To intervene, or not to intervene?

(1)

To intervene, or not to intervene?

Developing an understanding on the relationship between international intervention and the ethnification of politics

BIANCA BUCEC Master's Thesis

Spring 2021

Word Count: 21459 words

Department of Peace and Conflict Research, Uppsala University

Supervisor: Johan Brosche

(2)

ABSTRACT

This study explores the relationship between international intervention and the ethnification of politics and seeks to explain the degree to which international involvement affects the degree of ethnification of politics. Deriving from explanations that suggest that the ethnification of politics is attributed to the greater interethnic trust facilitated through credible institutions, this study argues that the ethnification of politics is lower in cases where the greater involvement by the international intervention in local institution-building can be observed. Using the method of structure focused comparison, the suggested hypothesis is tested on two cities in Bosnia and Hercegovina – Mostar and Sarajevo. Data was collected through a manual empirical analysis and the tracing of historical institution-building actions by the international community in each city.

The main finding shows relative support for the causal relationship; however, the observed causal mechanism is different to the expected one. This signals that the degree of international involvement in local institution-building processes cannot, in isolation, explain variation on the ethnification of politics. Thus, further research is needed to both identify additional causal factors and build the interaction effects that can explain the observed variation.

Keywords: post-war politics, international intervention, ethnification of politics, Mostar, Sarajevo

(3)

ACKNOWLEDGEMENTS

This thesis would be incomplete without the support, understanding, and compassion given to me by many along the way. Firstly, to my supervisor Johan Brosche – thank you for your patience, your understanding, and your guidance throughout this process. Thank you also for providing clarity and understanding throughout the emotional rollercoaster of this thesis. To Roland Kostić, duboko se zahvaljujem for your mentorship over the last two years, for your endless words of advice and for your confidence in my abilities. An indispensable amount of gratitude goes to Cathinca, Katrina, Louis, Kamil, Mathilda and Lars – even through a pandemic and a distance of over ten thousand kilometres, I’ve had your support, encouragement, and positivity. Thank you for being the best friends throughout this programme. To my JP – without your warmth, your kindness, and your endless understanding, this thesis would not have been possible. Thank you for your attentiveness, your love, and your patience with me throughout this programme, and especially over the last six months. To Pete – thank you for offering me sanctuary, a place to write, space to speak, and an ear to listen. To all my friends and family, wherever ‘here’ is for you – my gratitude, my love, and my sincerest appreciation.

And lastly, most importantly to my incredible parents Sanja and Zoran, to whom I owe everything;

I dedicate this work to you.

(4)

List of Abbreviations

BiH Bosnia and Hercegovina

EU European Union

DPA 1995 Dayton Peace Agreement 1995

FBiH Federation fo Bosnia and Hercegovina

HDZ Hrvatska Demokratska Zajednica (Croatian Democratic Union)

ISM Interim Statute of Mostar

OHR Office of High Representative in Bosnia

SDP Socijalna Demokratska Partija BiH (Social Democratic Party of Bosnia)

US United States

UN United Nations

UK United Kingdom

(7)

Chapter One: Introduction

One of the most significant peace initiatives to mark the end of an ethnic conflict is the General Framework Agreement for Peace in Bosnia and Hercegovina. Also known as the Dayton Agreement 1995, it marks one of the most profound restructures of a post-conflict society. Not only did the agreement bring about the cessation of ethnic violence in former Yugoslavia, it also deeply entrenched the involvement of international actors in the post-conflict reconstruction and ethnic reconciliation efforts of the reunified Bosnia and Hercegovina (BiH) (Kostić 2008). Nearly twenty-five years on, the Dayton Peace Agreement 1995 stands as the fundamental ground upon which modern Bosnia and Hercegovina is founded. To supplement and reinforce the DPA 1995, there have been a plethora of staggered mandates and nation-building initiatives. The relative success and failure of these in Bosnia and Hercegovina is starkly evident in its national political instability; a tension that is debated along ethnonationalist lines. For a conflict that was (by all common accounts) fought along ethnic lines, Bosnia and Hercegovina remains nationally divided.

Recent national election polling shows that the presence, and popularity of ethnonationalist political parties are strong, with political rhetoric reflecting interests that mimic ethnic differentiation (Whitt 2010). This trend suggests at the very least that at the national level of BiH, ethnicity plays a large role in steering and guiding the political agenda of the nation, and that policy is informed by ethnic differentiation. At the sub-national level recent cantonal elections held indicate that the local municipalities share a similar narrative. Cities such as Sarajevo tell a story of growing support for interethnic communication and cooperation, and of a reduction in support for ethnonationalist political parties (Izbori 2020). In contrast, cities such as Mostar indicate a deepening of divisions, resulting in political stalemate, widespread defunding, and a turn to conservative ethnonationalist parties that breed ethnic division (Balkan Insight, 2020).

Given this complex context, the international community has, time and time again, maintained that the nation is considered a ‘successful’ model of both power-sharing and nation- building by an international intervention (Merdzanovic, 2015). It has continued its year-on-year scale back of involvement, receding its role institutionally, politically, and socially (Kostić 2008).

This reveals an empirical puzzle. The paradox between the supposed successes of the actions of the international intervention and the disparity in popularity of ethnonationalist political parties

(8)

demands further investigation. Building on this empirical puzzle, this thesis investigates the following question:

How does international intervention affect the ethnification of politics?

A thorough investigation of previous research on the question asked reveals that very little interest has been paid at testing this question systematically. There are, within existing research, three main themes that emerge studying the presence and popularity of ethnified political parties. The first is the structural argument, which suggests that internal or social structures within society influence the degree of ethnification of politics (Horowitz 1985; Forsberg et al 2008). Secondly, the consequences of conflict argument posits that the ethnification of politics varies due to the effect that violence has on inter-ethnic cooperation (Gurr 1993; Sambanis and Shayo 2013). Lastly, the historical identity argument suggests that the ethnification of politics in a state will be higher in circumstances where ethnic identity is historically entrenched (Fenton 1999; Baumann 2004).

Variation in the degree of ethnification of politics as a result of external factors has rarely been studied.

With the aim to address this deficiency in previous research, and offer an explanation to the empirical puzzle, this study investigates how international intervention affects the ethnification of politics. From a theoretical perspective, it builds on a theoretical model suggested by Bogdandy (2005), Hippler (2005), and Kostić (2007), I investigate the extent to which the presence and degree of involvement by the international intervention has an effect on the degree of ethnification of politics. Precisely, I hypothesise that the greater presence and greater involvement of the international intervention will decrease the ethnification of politics, because it creates credible institutions that foster inter-ethnic trust.

From an empirical perspective, the focus of this study is two comparatively under researched cities in Bosnia and Hercegovina, the cities of Mostar and Sarajevo. Two of the largest cities in the Federation of Bosnia and Hercegovina (FBIH), Sarajevo and Mostar comprise of a significant number of Bosniaks and Bosnian Croats. While many previous studies focus on the relationship between the three major ethnicities, or the relationship between Bosniaks and Bosnian Serbs, the study of the intervention and the presence and popularity of ethnonationalist parties

(9)

within these two ethnicities, and these particular local regions is rare and the value often overlooked. The unit of analysis in this study is also empirically novel, as studies of ethnification of politics have largely had a national-level focus, whereas this study employs a sub-national lens to investigate the question from a local perspective. Both contributions will help to shed light on an under-researched theoretical relationship, and new perspective on the nation-building in Bosnia and Hercegovina.

To test the theoretical argument, I use the method of a structure focused comparison, and take two case studies - the cities of Mostar and Sarajevo. The same set of questions will be asked of the two cases to establish the extent to which the hypothesis exists. An analysis both within, and between the cases will analyse the extent to which variation can be observed. The main findings of this thesis suggests relative support for the causal relationship. It suggests that the presence, and a high degree of international involvement can be correlated with a decrease in the ethnification of politics. While this support exists, causality cannot be fully captured by the study of only presence and degree of involvement in institution-building. Additional observations between the two cases reveal that the addition of the type and priority of the institution-building by the international intervention could add to the causal explanation the theorised between international intervention and the ethnification of politics better.

The thesis proceeds as follows. Chapter Two provides a thorough review of the existing literature written on the linkages between ethnicity and politics. As a result, it provides a definition of ethnification of politics. Chapter Three develops the theoretical framework of this study, by outlining the conceptualisation of international intervention and outlining the hypothesis and causal argument. The research design and relevant case selections are described in Chapter Four.

Chapters Five and Six. Chapter Four serves as a contextual background, while Chapter Five and Chapter Six represent the main empirical discussion of this study. The empirical discussion focuses on investigating the variables within each case. Chapter Seven provides the analysis of the thesis, starting with a brief within-case analysis, before providing the substantial cross-case analysis. It includes a subsection that assesses alternative explanations that could affect the outcome of each case study, while also noting the theoretical and empirical limitations. To conclude, Chapter Eight

(10)

provides an overview of the study conducted, highlighting the strength of the theorised relationship, and outlining the broader implications of this study for further research.

(11)

Chapter Two: Previous Research

This chapter builds an overview of the existing literature that has examined the ethnification of politics. A systematic understanding of how this has been studied is imperative, as it builds a research gap within which this thesis is situated. This chapter should not be taken as an exhaustive summary of literature. Rather, it is an overview of the scholarly work that has contributed to such an intensely studied topic across many branches of social science including international relations, peace and conflict research, sociology, and anthropology.

Within these, there are three theoretical positions that explain what might affect the degree of ethnification of politics in a state: the structural argument, the consequence of conflict argument, and the historical identity perspective. To demonstrate how these debates in literature omit the importance of external intervention on the salience of ethnicity in a nation's political process, this chapter proceeds as follows. Firstly, it offers a discussion of the definition of ethnicity, positioning key debates on ethnic identity. This then leads to a definition of ethnification of politics, signposting three major theories within previous literature. The first theme identified is the structural argument, which suggests that the level of ethnification of politics is directly affected by relative social and institutional structures within a state. The second argument – labelled the consequences of war argument – posits that the ethnification of politics is greater within nations that have borne witness to violent inter-ethnic conflict. The last historical identity perspective argues that ethnicity is more salient in nations when the relationship between ethnic identity and national identity is historically rooted. The chapter then moves to outline the theoretical background that will ground the argument of this study.

Understanding Ethnification of Politics

To begin, ethnic identity is the quality of identifying oneself (or a group of people) by their ethnicity. Conceptually defining ethnicity is inconsistent in literature, indicating that the definition is one of the most exhaustively debated, discussed, and unilaterally complex definitions within scholarship (Forsberg et. al., 2017). According to the Concise Oxford Dictionary of Politics (2009:

117) ethnicity is considered the ‘the quality of belonging to an ethnic group’. Within comparative politics, ethnicity encompasses several identity categories: a shared religion, sect, language, tribe, race, nationality, region, and caste (Chandra, 2012: 51). Cultural anthropology characterises

(12)

ethnicity as an individual process of social classification that includes aspects of race, class, kinship, age, wealth, and gender (Baumann, 2004: 12). According to Hutchinson and Smith (1996) ethnic identification is most often displayed across six markers: a common name, ancestry, memory of the past, culture, symbolic attachment to ‘homeland’, and solidarity among the wider group (p. 6-7). Traditional conceptions of ethnic identity have been shaped by primordial and instrumentalist debates. Contemporary literature conceptualises ethnic identity as a malleable construct that changes and moves according to two fluid concepts; the interest of the community from within, and the external political and institutional strategies (Birnir 2006: 21; Chandra 2012).

Consistent with this previous research, ethnic identity is an ascriptive identity both for the individual, and at a group level (Horowitz 1985; Varshney 2001) that shares some of the categories of identity, and rests on setting apart groups based upon perceptions of cultural or ancestral differentiation (Hutchinson and Smith 1996; Baumann 2004: 13; Chandra 2012: 51).

The expression of ethnic identity within politics can be incredibly persuasive for a number of reasons. Primarily it provides political elites the space to mobilise political support from a targeted group, while it also informs political priority through ethnic representation (Huber and Suryanarayan, 2016: 149). With this in mind, politics and ethnic identity can have a reciprocal relationship, where priorities motivated by ethnic identity informs political action, and paradoxically, where political action can change and challenge ethnic identity (Chandra, 2012).

Building on this previous literature, the theoretical definition of ethnification of politics refers to the national and sub-national ‘ethnification’ of party systems and electoral processes. Within this,

‘ethnification’ refers to the prominence of ethnic identity to a political party, and the unified ethnic support for the party (Fenton 1999, 2010; Huber and Suryanarayan, 2016: 150). Previous literature has contributed to establishing the link between ethnic identity and its prominence in politics through three broad arguments and perspectives: the structural argument, the consequences of conflict argument, and the historical identity perspective.

Structural explanation

The first approach to explaining the degree of ethnification of politics is labled here as the structural explanation. It suggests that internal state or social structures that could influence and affect the level of ethnification of politics in a state. The first of such explanations refers to the

(13)

effect of group polarization on the ethnification of politics. Ethnic polarization is theorised as a form of societal division based on a small number of ethnically homogenous groups. This argument positions ethnic identity at the forefront of politics in societies where there is a higher degree of ethnic polarization (Horowitz, 1985: 12). Forsberg theorises that ethnically polarized societies are more likely to form ethnically delineated party systems (Forsberg, 2008: 285).

Carment and James (2000) note that ethnic dominance will heighten ethnic polarization, inevitably leading to political confrontation and ethnic hostility (p. 197). The second structural argument refers to the system of government and the relative access of ethnic groups to this political system.

They have argued that the combination of a system of government, and relative access to political representation, will affect the degree that ethnic identity is politically represented (see Birdir, 2006). By this rationale, democratic systems and authoritarian systems would vary in their degrees of ethnification of politics. As a result, the structural argument emphasises the importance of reinforcing social and institutional ethnic representation and ethnic plurality in order to decrease the ethnification of politics.

Consequences of Conflict

The second approach posits that violent ethnic conflict, and the consequences of this conflict, are pivotal in explaining the variation in the ethnification of politics. It contends that this is due to the negative impact that violent conflict has on inter-ethnic cooperation. According to this approach, violent conflict that exhibits systematic discriminatory treatment against, or in the benefit of one ethnic group will result in the expression of greater political mobilisation of the interests of ethnic groups (Gurr, 1993). This is supported by the work of Sambanis and Shayo (2013), who argue that concern and care for an individuals’ ethnic group is greater because of the individual or group experience of intense ethnic conflict, and as a result, is predicted to lead to a greater desire for political representation (p. 319). In brief summation, the consequences of conflict argument focuses on the studying the variation of ethnification of politics as a result of consequences; either through its effect on eroding inter-group trust, or consolidating intraethnic solidarity.

Historical Identity perspectives

A review of previous research would be incomplete without acknowledging the historical identity perspective. I call this a perspective, as opposed to a theoretical argument, as there is a notable

(14)

absence in previous research of systematically testing this relationship. This perspective posits that ethnic identity is indivisible from individual identity, stressing that this is because it is fostered within family, within community, and within ancestral roots (Geertz, 1963; Fenton, 2010: 73).

Reflecting key attributes of the Primordial position on the nature of ethnicity, ethnic identity, from this perspective, fosters a sense of unity and community among its members because of this shared ancestral identity (Fenton, 1999). According to Fenton (1999: 173), the historical transformation of national systems has developed alongside ethnic identity and has come to be defined by a shared ancestry, religion, and/or language. Therefore, the historical identity perspective suggests that the ethnification of politics would occur more in cases where ethnic identity is historically indivisible.

This perspective is problematic for a few reasons. This makes it difficult to distinguish whether variation in the ethnification of politics of a state can be explained by the historical argument perspective, particularly when it is so intrinsically linked to the very political system it seeks to explain.

Situating the study

As this short overview indicates, the study of the ethnification of politics has for long been a study of linking ethnic identity to politics. Only recently has study developed in investigating causal relationships that could affect the variation in the ethnification of politics. Previous research has largely focused on historical, structural, and conflict-driven arguments that explain variation on the ethnification of politics. All three themes reflect an assumption that ethnic identity is hardened, however there is very little account for whether this ethnic identity transforms. I argue that this view of ethnic identity is static and has limited the development of theory within the study of ethnification of politics. Mainstream political science, international relations and peace and conflict research does not account for the ability of ethnic identity to transform, thereby taking ethnic identity to be a unitary, static frame of analysis (Chandra, 2012). The development of theory to explain variation in ethnification of politics is consistent with Sambanis and Shayo (2013) and Huber and Suryanarayan (2016). Both reveal a renewed interest in observing, and measuring, the process of ethnification of politics from more nuanced perspectives, particularly noting the transformative nature of ethnic identity.

(15)

The study field of ethnification of politics, and in particular relevant variables that affect the degree to which this is both present, and measured, remains underdeveloped¹. To the authors knowledge, there has been a notable absence of developing and testing the variation of ethnification of politics over a longer term, particularly within the post-conflict context. There has also been no notable systematic study of external, or institutional factors within this post-conflict context that have been tested for their role in affecting the ethnification of politics. Previous studies have established that the level of ethnification of politics in a state is reflective of the salience of ethnic identity to the members of the ethnic group. In the following section, I argue that the institutional structures that influence ethnic identity will affect the ethnification of politics. Additionally, the study of external influence on the institutional processes of a state in conflict resolution will develop a long-term understanding of how ethnification of politics can change, develop, diminish or strengthen over time.

The ethnification of politics in the aftermath of ethnic violence

Ethnic identity rests on identification based on difference, and thereby inherently creates an ‘in’

group and ‘out’ group. In the aftermath of violent ethnic conflict, ethnic identity is incredibly salient, and the ‘in’ and ‘out’ group environment reinforces the perception of threat and competition by members (Kostić 2007: 23; Hewstone 2008: 212). Across all disciplines, it is evident that ethnic identity has become a significant factor of political mobilisation, particularly in the aftermath of ethnic conflict. Ethnic leaders are, according to theory, increasingly more involved in the post-conflict process (Pfaff-Czarnecka 2005: 28-29). Within a context of distrust and division, political mobilisation on the base of ethnic identity can pose significant risks to the cohesiveness of a state.

To adequately respond to the dangers of intra-ethnic political stalemate, policy must respect and understand the importance of the existence and meaning of ethnic identities (Wieters 2006).

The greater the feeling of exclusion, the more profoundly detrimental the effect on national constructions of shared culture, cohesion, and even nationality (Fenton 1999: 183). According to social psychology research, building trust and cooperation socially and institutionally is incredibly important for reducing suspicion and hostility among ethnic groups (O’Loughlin 2010: 27;

1 This has been reiterated by Forsberg et. al. (2017) who note that the study of the relationship between ethnic identity, conflict and politics is severely underdeveloped

(16)

Varshney, 2001). This is particularly relevant in the aftermath of ethnic violence, where identities can be hardened. Trust is a means by which the tension between vulnerability in uncertainty is presented; it relies on an individual overcoming the relative uncertainty regarding the motives, intentions, and agendas of others (Hewstone, 2008: 211). According to Whitt (2010) credible institutions are an integral factor in the (re)building and reconciling of trust in this context.

Hewstone (2008) suggests that institutional cooperation is integral in reinforcing a positive intergroup perception. Therefore, the encouragement and establishment of institutions conducive to (re)building trust has a number of positive outcomes, including increased information exchanges, cooperation, and reconciliation (Hewstone, 2008: 212).

Building on this research, I argue that facilitating trust through credible institutions is a condition that leads to a lower degree of the ethnification of politics. Institutions, on the other hand, do not exist in isolation; they are created and shaped by political actors. Due to this, I take credible institutions and trust to be two interconnected conditions that explain variation in the degree of ethnification of politics. This, then, highlights a two key questions: firstly, if credible institutions and trust are two conditions for why the ethnification of politics varies, then what induces credible institutions? And second, if credible institutions are constructed, what actor is responsible for the construction? The following chapter answers these two questions.

(17)

Chapter Three: Theory

As stated in Chapter Two, previous studies of the ethnification of politics have notably lacked in accounting for the effect of international intervention in two main ways. Firstly, they fail to investigate changes in the ethnification of politics in correlation with the actions undertaken by an international actor. Secondly, they fail to conduct a systematic study in variation of the ethnification of politics within a post-conflict context over a longer time-period.

To investigate this relationship, the following theoretical framework is developed. Derived from theory on international intervention, I argue that institution-building by an international intervention fosters trust among groups. This is positioned within theory that postulates that the peacebuilding agenda fosters an environment of social and institutional cooperation, reducing uncertainty and threat. This should, by this theory, lead to a lower degree of ethnification of politics. To build this argument, the chapter proceeds in the following parts. Firstly, it conceptualises international intervention by providing a theoretical definition. From this, it provides a typology of actions and processes that constitute the international intervention for this study. Lastly, it outlines the theoretical relationship drawn between international intervention and the ethnification of politics, providing the casual mechanism and argument that guides this study.

Conceptualising international intervention

To begin, an international intervention is, fundamentally, a strategy employed by the international community to achieve a purpose. This thesis’ theoretical conceptualisation of international intervention exists beyond the bounds of what classical conflict resolution literature would deem

‘third-party intervention’ (Ramsbotham et. al., 2016). Where previously viewed as strictly as a strategy employed for the cessation of violence, and as an enabler of resolution by way of bringing parties to the negotiation table, a contemporary approach views international intervention in a much broader scope, both by timing and by nature. Contemporarily, the purposes of an international intervention extend into the peacebuilding phase of the conflict resolution process and encompasses activities that fall within the bounds of peacebuilding, state-building, and nation- building.

To understand this contemporary conceptualisation is to understand how the liberal peace theory informs international intervention. The contemporary international intervention was bolstered in

(18)

the wake of the Cold War and the emergence of a multi-lateral system of competition for power, resources, and identity. Liberal Peace has since been exemplified in half a century of policy practice and development (Kostić 2007: 36). The application of international intervention evolved beyond traditional peacekeeping, and began to encompass new models building peace, guided by the principles of democracy, free elections, and market liberalisation (Ibid., 39). Based on this theory, an international intervention is defined by an agenda of peacebuilding that consists of a singular, or dual priority of, nation building, and state building. Each priority engages liberal principles of social, political, and economic organisation as a strategy of conflict resolution (Paris 1997; Ramsbotham et. al. 2016: 236). To draw out one sole purpose of an international intervention would require a systematic comparison of all past interventions. However, particularly when used in the aftermath of ethnic conflict, the contemporary international intervention, grounded within this liberal peace theory, is driven by the purpose of the creating and fostering a space of cooperation and compromise between groups that is founded on democratic principles (Carment and James 2000; Whitt 2013)

If contemporary international intervention is a strategy, theory suggests it is largely constituted by two processes. The peacebuilding agenda refers to the “project of overcoming structural and cultural violence…in conjunction with peacemaking between conflict parties… and peacekeeping” (Ramsbotham et. at., 2016: 237). Peacebuilding is guided by the processes of nation- and state-building. Within this study, they are two distinct, but interrelated processes integral to the peacebuilding process. State-building is a political process through which political and administrative institutions that enable the state to function efficiently are created. It emphasises the importance of the state’s ability to govern independently, provide public goods to its citizens and maintain a monopoly on violence through the governance of police and military forces (Bogdandy et al.,2005: 584; Kostić 2007: 40; Ramsbotham et al., 2016: 238). Where state-building focuses on the structural factors that enable a state to function efficiently, nation-building is a strategy utilised by external actors to promote the collective identity formation with the purpose of legitimising a constructed state authority (Bogdandy et. al 2005: 586; Kostić 2007). Evidently nation-building and state-building processes are mutually reinforcing, and crucial for the success of peacebuilding in a nation. Hippler (2005) clarifies three core elements of nation-building; the persistence of an integrative ideology; the cohesive social integration of all groups in society; and

(19)

the development of a functional state apparatus that controls its relevant territory in all aspects (Ibid., 8-9).

Derived from these arguments by Bogdandy et al. (2005), Hippler (2005), and Kostić (2007), the nation-building and state-building actions and processes are visualised in Table 1.

The Peacebuilding Agenda

Nation- and State-Building Priorities and Measures Nation-Building

Priorities Examples of measures Integrative Ideology

- The formation of a national identity that supersedes other identification markers such as ethnoreligious identity, clan, or tribe

Social Integration

- Promoting a sense of ‘common nation’ founded in social reality - Social reality reflects a ‘common nation’ where communication

between groups is rife

- Social infrastructure such as media and transport are key variables for achieving this priority, as they establish a national political and cultural discourse

State-building Functional State

Apparatus - Creation of an effective financial system - Creation of police forces

- Creation of a representative legal system

- Creation of an administrative system that is trusted and accepted by all

- Creation and promotion of functional political system

Table 1: The Peacebuilding Agenda (based on Hippler, 2005: 7-10, Bogdandy,et. al 2005:584-586; Kostić, 2007:

39-40)

This visual representation may suggest a linear process, with measures and priorities implemented in an ordered fashion. Practice contradicts this claim; it reflects a process that is overlapping, muddled, and often broken down into sub-national prioritisations based upon needs and responses.

To better visualise the complexity of relationship, I suggest the following:

(20)

Table 2: Conceptualising the processes of International Intervention

Drawing on this theory, I argue that contemporary international intervention thrives in the peacebuilding phase of the conflict resolution process. It is conceptualised as a peacebuilding

The Peacebuilding Agenda

Nation- and State-Building Priorities and Measures

Nation-Building

Priorities Examples of measures Integrative

Ideology - The formation of a national identity that supersedes other identification markers such as ethnoreligious identity, clan, or tribe

Social

Integration - Promoting a sense of ‘common nation’ founded in social reality - Social reality reflects a ‘common

nation’ where communication between groups is rife

- Social infrastructure such as media and transport are key variables for achieving this priority, as they establish a national political and cultural discourse

State-building Functional State Apparatus

- Creation of an effective financial system

- Creation of police forces

- Creation of a representative legal system

- Creation of an administrative system that is trusted and accepted by all

- Creation and promotion of functional political system

Achieved by

Exhibits

(21)

strategy that is composed of activities that fit within the agendas of peacebuilding, state-building, and nation-building. It is a reciprocal process between intention and action and is exhibited most in the institutions within which the agenda plays out. Institutions are, by this theoretical reason, an international intervention assumes the abandonment of old structures and institutions in favour of redistributing resources and power through the creation of new institutions (Hippler, 2005: 12). I define institutions to be both a physical structure, and a specific collection of practices and rules that define appropriate behaviour for specific groups (according to Finnemore and Sikkink 1998:

891). Table 2 reaffirms this definition, showing that institutions are both the forum within which an international intervention implements actions, and the space within which an international intervention exhibits the objectives. This highlights an explicit, and implicit dimension of international intervention; the institution, and the actors (respectively). I argue that investigating the presence and changes to actors and institutions within international intervention are fundamental to understanding any effect that international intervention (and its relevant processes) may have. It is important to note that while international interventions guided by this liberal peace strategy have also undergone several transformations and critiques², they remain the most salient

‘form’ of international intervention contemporarily.

3.4 Causal Mechanism and Argument

In accordance with previous research, this thesis makes the following argument. It starts from the position that the degree of international intervention will affect the degree of ethnification of politics in the immediate, and long-term aftermath of ethnic war. More precisely, I argue that it is the degree to which actors are indirectly or directly engaged in the facilitation, implementation, and endorsement of credible institutions that determines the degree to which ethnification of politics is rife. I argue that this causal mechanism applies because greater involvement by an international intervention, particularly in institution-building, raises trust not only within ethnic groups, but among ethnic groups, as it fosters cooperation and compromise. It is expected for this to significantly lower the need for the political elite to mobilise ethnic identity, thereby decreasing the rate of the ethnification of politics.

2 For a comprehensive discussion of the transformation of the liberal peace strategy, see Kostić 2007, and Erickkson and Kostić, 2013; for a discussion on the success of the liberal peace strategy of international intervention, see Wieters 2006

(22)

Following this theoretical reasoning, the causal mechanism is as follows:

Figure 1: Causal Mechanism

From this causal mechanism and theoretical proposition, I draw one core hypothesis that will be empirically tested.

H1: the greater the degree of international intervention, the lower the ethnification of politics

Unfortunately, theory provides no clear pathway to measuring the application of an international intervention by degrees. This is largely because theory acknowledges the contextual fluidity of international interventions; each violent conflict demands a unique, contextually relevant response by the international community, that encompasses a variety of actions and processes, applied at a varied length of time. Regardless, this thesis endeavours to test the hypothesis in a method that measures the presence and relative degree of involvement of actors within the international intervention by examining this twofold. Firstly, through the presence/absence of actors (indicating an intensification, or drawback of local and sub-national support), and secondly, by looking at the actions of international actors in the international intervention. A change to either of these is measured in relative terms, taking a contextual temporal comparison to determine relative change.

international intervention

credible institutions greater trust

degree of ethnification of politics

(23)

I distinguish that this hypothesised relationship applies only under certain scope conditions.

Firstly, it applies only in cases where there is a notable presence of the international community in the post-conflict nation-building process of a state. I recognise that it is in many cases difficult to define the start and end date of such a process, because oftentimes, international withdrawal or departure does not necessarily indicate the conclusion of the international intervention. Secondly, within this study it applies to only cases where ethnicity is an important and identifiable aspect of both political, and civil life, and is an explanatory variable for the cause of conflict. This is a condition that carries with it a plethora of contestation. While this thesis does not have the scope to discuss the different applications of ethnicity within the causes of war, it recognises that in some conflicts it is more salient than in others. Lastly, it should be noted that there is contention in whether this is the primary explanatory variable for the degree of ethnification of politics in a state.

Rather, I argue that with the prevalent use of international intervention within contemporary peacebuilding, the impact of studying its effects on sub-national and national level processes is important.

(24)

Chapter Four: Research Design

The following chapter outlines the method, structure, operationalisation and research design of the empirical study to follow. It proceeds in three parts. Firstly, it will outline the chosen method, while providing the structure of the research. Secondly, it will justify and motivate the chosen case studies, outlining their similarities, differences, and relevance to this type of scientific study.

Thirdly, it will operationalise the variables, and provide indicators for which measurements will be taken. Lastly, it will provide the research questions and designs that will be used within the following two comparative case studies.

Method

While there are a multitude of different types of research designs that could be employed to test the hypothesis, this thesis chooses to use a qualitative study method called a structure focused comparison. Understanding that the purpose of scientific study is the production of systematic and cumulative knowledge, I believe the chosen method will garner conclusions that will satisfy these conditions. A qualitative research design has been chosen as opposed to a quantitative study, as it will allow the thesis to deeply explore the nuances and relationships within a single case study, and test the causal mechanism between the cases accounting for contextual confounders.

A structure focused comparison (SFC) has features that distinguish it from other comparative case study methods. A SFC is structured, in that the same research objective guides, and research questions ask the same of all cases; focused, in that has a clear and applicable research objective and theoretical framework; and comparative, in that it has a comparative angle either across time, or across units of analysis (George and Bennett, 2005: 70). These two features allow a SFC study to test the casual mechanism in a structured manner, allowing for scientific replicability, and generalisability. Moreover, the focused nature of the study will draw conclusions that will allow for both the data and theoretical propositions to be cumulative and scientifically applicable. In particular reference to the question of this thesis, there are two main reasons for the method chosen.

As it is interested in understanding the variation in the ethnification of politics within various contexts, a comparison between two or more cases is necessary to build this understanding.

Secondly, as the study concerns itself with measuring degrees of international intervention and conducting an observational study to measure the degree of ethnification of politics, the absence

(25)

of data lowers the applicability of any large-n quantitative study. With that in mind, an in-depth qualitative study will not only test the relationship but explore the contextual and social complexities of the studied concepts in more thorough detail.

Case Study Selection and Motivation

Any structure focused comparison that investigates the viability of a casual mechanism requires an explanation for which cases were chosen, and what these chosen cases are cases of. Case selection is especially important in qualitative studies with a limited number of studies, as random sampling is abandoned in favour of a deliberate choice of cases based upon the research question or empirical puzzle of interest (Seawright and Gerring 2008). This process does not come without significant risk of selection bias – the risk that the conclusions are based upon inferences that suffer from systematic error (Collier and Mahoney 1996: 59). In order to avoid unintended selection bias, this study choses its cases based upon two factors. Firstly, it has chosen cases based upon an observed variation on the dependent variable. This means that both sub-national areas of interest have relatively high, or relatively low, levels of ethnification of politics. Secondly, this study has chosen two cases that are as similar as possible except on the independent variable, in order to be able to control for alternative explanations. On the basis of these factors, I have chosen two cases following the most-similar case study design.

The two cases I have chosen are two cities in the Federation of Bosnia and Hercegovina – Mostar and Sarajevo – and the measurement of the ethnification of politics for political parties associated with two ethnicities – Bosniaks and Bosnian Croats. These two cities provide a good basis for cross-case comparison, as they are similar across many values that can be kept constant: their economic size, multi-ethnic character, and experiences of the destructiveness of war. They do, however, differ across two values: the proportion of ethnic demographics, and their measurement of ethnification of politics. The ethnic composition of Mostar and Sarajevo differs greatly, with Mostar representing an almost even split between Bosniaks and Bosnian Croats, while over 75%

of inhabitants in Sarajevo identify as Bosniak (Census of Bosnia and Hercegovina, 2013). I recognise that ethnic demographics may influence the degree of the ethnification of politics.

Unfortunately, this is not a confounder that I was able to control for, and the choice of two imperfect case studies could cast doubt on the validity of the findings. Regardless, I argue that

(26)

examining the relationship between the degree of international intervention and the ethnification of politics is still a worthwhile endeavour between the cases, as previous studies cannot fully capture the variation of ethnification of politics in these cases.

The empirical novelty of this study is threefold. While the DPA 1995 was a nationally reformative agreement, any peacebuilding activity carried out by the international intervention used to implement and reinforce the agreement has an indispensable local component. It relies on local cooperation to enforce the action, and local conditions to signal the relative success or failure of this action. Broadly, studies of Bosnia and Hercegovina predominantly focus on national-level processes of international intervention, and only a few delve deeper at understanding the local effects of these actions. For this reason, the investigation of sub-national electoral trends is an empirical novelty in studies of Bosnia and Hercegovina, as builds a more nuanced understandings from a different frame of analysis. More broadly, previous research conducted on either ethnonationalism in politics, or on the involvement of the international community in nation- building projects, the comparative analysis is overwhelmingly focused on the comparison of nations. To the authors knowledge, there are only a few studies that trace mandates, actions and processes at the sub-national level. There are also very few studies that provide a systematic analysis of the ethnification of politics. Such a preponderance of national-level studies has meant that the interesting observations that could be gathered from a more micro-level study has often been overlooked

In addition, the choice of these cities, and these political parties, comes from the novelty of studying the relationship between two ethnic groups that has not been as readily tested. The empirical variation that has been observed has shown variation on the dependent variable, and this motivates the study to question the extent to which the chosen independent variable has effect.

With this in mind, this study uses this sub-national unit of analysis in two ways. Firstly, it focuses on two cities within the Federation of Bosnia and Hercegovina – Mostar and Sarajevo. Secondly, it examines the dynamics of a nationally ratified international intervention project at the more micro-level, and how the mandates, processes and actions are translated and affect the local area of study.

Accounting for alternate, mainstream explanations

(27)

I recognise that this theoretical explanation and chosen case study is studied within the context of various other, more common explanations for why the variation could exist on the dependent variable. It is important to stress that my argument lies precisely within this argument, rather than against or alongside this argument. There has been an exhaustive number of studies³ that have been conducted investigating the causes of, and effects of, the Yugoslav War 1991 on the nation of Bosnia and Hercegovina. Often, the discussion focuses on the institutionalisation of ethnic differences within the Dayton Agreement, particularly as the common justification for the waging of war was the historically entrenched ethnic hatred. This thesis respects and understands this argument. It acknowledges that the Dayton Agreement has had a profound effect on the ability of Bosnia and Hercegovina to transition into a form of governance that does not heavily predicate itself on ethnic difference. It also recognises that there are numerous studies that debate why the Yugoslav War 1991 was waged in the first place. Despite all these arguments, I believe that there is novelty and importance at investigating the research question within the context of such a profound peace agreement, and within the historical context of these ethnic differences.

Structure of Empirical Analysis

The empirical analysis will follow a very systematic, general structure. The first empirical chapter provides a historical and contextual overview of Bosnia and Hercegovina. This is incredibly important, as it contextually situates the analysis, and narrows the scope of study. It also provides a thorough review of the major political parties in the Federation of Bosnia and Hercegovina (FBiH) based on data extracted from the Electoral Commission of Bosnia and Hercegovina (Izbori BiH, 2020). This data is used to present a classification system of ‘ethnification’, one that will determine the degree of the ethnification of politics in Sarajevo and Mostar. The second two chapters are a focused in-depth study of the two chosen cities – Mostar and Sarajevo. Within these in-depth chapters, the same structure will reflect the questions that will be asked in order to present the data in a concise and consistent manner.

With any structure focused comparison, the fundamental task is outlining which questions will be asked, and which data will be used to measure each variable. Explaining the use of this data as part

3 This thesis has neither the scope to investigate, nor the space to list the main arguments of all these studies. For an overview of literature that illustrates this argument, see: Gibbs 2009

(28)

of the measurement is a crucial component of justifying the use of this method, as without it, the comparability and systematic investigation would be difficult. Moreover, it would decrease any significance of the results (George and Bennett 2005: 250). Therefore, I have chosen several questions that will be asked of each case study. The questions themselves are general in nature to allow comparability but focused enough that they will encapsulate the complexities of an in-depth study of each case study. The answers to these questions will be presented in each individual case study and will then be investigated further in the comparative analysis chapter.

These answers will be presented in a manner that follows the four-year interval between local and cantonal elections in Bosnia and Hercegovina. It begins by presenting the data for each indicator as of 1997, and then following by the years 2000, 2004, 2008, 2012, 2016 and 2020. These seven divisions follow the results of each local election held in Bosnia since the signing and ratification of the Dayton Peace Agreement. The measurement of the data at these intervals will also allow the tracing of significant local level dialogues, mandates and processes undertaken, implemented or ceased by the international community within the chosen sub-national unit of analysis.

Measurement of Data

The aim of this thesis is build causation between the degree of international intervention and the ethnification of politics. It measures the degree of international intervention by the relative change in the facilitation of dialogue, and institution-building programs of the international intervention across a number of actors. It gathers the salience of actors from the case studies, as it recognises that every international intervention hosts different actors from different sources. It also accounts for major changes in the involvement of international institutions such as the United Nations and the European union, and all their sub-structures. This does not mean that it is an exhaustive overview of the actions, processes, and priorities of all international actors within the purview of the international intervention in Sarajevo and Mostar.

In contrast, it measures the ethnification of politics by conducting a manual empirical analysis of election data extracted from the electoral archive of the Central Electoral Authority of Bosnia and Hercegovina (Izbori BiH, 2021). I constructed this empirical analysis in the following way. Firstly, it combines all of the data across all seven elections, in all four municipal districts in Sarajevo and in all the variations of the city of Mostar, to build a year by year understanding of the average

(29)

voting patterns across Sarajevo and Mostar. It classifies ‘Sarajevo’ to be an average yearly measurement of party support across the four municipalities of Sarajevo – Novo Sarajevo, Novi Grad, Stari Grad and Centar Sarajevo. These four municipalities are relevant as candidate and party popularity from these four municipalities represent the political construction of the Sarajevo City Council (Gradsko Vijece Sarajevo). Similarly, the same average is drawn across the municipalities of Mostar, to achieve the same. The aggregate for each year is based upon how the city of Mostar has voted across all municipalities, regardless of the political constitution of the City Council. I measure the average vote by party, by year by taking the sum of all votes cast for that party across all municipalities, and then divide this by the sum of all votes cast in that election.

After converting this statistic into a percentage, I then sort the results based upon the percentage of total vote per party, indicating the relevant popularity of each political party in that cycle. I extract the top 5 parties (which by average, constitute over 80% of the local vote in both Mostar and Bosnia).

This choice affects the validity of data that I extract, as it leaves a proportion of data unaccounted for. This is represented in the respective tabulations in Chapter 6 and Chapter 7 as

‘other’. With this choice, I do not suggest that those political parties do not follow a classification of ethnified/non-ethnified, nor do suggest that their measurement and classification would yield a more accurate measurement of the ethnification of politics. The intention of this empirical analysis is to gauge the relative change in popularity over time which requires a mere approximation of the data. As I use over 80% of the total votes each year in my analysis, I argue that the conclusions drawn can be made with relative certainty. A more accurate statistical measurement would require an improved, two step research design: a more systematic classification of all political parties at each election cycle, and a subsequent quantitative analysis for which this thesis did not have scope.

According to a classification system built in Chapter 5, I classify each political party, for each year, by the ethnified classification. This allows me to construct a tabulation of each other party classifications in Chapter 6, Chapter 7 and Chapter 8. A closer look at the cross-party tabulation (and relevant colour key) for Sarajevo can be found in Appendix 1, and for Mostar in Appendix 2 (with a minimized version provided in the respective case studies). The full data-set sorted by popularity can be found in Appendix 3.

(30)

Operationalisation and Research Questions

The ability to test the hypothesis rests on translating theoretical definitions of international intervention and ethnification of politics into operational definitions. These operationalisations will then be used to empirically capture the concepts. The following sections outlines the referent operationalisations, while also establishing a set of indicators of each variable from which research questions will be developed. These research questions will furthermore guide the empirical analysis in further chapters.

Ethnification of politics

The ethnification of politics variable will be measured by the study of electoral behaviour over a twenty-five year time frame. The measurement of voting behaviour is only one way of measuring the ethnification of politics, but I stress it is an important one for two reasons. One the one hand, it provides a crucial measure of the relevance of a political party to sub-national ethnic identities, where voters are expected to vote with a party they feel is best representative of their ethnic identity (Chandra, 2012; Huber and Suryanarayan, 2016:153). On the other hand, it provides an indication of the salience of ethnic identity within the region itself; where the greater support for ethnified political parties reflects the individual importance of one’s own ethnic identity. The time period of twenty-five years has been reflected in other studies as one necessary in order to evaluate the impact of nation-building projects in a post-war state (Whitt, 2010). This is because twenty-five years captures the sentiment immediately post-war, but after a significant time-period of international involvement increasing, or decreasing, in the reconstruction and nation-building process. As mentioned, the data used to measure the voting behaviour that constitutes the ethnification of politics in a local region is the Izbori BiH (2020). This is a comprehensive overview of election statistics published from 1997 at four-year intervals. Election polling results are the most reliable indicator for measuring both the presence and the temporal increase/decrease of support for a political party. Using this measurement will also make the measurement of variation in the dependent variable straightforward.

The analysis itself will be guided by the established definition developed in the Previous Research and Theory chapter and is operationalised to measure a) the presence or absence of ethnonationalist political parties and b) the relative popularity of these parties. While it excludes a measurement of

(31)

‘minor’ parties, i.e. parties that do not feature as strong contenders in the election year, it will look to see the increase/decrease of popularity for major parties over time, and will investigate whether a decrease in popularity for one party indicates a shift to another.

Based on these two indicators, the following questions will be asked of each case in order to measure the degree of ethnification of politics:

• What are the ethnonationalist political parties present at each election?

o What are the Bosniak political parties?

o What are the Bosnian Croat political parties?

• What was the level of support of each of these parties between 1995 and 2020?

• To what extent does popularity for ethnified parties change over time?

International Intervention

Indicators for measuring the degree of involvement in international intervention are guided by the two dimensions of concern to this study; the actors, and the institutions. For each dimension, a set of indicators has been developed, which will allow the measurement of relevant actors, and institutions. The measurement of the degree will come from a within-case temporal comparison, where greater involvement is operationalised by a higher presence of actors and higher volume of institution-building mandates and actions. The measurements are based on data gathered from a variety of sources, including official transcripts, statistics, reports, and previous research. It recognises the impossibility of measuring every policy, mandate, action or process undertaken by one or all international actors within Bosnia at any given one time. That is why there will be a particular focus on mapping three core actors; the OHR, the UN (and its sub-sections) and the EU (and its subsections).

(32)

Indicators for Independent Variable

Research Question Indicators:

Has the presence and involvement of international actors changed over time?

- Activity and presence of key international actors over the 25 year period

To what extent are international actors involved with local institutions?

- Facilitation of Meetings

- Developed mandates/decisions/memos in conjunction with local leadership Policy briefs

- Media reports

- External mechanisms of cooperation

To what extent does the involvement of international actors in local institutions change?

- Policy briefs - Media reports

- External mechanisms of cooperation

Table 3: Indicators for Independent Variable

(33)

Chapter Five: Bosnia and Hercegovina; an Overview

In this chapter, a brief contextual overview is provided of Bosnia and Hercegovina (BiH). It discusses the ethnic composition of pre-war and post-war Bosnia, and provides a brief, yet necessary summary of the impact of the Yugoslav War 1991. Thereafter, the case studies in Chapter Six and Chapter Seven will present an investigation of the dependent variable (ethnification of politics), the independent variable (international intervention) and causally map the relationship between the two.

Background: the Yugoslav War, Dayton, and ethnic identity

Historically, Bosnia and Hercegovina⁴ was, and remains to be, one of the most ethnically diverse states in the Balkan region (Whitt, 2010: 275; Glenny, 2012: 688). Until the early 1990s, Bosnia was part of Yugoslavia; a multi-ethnic state with no majority ethnic group (Kaufmann, 2013: 273).

Previous work is exhaustive in debating the various narratives about the reasoning for the Yugoslav War⁵. Some say it started because of historic ethnic tensions, others because of the power vacuum that Tito’s death left, and even some that attribute it to the declining Yugoslav economy in the last 1980s.⁶ Regardless, there is one commonality that emerges out of both fact, and debate;

the importance of ethnic identity to the political process of Bosnia and Hercegovina.

From the war, Bosnia emerged as a divided state. After the signing of the General Framework Agreement for Peace in Bosnia and Hercegovina 1995 (more commonly known as the Dayton Agreement 1995), it was the only republic out of the former Yugoslav states that did not emerge with one ethnic majority government (Gibbs 2011, 68). Instead, it emerged with a political system that was intended to account for, and recognise, the existence and prevalence of three main ethnic groups – the Bosniaks, the Bosnian Croats and the Bosnian Serbs. Within this system, it would feature two ethno-nationalist sub-state entities (Republika Srpska and the Federation of Bosnia and Hercegovina), 10 cantons within the Federation of Bosnia, and 142 local municipalities across both entities. This is best visualised by Figure 2.

4 Thereafter referred to as either BiH or Bosnia

5 Gibbs 2009

To intervene, or not to intervene?