January 27th, 2017
University of Amsterdam
The impact of a MNEs value orientation, CEO orientation and ownership on conflicts with indigenous communities and the moderating effect of CSR practices.
Joe Duell – 11112298
MSc. In Business Administration – International Management Business School, University of Amsterdam
Supervisor: Dr. I. (Ilir) Haxhi
Statement of originality
This document is written by Joe Duell 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 Economic and Business is responsible solely for the supervision of completion of the work, not for the contents.
Abstract
Exploiting local resources, Multinational Enterprises (MNEs) face resistance of the local communities who are highly depended on the environment. Current literature on conflicts between the MNEs and indigenous communities around the world leaves us somehow without any indication of what the global drivers of conflicts are. This study therefore explores conflicts between MNEs and indigenous communities from a more global perspective, opposed to the current existing literature, which takes a case-by-case approach. Since current literature fails to identify any global drivers of conflicts, we start with analyzing corporate governance, a predominant driver of MNE outcomes in the 21st century. We identify the effect of the value
orientation, Chief Executive Officer (CEO) origination and ownership of the MNE, as corporate governance proxies, on the duration of conflicts with indigenous communities and degree of violence in one. Additionally, we analyze the integration of Corporate Social Responsibility (CSR) practices in the MNE as moderator to contribute to the ongoing discussion if CSR practices influence organizational outcomes. Our results support for our following prediction: the odds that MNEs with an inside-originated CEO have a long rather than short conflict is almost a factor two higher, than for MNEs with an outside-originated CEO. The odds that MNEs with an inside-originated CEO have a long rather than short conflict is almost a factor two higher than for MNEs with an outside originated CEO. Consequently, our study enriches current literature with the identification of, to our knowledge, the first global driver of conflicts from the MNEs side. Future research should focus on further explaining this phenomenon and expanding our knowledge by identifying other drivers to decrease, or even stop these conflicts.
Table of Contents
ABSTRACT ... 3 TABLE OF CONTENTS ... 4 INTRODUCTION ... 5 LITERATURE REVIEW ... 10 INDIGENOUS COMMUNITIES ... 10 CONFLICTS ... 12 CORPORATE GOVERNANCE ... 15 Value orientation ... 16 CEO origination ... 17 Ownership ... 19 CSR ... 20 THEORETICAL FRAMEWORK ... 23 THE DEGREE OF VIOLENCE AND DURATION OF THE CONFLICT ... 24 THE INFLUENCE OF VALUE ORIENTATION ... 24THE INFLUENCE OF THE CEO ORIGINATION ... 26
THE INFLUENCE OF OWNERSHIP ... 27 THE MODERATING EFFECT OF CSR ... 29 CONCEPTUAL FRAMEWORK ... 30 METHODS ... 32 SAMPLE ... 32 VARIABLES ... 33 Dependent variables ... 34 Independent variables ... 35 MODERATING VARIABLE ... 35 CONTROL VARIABLES ... 36 MODEL SPECIFICATION ... 37 RESULTS ... 40 DESCRIPTIVE STATISTICS ... 40 ASSUMPTIONS FOR THE REGRESSION ANALYSES ... 42 REGRESSION ANALYSES ... 46 MODERATION ... 52 DISCUSSION ... 54 THEORETICAL IMPLICATION ... 61 PRACTICAL IMPLICATION ... 61 LIMITATIONS ... 62 FUTURE RESEARCH ... 64 CONCLUSION ... 65 REFERENCES ... 67 APPENDIX I ... 80 APPENDIX II ... 85
Introduction
The relationship between indigenous communities and foreign Multinational Enterprises (MNEs), which operate mainly in extraction industries, often result in conflicts with these local communities (Barber & Jaxkson, 2012; Davis & Franks, 2011). In search of resources and efficiency, the MNEs exploit local resources by expanding strategically to foreign markets; which can harm the natural and social environment of these indigenous communities (Dunning, 1991). Indigenous communities tend to resist the activities of MNEs (Lertzman & Vredenburg, 2005), since they are highly depended on their local environment (Calvano, 2008; Lertzman & Vredenburg, 2005). Resistance can result in conflicts between these indigenous communities and the MNE (Barber & Jackson, 2012; Fontana, Sastre-Merino & Baca, 2015; Mutti et al. 2012).
In this context, conflict is an interaction between a MNE and an indigenous community concerning a disagreement. These conflicts externalize in e.g.: peaceful public protests, destruction of property and all sorts of violence, which in worst cases resolves in deaths (Davis & Franks, 2011; Oetzel, Getz & Ladek, 2007).
Despite growing interest in this subject, previous literature (e.g., Fontana et al., 2015) has been unable to identify global drivers of conflicts between MNEs and indigenous communities, as it mainly focuses on specific cases offering a myopic and fragmented view. Understanding the global drivers of conflicts could lead to generalizable counteractive measures, universally applicable in generalized governance types (Brunninge, Nordqvist & Wiklund, 2007).
Since current literature fails to identify any global drivers of conflicts (Fontana et al., 2015), we turn to corporate governance, a predominant driver of MNE outcomes in the 21st century (Tricker, 2015, p. 3). The corporate governance of an MNE can be explained as the manner in which an organization deals with uncertainty of transactions (Aguilera & Jackson,
(Alves, & Mendes, 2004) and agency problems (Core, Holthausen & Larcker, 1999). Acknowledge abovementioned, comprehension of the relationship between corporate governance of a MNE and their conflicts with indigenous communities therefore enhance the probability of counteractive measures to conflicts.
Nordqvist & Wiklund (2007) argue that corporate governance covers all issues related to the dimensions’ value orientation, board composition and ownership of a MNE. In 2003, Aguilera and Jackson propose an actor centered, broadly adopted, model, in contrast to what they claim mainly over socialized corporate governance theories. They conceptualize capital, management and labour as critical stakeholders. Cross-refering their model with Nordqvist & Wiklund (2007), we can connect value orientation with management, board composition with management and ownership with capital as domains of corporate governance.
Scholars generally argue that MNE characteristics influence conflict’s outcome (Banerjee, 2008). More specifically, Calvano (2008) provides a link between the value orientation of a MNE and conflicts with indigenous communities. Although MNE’s are increasingly interested in finding common ground with external stakeholders, actual relevance still highly depends on how interesting the group is to a MNE (Mitchell, Agle & Wood, 1997).
Literature falls short in directly linking board composition of a MNE with these conflicts. Board composition is however a broadly discussed topic in the literature (Daily, Dalton & Cannella, 2003). Most value is in considering the Chief Executive Officers (CEOs) origination if it comes to effect on MNE outcomes, for the CEO is responsible for the operations and the questionable effect of the board (Hermalin & Weisbach, 1991). Therefore, CEO origination is commonly used in research related to MNE outcomes (Kuang, Qin & Wielhouwer, 2014; Zhang et al., 2013).
Like CEO origination, ownership of a MNE is, to our knowledge, not studied in relation to the conflicts between MNEs and indigenous communities. Current literature does however connect ownership to other MNE outcomes, e.g. performance (Boardman & Vining, 1989; Mehran, 1995) and CSR activities (Barnea & Rubin, 2010; Johnson & Greening, 1999). Moreover, scholars argue that CEO origination and ownership have a broader effect on stakeholder involvement and relationships with external parties (Prado-Lorenzo, Gallego-Alvarez & Garcia-Sanchez, 2009; Hillman, Keim & Luce, 2001; Rashid & Lodh, 2008). Hence, CEO orientation and ownership of a MNE should have a direct effect on conflicts with indigenous communities, considering these communities as important stakeholders (Freeman, 2010).
Thus, by building on previous research, this study aims to bridge the gap in the current literature and extend our understanding of corporate governance of MNEs in relationship with the conflicts with indigenous communities. We therefore aim to develop partial insight in drivers of the duration and degree of violence in a conflict. Considering the discussed dimensions of corporate governance of MNEs and the conflicts, the first research question addresses it as followed:
RQ1: How do the value orientation, CEO origination and ownership of a MNE effect the degree of violence and duration of conflicts with indigenous communities?
MNEs currently try to improve their relationship with indigenous communities with programs under the banner of Corporate Social Responsibility (CSR) (Mutti, Yakovleva, Vazquez-Brust & Di Marco, 2012). The incorporation of CSR in governance practices has, among others, been researched by Kolk (2008) as well as Jamali, Safieddine & Rabbath (2008).
The latter even arguing that corporate governance is not entirely effective without a sustainable CSR. Porter and Kramer (2006) explain that without strategic integration of a CSR program which fit with the core-business, these programs fail to accomplish what they are designed for.
Scholars disagree on the influence of CSR on conflicts between MNEs and indigenous communities. Kapelus (2002) and Jenkins (2004) for example argue that CSR has a positive influence on the conflicts, whereas Banerjee (2008) and Slack (2012) argue against such claims. To our knowledge, no study has been conducted on the moderating effect of CSR on conflicts between MNEs and indigenous communities. Therefore, leaving this discussion inconclusive. Extending on case studies and previous research, this study addresses the second research question as followed:
RQ2: To what extent does CSR play a moderating role in the relation between value
orientation, CEO origination and ownership of a MNE and the duration of a conflict and degree of violence with indigenous communities?
This study takes a quantitative approach with the purpose to empirically establish which types of corporate governance are or are not related to a higher degree of violence and longer duration of a conflict. As well as to establish the same support for the moderating effect of the CSR practices of the MNEs involved. To confirm this empirically, this study uses exclusively secondary data, namely approximately 700 cases from 51 countries and 5 continents. Firstly, this study tests the mean differences between the MNE value orientation, CEO orientation and ownership, used to represent corporate governance, and the degree of violence and the duration of the conflicts. Secondly, the effect of CSR practices of the MNEs is divined analysing if the
results of degree of violence and duration of the conflicts differs significantly if a MNE incorporates CSR practices more or less than average.
This study contributes to the current literature by creating a foundation on which future research can build while analysing the global drivers of conflicts between MNE and indigenous communities. The influence of CEO origination is supported in this study, whereas MNE orientation and ownership are not considered prominent drivers of conflict outcome. Secondly, this study provides insight in the effect of CSR practices of a MNE in conflict situations. For practitioners, this study embodies a summary of current literature on conflicts, CSR practices and the influence of corporate governance on an organisation. For practitioners, this study embodies a summary of current literature on conflicts, CSR practices and the influence of corporate governance on an organisation. More actionable however, the evidence that origination of the CEO of the MNE has a direct effect on the durations of conflicts.
This study starts with a review of current literature concerning this matter, namely indigenous communities; conflicts; influence of the dimensions of MNE governance: orientation, CEO origination, ownership and CSR practices. Moving onwards, introducing the theoretical framework and hypothesis. Then explaining the research design, briefly discussing the variables as well as the used sample. Leading to the analyses of the results, discussion of these results and the implications. Lastly, this study concludes with our main claim, answering the research questions and a higher level of abstraction, opening new positional ways to tackle the issue.
Literature review
Several concepts are of importance studying the relationship between MNE governance characteristics and the degree of violence, and duration of conflicts with indigenous communities
(RQ1), and the effect of CSR practices of MNEs on this relationship (RQ2). To analyze and
explain this relation one needs understanding in underlying theory of the indigenous communities, the conflicts that conspire and the MNE governance characteristics: value orientation, CEO origination, ownership and CSR practices. The literature review elaborates on these concepts exhaustively. Firstly, we discuss the indigenous communities and the conflicts. After which the MNE governance characteristics: value orientation, CEO origination and ownership are further illustrated, and lastly CSR practices as well.
Indigenous communities
“Land is our physical life—food and sustenance. Land is our social life; it is marriage; it is status; it is politics; in fact, it is our only world” (Denoon, 2000, in Banerjee, 2001; p:54).
The UN describes indigenous communities as pre-settler’s societies with a strong links to territories and surrounding natural resources with a distinct social, economic or political system and distinct language, culture and beliefs (United Nations, w.d.). This is a working definition, provided by Jose R. Martinez Cobo, for the practical purpose of a commonly accepted understanding of the term (NCIV, 2010). To avoid forced classification, the UN created Article 33: The United Nations Declaration on the Rights of Indigenous Peoples (United Nations, 2008). This article introduces self-identification as basis for defining oneself as an indigenous person (United Nations, 2008). Netherland Centre for Indigenous People (NCIV) (2010) provides us with an explanation of this phenomenon: “It is sometimes argued that all Africans are
groups, separate classes of citizens are being created with different rights. The same argument is made in many parts of Asia.” (NCIV, 2010). By instating self-identification, to the format of
Cobo’s definition, the UN avoids creating a strict definition and all conflicts that would follow between them and communities, and communities amongst themselves.
Approximately twenty percent of the world population, spread across 70 different countries, are indigenous people (United Nations, w.d.). These communities cumulative speak more than 4.000 languages and all have their own culture base of unique merits (NCIV, 2010). Even though, these communities share common attributes and comparable history. A popular quote supporting this was made by Linda Tuhiwai Smith: “share experiences as peoples who
have been subjected to the colonialization of their lands and cultures and the denial of their sovereignty, by a colonizing society that has come to dominate and determine the shape and quality of their lives (Smith, 1999 p. 7 in: Murphy & Arenas, 2010, p. 103)”. Group
identification becomes more important when a community is threatened, and indigenous communities are especially dependent on the ecosystem for their survival (Calvano, 2008; Lertzman & Vredenburg, 2005). As such, indigenous communities share common perspectives as land is life, strong collectivism, strong cultural expressions, dependence on traditional knowledge and high biological diversity (NCIV, 2010). Moreover, O’Faircheallaigh (2013) emphasizes that indigenous communities’ relationship with nature is of importance for their cultural identification and economical and social survival.
From the MNE perspective, indigenous communities are an important stakeholder, especially in the extraction industry (Lertzman & Vredenburg, 2005). The interest of MNE in a stakeholder group depends on the cumulative of three characteristics: power, legitimacy and urgency (Mitchell, Agle & Wood, 1997). Indigenous communities are a legitimate stakeholder,
considering the high dependence of indigenous communities on their ecosystem (Calvano, 2008; Lertzman & Vredenburg, 2005). Yet from the perspective of MNEs, indigenous communities have historically been seen as “fringe” stakeholders, meaning “typically disconnected from or
invisible to the firm because they are remote, weak, poor, disinterested, isolated, non-legitimate, or non-human” (Hart and Sharma, 2004, p. 10; in Murphy & Arenas, 2010). The communities
also benefit partially from globalization, since globalization of NGO ensures more international attention for the conflicts and therefore increasing their urgency and power (Kapelus, 2002; Murphy & Arenas, 2010).
Conflicts
Conflicts have financial and reputational consequences for the MNE and are harmful for the indigenous people (Calvano, 2008; Kemp et al., 2010; Murphy & Arenas, 2010). Indigenous communities are highly depended on their geographical location (Calvano, 2008) and therefore tend to actively defend their spatial location. Consequently, conflicts between MNEs and indigenous communities arise due to utilization of opportunities by MNEs to exploit or extract resources on which indigenous communities are highly depended (Calvano, 2008; Lertzman & Vredenburg, 2005). Because of the high dependence on these resources or land, indigenous communities resist the presence or activities of MNEs, what can result in a conflict between these indigenous communities and the MNE (Barber & Jackson, 2012; Bebbington & Bury, 2009; Calvano, 2008; Kapelus, 2002; Mutti et al. 2012; Fontana, Sastre-Merino & Baca, 2015).
Conflicts in the broadest definition is “a serious disagreement or argument, typically a
protracted one” (Oxford dictionary, 2016). This definition is supported in the application of the
term in the relationship between MNEs and indigenous communities by Rahim’s (2002) who describes conflicts as a disagreement within or between social entities, such as individuals,
indigenous communities are a dynamic process. Conflicts generally conspire when the costs and benefits are experienced differently or the development is not compatible with the interests of different stakeholders (Kemp, Owen, Gotzmann & Bond, 2010).
MNEs are increasingly confronted with opportunities to pursue activities as exploitation and extraction of resources as a result of increasing international integration of markets, goods, and capital (Garrett, 2000). Driven by technological innovation, growing international activities and liberalization of foreign economic policy, these activities are increasingly pursued in foreign and previously untouched areas of the world (Garrett, 2000). Unsurprisingly, since exploitation and extraction of resources can harm the natural and social environment, relationships between indigenous communities and MNEs who entered foreign countries in search for exploitation and extraction of resources are characterised by conflicts (Barber & Jaxkson, 2012; Calvano, 2008; Davis & Franks, 2011; Franks, 2009; Kapelus, 2002; Lertzman & Vredenburg, 2005). Besides harming the environment by exploitation or extraction of resources, the UN identifies six causes as main causes of conflicts: poor engagement of communities and stakeholders, inadequate benefit sharing, an extreme impact on the economy, society and the environment, mismanagement about funds and financing war, inadequate institutional and legal framework and unwillingness to address the natural resources question in peace agreements (United Nations, 2012). These causes match with the challenges described by Sawyer & Gomez (2008) about missing institutions in the extraction industry and Murphy & Arenas (2010) cross-sector collaboration value creation model.
The reason for conflicts between MNEs and indigenous communities are relocation of an indigenous community or environmental, cultural, welfare or human right diminishment. Conflicts are most likely to occur where resource exploitation or extraction contradict with local
social, cultural, religious and environmental norms, a group is systematically excluded from the decision-making processes and when the benefits are concentrated to a certain party (UN, 2012). Examples of conflicts are a MNE using force to drive indigenous communities from their land or destructive retaliations of indigenous communities after chemical spills in their environment.
Conflicts externalizes in peaceful public protests, destruction of property and all sorts of violence, which in worst cases resolves in deaths (Davis & Franks, 2011; Oetzel, Getz & Ladek, 2007). The Oxford dictionary (2016) defines violence as “behaviour involving physical force
intended to hurt, damage, or kill someone or something”. Davis & Franks (2011) argue that
more than 90 percent of the conflicts revolve in physical protests as demonstrations and 40 percent of all conflicts entail damage to properties, injuries to people and deaths. Rather than being discrete events, these conflicts progress over time (Calvano, 2008; Kemp et al., 2010), in which the conflicts flows through a lifecycle with various intensities of violence (Kriesberg, 1998). Based on Kriesberg’s (1998) theory, Oetzel, Getz & Ladek (2007) argue that solutions for conflicts have variant effectiveness over time.
From a MNEs perspective consequences of a conflict can be of financial and reputational nature (Calvano, 2008; Kemp et al., 2010; Murphy & Arenas, 2010). Resolving conflicts can be increasingly difficult, time consuming and costly (Davis & Franks, 2011). Davis & Franks (2011) identify risk & conflict management and disruption of the process as most commonly cost factors for MNEs in conflict situation. An example given by David & Franks (2011) entails a mining operation with a capital expenditure of US$ 3-5 billion, suffering a US$ 20 million a week due to disruption of the process, whereof only US$ 10.000 were wage costs. They also identify fine, legal and remediation costs, yet in a significantly lower number of cases relative to the abovementioned costs. Reputational consequences are strongly argued from the CSR
literature (Network for Business Sustainability (NBS), 2010). Management of the relationship with indigenous communities can have both positive and negative consequences for a MNE, depending on their positive or negative performance. Positive reputational effects are increased willingness to pay (NBS, 2010) and decreasing operational cost (Porter & van der Linde, 1995). Negative reputational effects are, besides the direct opposites of the proficiencies, difficulty raising new capital and high cost on public relations (David & Franks, 2011).
From the perspective of the indigenous communities these conflicts threaten their way of life (Calvano, 2008; Lertzman & Vredenburg, 2005). Calvano (2008) and Saywer and Gomez (2008) agree on the existence of a paradox where conflicts between MNEs and indigenous communities become increasingly known internationally and more institutions focus on protecting the rights of these communities, yet the majority of communities still find themselves victim to discrimination and exploitation. Due to their high dependence on their environment, relocation or environmental, cultural, welfare or human right diminishment have disastrous consequences (Calvano, 2008; Lertzman & Vredenburg, 2005).
Corporate governance
Corporate governance is a framework of formal and informal institutional which dictates the way business should be done to avoid the uncertainty that naturally comes with transactions. Daily et al. (2003) speaks of the blind leading the blind in studies concerning corporate governance, because they argue that there is more that we do not know about corporate governance then that we do. Corporate governance has in early comparison been divided into the Anglo-Saxon-model/shareholder orientation and the Rhineland model/stakeholder orientation (Aguilera & Jackson, 2010). Yet, argued in the same study, due to increased globalization this theory might not be comprehensive any more.
Driven by technological innovation, liberalization of foreign economic policy and increasing international integration of markets, goods, and capital, MNEs are more and more confronted with opportunities to pursue these activities (Garrett, 2000). However, Garrett (2000) argues this phenomenon as globalization, one should disassociate the definitions of globalization and internationalization. The latter suggesting cross-border relationships, whereas the former is a more inclusive concept, or at least is used as such in present context (Robertson, 2012). Scholars tend to disagree about the intensity of globalisation (Ferguson, 2005) and whether globalisation in reality is a new phenomenon (Ghemawat, 2003; Hopkins, 2011). Even though this discussion concerning globalisation is ongoing, continental models of corporate governance are no longer deemed sufficient (Aguilera & Jackson, 2010; Khanna, Kogan, & Palepu, 2006).
Brunninge et al. (2007) more extensively argue that corporate governance covers issues which are related to value orientation, CEO origination and ownership of a MNE. Although this, nor any other, definition is comprehensive (Aguilera & Jaxkson, 2010), Brunninge et al. (2007) theory is in line with Aguilera and Jacksons’ (2003) dimensions and incorporate multiple angles often discussed in the literature (Anderson & Reeb, 2003; Barnea & Rubin, 2010; Baysinger & Hoskisson, 1990; Bhagat & Black, 1999; Boardman & Vining, 1989; Daily & Dollinger, 1992; Hermalin & Weisbach, 1991; Johnson & Greening, 1999; Mehran, 1995; Webb, 2004; Zhang, Zhu & Ding, 2013). These different characteristics are exhaustive discussed below.
Value orientation
As introduced, corporate governance has in early comparison been divided into a two broad systems, namely the Anglo-Saxon/shareholder orientation and the Rhineland model/stakeholder orientation (Aguilera & Jackson, 2010). Per Letza, Sun & Kirkbride (2004), stakeholder oriented organisations are more focused on the diverging interest of different stakeholders, whereas a shareholder oriented organisation focusses on the wealth maximization
of its owners and agents. This represents the discussion between Friedman (1970) and Freeman (1984), stakeholder versus shareholder theory. Lazonick & O'sullivan (2000) explain the widespread believe that shareholder orientation leads to economic benefits. Whereas Webb (2004) argues stakeholder oriented firms are more likely to have an effective governance structures than shareholder oriented firms.
CEO origination
CEO and other executives’ origination, being from inside or outside the firm, might be the most central governance mechanism. It is also one of the most often mentioned ones in the literature of corporate governance, under the banner of inside or outside CEO or board composition (Daily et al., 2003). The CEO is the top executive, responsible for the operations and management of the organisation and usually a member of the board of directors. The board of a firm is the link between the owners and the agents, responsible for developing strategy and the execution of the firms’ daily activities (Brunninge, Nordqvist & Wiklund, 2007; Ozer 2010). Hereby its first and foremost task is monitoring (Hermalin & Weisbach, 1991).
Economists generally disagree over the importance of a board and whether an outside or inside originated CEO is preferable (Rosenstein & Wyatt, 1990). Moreover, Bhagat and Black (1999) summarise the evidence of previous studies into the relationship between outside management and firm performance as well as introducing their own. They find that a high proportion of outside executives on the board correlates with slow growth, or in other words, significant evidence that outside executives on the board is directly negative related to performance. Bhagat and Black (1999) additionally suggest it may therefore be valuable to have inside representation on the board. Hermalin and Weisbach (1991) studied the direct relation between specifically the CEO origination and firm performance and argue against a direct relation between inside or outside organizational leadership and performance of the firm, yet
admit this might be just because they used an insufficient test. Another explanation for their findings is that CEO origination simply does not matter, because inside and outside CEOs might both be equally skilled in representing shareholder interests (Hermalin & Weisbach, 1991). Dalton, Daily, Ellstrand, & Johnson (1998) support Hermalin & Weisbachs’ (1991) findings arguing against a relationship between CEO origination and firm performance. Although Hermalin & Weisbach (1991) deny the influence of CEO origination on firm performance, they do argue boards generally fail to accomplish their task in monitoring and guiding companies, and call for regulations requiring outside presence on the board. Comparing Hermalin & Weisbachs’ (1991) and Bhagat and Blacks’ (1999) suggestions on insider and outsider presence, Hermalin & Weisbach (1991) argue against influence on performance and steer towards better governance, whereas Bhagat and Black (1999) claim a negative relation with outside board members and performance for which they suggest to keep focus on wealth maximization.
Besides considering board composition in relation with firm performance, Baysinger and Hoskisson (1990) for instance formulate proposition for evaluating the strategic implications of board compositions, requiring empirical analysis. More recently, Hillman, Keim and Luce (2001) find evidence for the link between board composition and stakeholder relations. Whereas Webb (2004) argues that board composition contributes to continued investors interests. Considering more trending subjects, Zhang, Zhu and Ding (2013) examine that increased presence of outside and women CEOs is linked to better CSR performance within a firm’s industry.
Thus, considering the abovementioned literature one can state, as Daily et al. (2003), our knowledge about board effectiveness and influence is rivaled by what we do not know. Daily et al. (2003) therefore uses the phrase “the blind leading the blind” (p. 371). What scholars do
agree on is the influence of CEO origination on corporate governance, and the need for future study (Daily et al., 2003).
Ownership
Ownership of an organisation is a broadly discussed subject in organisational literature and Hart & Moore (1990) explain that ownership influences firm behaviour. Scholars collectively argue that a certain type of ownership has a significant effect on performance, strategy or CSR practices. Mehran (1995) for instants argues that private ownership has a positive effect on firm performance. Anderson & Reeb (2003) complements Merhan (1995) by stating that not only privately owned firms perform better, but family owned firms outperform non-family owned firms. Boardman & Vining (1989) however, argue against this assumption and claim the empirical evidence supporting it to be weak at best. On contrary, Daily & Dollinger (1992) argue that unification of ownership and control positively effects performance, thanks to better structure and proses management due to interest alignment. Considering CSR, Barnea and Rubin (1992) argue that dispersed ownership has a positive effect on disclosures of CSR information.
Following to Brunninge et al. (2007), this study divides ownership in publicly, private and state owned. Public ownership involves publicly traded shares, diffused ownership and control, and public corporate governance. Private ownership however involves no publicly trades shares, usually with concentrated ownership and control (Boot, Gopalan & Thakor, 2006). State ownership is also characterized by no market listing and concentrated ownership and control (Dewenter & Malatesta, 2001). In contrary to public and private ownership, Dewenter and Malatesta (2001) proclaim that state ownership has less agency problems, display inferior profitability, generally uses more leverage and tend to display greater labor intensity. Although publicly and privately owned firms seem to perform better, Dewenter and Malatesta (2001) state that this is not evidence that privatization itself increases firm profitability. Something in
contradiction with the Washington Consensus, which proclaim privatization as a means for better performance (Williamson, 2000). Another difference between publicly/privately owned firms and state owned ones is that state owned firms are obligated to strife for social and political objectives, such as wealth redistribution, instead of perusing maximum profit (Dewenter & Malatesta, 2001).
Another difference between the different ownership types is the governance and control. For example, privately owned firms have less concern about reputation/public opinion, since there is no stock value to worry about or political pressure (Boot et al., 2006). Moreover, the more dispersed control of publicly traded firms increases its agency problems in comparison to privately owned firms (Boot et al., 2006; Brunninge et al., 2007).
CSR
Till today there is not one standard definition of CSR, and despite numerous efforts there is still a lot confusion on how CSR should be defined (Dahlsrud, 2008). In this study CSR is defined as a concept whereby companies integrate social and environmental concerns in their business operations and in their interaction with their stakeholders on a voluntary basis and in a context specific way (Aguinis & Glavas 2012). Crane, Matten and Spence (2013) expand on this definition by underling six main ideas of CSR (the focus on triple bottom line): economic, social and environmental, the importance of all stakeholder, that these practices are voluntary, context-specificity, managing externalities to sustainable development and that these practices should be rooted in values and ethics of the organization. Using this definition, CSR should not be confused with corporate sustainability, meaning “meeting the needs of the present, without compromising the ability of future generations to meet their needs” (Brundtland Commission, 1987). Both, closely related practices however, concern business decision making linked to
ethical values, compliance with legal requirements, and respect for people, communities and the environment (Business for Social Responsibility; in Crane, Matten & Spence, 2013).
The incorporation of CSR in governance practices has, among others, been researched by Kolk (2008) and Jamali, Safieddine & Rabbath (2008). The latter even arguing that corporate governance is not entirely effective without a sustainable CSR. On the other hand, Porter and Kramer (2006) explain that without strategic integration of a CSR program that fit with the core-business, these programs fail to accomplish what they are designed for. Together, Jamali et al. (2008) and Porter and Kramer (2006) describing co-dependence between CSR practices and corporate governance.
Like the definition, there is no universal CSR measurement for companies either. Despite the growing interest of actors and scholars in CSR (Williams & Aguilera, 2008), the world has not yet adopted one specific measurement used by firms all over the world. Arguably, Aguilera, Rupp, Williams & Ganapathi’s (2007) statement that CSR practices around the world are increasingly diffusing, contradictory to DiMaggio & Powell’s (1983) isomorphism theory, can explain the absence of such model. Currently there are numerous different indexes, the FTSE4Good of, among others, the SIRI group being the largest in Europe and the KLD index representing the most of firms in the U.S.A. (Márquez & Fombrun, 2005).
FTSE4Good identifies itself as an index designed to measure the performance of companies demonstrating strong environmental, social and governance practices. They proclaim to be transparent and that their clearly-defined criteria make FTSE4Good indices suitable tools to be used by a wide variety of market participants when creating or assessing responsible investment products (FTSE, 2016). FTSE4Good indexes more than 2000 companies in Europe and Amerika (FTSE, 2016; Márquez & Fombrun, 2005). KDL on the other hand is a data set
with annual snap-shots of the environmental, social, and governance performance of companies rated by KLD Research & Analytics. The annual statement contains Identifying Company information, strength and concern ratings for multiple indicators within seven qualitative issue areas, concerns for six controversial business issues and a summary for each of these 13 areas it covers (KDL STATS, 2010). Since 2003, KDL covers the largest 3100 US companies (KDL STATS, 2010; Márquez & Fombrun, 2005).
Theoretical framework
Current literature fails to present a generalizable methodology to design a strategy for avoiding conflicts, even though methodology for designing a strategy is universally applicable in generalized governance types (Brunninge, Nordqvist & Wiklund, 2007). Moreover, as comprehensively discussed in the introduction, current literature leaves us with a gap in insights of general influence of certain governance types on conflicts with indigenous communities
(RQ1). Previous research has placed emphasis on single case studies of conflict between MNEs and indigenous communities, which provides us with insight in the specific situations and therefore underlying information to analyze the overarching framework. Comprehension of the relationship between corporate governance of a MNE and their conflicts with indigenous communities therefore enhances the probability of creation of a partial generalizable counteractive methodology. Hence, this study core focus is on establishing a link between corporate governance and conflicts. Hypotheses concerning proxies for governance types and their relation with conflicts are presented below. The results of these hypotheses are the next phase in the literature towards developing counteractive measures to the engagement of conflicts and their duration and degree of violence.
Additionally, this study entails the moderating effect of CSR. As Calvano (2008) argues: “I uncovered several trends that may bode well for a de-escalation of tension in the future. The
first is the increasing connection between CSR and the international development goals of poverty relief and sustainability”. Yet besides case studies, no overarching study has been done
into the relationship between CSR practices and conflicts between MNEs and indigenous communities. Therefore, this study also incorporates the effect of CSR practices of the MNEs on their relationship with conflicts with indigenous communities (RQ2).
The hypotheses discussed below are not meant to exhaust the universe of predictors, but merely to suggest several hypotheses that may be pursued using data on conflicts between MNEs and indigenous communities. Directly below we first discuss the conflicts. The degree of violence and duration of the conflicts represent the conflicts in this study, since they are clear portray of the disastrous effect of these conflicts.
The degree of violence and duration of the conflict
As discussed in the literature review, conflicts externalize in procedure-based and physical protest, violence to property and violence to people (Davis & Franks, 2011; Oetzel, Getz & Ladek, 2007). Davis & Franks (2011) argue that more than 90 percent of the conflicts revolve in physical protests as demonstrations and 40 percent of all conflicts entail damage to properties, injuries to people and deaths. Rather than being discrete events, these conflicts progress over time (Calvano, 2008; Kemp et al., 2010), in which the conflicts flows through a lifecycle with various intensities of violence (Kriesberg, 1998). Based on Kriesberg’s (1998) theory, Oetzel, Getz & Ladek (2007) argue that solutions for conflicts have variant effectiveness over time. Kriesber’s (1998) theory entails that throughout a conflicts tensions rise, violence increases and over time decrease again due to decreased tension. This circle repeats itself until the conflict is resolved. Therefore, increased duration of a conflict has a direct impact on the suffering of the MNEs and indigenous communities. Consequently, this study aims on relating these outcomes with corporate governance proxies and CSR practices.
The influence of value orientation
Although MNE’s are increasingly interested in finding common ground with external stakeholders, actual relevance of an external stakeholder still highly depends on how interesting the group is to a MNE (Calvano, 2008). Mitchell et al. (1997) argue that the interest of MNEs in a stakeholder group depends on the cumulative of three characteristics: power, legitimacy and
urgency. Considering the high dependence of indigenous communities on the ecosystem, their legitimacy as stakeholder is argued for (Calvano, 2008; Lertzman & Vredenburg, 2005). From the MNE perspective however, indigenous communities have little to no power and are interpreted as illegitimate and non-urgent stakeholders (Hart and Sharma, 2004). Accordingly, acknowledging a stakeholder or even acting upon its interest is not directly implied with the theory of Mitchell et al. (1997). Concluding, the value orientation of a MNE being relatively more oriented towards stakeholders, opposed to more shareholder oriented MNEs should affect the interaction between a MNE and an indigenous community.
Letza, Sun & Kirkbride (2004) elaborate in their study on this Freeman (1970) versus Friedman (1984) discussion by stating that stakeholder oriented organisations are more focused on the diverging interest of different stakeholders, whereas a shareholder oriented organisation focusses on the wealth maximization of its owners and agents. Considering a firm’s own priority, a stakeholder oriented firm is therefore more open to recognising the interest of indigenous communities as stakeholders, than a shareholder oriented firm. Contradictory however, following the theory of Porter & van der Linde (1995), it would be in a firms’ best interest to manage its operations without conflicts with indigenous communities. So, considering Davis & Franks (2011) calculations of costs, mainly derived from risk & conflict management and disruption of the process, a firm can create a lot of value for its shareholders by avoiding confrontations. Nonetheless, most conflicts conspire due to access to resources otherwise out of reach for firms (Barber & Jaxkson, 2012; Calvano, 2008; Davis & Franks, 2011; Franks, 2009; Lertzman & Vredenburg, 2005; Kapelus, 2002). Firms solely focused on their own wealth maximization would arguably be less compelled to comprehensively take local communities into account whilst designing its operation.
Considering this literature, a stakeholder orientated MNEs would generally be more interested in considering the indigenous community interests, than a shareholder orientated MNEs. Therefore, the duration of conflicts is expected to be lower with a stakeholder oriented MNE, to decrease the suffering of the indigenous communities. Hence the following hypothesis:
H1a: The duration of conflicts will be significantly longer with shareholder oriented MNEs, as opposed to stakeholder oriented MNEs.
Haxhi and Aguilera (2016) argue a firm is more interested in the degree of violence, due to increased publicity with higher degree of violence in a conflict. Considering the same argumentation as for H1a, specifically Mitchell et al. (1997), we expect the degree of violence in a conflict to be lower with a stakeholder oriented MNE. These firm’s already have relatively increased interest in their stakeholders and the power of these stakeholders only grows due to increased media attention (Mitchell et al., 1997).
H1b: The degree of violence of conflicts will be significantly higher with shareholder oriented MNEs, as opposed to stakeholder oriented MNEs.
The influence of the CEO origination
The CEO is responsible for guiding the operations of the organisation whereby he or she should consider the interest of the stakeholders. Scholars disagree whether an inside or outside originated CEO is better suited for representing these interests (Bhagat & Black, 1999; Hermalin & Weisbach, 1991). Scholars do however agree on the fact that the origination of the CEO does has an impact on the corporate governance of a MNE (Daily et al., 2003), and therefore should affect the MNEs relation with stakeholders (Banerjee, 2008). Concluding, the origination of the CEO, being it from inside or outside a MNE, should affect the relation between MNEs and
Hermalin and Weisbach (1991) classify CEO based on his or her background into insider or outsider, hereby proclaiming that outside CEOs are more stakeholder interest oriented than inside CEOs. More recently, Liang & Hendrikse (2013) explain in their paper that outside CEO origination often suggests better performance of the CEO. Possible due to the fact they have a stronger desire to demonstrate superior performance and receive more pressure from the board and labour market. A better performance of the CEO, or in other words, better representing the interest of the stakeholders, should positively affect the conflicts between MNEs and the indigenous communities. Following abovementioned literature, and considering indigenous communities as stakeholders, an outside originated CEO should have a direct positive effect on duration of a conflict, hence the following hypothesis:
H2a: The duration of conflicts will be significantly longer with an inside originated CEO, as opposed to an outside originated CEO.
Comparable to H1a and H1b, we expect the same relation between CEO origination and degree of violence as we do for CEO origination and duration of a conflict. As explained, CEO origination is directly linked to other MNE outcomes as better incorporating stakeholders interest (Liang & Hendrikse, 2013). A better relation with the indigenous communities should decrease the degree of violence, and therefore we expect this to be lower with an outside originated CEO. Thus, the following hypothesis:
H2b: The degree of violence of conflicts will be significantly higher with an inside originated CEO, as opposed to an outside originated CEO.
The influence of ownership
outcomes as performance (Anderson & Reeb, 2003; Mehran, 1995), strategy (Daily & Dollinger, 1992) and CSR activities (Barnea & Rubin, 2010). Scholars have also argued that ownership has a broader effect on stakeholder involvement and relationships with external parties (Hillman, Keim & Luce, 2001; Prado-Lorenzo, Gallego-Alvarez & Garcia-Sanchez, 2009; Rashid & Lodh, 2008). Following this logic, ownership should have a direct effect on stakeholder interaction like conflicts with indigenous communities (Freeman, 2010).
Difference between privately owned firms, publicly traded and state owned ones is that private firms are more or less self-sufficient, due to concentrated control and ownership (Boot et al, 2006). Publicly traded firms however are more concerned about their reputation/public opinion due to sensitivity of their stock and outside control (Boot et al., 2006). Moreover, the more dispersed control of publicly traded firms increases its agency problems in comparison to privately owned firms (Boot et al., 2006; Brunninge et al., 2007). State owned firms on the other hand are obligated to strife for social and political objectives, such as wealth redistribution, instead of perusing maximum profit (Dewenter & Malatesta, 2001).
Considering these dependencies and presumable owners’ interests, privately owned MNEs should be related to a longer duration of conflicts. Ergo the following hypothesis:
H3a: The duration of conflicts will be significantly longer with privately owned MNEs, as opposed to publicly and state owned MNEs.
In line with the previous hypotheses, we expect the influence of ownership of a MNE to be the same for the degree of violence in a conflict as its influence on the duration of one. We hereby follow the same argumentation as for the duration of the conflict. Additionally, we consider the increased media attention for conflicts with a high degree of violence (Haxhi & Aguilera, 2016),
what will be relatively of more importance to publicly traded and state owned MNEs (Boot et al., 2006; Brunninge et al., 2007). Conflicts between state owned a privately owned MNEs and indigenous communities are expected to have a relatively low degree of violence. Consequently, the following hypothesis:
H3b: The degree of violence of conflicts will be significantly higher with privately owned MNEs, as opposed to publicly and state owned MNEs.
The moderating effect of CSR
“I have argued that increasing conflict between MNCs and local communities is a result of the
confluence of three interrelated factors including stakeholder power inequality, stakeholder perception gaps and cultural context. In addition, in developing my framework for understanding MNC and conflict, I uncovered several trends that may bode well for a de-escalation of tension in the future. The first is the increasing connection between CSR and the international development goals of poverty relief and sustainability” (Calvano, 2008, p. 801).
CSR practices are a way in which organization integrate social and environmental concerns in their business operations and in their interaction with their stakeholders on a voluntary basis and in a context specific way (Aguinis & Glavas 2012). Integrating these concerns into business decision creates a linked to ethical values, and respect for people, communities and the environment (Business for Social Responsibility; in Crane, Matten & Spence, 2013).
CSRs influence on the conflicts between a MNE and indigenous communities has become increasingly popular (Banerjee, 2008; Jenkins, 2004; Kapelus, 2002; Slack, 2012). Mutti et al. (2012) collaborate that MNEs try to improve their relationship with indigenous communities, amongst which decreasing the violence and duration of the conflicts, using the
2011; Kemp, Owen, Gotzmann, & Bond, 2011) or ethical reasons (Carroll, 1999). Yet to our knowledge, there has been no study on the general influence of CSR practices on conflict situations. Based on Calvano’s (2008) findings and the rooted nature of CSR practices this study hypothesis that CSR practices of a MNE have a positive effect on the conflicts between MNEs and indigenous communities. However, without insight in which factors have a significant effect on the length of a conflict or its degree of violence this hypothesis can only be generally formulated. Therefore, the following hypothesis:
H4a: CSR practices of a MNE have a negative effect on the relation between a MNEs’ corporate governance dimension and the duration of a conflict.
Repetitive to the previous hypotheses, the moderation effect of CSR is expected to be the same for degree of violence in a conflict as the duration of one. The argumentation for the effect on the degree of violence is identical to the argumentation of H4a. In line with H4a, we miss critical insight which factors have a significant direct effect on the degree of violence in a conflict, what restrains us from defining distinct hypotheses. Ergo the following hypothesis:
H4b: CSR practices of a MNE have a negative effect on the relation between a MNEs’ corporate governance dimension and the degree of violence in a conflict.
Conceptual framework
The figures below represent the conceptual framework in this study. The concept is the same for all three analyses. On the left in the figures we present the independent variables and on the right we have the depended variables. These are the direct relation we hypotheses presented in black outlined boxes with black arrows connecting them. In the middle between the two direct lines between the independent and depended variables we have a blue outlined box in which we
present the moderator variable. The blue lines represent the moderation effect on the direct relation. Figure 1 represents the framework concerning the independent variable value orientation. Figure 2 considers the independent variable CEO origination in the same setup, as is the independent variable ownership is in Figure 3.
Figure 1: Conceptual framework for value orientation
Figure 2: Conceptual framework for CEO origination
Methods
This chapter discusses the sample, the variables and research design used for this study. Moreover, it discusses the specifics of the dataset and describes the analyses. The analyses are of a quantitative nature and were executed using IBM SPSSStatistics, version 23.0.0.0 from 2016. All cases selected are collected from the database made available by Dr. Ilïr Haxhi, which entails empirical data of 709 cases. These cases, initially documented by third parties who study indigenous communities and the conflicts they encounter, are coded according to a coding scheme (Appendix I).
Sample
Using purposive sampling, the sample for this research entail 709 cases of conflicts between MNEs and indigenous communities in Argentina, Australia, Bolivia, Brazil, Cambodia, Canada, Cayman Islands, Chile, China, Colombia, Denmark, Dominican Republic, Ecuador, Ethiopia, Finland, France, Germany, Great Britain, Honduras, Hong Kong, India, Ireland, Italy, Japan, Jersey, Channel Islands, Kenya, Kuwait, Libya, Luxembourg, Malaysia, Mauritius, Mexico, Namibia, Netherlands, Nicaragua, Nigeria, Norway, Pakistan, Panama, Peru, Philippines, Poland, Republic of Korea, Russia, Saudi Arabia, Singapore, South Africa, Spain, Sweden, Switzerland, Taiwan, Tanzania, Thailand, United Arab Emirates, United Kingdom, United States of America, Venezuela, Vietnam and Zimbabwe. Due to missing values in the independent, moderator and control variables, 494 cases are excluded, ergo bringing the total sample for the analyses to 215. Checking for the assumptions of a regression, no more cases were dropped.
The data gathering was under the guidance of Dr. Haxhi with a group of seven undergrads, including myself. All of us followed the same coding scheme, which is added in Appendix 1. Considering the subject and available resources and time, the data was purposive
data, including but not limited to, the United Nations, NGOs; independent groups researching and/or monitoring indigenous communities, newspapers and other media sources. Although all variables in the coding scheme are collected for the cases, only the ones discussed below are used in this study.
Variables
Using existing literature, cases are selected complying with the abovementioned criteria. The independent variables are MNE value orientation, ownership and board composition. The depending variables are degree of violence and duration of the conflicts with the indigenous communities, and the moderating variable is the integration of CSR practices. The table below presents all variables, dimensions and coding, which are explained thoroughly below.
Table 1: Variable description
Variable Dimensions
Length of conflict A Case divided by sum of total Length of conflict B 0. short, 1. long
Degree of violence 1. low intensity, 2. low-medium intensity, 3. medium intensity 4. medium-high intensity, 5. high intensity
Value orientation 0. Stakeholder orientation, 1. Shareholder orientation CEO origination 0. Outsider,
1. Insider
Ownership 0. Privately owned,
1. Publicly and State owned CSR practices 0. Under average CSR practices,
1. Above average CSR practices Geographic isolation of
community
1. Highly isolated, 2. Low isolation Cultural Isolation of Community 1. Highly isolated,
2. Low isolation, 3. No isolation
Dependent variables
The dependent variables in this study concerns outcome variables of the conflicts. Specifically, the duration and degree of violence of these conflicts. The dependent variable of conflict length was coded with ongoing cases coded until September 2016, which is the time frame of the research. This study considers a conflict to have ended in case a settlement was reached, whether through legal action or without the involvement of the legal system. The length of conflict is gathered in number of months and therefore a continuous data. To improve the value of the outcome we recoded the length in two different variables, Length of conflict Continuous (C) and Length of conflict Dichotomous (D). The former to give insight in the contribution of a case to the total length of conflicts as a percentage. The variable is calculated by dividing the length of a conflict by the sum of the total duration of all cases in the study. The sum equals 26479 months. Next we multiplied the variable with one hundred to align the outcome value with the percentage on a scale of zero to one hundred. Moreover, this action will increase the readability of the results. Length of conflict D, cases are separated on basis of having a value above or below the mean of all cases included in the study. Hereby coding cases below the mean as short (0), and cases above as long (1).
The degree of violence ranges from low intensity, low-medium intensity, medium intensity, medium-high intensity to high intensity violence. Low intensity entailing road blocks, occupation of firm-owned property; occupation of public property; occupation of private property and occupation of other property. Low-medium intensity adding destruction or damaging these abovementioned properties, medium intensity adding injuries, medium-high intensity adding kidnapping, and high intensity additionally including deaths. Calvano (2008) classifies the intensity of the conflict as an important characteristic concerning the nature of the conflict. As Rhemtulla et al. (2012) state in their research, when a variable has more than five
categories, it is acceptable to treat such a variable as continuous variable. The degree of violence variable therefore is a continuous variable and thus a scale measurement was used.
Independent variables
The independent variables in this study are MNE value orientation, ownership and board composition, which are all nominal variables. The first, value orientation is divided in
shareholder and stakeholder orientation. The former indicating that a MNE is solely focused on
the common interest wealth maximization of its owners and agents, the latter implying that the MNEs focusses on the diverging interest of different stakeholders (Letza, Sun & Kirkbride, 2004). Following García-Castro, Ariño, Rodriguez & Ayuso (2008) example, this study uses Porta, Lopez-de-Silane, Shleifer & Vishny (1996) paper to assumes value orientation of an organization based on their home country.
The second independent variable, CEO origination, is divided into two categories, based on Hermalin and Weisbachs’ (1991) study: insider and outsider. Using Orbis (2016) as source, insider or outsider CEO was determined on the bibliographic information in congruence with the year the selected case started.
The last independent variable, ownership, is categorised as privately owned, publicly
owned and state owned, following Brunninge et al. (2007) theory. To properly analyze
hypotheses 2a and 2b, privately owned is set of against publicly and state owned. Moderating variable
The moderating variable is CSR practices of the MNE. KLD and FTSE4Good scores are examples of measurement tools for CSR practices. As Clarkson, Richardson and Vasvari (2008) as well as Cho, Lee and Pfeiffer (2013) argue, the use of a general measure of CSR does not only increase the comparability by rooting out information asymmetry. It also increases the reliability of the CSR score, instead of using organisations own CSR disclosures. The measurements for
CSR practices, like KDL and FTSE4Good, base their score mainly on these disclosures, yet also factor in benchmarks and enforces the transparency.
For the analyses, the CSR tool CSRHUB (2016) was selected by means of availability and accessibility. This measure includes different third parties’ analyses and converts them into their own overall measurement scale. The mean of the CSR practices became the orientation for the dummy variable. Cases higher than the average were coded as high integration (1) and cases which integration was below average were coded as low integration (0).
Control Variables
This study controls for the possible isolation of the indigenous communities with two variables, geographic isolation and cultural isolation. The isolation of a community in geographical sense is expected to influence conflict, since firms would only go to isolated areas for very specific reasons as needed resources. The importance of the venture for the MNE may influence the duration or degree of violence in a conflict. Furthermore, isolated locations are expected to have less coverage in terms of media, NGO and political attention. Additionally, communication at all is probably more challenging overall with more isolated communities. This variable entails highly isolated communities, coded as 1, where members live within a community with little to no contact with the outside world. Moreover, communities where members live within broader municipalities with mixed populations are coded as 2, less isolated.
The latter, cultural isolation of the indigenous community is expected influence the situation as well. High isolation of a community can affect communication due to language barriers or cultural misconceptions. Moreover, high isolation is expected to influence the coverage in terms of media, NGO and political attention in the same manner as geographical isolation. Communities with an isolated culture, a different language and without being
are connected to a broader ethnic group then their own, but are isolated within these connections are coded 2, low isolation. Lastly, communities which are embedded within a broader mainstream ethnic group are coded as 3, no isolation.
Model specification
To analyze the different relationships between the three nominal independent variables and the three dependent variables, of which length of conflict C and degree of violence are continuous and length of conflict D is dichotomous, this study uses linear and logistic regression analyses. The regressions conclude which independent variables significantly predict a longer duration of conflicts and higher degree of violence between indigenous communities and MNEs. A stepwise approach is applied in both regression types, meaning that the DV and CV are put together first, after which the IV is added. Results are deemed significant with a p value of/below .05.
To test for validity and reliability of the analysis, we check for assumptions. The assumptions for a linear regression and a logistical regression are the same (Field, 2009). The first two assumption concern the multicollinearity and correlation of the variables. Considering the former, this study uses a variance inflation factor test (VIF-test) with a cut value of VIF = 5 and significance cut value of .05. The correlation of the variables is checked using the Pearson correlation presented in a cross table, using a cut value of p = .05. Thirdly, the errors should be independent and uncorrelated which is a matter of study design (Pallant, 2013; Pedhazur, 1997; Wright, 1995). Additionally to these assumptions, we considered The Hosmer-Lemeshow test for the logistical regression. This test is designed to assess the goodness of fit of the model and generally used in risk assessment analyses in predictive modelling and machine learning. The findings of the Hosmer-Lemeshow test is however very sensitive to sample size, chancing significantly of value and p-value when increased or decreased (Kramer & Zimmerman, 2007).
After testing for the direct relation between de IVs and DVs, this research tests for the moderation effect of CSR practices of a MNE for the IV. To do so we centralize each the IV’s and the moderation variable before multiplying them with each other. This product represents the interaction between the moderator and IV. The assumptions this test match the ones for the already executed regressions and are therefore already controlled for.
Table 2 on the next page shows a summary of the analyses of this study. This summery includes all direct relation tests and moderation analyses.
Table 2: Summery regression analyses
Control Variables Independent variables Moderator Dependent variables
Model Geographic isolation of community Cultural Isolation of Community Value orientation CEO origination
Ownership CSR practices Length of conflict continuous Length of conflict dichotomous Degree of violence 1 x x x 2 x x x x 3 x x x x 4 x x x x 5 x x x 6 x x x x 7 x x x x 8 x x x x 9 x x x 10 x x x x 11 x x x x 12 x x x x 13 x x x x x 14 x x x x x 15 x x x x x 16 x x x x x 17 x x x x x 18 x x x x x 19 x x x x x 20 x x x x x 21 x x x x x
Results
This chapter presents and discusses the analyses and results of the expected relations. The first section presents the descriptive of the dataset and discusses the validity of the regression analyses by checking the assumptions for multicollinearity and correlation of the variables. Moreover, this section concurs the errors are independent and uncorrelated. The pre-tests conclude the validity of the regression analyses and therefore the second section will discuss the findings of this analysis. The results section covers the analysis itself and confirms one significant finding concerning the stated hypothesis. CEO origination has a significant effect on the length of a conflict, if this length is measured in short and long (Length of conflict D). All other expected relations, including the moderations, are found to be insignificant. This chapter will however present the analyses fully in the latter paragraphs.
Descriptive statistics
Table 3 presents the descriptive data of the variables used in this study. To begin with the depended variables, the mean length of a conflict is approximately 10 years, and 68.2% of all cases have a duration between 7 months and 19 years. We coded the length in a continuous (C) and dichotomous (D) version to analyse the relative length between the cases (C) and a more general approach divided in long and short (D). The average contribution of a case to the total duration of all conflict is .4651 (SD: .41257), or .47 percent since we already multiplied the variable with a factor 100. Recognizing there are 215 cases, an average contribution close to .5 percent is in line with expectation. The average is .4047 (SD: .49197) if we divide the length of a conflict in short and long, indicating the sample contains more short than long cases. The average degree of violence is 2.535 (SD: 1.9641), however the average has a low value for a categorical variable. With a closer look at the data we can however state this variable has two
