Conflicts between Indigenous Communities and
MNEs
The effects of country-level institutions on conflict resolution and the
mediating effects of autonomy and isolation.
Master Thesis University of Amsterdam
Msc. Business Administration - International Management Name: Luc van Veldhuizen
Student number: 6160786 Date: 26-01-2018
Thesis Supervisor: dr. Ilir Haxhi
Statement of originality
This document is written by Luc van Veldhuizen 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.
Table of Contents
Abstract ... 4 1. Introduction ... 5 2. Literature review ... 11 2.1. Indigenous communities ... 11 2.1.1. Indigenous communities’ rights ... 12 2.1.2. Indigenous communities’ identity ... 14 2.2. Conflicts ... 15 2.3. Country-level Institutions ... 18 2.4. Autonomy ... 20 2.5. Isolation ... 21 3. Theoretical Framework ... 23 3.1. Conflict Resolution ... 23 3.2. Country-level institutions and their effects on communities’ characteristics ... 25 3.3. Internal institutional effect on autonomy and isolation ... 26 3.4 Conceptual Framework ... 28 4. Methodology ... 30 4.1. Data collection and sample ... 30 4.2. Dependent variable ... 31 4.3. Independent variable ... 32 4.4. Mediating variables ... 34 4.5. Control variables ... 35 4.6. Modelling and analysis ... 36 6. Analysis and Results ... 40 6.1. Descriptive Statistics ... 40 6.1. Correlation and multicollinearity ... 41 6.1.1. About multicollinearity ... 46 6.2. Regression Analyses ... 47 6.2.1. Mediator variables ... 47 6.2.2. Type of violence ... 49 6.2.3. Degree of violence ... 51 6.2.4. Duration of the conflict ... 53 7. Discussion ... 56 7.1. Findings ... 56 7.2. Theoretical and practical implications ... 60 7.3. Limitations ... 62 7.4. Directions for future research ... 63 8. Conclusions ... 66 References ... 69Abstract
Due to increasing globalization pressures, MNEs have expanded their business activities to low-cost countries, in areas inhabited by indigenous communities. Especially in the extraction industry, these activities have often led to conflicts between MNEs and indigenous communities. Despite the increasing interest in this topic, limited research has been conducted concerning the internal effects of country-level institutions on conflict resolution. Previous studies, have explored this relation, from a qualitative perspective, omitting the broader cross-country international perspective. This study focuses on conflict resolution between MNEs and Indigenous communities, by combining country-level institutions, isolation and autonomy of communities in a single framework. A sample of 706 cases across 40 countries was used to examine the effect of country-level institutions on conflict resolution, and the mediating effect of degree of indigenous communities’ isolation and autonomy on this relationship. Our results indicate that first, strong country-level institutions positively affect conflict resolution. in addition, community autonomy, was positively related to both the conflict duration and severity. Finally, it was also indicated that isolation was positively associated with conflict. The present study expands previous literature, as it is the first to examine the elements that lead to conflict resolution in indigenous communities, in terms of the duration and severity of the conflict and its subdivision into the degree and the type of violence. Practically, it provides a combination of strategic advice and operational guidance for conflict resolution.
Keywords: Conflict, resolution, MNE, government, indigenous, communities, country-level
1. Introduction
In the last decade, Foreign Direct Investment (FDI) has increased rapidly, reaching a peak of over three trillion US Dollars in 2007 and is currently estimated at approximately 2 trillion dollars in 2016 (World Bank, 2017). Aligned with FDI, an increasing number of MNEs has internationalized their activities around the world, in areas inhabited by indigenous communities. Especially MNEs, that are operating in the extractive industry, are often involved in conflicts with these communities. The majority of them are gas, oil, mining and forestry firms, which often cause severe damages in the regions of their operations, such as erosion, deforestation, possible threat of oil leaks and other environmental tragedies (Acuña, 2015). Due to greater demand and scarcity of the available resources, MNEs tend to operate in more distant areas, that, for decades, have been inhibited by local communities practicing their beliefs and culture. MNE’s operating in these areas are often accused of causing social and environmental problems, due to the extraction on indigenous’ living grounds. As a result, the culture, as well as the natural environment of indigenous communities is disrupted (Calvano, 2008; Foster, 2012; Murphy & Arenas, 2010).
According to Calvano (2008), MNEs often intend to pursue only their own interest, their aim is to maximize their profits, that are imposed by their shareholders, who rarely consider the communities’ interests. This results in heavy and long-lasting conflicts with Indigenous communities, who are trying to protect their rights and living grounds from the intruding MNE’s. During these conflicts, the damage that occurs for MNE’s is mostly due to project delays, legal costs, damaged infrastructure and shut-down time (Kemp et al. 2011). In one of their articles Amnesty International (2014) uses the following quote, to describe the consequences for the indigenous communities:
“We Indigenous People cry only when we have achieved our freedom. Today, it is like we are coming out of a prison, so many of us are crying because it is so emotional.” (Amnesty
International, 2014).
How should MNEs manage and react to these conflicts that involve indigenous communities, in order to minimize the damage on both sides? Recent research with regards to companies in the extractive industry, shows that their social and environmental practices have been improved. However, despite the improvements, results also show an increase and intensification of community and MNE-related conflicts. Calvano (2008) considers the grounds of indigenous communities as the “new battlegrounds” of modern society (Hodge, 2014). These battlegrounds, in terms of their intensity and duration, cause severe reputational and financial damage to each actor involved. Therefore, the various factors that influence conflicts are of high importance.
Consequently, previous research has focused on investigating the several factors that are associated with and influence conflict resolution. Research has shown that, it is of high importance to acknowledge the rights of indigenous communities and have the ability to end a conflict (Bellier & Préaud, 2012; Foster, 2012). Another step forward for the indigenous can be seen in the research of Kemp et al (2011) and Klok & Lenfant (2013), who both address the increasing importance of community autonomy as it can help in starting partnerships and building relationships between communities and MNEs operating in their territories. In addition, Lertzman & Vredenburg (2005) who investigated the role of local institutions, stated that sharing the power to overcome conflicts has positive outcomes, when forming strong relationships between both actors and being guided by the local authority. The researchers clearly showed the importance of indigenous rights, autonomy, relationship building and the
role of local institutions for indigenous communities, in order to discover how these can be utilized, under certain conditions, to overcome conflicts.
Prior literature has supported the fact that strong country-level institutions influence conflict resolution (Bellier & Préaud, 2012; Hodge, 2014). However, the duration and the degree of violence exerted in these conflicts have not been examined before. The present study aims to explore this relationship. The fact that indigenous communities have different culture, beliefs, worldview and are strongly associated with their land, leads MNEs to regard them as minority groups with weak institutions. As a result, they do not frequently consider them as factors in their decision-making process (Murphy & Arenas, 2010; Foster, 2012). This lack of recognition of their land, identity and rights, of course, leads to long-lasting and intense conflicts. As a result, the United Nations (2008) have developed legislations to defend indigenous peoples. Therefore, it could be suggested that strong country-level institutions assist in conflict resolution between indigenous communities and MNEs.
However, it is still unclear what are the mediating effects of community isolation and autonomy in conflict resolution. According to Calvano (2008), community autonomy is an essential part of conflict resolution, as indigenous communities who have a certain degree of autonomy can negotiate better with MNEs. Community autonomy is critically important for indigenous communities who want to protect their rights and gain recognition by local governments and MNEs (Kolk & Lenfant, 2013). On the other hand, isolation (measured as cultural/geographical distance) has detrimental effects on conflict resolution. It has been argued that isolated indigenous communities face great difficulties, in terms of communication with MNEs (Bellier & Préaud, 2012). Nevertheless, it has not been examined yet if there is a direct association between strong country-level institutions, autonomy and isolation in indigenous communities that assists in conflict resolution. In addition, it has not been investigated how country-level institutions might have affected conflicts in their duration and intensity.
Scholars do argue that the strength of a country’s institutions is related to the poverty and the level of country development, which in addition, is argued to be related to the degree of isolation and autonomy of indigenous communities (North, 1991). However, no research has been conducted in order to examine how exactly country-level institutions may influence other factors, such as the degree of community isolation or autonomy, and how they are associated with conflict resolution. According to Kemp et al (2011) and Klok & Lenfant (2013), autonomy can tremendously improve conflict resolution. It is, therefore, of high importance to examine the existing gap in literature, starting with country-level institutions and their impact on conflict resolution, followed up by how country-level institutions may influence the degree of isolation and autonomy of indigenous communities.
This study investigates the relationship between indigenous communities’ country-level institutions, via the mediation of community autonomy and isolation and their influence on conflict resolution. The reason is that the national context in which indigenous communities are located matters, institutions need to be strong to succeed in the enforcement of legislations to have a positive effect on the position of the communities (Acuña, 2015; Bellier & Préaud, 2012). Herewith we cover three existing gaps in literature, namely first the gap, of country-level institutions and how they affect conflict resolution, which is research by the first question. Secondly and third, the gap concerning the mediating effect of community isolation and community autonomy, respectively, on conflict resolution and how they might have been influenced by country-level institutions.
The aim of this study is to explore what implications this has for indigenous communities versus MNEs in conflict resolution and to help broaden the existing literature and improve practices on conflict resolution. To address the mentioned gap in literature, this research aims to answer the following research question, which is twofold:
RQ1: What is the effect of country-level institutions on conflict resolution (i.e. duration
and degree of violence).
RQ2: To what extend is this relationship mediated by the community characteristics
(i.e. isolation and autonomy).
The purpose of this study is to investigate if there is a relation between countries institutions’ and the degree of autonomy or isolation of Indigenous communities, who reside in the country. Whether this relation can be initially determined, based on a country’s institution and if the outcomes have implications for conflict resolution in these conflicts is examined. The current study broadens previous research as it is the first study to examine the mediating effects of community isolation and autonomy in conflict resolution. To stop the negative effects of these conflicts on either side, solutions and insight in these conflicts are needed now, more than ever. Therefore, this research aims to contribute to existing literature on conflict resolution. Moreover, it intends to form a clear analysis of relations between the aforementioned factors, and their impact in conflict resolution. This study also investigates firstly how they impact conflict, both independently and when combined, broadening our knowledge, in order to assist in resolving (future) conflicts.
This study employs a quantitative approach, with a cross-sectional research design. The analysis of the factors is conducted using a sample of over 700 cases of conflicts worldwide. Conflicts in the sample are selected based on the discussed variables. For this study, country-level institution is the independent variable, conflict resolution is the dependent variable, and isolation and autonomy act as mediating variables, of which the dependent variable is measured according to the degree of violence and the duration of the conflict. This research aims to find linkages between relevant factors that are discussed when it comes to conflict resolution. Most importantly, this study attempts to establish a linkage between the effect of country-level
institutions on the degree of isolation and autonomy of indigenous communities and how, when the effects are supported it can lead to either positive or negative outcomes in conflict resolution. The hypothesis is that the communities’ degree of isolation and autonomy are influenced directly by country-level institutions and supporting that their association can assist in conflict resolution.
The theoretical contribution of this study consists of its attempt to describe the direct impact that country-level institutions have on conflict resolution and hereby contributing to the current ambiguous literature. Secondly, the present study aims to explore how a country’s institutions are associated with the degree of autonomy or isolation of a community in a country and what implications this has for conflict resolution. Practically, this paper can be useful for policymakers, as well as managers, since current literature has not addressed the relation between these variables yet. As a result, this research aims to broaden previous literature on conflicts so that the insights can be used to minimize the chance of arising (or to assist in managing) conflicts. Moreover, it is of high importance to determine, for current and future projects in the extractive industry, how each of the parties is involved and should interact with one another. To conclude, the paper intends to provide information, that can help to develop certain policies and help to apply or align them in various situations of (arising) conflict.
To provide a clear insight and understanding in the described research topic, the concepts of conflicts and indigenous communities are broadly explained in the following chapter, the literature review. Thereafter, the various relations and linkages between the variables, such as autonomy and isolation are discussed, and several hypotheses generated after a conceptual model is introduced in the theoretical framework, which is followed by a methods-and-data-collection part and complemented with sections to present and discuss the research findings.
2. Literature review
In the following chapter, the main concepts concerning the research topic are introduced and defined and previous research on these topics is discussed. First, a broad analysis of Indigenous communities, their identity and rights is presented. Second, the concept of conflicts is discussed and the major causes of arising conflicts and their lengths and degree of violence are mentioned. The following part is dedicated to countries’ level of institutions and degree of development and what implications this has for communities and conflicts within those countries’ boundaries. In the end, a clear description with regards to degree isolation and autonomy is suggested and the ways those two factors are being influenced are discussed.
2.1. Indigenous communities
In contrast to the previously discussed possible negative impacts for MNEs, stands the damage done to the so-called indigenous communities. To identify Indigenous communities’, scholars have proposed varied definitions, hence not agreeing on a unanimous definition to describe Indigenous peoples. However, Indigenous communities are characterized mostly by their distinct culture, traditional lifestyles, laws and customs; compared to the national population of the country wherein the community resides. That results in a community living within their unique social environment, geographical territory and being protected by international law (Calvano, 2008; Coates, 2005; Lertzman & Vredenburg, 2005). Consequently, the term Indigenous communities or peoples is used within this research to refer to these groups.
2.1.1. Indigenous communities’ rights
It is estimated by the UNPFII (2015) that the Indigenous reside in around 90 countries worldwide and that their numbers reach approximately 370 million. The individuals living in the indigenous communities are known to belong to the most isolated in the world, socially as well as politically, due to their geographical location and accordingly their unique language, traditions, cultures and history. A direct effect of isolation leads them to be amongst the relatively seen, poorest in the world, compared to the non-indigenous peoples (UNPFII, 2015). Their poverty results in a low quality of life and in an inability to have access to proper health care. The accessibility to healthcare for everyone is one of the main tasks of the World Health Organization (WHO) and one out of the six mandates stated by the United Nations Permanent Forum on Indigenous Issues (UNPFII, 2015; United Nations, 2008). In addition, Indigenous communities’ access to education is marginal and protection of their culture, environment and human rights is lacking, just like their economic and social development is. It is therefore, the reason why the UNPFII (2015) sees Indigenous communities as the most vulnerable in the world and the protection of their human rights and improvement of other mandates as the most challenging areas for indigenous communities to overcome.
As a result, Indigenous communities’ collective rights got strengthened in 2007 by the United Nations Declaration on the Right of Indigenous Peoples; UNDRIP (United Nations, 2008). That was initiated to help Indigenous communities to obtain autonomy, live collectively, apply their own institutions and become a political actor at an international level (Bellier & Préaud, 2012), recognizing the importance for humans living in indigenous communities, to have collective rights. More specifically, one of the rights is noted within the declaration and states that Indigenous communities can preserve their lands as actions that result in the communities’ displacement are forbidden. Nevertheless, scholars state that practice shows that the enforcement of these laws by governments is lacking and major differences of enforcement
are found within and between several countries. That suggests that the position of Indigenous communities remains to be one of the minorities in social, economic and political processes (Bellier & Préaud, 2012; Wiessner 2011). To change this vulnerable position, Garvey and Newell (2005) stated that, governments can enforce these rights and protect the indigenous communities using legal frameworks and policies. However, in order to enforce and protect these rights, it is necessary for the community to be recognized by the state (Garvey & Newell, 2005). Apart from recognition, the relationship between state and community is essential to be good in order to determine the degree of political support and protection gained by the state. Often marginal communities, in both developed and developing countries do not receive state support, instead the government focusses on developing policies which are in favor of the MNEs desires, MNEs who are willing to invest and that bring FDI were the state is depending on (Garvey & Newell, 2005; O’Faircheallaigh, 2012). Previous research did show that, historically seen, western governments had already enforced indigenous people’s rights more effectively in comparison to governments from less developed nations in Africa and Asia since these latter mentioned governments pursued mostly the interest of the majority of their nationals (Coates, 2005). However, the author stated that, western governments do not have to improve the enforcement of their policies and legal frameworks.
As mentioned, previous research has emphasized on the relationship between state and community to protect indigenous rights. This relation is accompanied by MNEs, as they are the major reason for conflicts, because their urge is to extract natural resources from the indigenous communities’ living grounds and the financial dependency of states on their FDI and thus, the relation becomes complicated (Valvano, 2008; Foster, 2012; Garvey & Newell, 2005; O’Faircheallaigh, 2013). The presence of MNEs on the indigenous living grounds and their financial power, concerns the communities for negative impacts, even though it is argued that not all FDI in indigenous regions will only neglect their rights, but they can also contribute
to new opportunities for the inhabitants (Lertzman & Vredenburg, 2005; Foster, 2012). Despite the potential benefits FDI can bring, still communities stay concerned with regards to the protection of their rights and living grounds, since most projects start as long-term, however history has shown that business can rapidly change. Therefore, this results in projects to be just as quickly abandoned, just as they started (Acuña, 2015).
2.1.2. Indigenous communities’ identity
Whiteman and Mamen (2002) argue that, identity is one of the core concerns Indigenous communities deal with aside from protecting their natural habitat. Communities strongly rely on both; however, the original living grounds are known to contain several natural resources, which provides the communities apart from their cultural and social benefits, also with economic benefits. Since indigenous are very reliant on these grounds, damage done to this environment will cause direct effects to the communities’ wellbeing (Bruijn & Whiteman, 2010). This is illustrated by Bruijn & Whiteman (2010) who state that, especially when MNEs in the extractive industry claim grounds and start extracting, indigenous heritage is being removed. To emphasize their findings the explanation which is given shows that the environment is not only heritage but part of an indigenous communities’ culture.
To understand this phenomenon, Indigenous communities culture’ and their construct of identity is analyzed by Weaver (2001), who argues that identity formation of the communities consists of three conditions, namely community identification, external identification and self-identification. The first and the latter are closely connected to the communities’ natural environment namely; their place of residence, traditional homelands and the territory surrounding it combined with having a shared history (Weaver, 2001). As a result, Bruijn & Whiteman (2010) highlight in their study the excessive negative impact MNEs from the extraction industry have on Indigenous communities which can be directly linked to their
culture and identity. Their findings emerged from a study on the Machiguenga community which is historically located in Peru, where MNEs operating in the gas extraction industry are conducting drills in search of gas on the historical living grounds of the community, therewith harming the preservation of the natural habitat. That is concluded to form a direct danger to the survival of the indigenous identity (Bruijn & Whiteman, 2010).
Because of the threatening loss of the indigenous living grounds and identity, activism against MNE’s operating on Indigenous grounds is growing in numbers and intensity (Calvano, 2008). Accordingly, due to the threat to their community, Lertzman and Vredenburg (2005) found that at these moments identification with the community becomes most essential. To conclude, during these conflicts with MNE’s the historical identity of the indigenous becomes a powerful instrument and by highlighting its importance, international attention and support are gained (Bruijn & Whiteman, 2010).
2.2. Conflicts
As mentioned in the previous chapters, several studies have demonstrated the major cause of conflict is to interfere within the natural habitat of indigenous communities, mostly without prior consultation or recognition by the MNEs operational in these regions. That is supported by the argument of the high dependency of indigenous peoples on their environment and accordingly the importance of their surroundings for their identification as a community (Bruijn & Whiteman, 2010; Lertzman & Vredenburg, 2005). Scholars note that, from a historical perspective, mostly western MNEs have neglected the value, identity and distinctive cultures of the communities and have mostly considered their own norms, values and culture as superior compared to the indigenous. As a result, indigenous communities were hardly considered as stakeholders which made them keep their vulnerable position in protecting their
lands and therefore made them involved in conflicts for decades (Laplante & Spears, 2008; Calvano, 2008; Amnesty International, 2014; Acuña, 2015).
The nature of conflict can be found when two or more organizations or groups interests are not corresponding. Conflict is part of society and usually leads to social change and development, mostly when it is non-violent (United Nations, 2012). When a country’s institution is stable and law and order are being enforced, different interests are usually managed with trust amongst both parties involved. However, conflicts can turn violent when societies have weak institutions and their political systems are insufficient to protect rights of weaker groups in conflict situations (United Nations, 2012). Authors note that it is in the interest of the international community to prevent violent conflict to arise and to work towards peaceful resolutions and recognition for the indigenous communities. Since numerous political and economic powers are affected in conflicts (Bellier and Préaud, 2012). Moreover, they can cause severe damages, for example dislocation, disruption of culture and loss of the previously described identity from the communities living on these grounds (Acuña, 2015; Laplante & Spears, 2008).
Conflicts especially arise in the extractive industry, when it comes to exploitation of natural resources, like minerals, gas and oil, since the consequences of these extractions are the most devastating for the preservation of the indigenous peoples living grounds (Calvano, 2008; United Nations, 2012). Foster (2012), for instance, argues the literature focused on conflicts shows indigenous strive for recognition, bargaining power and sharing in profit flows to make it possible to serve as protectors of their natural environment, towards MNEs as well as national governments. While several authors agree upon the necessity of recognition and benefit sharing in conflicts, literature shows a lack of mutual understanding (Calvano, 2008; Hodge, 2014). As a result, still a high number of conflicts in this industry turn violent since the indigenous communities living on the extracting grounds are not given the recognition they need in the
process ahead of the extractions. It turns out, MNEs are forcing communities to leave their historically determined living grounds to extract the non-renewable resources without substantial benefits for the communities; which only increases the dissatisfaction and therefore the chance of conflict (Laplante & Spears, 2008). The example of increasing danger for conflict given by Laplante & Spears (2008) complies with four out of six drivers for conflict that are identified by the United Nations (2012) and explain the existence and intensification of conflicts. Namely: Firstly, inadequate benefit sharing, secondly extreme impact on the economy, society and environment, thirdly poor engagement of communities and stakeholders and finally, most important in this research, inadequate institutional and legal frameworks. The remaining drivers are mismanagement about fund plus financing war and lastly, unwillingness to address the natural resources questions in peace agreements (Laplante & Spears, 2008; United Nations, 2012).
Two other major factors that are increasing the chance of conflict and impose serious risks on both MNEs and communities are the following: firstly, the presence of MNEs in the extraction industry on a certain location is rarely being reversed, since the nature of investments in this industry is long-term and the financial costs are high. Secondly, the operations are bound to a certain location and therefore cannot be easily moves elsewhere (Acuña, 2015; Kolk & Lenfant, 2013; Hodge, 2014). Consequently, scholars argue that given the dominance of MNEs over Indigenous communities’ conflicts increase and intensify. Therefore, managers of these companies should recognize the position and needs of their vulnerable stakeholder who needs a more deliberate approach of certain MNE practices (Kolk & Lenfant, 2013). As a result, several MNEs have developed new practices towards indigenous communities and other stakeholders in their Corporate Social Responsibility (CSR) guidelines (Kolk & Lenfant, 2013; Calvano, 2008; O’Faircheallaigh, 2013). Yet, multinationals regularly do not live up to their Corporate Social Responsibility (CSR) guidelines; it is therefore needed that MNEs will be
held responsible when their activities result in negative effects and spillovers for the local community. To achieve this, more insight in the development of conflicts and their resolution across various countries and their local context, is needed (Calvano, 2008).
2.3. Country-level Institutions
In literature on conflicts multiple authors argue that it is a great necessity to consider the role of the national context within a country when analyzing the existence of conflicts, instead of focusing only on acts of communities and MNEs and governments (Ikelegbe, 2005; Kolk & Lenfant, 2013; North 1991). The national context of a country can be described using several factors such as national governance, degree of development or amount of natural resources. In order to increase our understanding of Indigenous communities’ rights and their enforcement across countries this research will focus predominantly on country-level institutions. This is one out of six main drivers of conflict situations (United Nations, 2012; Laplante & Spears, 2008). A definition is proved by North (1991) who considers institutions to be the ‘rules of the game’ played in a specific context. The rules can be separated in formal and informal institutions, where formal institutions consist of legislative executive, bureaucratic and judicial measures of governments to create social stability while informal institutions are customs, norms, taboos, codes of conduct and traditions (Williamson, 2000). To study the specific national context Hodgson (2006) notes that institutions are crucial to bring structure in the social sphere and names them a system of rooted and ruling social norms that shape interaction of humans. As a result, a strong institutional base comes with expectations of thought-out frameworks for social interactions which can either facilitate or restrain conflict situations (Hodgson, 2006).
The importance of strong country level institutions is shown by Ikelegbe (2005), who concludes that the chance of arising conflict can have major differences per country based on
regulations, formal laws, governmental recognition and protections of communities. Ikelegbe (2005) complies with North’s (1991) research who states, from an institutional perspective, the impact of underdeveloped countries, characterized by fragile social and legal structures with limited possibilities for enforcement, is experienced by MNEs in the form of an increasing chance for conflict. Another factor is the paradox brought by the ‘resource curse’ mentioned by Bebbington (2008) who argues underdeveloped countries tend to have a high number of natural resources, however, instead of bringing development, the abundance of resources has the opposite effect on societies wellbeing. As an example, these governments often outweigh social concerns, like protecting indigenous rights, over (personal) financial benefits that FDI from MNEs brings, which results in more freedom for MNEs when extracting resources (Ikelegbe, 2005; Foster, 2012; Kolk & Lenfant, 2013). While other authors argue that apart from threats, FDI can bring potential benefits to a community, the previous example shows uncertainty of outcomes and duration of projects are still a major concern for the indigenous groups, especially when it comes to those living in countries with weak institutions (Lertzman & Vredenburg, 2005; Foster, 2012; Hodge, 2014; Acuña, 2015).
Based on previous studies and their own research, Garvey & Newell (2005) argue that there is a necessity of building relationships between states and communities to protect and enforce indigenous rights. That is supported by Foster (2012) however, the author argues, the indigenous communities can not solely depend on the implementation and enforcement of these laws since strong national institutions are exposed to certain political pressures as well. Accordingly, the international commissions responsible for UNDRIP and FPIC lack the resources to measure the compliance of national governments which highlighted by scholars who claim major differences in enforcement of those legislations (United Nations, 2008; Wiessner, 2011;Bellier & Préaud, 2012). Even though, several scholars state the recognition of indigenous peoples becomes more advanced, however, these conclusions are drawn on
research conducted in mostly developed countries; with strong legal frameworks involvement of these governments in conflict mostly leads to positive outcomes (Foster, 2012; Calvano, 2008). The conclusion that there still is a high contrast when it comes to government involvement is strengthened by the statement that in certain situations the combination of weak regulations and corrupt governments can lead to even worse conflicts (Kolk & Lenfant; 2013; Ikelegbe, 2005). To conclude, the different findings of the authors in literature do not show clear consensus on which factors or institutions have best outcomes when it comes to conflict resolution (Bellier & Préaud; 2012; Acuña, 2015).
2.4. Autonomy
As previous research shows, in most cases, the needs of indigenous communities are not given any recognition (Kolk & Lengant, 2013; Calvano, 2008). The communities that have inhabited the land until the MNEs seized their grounds needs to be considered as a partner which can possibly have a positive impact on durations and intensity of conflicts (Davis & Franks, 2014). The importance of community identity, recognition and rights in conflicts are further addressed by Bellier and Préaud (2012) who argue that marginalization and a different worldview are the major reasons why indigenous communities often do not identify with a state, and therefore want to be autonomous to conserve and protect their territories and cultural habitat. It shows that once a community is given certain rights and autonomy, it has a better negotiating position towards MNEs and governments (Calvano, 2008). Once again,it is argued that within these conflicts national context matters although it is not proven if the resolution of conflicts is better in developed compared to undeveloped countries. (Acuña, 2015; Bellier & Préaud, 2012).
In the current situation, with regards to conflicts research of Laplante & Spears (2008) shows apart from recognition, poor engagement of MNEs with other stakeholders is a common
phenomenon. Once more, it is shown that recognition is a first step towards a better bargaining position and increases political capacities for indigenous communities (Calvano, 2008; O’Faircheallaigh, 2013). In addition, scholars argue that when communities are granted with a certain degree of autonomy this can help them to get acknowledgement from MNEs and as a result will start the possibility of building partnerships (Kemp et a., 2011; Kolk & Lenfant, 2013). Furthermore, Lerzman & Vredenburg (2005) note the formation of relationships between several actors in conflict areas, with different interest can result in positive outcomes for all involved. It shows that there is a clear consensus amongst scholars, that the minimum to resolve conflicts is creating dialogue, use FPIC and build relationships between actors involved. In this way contrasting values can be discussed, underlying causes of conflicts can be reduced and negative outcomes such as reputational or financial damage for the firm and quality of life for indigenous groups can be minimalized (Garvey & Newell, 2005; Foster, 2012; Hodge, 2014). To conclude, when applied correctly, granted autonomy can give indigenous communities the power of control which diminishes one of the main causes for community vs MNE conflict and grants them the ability to cancel, delay or disrupt MNEs activities based on legislation instead of violent conflict (Bellier & Préaud, 2012; O’Faircheallaigh, 2013).
2.5. Isolation
The last factor to complete the framework is isolation. It is argued that a community’s degree of isolation influences the difficulties MNEs face when trying to establish themselves on new business grounds. According to Ghemawat (2001), isolation can be explained using the concept of distance divided in geographical and cultural distance. Cultural distance occurs between countries or groups and is based on differences in traditions, norms and values. As
explained with the differences in country-level institutions, isolation is an important factor that leads to different behavior and is therefore used in this study.
The impact of a communities’ isolation, based on the degree of geographical and cultural distance, is argued to have negative implications when it comes to conflict resolution. Gupta & Ferguson (1992), note that when an indigenous community is isolated it can lead to an increase of cultural distance towards for example MNEs or governments. It is argued that this might lead to increasing difficulties when it comes to communication and relationship building, which is crucial for smooth conflict resolutions (Bellier & Préaud, 2012). To capture the concept of culture and determine cultural distance, Hofstede (2011) developed a model of cultural dimensions. The model is a rather standardized and relative simple tool for measuring separate cultures or the difference or distance between them. In the model a distinction on several factors that, according to Hofstede (2011) implicate the differences between cultures. The factors used are named Uncertainty Avoidance, Masculinity vs Femininity, Power Distance, Long term vs short term orientation and indulgence vs restraint.
To measure geographical distance, Bebbington & Bury (2009) focus in their study on the right and possibility of indigenous communities to get consulted ahead a certain change in their environment, for example when MNEs are planning to start extracting natural recourses from their grounds. The research shows a possible relation between geographical distance of a community and whether consultation is applied.
3. Theoretical Framework
In this chapter, previous research findings are discussed, with the aim to highlight the gaps and show the associations between different factors, that occur in the process of arising and resolving conflicts. First, the most relevant aspects for this research, with regards to conflict resolution, are discussed. Hereafter, these aspects and their direct influence between one another and conflict resolution are elaborated. Afterwards, the mediation effects are added to form the basis, in order to explore the research hypotheses. The outcomes of this study aim to provide a combination of strategic advice and operational guidance, by uniting several factors under a single framework, which is represented in a conceptual model.
3.1. Conflict Resolution
For several years, previous research has addressed conflicts involving MNE’s and indigenous communities and discussed numerous factors that are influencing conflict resolution. Research has been conducted to determine the influence of various variables, such as country-level institutions, stakeholders, relationship-building, concentration of power within indigenous communities, as well as community isolation and autonomy. Research findings indicate that, when recognition by local governments and MNEs is granted, indigenous communities’ bargaining position and political capacity are strengthened (Calvano, 2008; Bellier & Préaud, 2012; O’Faircheallaigh, 2013). The findings have supported that the empowerment of local communities can lead to the acknowledgement of their unique identity, which in turn will help them to act as an autonomous stakeholder in the decision-making processes, when it comes to extracting natural resources from their lands (Bruijn & Whiteman, 2010).
Apart from recognition, Free, Prior and Informed Consent (FPIC) can serve as a second step in the empowerment of indigenous communities. FPIC stands for the practice, in which communities receive consultation regarding the extracting practices of MNEs in their region (Laplante & Spears, 2008). However, the frameworks are not legally binding, which results in implementation within countries in a way that does not comply with the country’s own institutions. The framework of FPIC and its application by several national governments goes alongside the application of the earlier-mentioned declaration on the rights of indigenous peoples (UNDRIP) by the United Nations (2008). Even though, these two frameworks are an impressive step-forward in protecting indigenous people’s rights and improving their position in conflicts and society, national governments and MNEs still often do not comply with them fully. The main reason is the discrepancy between the governments’ national political interests (which, in this case, are wealth and economic growth) and the FDI’s aim to protect indigenous communities to achieve the preservation of their environment (Garvey & Newell, 2005; O’Faircheallaigh, 2013). For this reason, Foster (2012) indicates that, governments are still indecisive whether to support investing in MNEs and herewith do not live up to the standards of the imposed frameworks.
Nevertheless, even some improvements have been made and states work on the implementation of indigenous people’s rights, the enforcement of these rights is often lacking. As prior research has indicated, when MNEs offer substantial financial benefits for states, developing countries often do not enforce the FPIC and do not support the rights of indigenous peoples UNDRIP (United Nations, 2008; Foster, 2012). This shows a direct influence of a country’s institutions and rate of development on conflict situations. Therefore, the presence of strong institutions and political arrangements is vital. However, even if strong country-level institutions exert a major influence in conflict resolution (Hodge, 2014), the degree of violence
and its duration have not been investigated before. The effects of country-level institutions on conflict resolution are explored by utilizing the following hypothesis:
H1: Strong country level institutions will positively affect conflict resolution
3.2. Country-level institutions and their effects on communities’
characteristics
Research on conflict resolution has showed the possible influence of country-level institutions on other factors associated with it. Protecting indigenous groups’ rights and continuity of their practices, identity and culture are clearly explained by the research of Bruijn & Whiteman (2010). They noted that indigenous groups, especially in regions with weak institutions, are quite vulnerable. Therefore, authors argue that, recognition is essential for indigenous peoples, who wish to maintain their identity and initiate the process of relationship building and mutual knowledge sharing with MNEs (Bellier & Préaud, 2012; Hodge, 2014). Calvano (2008) indicates that this is the first step in granting communities a certain degree of autonomy, so that they can have a better negotiating position with MNEs, when conflicts arise.
Theoretically, community autonomy is a clear step-by-step plan to improve the position of indigenous communities. However, in reality, the majority of them is not given any form of recognition by local governments or MNEs (Kolk & Lenfant, 2013). The reason is that the national context, in which indigenous communities are located, matters and institutions need to be strong to succeed in order to enforce the legislations and have a positive effect on the position of the communities (Acuña, 2015; Bellier & Préaud, 2012). Specifically, the enforcement of earlier discussed UNDRIP and FPIC varies greatly amongst different countries. Authors suggest that, one of the reasons for this variance can be found in the strength of a country’s institutions (Lertzman & Vredenburg, 2005; Foster; 2012; Hodge, 2014). This
indicates a possible internal effect of a country’s institutions on the autonomy of indigenous communities who reside in these countries.
Autonomy can improve conflict resolution tremendously (Kemp et al., 2011), as it is critically important for the recognition of the indigenous communities’ rights. In addition, the possibility for building partnerships between MNEs or local governments and indigenous communities are associated with their autonomy. Autonomous communities have the power to control the main causes for conflicts, as well as the ability to defend themselves democratically, as they are aware of the legislation. This study is a step-forward, compared to previous research, as it examines the mediating effect of autonomy on conflict resolution. Therefore, it argues that autonomy is associated with strong country-level institutions and minimizes the presence of conflicts. Therefore, the following hypothesis is explored in this study:
H2: The mediation effect of autonomy will positively affect conflict resolution.
3.3. Internal institutional effect on autonomy and isolation
Based on previous literature, indigenous communities are the most vulnerable actors in conflicts with MNEs, as they are often described as minority groups that are isolated, because they are highly dependent on their current living grounds, which they have inhabited for several decades. Their different worldview, culture, beliefs, practices and their minority status leads MNEs and national governments to hardly consider them as factors, when starting projects on their grounds. (Coates, 2005; Foster, 2012; Kolk & Lenfant, 2013; Murphy & Arenas, 2010). The lack of recognition of their rights and identity leads to MNEs starting their projects on the indigenous communities living grounds without any consultation, which results in the destruction of the indigenous habitat. This results in longitudinal and violent conflicts with
these conflicts, the United Nations (2008) have developed the so-called UNDRIP and FPIC to strengthen the position of the indigenous. However, these legislations are being enforced in different ways across countries, as the rights or the institutional capacity to implement the legislations of countries with more weak institutions are not recognized (Ikelegbe, 2005; Wiessner, 2011; Bellier & Préaud, 2012). As a result, the position of indigenous communities in conflicts continues to be marginalized.
Another factor that influences conflict resolution is community isolation as a visible increase in difficulties towards conflict resolution is noted, both when cultural and geographical distance is witnessed (Bellier and Préaud (2012; Gupta & Ferguson, 1992; Bebbington & Bury, 2009). The difficulties in communication and relationship-building leads to more conflicts between MNEs and indigenous communities (Bellier & Préaud, 2012). However, the possible internal influences of country-level institutions are again neglected. In addition, isolated indigenous communities are less likely to protect their rights against MNEs and/or local governments and this may result in more intense and longitudinal conflicts.
It is argued that, partly, isolation is the result of weak institutions combined with a lack of autonomy for indigenous communities living in these countries. However, scholars do not mention whether low autonomy is influenced by a country’s weak institutions. This is the first study conducted, that investigates the association between weak country-level institutions and isolation, as well as the mediating effect of community isolation in the resolution of conflicts. Nevertheless, their findings show a certain correlation between the autonomy and community institutions (Kolk & Lenfant, 2013; Acuña, 2015). For the aforementioned reasons, the role of isolation in conflict resolution is still unclear within the process of conflict resolution. Therefore, this study attempts to extend our current knowledge, gained from previous studies, regarding conflict resolution by investigating the internal effect of country-level institutions on
the degree of community autonomy and their effect on the resolution of conflicts. To test this association, the following hypothesis is formed:
H3: The mediation effect of isolation will negatively influence conflict resolution.
3.4 Conceptual Framework
The figure below displays the variables used in this study. The relationships between the different variables are demonstrated in Figure 1. The independent variable, country institutions, is measured based on six development indicators such as, government effectiveness, regulatory quality, political stability and absence of violence, voice accountability and rule of law. The dependent variable conflict resolution, can be measured in the degree of violence and length of conflicts. Furthermore, the two mediating variables are community isolation and community autonomy. To determine their mediation, the results of their mediating effect will be compared against the direct effect (that country-level institutions have on both factors of conflict resolution). This approach is proposed to be solved using a four-step multiple regression analysis, proposed by Baron and Kenny (1986). The control variables in this research will be community size, the MNE Experience in the country and the HDI Score given to each country.
4. Methodology
This study uses a quantitative approach, first it aims to determine the relation between country-level institutions, measured by government effectiveness, regulatory quality, political stability and absence of violence, voice accountability, and rule of law; and conflict resolution measured by the length of the conflict and the degree of violence. Hereafter the mediating effect of isolation and autonomy on this relation will be analyzed, all following a cross-sectional research design. The relations and mediating effects will be examined with multiple regression- and correlations analysis using SPSS statistics. Other complex methods and approaches for mediation modelling such as Structure Equation Modelling (SEM) will be discussed in chapter 7. To give a clear insight in how these analyses were established the following paragraphs are dedicated to explanation of the data collection, depending the sample size, the independent-, dependent-, control variables and how the mediators are determined.
4.1. Data collection and sample
The overall design of this study will be based on a database research. The database used to conduct this research consists out of 706 cases of conflicts throughout the world, reaching from Oceania, Asia and Africa to whole of continental America. However, only completed conflicts are considered, so the sample is reduced to 269. Therefore, there are fewer countries within the analysis, but it is a sufficient sample to perform a satisfactory multiple mediation regression analysis. This is largely supported in the literature and specifically in the article by Gregory Knofczynski (2008) on the size of samples using multiple linear regression. In this article, the author emphasizes that a reliable sample for this type of models must have a minimum of 100 data, but if such sample has more than 200 data, the results will be more reliable.
The secondary data was selectively gathered via third parties such as, MNEs public data, non-profit reports and databases on conflicts involving MNEs and Indigenous communities, namely Environmental Justice Organisations Liabilities and Trade (EJOLT), Minewatch, Environmental Justice Atlas and Intercontinentally Cry. For most of these conflicts, data relevant for this study such as, country institutions, community isolation or autonomy are present. However, in certain cases, little information was presented in the previous named databases, therefore, in these cases, extra data is found using other sources such as articles found in newspapers. Accordingly, a group of students was assigned to extend the data for each of the presented conflicts, with information on stake- and shareholders involved in each conflict to create the possibility to conduct future research on these variables. To ensure the quality, comparability and completeness of the gathered data, all data inserted in the database is conducted following a uniform coding manual. To keep the validity and reliability of this research high, only complete data on the variables necessary for this study will be used for the sample. Also, a replacement of the possible outliers will be applied by using the median or the mean for categorical and continuous variables respectively.
4.2. Dependent variable
In this research, the dependent variable is conflict resolution, which is determined and measured first by types and degree of violence and later by the length of a conflict. These variables are used since several scholars demonstrated it clearly reflects the existence of conflicts and their resolution (Calavano, 2008; Whiteman & Mamen, 2002; Kemp et al., 2011).
The intensity of violence will be determined first, using the degree of violence, measured on a six-point scale starting from (0) none (i.e. violence free) towards (5) heavy (meaning involving death) and complemented by (1) destruction, (2) occupation of property, (3) roadblock and (4) kidnapping. Second, the types of violence will be measured since the
different types used are known to be a good indicator when it comes to conflict severity (Foster, 2012). The types are measured on a five-point scale and are reaching from low levels of violence like peaceful protest (1), negotiations and court actions (2), towards higher and heavier levels of violence from both sides, like intimidation (3), physical damage (4), kidnapping and violence including death (5).
Hereafter the duration of conflicts will be used to form a clear degree of conflict resolution. However, since in several cases in the database conflicts are still ongoing, the durations of these cases will be impossible to determine, this could result in a low validity of this research plus incomplete conclusions, therefore the sample studied in this research will only involve conflict cases which are ended. In appendix 1, an overview is given on the variance of the 40 different countries in this sample of 269 ended conflict cases world-wide.
Hereafter, the durations of conflicts will be specified by the creation of a binary variable, where a conflict is either displayed as long (1) or short (0). To establish a variety in this variable, the difference in long and short conflicts is determined by the mean of all durations of ended conflicts. This mean in the sample equals 76 months, therefore, in this research conflicts that lasted 76 months or longer are coded as long and conflicts shorter than 76 months are coded as short conflicts. In studying only ended conflicts the author tends to present a clear analysis and excludes incomplete conclusions.
4.3. Independent variable
Country-level institutions, are found to be the independent variable in this research. The level of a country’s institutional strengths will be determined using the following variables from the database: government effectiveness, regulatory quality, political stability and absence of violence, voice accountability and rule of law.
The importance of institutions and the national context to avoid or shorten conflict situations are highlighted multiple times in literature (Ikelegbe, 2005; Lertzman & Vredenburg, 2005; North, 1991). Furthermore, North (1991) argues that a country’s degree of development is closely related to the strength of a countries institutions and can be expressed in variables such as government effectiveness and regulatory quality. These two variables are used in this research and are be complemented by rule of law, control of corruption, political stability and absence of violence, and voice and accountability. The scores of each variable is, for each country, determined based upon the World Governance Indicators (WGI) set by the world bank, which reports individual and aggregate indicators to determine a country's institutional strength (The World Bank, 2017). An average of all 6 scores will be taken to serve, as the countries institutional strength in this research. Each score is measured on a point-scale from 1 to 6 being coded as (6) 90 - 100th percentile; (5) 75 - 90th percentile; (4) 50 - 75th percentile; (3) 25 - 50th percentile; (2) 10 - 25th percentile; and (1) 0-10th percentile. Hereafter, the average score will be continuous from 1 to 6, being 1 the closest to a country’s poor institutional strength, and 6 the closest to a country’s excellent institutional strength.
Moreover, this research argues, based on literature, that country-level institutions are not only influenced by the degree of the level of countries institutions, but can also internally influence other factors relevant for conflict resolution, such as isolation or autonomy of Indigenous communities (Acuña, 2015; Kolk & Lenfant, 2013). As a result, these factors will be implemented and tested in this research as mediating variables, displayed in the conceptual framework.
4.4. Mediating variables
According to Kemp et al. (2011) an indigenous communities’ degree of autonomy is crucial for the improvement of conflict resolution and their findings show a high variety in degrees of autonomy granted to indigenous communities by governments. Therefore, the first mediating variable used in this research will be community autonomy, which measures the amount of recognition granted to a community by their local government. The higher the degree of autonomy the more a communities’ rights (UNDRIP) are strengthened and the higher are the levels of political influence (Hodge, 2014). The variable is displayed based on a five-point scale ranging from (1) no autonomy granted whatsoever towards (5) where almost full recognition of the community is displayed by a government and is thus the highest level of autonomy used in this research. This scale is complemented with (2) cultural recognition, (3) limited recognition of rules, meaning that they can issue some minor rules, like marriages, recognized as valid by government, and (4) partial recognition of rules or right of consultation. The second mediating variable is the degree of isolation which is measured in cultural and geographical distance. The role of isolation in conflict situations has been studied by Bebbington & Bury (2009) who concluded that the higher the degree of isolation the less likely it is for indigenous communities to get their rights (UNDRIP) enforced, thus increasing chance of conflict and decreasing the chance of smooth conflict resolution. Geographical isolation is coded in (1) highly isolated, which stands for community members living within their community and with little contact to the outside world, and (2) where most members live within broader municipalities with a mixed population. Cultural isolation is coded based on a three-point scale namely (1) highly isolated, meaning a community from an isolated culture, own language and not connected to a broader ethnic group, (2) a community which is connected to a broader ethnic group, however still isolated within own group, and (3) when a community is embedded within a broader mainstream ethnic group. The cultural transmission of one human
group to another is inversely related to the distances between those groups (Bentley, Caiado, & Ormerod, 2014; Crema et al., 2014; Premo & Scholnick, 2011). This implies that, the greater the distance the less cultural correlation will be. This is explained from the social evidence and also from the biological behavior as far as the genetic similarity between different peoples is concerned (Wright, 1943). Due to this, isolation will be analyzed as a single variable, as suggested from the theoretical and conceptual frameworks, being the average of the cultural and the geographic factors. The result will be isolation being measure as a continuous variable from 1 to 3, being 1 the highest isolation possible and 4 the less isolated possible.
4.5. Control variables
There are three control variables used to increase the legitimacy of this research. All of the following variables can impact the results of a conflict resolution, so this research seeks to ensure them to not affect the outcomes by control them while evaluating the effects of the dependent and mediator variables.
Firstly, on country-level the Human Development Index (HDI) will be used to eliminate possibly external factors that are influencing the outcomes. The HDI is based on the degree of development of a country, which can negatively influence the chance of arising conflict and conflict resolution (Lertzman & Vredenburg, 2005; Calvano, 2008). It is measured on a 0,000 to 1 scale; the closer to the number 1, the more developed the country.
Secondly, at community-level, this research uses community size as control variable to test whether external factors are influencing the data of our mediating variables. However, the HDI is presumed to be highly related to the independent variable country-level institutions since its measurements have a similar provenance and both reflect similar aspects in terms of a country's ability to be efficient in the management of its allocated resources to the well-being of its inhabitants (Eweje, 2006) Therefore, its relationship with the independent variable must
be quantified and confirmed according to whether or not it can statistically control the model to be proposed. Collinearity between those two will be tested.
Thirdly, the control variable, at firm level, will be MNE experience in a host country, since a firm’s experience and knowledge of the local context enables the MNE to improve interaction with actors within the borders of the country, which in the end can contribute in the process of conflict resolution (Kolk & Lenfant, 2013). MNE experience is displayed in the database by the year in which the firm first started to be active in the country where it now has a conflict and will measured in years.
4.6. Modelling and analysis
The research aims to study the behavior of a concept: conflict resolution, by measuring it in different units. So, for statistical modeling, three dependent variables are proposed, therefore, three analyzes must be performed separately. The multiple mediation regression analysis is used to test the hypothesized relationship between the variables, which are described in the following equations:
!
"= $
%+ $
'()* + ,
"-
"+ .
("(1)
In the equation (2),
!
"represents the dependent variable conflict resolution. This can take three values: violence degree, violence type, and conflict duration (2 = 3). The independent variable country-level institutions, is denoted by
X
. $'()* represents the interaction between * and ! The control variables are represented by -" (2 = 3) and ," indicates their interaction on the outcomes. ..
52 represents the errors for the dependent variables.6
"= 7
%+ 7
'8)* + .
8"(2)
Here Z represents each mediator variable and i represents the number of mediators. In this case, 6:will be representing the community autonomy, and 6; the community isolation. *
represents the independent variable and
.
8" the residual errors for each mediator.According to the four-step method, equation (1) must be tested first to see the direct
effect of the country institutions effects
($
'())
on conflict resolution while controlling the mediators. The second step is to conduct the regression analysis on equation (2). This equation will measure the effects of country-level institutions on both community autonomy and isolation degree.7
'8) will be representing those effects. These interactions will evaluate the second group of hypotheses.!
"= $
%+ $
8)()6
"+ ,
"-
"+ .
("(3)
!
"= $
%+ $
8)()6
"+ $
'()* + ,
"-
"+ .
("(4)
The third step is to conduct the regression analysis on equation (3), having
$
8)(),
as the coefficient representing the effects of the mediators on the conflict resolution dependent variables, thus testing the hypotheses H1b and H1c. The final step is to conduct the multiple regression with the mediators and the independent variables all together. Control variables will be included when evaluating the dependent variable.For the complete model presented in step four, some form of mediation is validated if the effect of mediators ($8)()) remains significant after controlling for *. Full mediation can
be concluded, if * is no longer significant ($'()) when the mediators are controlled. On the other hand, if * is still significant ($'()), the finding supports partial mediation.
As an additional procedure, the product of $8)() and 7'8), proposed by Sobel (1982), will dictate the indirect effect from the independent variable to the dependent variable through the mediator. The total effect given by the sum of the direct effect and the indirect effect will serve to conclude on the H1a hypothesis.
To conclude, . represents the error term which corrects the variations in the value that the dependent and the mediator variable takes that are not captured by the model. The researcher assumes the error terms (
.
(" and.
8") are uncorrelated since this is of high importance when making assumptions for causal inference performing mediation analysis (Imai, Keele & Tingley, 2010). The different combinations of variables used in the regression models to test the hypotheses are illustrated in Table 1.As a final note, the dependent variables, degree of violence and type of violence are categorical while conflict duration is binary. Mediator variables are categorical and the independent and control variables are continuous. Authors such as, Rhemtulla, Brosseau-Liard & Savalei (2012) have found that better results are obtained when using linear methods instead of using categorical alternative methods, this is why categorical variables are run under multiple linear regressions. The conflict duration is a binary variable and therefore a logistic regression is used in order to prove the hypothetical relationships related to the conflict duration (Field, 2009)
