation of the firm and are important as they present the business’ willingness to align practices with global stakeholders’ norms and expectations (Bansal & Kistruc, 2006; Young & Marais, 2012).

CSR reporting can be mandatory, solicited, or voluntary (Van der Laan, 2009). Voluntary CSR reporting is most used and refers to reports that favorably present the firm. In 1999, the Global Reporting Initiatives (GRI) introduced the initial voluntary disclosure guidelines, which represent the first global framework for sustainability reporting, including the triple bottom line of economic, social, and environmental issues. The introduction of the GRI made the triple bottom line the dominant way of voluntarily reporting CSR to stakeholders (Milne & Gray, 2013; Raar, 2002). Comparative research into CSR reporting often focuses on how intensive firms report on the triple bottom line of CSR (Fortanier, Kolk, & Pinkse, 2011). The sum of the extent and depth to which firms report about aspects of the triple bottom line is defined as CSR reporting intensity.

The literature review highlights that voluntary CSR reporting is usually used by firms to enhance their legitimacy (Golob & Barlett, 2007). 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” (Suchmann, 1995, p. 574). Historically, a firm’s legitimacy was viewed from a macro perspective, where firms are impacted by socio-political approval and cognitive approval from society (Ahlstrom, Bruton & Yeh, 2008), which argues that legitimacy stems from the approval of a leader, the government, or the public. According to the latter, legitimacy is based on the spread of knowledge and the general acceptance of a firm and its culture at large.

11 Previous research indicates that firms can strengthen their legitimacy and reputation through extensive CSR disclosures because these show the firms induce desirable outcomes for their stakeholders (Bachmann & Ingenhoff, 2016). Attaining legitimacy can improve a firm’s ability to compete for resources, gain stakeholder approval, and provide goodwill in times of crisis (Godfrey & Hatch, 2007). Unerman (2008) calls this motive for using CSR reputation risk management (RRM). RRM explains that firms use CSR reporting as a powerful medium to influence the perceptions of firms. Social and environmental reputation has economic value; reporting to stakeholders how well the firm has built and maintained these values may help in building reputation (Underman, 2008).

Although research often regards CSR to strengthen legitimacy, stakeholders can also become skeptical and distrust CSR disclosures (Chen, Patten, & Roberts, 2008; Deegan, 2002) as CSR disclosures may not always reflect a firm’s actual CSR activities (Bansal & Kistruck, 2006; Dhaliwal, Li, Tsang, & Yang, 2011). Firms can also use CSR practices to create a symbolic impression. CSR practices are often symbolic rather than genuine (Bansal & Kistruck, 2006). Firms may seem to engage in CSR, but mostly, these initiatives are intended to appease stakeholders’ demands or to meet the minimum requirements of standards (Marano et al., 2016). Bachmann and Ingenhoff (2016) examined whether CSR strengthens or weakens legitimacy and found that firms are not penalized if they use extensive CSR disclosure. The legitimacy advantage of this extensive use outweighs the disadvantages of stakeholder skepticism (Bachmann & Ingenhoff, 2016). Even though stakeholders are highly skeptical in terms of the perceived persuasion intent, the extent of CSR communication still has a positive effect on corporate legitimacy.

The literature review so far highlighted that CSR reporting is mainly voluntary and can improve a firm’s legitimacy and reputation. This study considers emerging and developed countries with different political and economic climates and explores how CSR reporting is

12 managed by MNEs in these countries. Possibly, not only regulatory demands differ, but also how voluntary reporting occurs and the drivers of reporting are of interest. To consider these questions, institutions and their influences on legitimacy and CSR reporting are now explored.

2.3 Developed and emerging country institutions

Literature often makes a distinction between emerging countries and developed countries. Developed countries have highly developed economies and advanced technological infrastructures. Conversely, emerging countries do not meet the high standards of developed countries, although they have some shared characteristics (Morgan Stanley Capital International [MSCI], 2016). Several measurements have been developed to further specify the difference between emerging and developed countries. One of the most commonly used measurements is developed by the World Bank (2013), which distinguishes emerging countries from developed countries based on the country’s gross national income per capita (GNI) and the available monetary resources. Developed countries are countries with high incomes, while low/middle income countries are defined as emerging countries. Morgan Stanley also developed a classification, the MSCI, which is based on economic development, market accessibility, and size and liquidity (MSCI, 2016). A country is classified as developed if the following criteria are met: (1) The country’s GNI per capita is 25% above the World Bank high income threshold for three consecutive years; (2) the country has a minimum number of companies with satisfied minimum size and liquidity requirements; and (3) if there are convenient capital inflows/outflows, stable institutions, and a highly efficient operating framework. Emerging countries score lower on these three criteria (Baston & Caiado, 2011; MSCI, 2006). Based on the MSCI classification, countries in Eastern Europe, South America, Africa, the Middle East, and Asia are defined as emerging countries and countries in Western Europe, North America, and Southeast Asia as developed countries.

13 The classifications indicate that the distinction between emerging and developed countries is often based on quantitative economic measurements. However, economic differences are only a small part of the overall differences between the countries. A good way to compare these countries is by looking at their institutions. Institutions can be defined as “the humanly devised constraints that structure political, economic and social interaction” (North, 1991, p. 97) and are formed by formal and informal constraints (North, 1991). Formal constraints include political rules, judicial decisions, and economic contracts (Peng, 2002), which provide incentives for competitive markets and decentralized decision-making. Informal constraints determine the everyday process of making decisions and result from cultures and norms of behavior recognized as standards of conduct or integrity (North, 1991). Because history determines the development of institutions, the institutional environment differs per country (Sokoloff & Engerman, 2000).

Developed countries are characterized by strong institutions, which consist of a good regulatory system and the existence of laws and property rights (North, 1991). These institutions support the voluntary exchange underpinning an efficient market mechanism, providing incentives for organizational learning, induce innovation, and encourage creativity (North, 1991). Strong institutions are viewed as legitimate or at least enjoy a strong presumption of legitimacy (Buchanan & Keohane, 2006). This contrasts with weak institutions where legitimacy is questioned (Henisz & Zelner, 2005). Emerging market countries often have institutional voids, or the absence or underdevelopment of institutions. Although, there are differences among emerging countries (Khanna & Palepu, 2010; Wright, Filatotchev, Hoskisson, & Peng, 2005), they share several characteristics that differentiate them from developed countries.

One key characteristic is that they are low-income, rapid-growth countries using economic liberalization as their primary engine of growth (Hoskisson, Eden, Lau, & Wright,

14 2000). The GDP growth of emerging market countries is, therefore, higher than the GDP growth of developed countries. However, the capital markets of emerging markets are less developed, making trading less liquid and financial exchanges more volatile (Marquid & Raynard, 2015). In addition to economic differences, legal and political factors differentiate the institutional environment. Emerging countries are characterized by less developed governments and regulatory infrastructures, resulting in underdeveloped market regulations, corporate governance, transparency, and intellectual property protection (Marquis, Zhang, & Zhou, 2011). Furthermore, the unstable political environment makes emerging countries susceptible to corruption and coups.

2.4 Institutions and legitimacy

Differences in institutions may result in differences in defining, addressing, and implementing firms’ responsibilities to society (Matten & Moon, 2008). Because of home country institutional voids, foreign actors can develop negative stereotypes about firms from emerging markets (Marano et al., 2016), which makes it harder for these firms to preserve legitimacy than for firms from developed countries. MNEs from emerging markets that invest abroad need to overcome the negative perceptions of the legitimacy in their home countries in order to be successful in host countries. MNEs engage in outward foreign direct investment to “exercise effective control and undertake value-adding activities in one or more foreign countries” (Luo & Tung, 2007, p. 482). EM-MNEs are MNEs that have emerging countries as their home countries.

The Edelman Trust Barometer illustrates that the location of a company’s headquarters matters when it comes to people’s perception of trust, showing that emerging countries score lower in levels of trust than developed countries (Kostava & Zaheer, 1999). Low levels of trust can undermine the legitimacy of firms and their capacity to fulfill task (Gilson, 2003). This corresponds to the results of the study by Matten and Moon (2008), which showed that

15 the characteristics of the National Business System influence legitimacy. Foreign stakeholders may engage in what they call “adverse institutional attribution”, which means that stakeholders discriminate against firms from certain countries because of their (perceived) conflict with the regulative, normative, or cultural-cognitive elements of the host country’s institutional environment (Pedersen, 2010). Host country consumers, for example, may perceive the products and services of EM-MNEs’ subsidiaries of lower quality because of home country institutional voids (Klein, 2002). Furthermore, the host countries’ governments may assume that the lack of transparency and corruption in the home country spills over to the EM-MNEs in these countries (Cuervo-Cazura, & Genc, 2008). The discrimination against foreign firms by the host country can be defined as a potential source of liability of origin, referring to the discrimination by host country consumers, governments, and stakeholders because of a MNE’s home country. Host country stakeholders may judge EM-MNEs based on negative perceptions about the weak institutional conditions in their home countries (Pedersen, 2010).

For EM-MNEs, it is important to overcome the liabilities of their origin and convince stakeholders of their trustworthiness and legitimacy (Marano et al., 2016). Ahlstrom, Bruton, and Yeh (2008) suggest that building legitimacy is vital for firms based in emerging countries, giving the weak protection for private property. As CSR reporting enhances firms’ legitimacy, EM-MNEs are expected to use CSR reporting as a way to escape from their home countries’ institutional constraints.

2.5 Institutional view on CSR

Until now, few studies have addressed the institutional determinants of CSR (Jackson & Apostolakou, 2010). Most studies emphasized CSR and its link with economic and social performance. The studies that did investigate CSR in relation to institutions were often focused on developed countries and not comparative in nature (Campbell, 2007; Jackson &

16 Apostolakou, 2010; Marano et al., 2016; Matten & Moon, 2008; Zhao, Tan, & Park, 2014). In general, two institutional views on CSR reporting can be distinguished.

In the first view, CSR reporting is a function of a country’s characteristics. Research often assumes that CSR is a function of economic, social, and political development. This follows from the observation that firms in developed countries started with CSR reporting because of more resources and greater awareness of CSR in these countries (Chan, Watson, & Woodliff, 2014; Kemp, 2001). For example, the United Kingdom was the first country to have a CSR minister in the department of Industry and Commerce. Another example is France, which has a mandatory law that requires firms with more than 300 employees to draft CSR reports (Wanderley, Lucian, Farache, & de Sousa Silho, 2008).

According to Jamali and Mirshak (2007) numerous factors prevent the adoption of CSR in lesser developed countries, such as the absence of pressures from the government and press and the unorganized civil society. This means that the institutions and standards, which are the foundation of CSR in developed countries, are often absent in lesser developed countries (Kemp, 2001). This corresponds to Campbell (2007), who argued that firms are likely to act in socially responsible ways if strong and well-forced regulations are in place, along with a healthy economic environment. In addition, conditions that provide options for monitoring firms’ behavior, such as the presence of independent organizations and normative standards, are considered favorable for ensuring socially responsible behavior (Campbell, 2007). Strong institutions thus put pressure on firms to adopt CSR practices, suggesting a positive relationship between home country institutional quality and CSR practices. However, applying Campbell’s (2007) assumptions to the international business context shows that strong institutions in the home country may not be necessary. MNEs that operate internationally are also exposed to external pressures from multiple host countries. These

17 external pressures can cause MNEs from weaker institutions to increase their CSR practices as well.

A second institutional view on CSR reporting, therefore, states that CSR reporting is not a function of the economic and social development of a country per se. The globalization causes environments to have less influence, and firms adapt themselves to the expectations of the global environment (Tempel & Walgenbach, 2007). An institution with weak regulations does not necessarily indicate the absence of institutional pressures. Indeed, institutional voids can be filled with other types of institutional pressures that also stimulate the use of CSR practices (Valente & Crane, 2010). According to the neo-institutional theory, there are three types of isomorphic pressures that can change firms’ behavior in order to gain legitimacy: (1) Coercive isomorphism, which is derived from pressures exerted by external powerful organizations; (2) mimetic isomorphism, where increased uncertainty causes firms to imitate other firms that are viewed as successful; and (3) normative isomorphism, resulting from professionalization (DiMaggio & Powell, 1983). Applying this theory to the international business context, EM-MNEs need to comply with institutional requirements to gain legitimacy in host countries (Park, Chidlow, & Choi, 2014). Pressures from external organizations, such as NGOs and activity groups, may cause EM-MNEs to increase their CSR reporting. In addition, EM-MNEs’ uncertainty about home country institutional voids may result in imitating CSR reporting from MNEs from developed countries to reduce uncertainty. Since home country institutional voids make it more difficult for EM-MNEs to prove their legitimacy, they even may report more intensively on CSR.

2.6 Institutional voids and CSR reporting

There is growing evidence that MNEs from emerging countries are more motivated to use CSR. Research indicates that, especially, firms that are already seen critically by the public audience can profit from CSR disclosures (Palazzo & Richter, 2005). This would count MNEs

18 from emerging countries that suffer from legitimacy challenges caused by their home countries’ institutional constraints. These MNEs use CSR to escape from institutional constraints in their home countries and these constraints’ negative impact on firm performance (Ahlstrom, Levitas, Hitt, Dacin, & Zhu 2014; Cuervo-Cazurra & Genc, 2008). According to Marano et al. (2016), CSR reporting can mitigate stakeholders’ negative perceptions about EM-MNEs. Because CSR reporting is globally established and legitimate, it helps EM-MNEs to dissociate themselves with negative home country perceptions and associate themselves with the global meta-institutional field (Kostova, Roth, & Dacin, 2008). Furthermore, CSR reporting provides stakeholders in host countries with useful information that can help them facilitate a less biased assessment of EM-MNEs. This will prevent stereotyping EM-MNEs based on their home country institutional voids (Choi & Wong, 2007; Kolk & Perego, 2010).

Jackson and Apostolakou (2010) provided support for the positive relationship between home country institutional voids and CSR reporting and examined the adoption of CSR standards of Western countries. Their results show that institutions that support stakeholder involvement do not strengthen firms’ CSR practices (Jackson & Apostolakou, 2010). Instead, firms from institutions with strong regulations are more likely to adopt minimum standards of CSR. This indicates that firms do not simply mirror institutional forms of behavior. Instead, Jackson and Apostolakou (2010) argued that firms develop strategic responses to institutions that involve pro-active attempts to fill institutional voids. If regulations are minimal, firms have more room to take CSR initiatives and greater pressure from stakeholders. This may cause firms from weaker institutional environments to adopt more extensive CSR practices because the low regulations give them a chance to differentiate themselves from their peers. In addition, Marano et al. (2016) examined the pervasiveness of the home country institutional voids in relation to the CSR reporting intensity of MNEs from

19 emerging countries. Their results indicate that EM-MNEs from home countries with high institutional voids report more intensively in CSR, which is explained by means of the higher liabilities of origin in these environments, making the firms more inclined to overcome these liabilities (Marano et al., 2016). These findings are interesting because they show that EM-MNEs adopt CSR practices not necessarily supported or mandated by their home country institutional environments. Thus, the presence of home country institutional voids is expected to result in more intensive CSR reporting. Therefore, the following hypotheses are proposed:

Hypothesis 1: Institutional voids in emerging countries are higher than the institutional voids in developed countries.

Hypothesis 2: There is a positive relationship between the home country institutional voids and the CSR reporting intensity of MNEs from emerging and developed countries.

However, a comparative study examining the CSR reporting of both developed and emerging countries is so far lacking. Most studies on CSR compared the CSR reporting among developed countries or among emerging countries (Baskin, 2006; Belal & Momin, 2009; Chapple & Moon, 2005; Golob & Bartlett, 2007; Jenkins, 2005; Wanderly, Lucian, Farache, & de Sousa Filho, 2008). Since developed countries have a strong institutional environment (Makino, Isobe, & Chan, 2004), the tendency for these MNEs to prove their legitimacy to stakeholders may be lower. As argued by Jackson and Apostolakou (2010), firms from strong institutions are more likely to adopt minimum standards of CSR. Conversely, EM-MNEs are more motivated to use CSR since home country institutional voids foster the use of legitimacy practices. This could indicate that MNEs from developed countries have a lower level of CSR reporting than MNEs from emerging countries. Therefore, these arguments are hypothesized:

Hypothesis 3: The CSR reporting of MNEs from developed countries is less intensive than the CSR reporting of EM-MNEs.

20 2.7 Internationalization

Internationalization likely influences the relationship between home country institutional voids and CSR reporting intensity. Internationalization can be described as “the relative importance of MNEs business activities conducted abroad, including foreign direct investment and international trade” (Marano et al., 2016, p. 6). Firm internationalization can be viewed as a strategy to increase a firm’s competitive advantage (Nachum & Zaheer, 2005). On the one hand, internationalization enhances the economies of scale and scope, diversification benefits, and access to new knowledge and resources. On the other hand, high levels of internationalization increase the exposure to liability of foreignness and hostile international environments.

Internationalization drives MNEs to use more CSR practices for several reasons. First, MNEs face an increased level of risk in foreign countries, as they are unfamiliar with societal and regulatory requirements. By using CSR reporting MNEs can reduce this risk and to establish reputations as good citizens in host countries. Feldman, Soyoka, and Ameer (1997) noted that the adoption of environmental practices is useful to reduce perceived risk. The use of CSR practices can enhance the commitment to the host country’s community and, consequently, reduce communication problems and the liability of foreignness (Zahra, Ireland, & Hitt, 2000). Second, higher levels of internationalization increase MNEs’ exposure to foreign customers, communities, investors, employees, regulars, and nongovernmental organizations (Attig, Boubakr, Ghoul, & Guedhami, 2014). Hence, MNEs need to consider the interests and expectations of a wider group of stakeholders. To deal with their expectations, MNEs may increase their CSR reporting (Attig et al., 2014). Yu and Choi (2014) examined the relationship between stakeholder pressures and CSR practices of Chinese firms. They expected the pressures of regulation and competition to require Chinese firms to comply with the international environment and CSR standards. The results of their

21 study affirm that CSR practices are positively influenced by stakeholder pressures. Internationalization increases MNEs’ exposure to global standards and global legitimating actors that monitor the social and environmental impact of MNEs (Chapple & Moon, 2005; Marano & Tashman, 2012). For MNEs, it is important to comply with global standards to be competitive and win orders in the global market (Yu & Choi, 2014). For EM-MNEs, especially, compliance with global standards is important since developed countries erect environmentally-related export barriers.

There are several studies that confirm the positive effect of internationalization on CSR practices (Attig, Boubakri, El Ghoul, & Guedhami, 2016; Chapple & Moon, 2005; Laudal, 2011). Chapple and Moon (2005) argue that internationalization is a driver for CSR developments and is one of the causes for Asian MNEs to develop new CSR practices. They state that the global presence of EM-MNEs can create incentives for CSR. Their presence results in an increase of global watchdogs, such as multilateral or international nongovernmental organizations (Chapple & Moon, 2005). More global watchdogs, in the end, cause an increase in global standards of business practices, including CSR. In addition, Attig et al. (2016) found that internationalization is positively related to U.S. firms’ CSR ratings. They argued that the positive effect of CSR goes together with internationalization. It would thus be likely that internationalization positively moderates the relationship between institutional voids in emerging countries and CSR reporting. Therefore, the following is hypothesized:

Hypothesis 4: The degree of internationalization positively moderates the relationship between home country institutional voids and the CSR reporting intensity.

22

23 3 Method

This section provides an overview of the research method, including the research strategy. In the first part, the research design is explained. This is followed by a description of the data frame and an elaboration on the methods of data analysis. Last, all the study’s variables are elaborated and their uses explained.

3.1 Research design

Quantitative database research was adapted to examine this research question. Quantitative research can be defined as research that explains phenomena by collecting numerical data that are analyzed with mathematical methods (Aliaga & Gunderson, 2000). According to the realist view, quantitative research is an objective research method that allows one to uncover an existing reality. Since quantitative research is objective, it minimizes the involvement of researchers in the study (Muijs, 2010). In addition, quantitative research allows for a large data sample, resulting in high generalizability. However, the positivism view argues that it is not possible to conduct research objectively and that the involvement of human subjectivity is important (Muijs, 2010). Qualitative research allows one to gain a deeper understanding of phenomena and can explain relationships. However, since this study tries to answer a research question about a relationship and the influence of variables, quantitative research is best suited (‘t Hart, Boeije, & Hox, 2009).

This study is deductive, meaning that theories are tested but no new theories are generated. The objective of this study is to explain the relationship between institutional voids and CSR reporting intensity and to determine whether this effect is different for MNEs from emerging and developed countries. Furthermore, internationalization was added as a moderator variable and industry type and firm size as control variables. Data about the CSR reporting of MNEs was obtained through the coding of CSR reports. This coding was

24 deductive, i.e., the measurements were based on previous research and determined before the coding process (Fortanier et al., 2011). The rest of the database was completed with the use of secondary data from the United Nations Conference on Trade and Development (UNCTAD) and the World Bank Governance (WGI). A limitation of using secondary data is that the data was originally collected to answer different objectives (Denscombe, 2008). Hence, some adjustments were made to make the data appropriate for this study.

3.2 Data frame

The sample consists of two lists of the largest firms from UNCTAD. These are firms from the world’s top 100 non-financial MNEs, ranked by foreign assets, and firms from the top 100 non-financial MNEs, ranked by foreign assets from developing and transition economies. The UNCTAD list is published annually in the World Investment Report (WIR), which is one of UNCTAD’s leading policy publications on global foreign direct investment trends. The list consists of data on each firm’s industry, degree of internationalization, and country of origin. Data from the years 2014 and 2015 was analyzed because this was the most recent publication of the UNCTAD. This study used the UNCTAD list, because it selects MNEs from emerging countries and developed countries based on the same criteria, making it possible to compare the MNEs of the two country types.

A content analysis was performed on the CSR reports of each firm in the sample. CSR reports from 2014 were analyzed because not all firms in the sample had published their 2015 CSR reports. All 200 firms in the sample had their most recent report scrutinized; this could be either a separate CSR report or a section in the annual report that contained non-financial information. The reports were collected from the Corporate Register, the Global Reporting Initiative’s Sustainability Disclosure Database, and the firms’ websites. In total, 17 observations were excluded: Nine due to missing English reports in 2014 and eight because they were present twice in the sample. The final sample consisted of 183 MNEs, of which 84

25

MNEs from emerging countries and 99 MNEs from developed countries. In total, 36 countries were included in the whole data sample: Eighteen developed countries and seventeen emerging countries. A full list of MNEs and their home countries can be found in Table 1. The distinction between emerging and developed countries is based on the MSCI ACWI which is explained in the next section.

Table 1. Frequencies and percentages of MNEs in a certain country

Developed Emerging

Country Frequency Percentage Country Frequency Percentage

USA 21 11.5% China 16 8.7%

UK 16 8.7% China, Hong Kong 14 7.7% Germany 13 7.1% India 8 4.4% Japan 11 6.0% South Africa 7 3.8% Singapore 10 5.5% Taiwan 7 3.8% France 8 4.4% Korea 7 3.8% Switzerland 5 2.7% Malaysia 5 2.7% Spain 3 1.6% Brazil 5 2.7% Italy 2 1.1% Mexico 4 2.2% Sweden 2 1.1% UAE 3 1.6% Australia 1 0.5% Russia 2 1.1% Belgium 1 0.5% Thailand 1 0.5% Denmark 1 0.5% Saudi Arabia 1 0.5%

Ireland 1 0.5% Qatar 1 0.5%

Luxembourg 1 0.5% Philippines 1 0.5% The Netherlands 1 0.5% Kuwait 1 0.5% Norway 1 0.5% Argentina 1 0.5% Israel 1 0.5%

26 3.3 Variables and measures

3.3.1 Dependent variable

The dependent variable, CSR reporting intensity, is defined as the extent to which businesses report on a set of CSR issues and the depth to which they report about these issues (Fortanier et al., 2011). A common tool to measure a firm’s CSR reporting intensity is by means of the triple bottom line: The environmental, social, and economic dimensions of a firms’ activities (López, Garcia, & Rodriguez, 2007). In this research, the coding approach of Fortanier et al. (2011) was used to measure the CSR intensity. Fortanier et al. (2011) used the triple bottom line approach but divided the social dimension of CSR reporting intensity into three different variables. This distinction was made because the social dimension of CSR reporting reflects a broad range of issues and is, therefore, difficult to measure with one variable (Chapple & Moon, 2005). The coding approach thus consists of five components: Three social (employment, rights, and community), an economic, and a environmental (Appendix, Table 2). Each of these five components consists of five items, totaling 25, and are measured using binary items. A (0) was reported when the CSR report did not include a typical CSR issues and a (1) was reported when it did. If the CSR report mentioned an issue without reference to the specific efforts, these issues were reported as a (0). The variable CSR intensity was created with a sum of all the items because it was measured this way in previous research (Marano et al., 2016). Thus, the variable varied between 0 and 25. The author coded all the reports. To test the reliability of the coding process, the intercoder reliability was tested by means of Krippendorff’s alpha. Two coders measured the CSR intensity of the same 15 CSR reports before the definitive coding. The Krippendorff’s alpha was 0.94, which indicates a high intercoder reliability. Additionally, the bootstrapping procedure indicated that there was only 0.23% chance that Krippendorff’s alpha would be below 0.80 if the whole population were tested. This indicates that the degree of intercoder reliability is high.

27

3.3.2 Independent variables

The sample of firms consists of data from the WIR, which lists 100 MNEs from developing and transition economies and the top 100 countries worldwide. The latter list consists mostly of developed economies. To make sure that the distinction between emerging and developed countries in the lists is right, the MSCI ACWI index was used. This framework distinguishes between emerging and developed countries based on three criteria: Economic developed, size and liquidity, and market accessibility. The MSCI ACWI framework is a good framework to use as it has an exhaustive equity market coverage for more than 75 countries in developing, emerging, and frontier markets (MSCI, 2016). Based on the MSCI ACWI framework, emerging countries are indicated with a (0) and developed countries with a (1).

The independent variable, institutional voids, is defined as the lack of institutional facilities, norms, and regulations needed for a well-functioning economy (North, 1991). Lack of institutional norms and regulations and lack of financial credit availability are important indicators of institutional voids (Chakrabarty, 2009). Institutional voids were measured using an index of World Bank Governance Indicators (WGI). The WGI consists of six measurements of the institutional quality, which are available on 215 economies over the period 1996–2014 (WGI, 2015). The six measurements, voice and accountability, political stability and absence of violence, government effectiveness, regulatory quality, rule of law, and absence of corruption (Appendix, Table 3), varied from -2 to +2. The WGI is a reliable measurement, as it is based on hundreds of variables that reflect the views of citizens, firm survey respondents, and experts worldwide (Marano et al., 2016). For each MNE, the intuitional voids were determined by referring to the MNE’s home country, which is available on the WIR. The institutional voids of the year 2014 were used, as this is the most recent year. Because the WGI measures institutional quality instead of institutional voids, the index needed to be reverse coded to make it appropriate for this study.

28

3.3.3 Moderator

Internationalization is the moderator variable in this research. Internationalization refers to the relative importance of MNEs’ business activities conducted abroad (Marano et al., 2016). Research indicates that there is no single way to measure the degree to which industries, firms, or countries are internationalized. The extent of internationalization depends upon which variables are considered most important (Ietto-Gillies, 1998). While some authors have developed single measurement variables, others developed multidimensional measurements of internationalization (Sullivan, 1994). This study emphasizes the internationalization of MNEs and uses three variables to measure internationalization: The proportion of foreign sales of total sales, the proportion of foreign assets of total assets, and the proportion of foreign employment of total employment. These three measures are modeled after the internationalization index of Sullivan (1994) and are commonly used to measure the degree of internationalization (Marano et al., 2016). The three variables are all found in the WIR and ranged from 0 to 1. The variable internationalization is the weighted average of the three variables.

3.3.4 Control variables

This study includes two control variables, industry type and company size, to ensure that the relationship between the countries, institutional voids, CSR intensity reporting, and internationalization was not confounded by other variables. Industry type was included because previous research indicates it might influence a firm’s sustainability and customers’ behavior (Artiach, Lee, Nelson, & Walker, 2010) because large organizations are accountable to various constituencies and need to undertake more CSR actions to ensure their legitimacy (Goodstein, 1994). In addition, most emerging markets are subject to largely different business groups and industries (Khanna & Yafeh, 2007), making it important to control for industry type. The original sample from the WIR was classified by means of 29 different

29 industry types. To limit the number of industry types, the standard industrial classification system (SIC) was used. SIC is widely used in government and private data and defines the broadest industry categories in the system. The 29 different industry types were classified using ten SIC codes: Agriculture, mining, construction, manufacturing, transportation, wholesale trade, retail trade, finance, services, and public administration. There were eight types of the ten industries represented in the sample, according to the SIC system.

In addition, Firms size was included because it is considered a fundamental firm characteristic commonly used as control variable. Because larger firms are often exposed to more public attention regarding social and environmental impact, firm size can have an influence on CSR reporting (Christmann & Taylor, 2001). In this study, firm size was measured using the total number of firm employees, which is often used as an important indicator of the size of a firm (Sharma, 2000).

30 4 Results

This section describes the results of the different analyses. First, the analytical strategy and the descriptive statistics are explained. This is followed by a correlation matrix of the different variables with signification correlations highlighted. Finally, the results of the t-tests and the regression analysis are provided.

4.1 Analytical strategy, descriptive statistics, and correlation

First, all variables were checked for missing values. Because the database was composed of primary data and secondary data, there were no missing values. A frequency test showed that the variable CSR reporting was wrongly reported a few times (11 instead of 1). These mistakes were corrected. Furthermore, all variables were checked for normality. All three variables were normal distributed and had a skewness and kurtosis within -1 and +1. Therefore, no transformation techniques were used to normalize the distribution. Second, all variables were checked for outliers, but no outliers were indicated.

Although the research used existing scales, a reliability analysis was performed to make sure that the formed scales were reliable. The analysis was executed on the variables CSR intensity, institutional voids, and internationalization. CSR intensity has a high reliability, with Cronbach’s alpha of 0.84. The corrected item-total correlations indicate that all the items have a good correlation with the total score of the scale. Also, none of the items would substantially affect reliability if they were deleted. The new scale CSR intensity was formed using the sum of all the items (M = 17.66, SD = 4.77). In addition, institutional voids also have a high reliability: α = 0.96. The corrected item-total correlations indicate that all the items have a good correlation with the total score of the scale. Also, none of the items would substantially affect reliability if they were deleted. The new scale institutional voids was formed using the weighted average of the six items (M = -0.93, SD = 0.74). The reliability

31 analysis of internationalization also resulted in a high reliability with a Cronbach’s Alpha of 0.82. The corrected item-total correlations indicate that all the items have a good correlation with the total score of the scale. Also, none of the items would substantially affect reliability if they were deleted. The new scale internationalization was formed using the weighted average of the three items that measured internationalization (M = 0.61, SD = 0.22).

In addition, a correlation analysis was performed to test whether there were pre-existing correlations between the variables (Table 4). To make industry type appropriate for the correlation analysis, seven dummy variables were created. The manufacturing industry was chosen as the baseline category because it was most represented in the sample, and thus was a good industry to compare with. The correlation analysis using Pearson Correlations shows that there is a significant negative correlation between the institutional voids and internationalization: r = 0.48. This might indicate that firms in home countries with high institutional voids are less internationalized. Furthermore, there is a significant negative correlation between size and internationalization: r = -0.21. This might mean that larger firms are less internationalized. Country type (emerging versus developed) correlates positively with internationalization r = 0.25 and CSR intensity r = 0.21. This illustrates that the more developed a country, the more MNEs are internationalized and report on CSR. Conversely, country type has a negative correlation with institutional voids r = -0.60, indicating that more developed countries have less institutional voids. Other significant correlations were found between industries and internationalization, institutional voids, and CSR intensity. This means that the internationalization, institutional voids, and CSR intensity differ between industries, making that industry type an important variable to control for.

32 Table 4. Correlation matrix.

Note. N=183.

** Correlation is significant at the 0.0 1 (2-tailed) * Correlation is significant at the 0.05 (2-tailed)

Variables M SD 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 1. Size 136395 235416 - 2. Internationalization 0.61 0.22 -0.21** (0.82) 3. Institutional voids -0.93 0.74 0.42 -0.48** (0.96) 4. CSR intensity 17.66 4.76 0.10 -0.04 -0.03 (0.82) 5. Mining 0.15 0.36 0.04 -0.13 0.18* 0.02 - 6. Construction 0.04 0.29 -0.06 -0.13 0.03 -0.12 -0.09 - 7. Agriculture 0.01 0.10 0.01 -0.06 0.03 -0.08 -0.05 -0.02 - 8. Transportation 0.12 0.33 -0.10 -0.01 -0.12 -0.15* -0.16* -0.08 -0.04 - 9. Wholesale trade 0.15 0.36 -0.08 -0.06 0.08 -0.12 -0.18* -0.09 -0.05 -0.16* - 10. Retail trade 0.13 0.34 0.22** 0.02 -0.12 0.19** -0.17* -0.08 -0.04 -0.15* -0.17* - 11. Services 0.08 0.27 0.02 0.08 0.02 0.09 -0.12 -0.06 -0.03 -0.11 -0.12 -0.11 - 12. Country type 0.49 0.50 0.73 0.25** -0.60** 0.21** .04 -0.14 -0.10 0.03 -0.11 0.11 0.01 -

33 4.2 Hypothesis testing

The first hypothesis (institutional voids in emerging countries are higher than the institutional voids in developed countries) was tested by means of a t-test for the equality of two means. The t-test indicates that there is a significant difference in institutional voids between emerging and developed countries: t(116.63) = 10.07, p = 0.000, 95% CI[0.70, 1.04], d = 1,19. The results show that the institutional voids in emerging countries were significantly higher (M = 0.-50, SD = 0.79) than the institutional voids in developed countries (M = -1.38,

SD = 0.28). The effect of size, Cohen’s d, indicates that the effect is large. On the basis of this

result, hypothesis 1 is supported.

A regression was performed to investigate hypothesis 2: There is a positive relationship between the home country institutional voids and the CSR reporting intensity of MNEs from emerging and developed countries. The ability of institutional voids to predict the CSR intensity was tested after controlling for industry type and industry size. In the first step of the analysis, the predictors of industry size and the seven dummies of industry type were entered. This model was significant (F(8, 174) = 2.27, p < 0.05) and explained 9.5% of the variance in CSR intensity (Table 5). In the final model, institutional voids were added and the results did not show a significant effect of institutional voids on the CSR intensity: F(1, 173) = 0.043, b* = -0.02, p = 0.836. This means that hypothesis 2 is rejected: the degree of institutional voids do not have an effect on the reporting on CSR. The final model shows that none of the individual predictors were statistically significant with the CSR intensity. This indicates that the control variables are only significant if they are combined and, individually, are close to significant (p <0.1).

34 Table 5. Predictors of CSR reporting intensity

CSR reporting intensity Variable Model 1 b* Model 2 b*

Constant 17.94 17.84 Industry size .03 .03 Mining -.01 -.01 Construction -13 -.13 Agriculture -.09 -.09 Transportation -.15 -.15 Wholesale trade -.12 -.12 Retail trade .13 .13 Services .07 .07 Institutional voids -.02 R2 0.10 0.10 F 2.27 2.01 ΔR2 0.00 ΔF 0.04 Note. N=183

Hypothesis 3 (the CSR reporting intensity of MNEs from developed countries is less intensive than the CSR reporting intensity of EM-MNEs) was tested with a t-test for the equality of two means. The t-test indicates that there is a significant difference in CSR reporting between emerging and developed countries: t(144.2) = -3.10, p = 0.002, 95% CI[-3.61, -0.80], d = 0.92. The CSR reporting intensity in the EM-MNEs’ reports were significantly lower (M = 16.69, SD = 5.30) than the CSR reporting intensity in the reports of developed country MNEs (M = 18.67, SD = 3.90). The effect size, Cohen’s d, indicates that the effect is large. On the basis of this result, hypothesis 3 is rejected. Emerging countries’ MNEs do not report more intensively on CSR than developed countries’ MNEs. In fact, developed countries’ MNEs report more intensively on CSR than EM-MNEs. To further

35 elaborate on this difference in CSR reporting, t-tests for the equality of two means was performed separately for the economic, social, and environmental dimensions of CSR reporting. The t-tests indicate that there is a significant difference in environmental CSR reporting (t(160.53) = 3.36, p = 0.001, 95% CI[ 0.21, -0.80, d = 0.50]) and social CSR reporting (t(145.37) = 3.23, p = 0.002, 95% CI[ 0.57, 2.37], d = -0.50) between emerging and developed countries’ MNEs. Hereby, emerging countries report significantly lower on the environmental (M = 3.75, SD = 1.14) and social (M = 10.40, SD = 3.51) dimensions of CSR than developed countries on the environmental (M = 4.27, SD = 0.93) and social (M = 11.89,

SD = 2.46) dimensions of CSR. There is no significant effect between the economic CSR

reporting from MNEs from emerging countries and developed countries: t(181) = 1.31, p = 0.191, 95% CI[ -0.12, 0.60], d = 0.62. This indicates that EM-MNEs (M = 3.30, SD = 1.29) and MNEs in developed countries (M = 3.53, SD = 1.21) report with the same intensity regarding economic CSR practices. A summary of the t-tests for the equality of two means can be found in Table 6.

Table 6. Contrast of developed and emerging country MNEs for CSR reporting intensity

(score range) Developed country MNEs Emerging country MNEs 95% CI Variable M SD M SD T p LL UL Cohen’s d Triple bottom linea 18.67 3.90 16.69 5.30 -3.10 0.002 -3.61 -0.80 0.92 Environmentalb 4.27 0.93 3.75 1.14 3.36 0.001 0.21 -0.80 0.50 Socialc 11.89 2.46 10.40 3.51 3.23 0.002 0.57 2.37 -0.50 Economicb 3.30 1.29 3.53 1.21 1.31 0.191 -0.12 0.60 0.62

Note. CI = confidence interval; LL = lower limit; UL = upper limit. Adopted from “American

Psychological Association. (2010). Publication manual of the American psychological association 6th

edition. Washington: American Psychological Association.

a

36 Lastly, hypothesis 4, the effect of institutional voids and internationalization on the CSR reporting intensity, was tested. This moderation analysis was performed by means of the program PROCESS, which allows to work with a moderator. The regression coefficient for XM is c3 = 5.83 and is statistically different from zero: t(183) = 2.86, p = 0.005. This means

that the effect of institutional voids on the intensity of CSR reporting depends on a firm’s internationalization. Moreover, this model accounts for 14% of variance in the CSR reporting (Table 7).

Table 7. Moderation effect of internationalization on the relationship between institutional

voids and CSR reporting intensity

Variable b* SE t Constant 18.57 0.66 27.93 Internationalization -3.26 1.99 -1.64 Institutional voids -0.06 0.61 -0.10 Internationalization 5.83 2.04 2.86** Industry size 0.00 0.00 0.94 Mining -0.03 0.90 -0.03 Construction -3.37 2.02 -1.67 Agriculture -5.03 10.92 -0.46 Transportation -2.76 1.44 -1.91 Wholesale trade -1.98 1.26 -1.57 Retail trade 1.28 0.98 1.30 Services 1.05 1.28 0.82 R2 0.04 F 2.37

37 A closer look at the effects indicates that the relationship between institutional voids and CSR intensity is only significant with the lowest level of internationalization: (-0.34),

effect = -2.04, p = 0.019 SE = 0.86, CI[-.3.75 to -0.37] (Table 8). The effect is negative,

meaning that institutional voids have a negative influence on CSR reporting intensity. However, as shown when probing the interactions (Figure 2), the slope linking institutional voids and CSR intensity becomes more positive when the level of internationalization becomes higher (0.28), effect = 1.56, p = 0.077, SE = 0.88, CI[-0.17 to 3.3]. Although not significant, under high levels of internationalization, institutional voids have an opposite effect and positively influence the CSR intensity. Hypothesis 4 is thus partly supported, internationalization moderates the relationship between institutional voids and CSR reporting intensity, although in a negative way.

Table 8. Conditional effect of institutional voids (X) on CSR reporting intensity (Y) at levels of

internationalization (M) Level of Internationalization b* SE t -0.34 2.04 0.86 -2.36* -0.13 -0.83 0.63 -1.31 0.01 0.00 0.61 0.00 0.16 0.91 0.73 1.24 0.28 1.56 0.88 1.78

38

Figure 2. The effect of home country institutional voids on CSR reporting intensity at high

and low levels of internationalization. CSR reporting intensity ranges between 0-25. The Figure is based on the highest and lowest level of internationalization from Table 8.

High levels of internationalization Low levels of internationalization Low institutional voids High institutional voids

C SR re port ing inte n si ty

39 5 Discussion

5.1 Institutional voids and CSR reporting in emerging and developed countries

Voluntary reporting about MNEs’ economic, social, and environmental contributions was shown to result in several benefits, such as strengthening reputation and legitimacy (Bachmann & Ingenhoff, 2016). Since EM-MNEs face legitimacy challenges due to home country institutional voids (Marano et al., 2016), they were expected to report more intensive on CSR than MNEs from developed countries. The main objective of this study is to provide empirical evidence for the relationship between institutional voids and CSR reporting intensity and to determine whether there is a difference between MNEs from developed and emerging countries. In addition, the moderation effect of internationalization was examined. Two hypotheses tested the relationship between the dependent, independent, and moderator variable. Another two hypotheses examined the difference between MNEs from emerging and developed countries concerning their home country institutional voids and CSR reporting intensity. The results of the study show that institutional voids only influence CSR reporting intensity when MNEs are low internationalized. Furthermore, MNEs from developed countries report more intensive on CSR than MNEs from emerging countries.

The assumption that institutional voids are more present in emerging countries was supposed by several previous studies (Henisz & Zeler, 2005; Marquis et al., 2011; Marquid & Raynard, 2015). According to Marquis et al. (2011), emerging countries are often characterized by underdeveloped capital markets, governments, corporate governance and legal systems. This makes them more susceptible to corruption and coups. This study distinguished emerging countries from developed countries based upon the MSCI ACWI index. Hypothesis 1’s confirmation proves the assumptions concerning the existence of

40 institutional voids in emerging countries and indicates that the distinction between emerging and developed countries in the sample is appropriate.

Second, this study reveals some interesting findings regarding the relationship between institutional voids and CSR reporting intensity. Hypothesis 2’s results show that the CSR reporting intensity does not directly depend on the institutional environment of a country. These results contradict the finding of Marana et al. (2016), which show that institutional voids have a direct positive effect on CSR reporting. When MNEs of developed countries are included in the sample, however, there is no effect found. This could be explained by the second institutional view which argues that CSR reporting in not a function of the economic and social development of a country. This view argues that increasing globalization causes that institutional environments have less influence and MNEs adapt themselves to the expectation of the global environment (Tempel & Walgenbach, 2007). The result might indicate that exposure to global pressures makes MNEs more inclined to use legitimacy practices, independently from the institutional quality in their home country.

The results of hypothesis 3, however, show that the CSR reporting intensity does differ between MNEs from developed and emerging countries. MNEs from developed countries report more intensive on the social and environmental dimensions of CSR than MNEs from emerging countries. Since the CSR reporting intensity does not directly depend on the institutional quality of a country, the difference in CSR reporting intensity among MNEs from developed and emerging countries must depend on another variable. In the literature review, the influence of the institutional environment on CSR reporting is mainly argued by the fact that strong home country institutions, such as governments, press, regulations, and independent organizations (Campbell, 2007; Jamali, 2007; Kemp, 2001), put pressure on firms to adopt CSR practices (Campbell, 2007). Since only the social and environmental dimensions of CSR reporting are higher for developed countries’ MNEs, it