REBEL WHISPERS
AN ISSUE-BASED APPROACH TO PEACE AGREEMENT SUCCESS AND CIVIL WAR RESOLUTION
TANIA M. ESTRADA CORPEÑO Master's Thesis
Spring 2020
Department of Peace and Conflict Research, Uppsala University Supervisor: Ralph Sundberg
Word count: 22,198
ii ABSTRACT
While issues remain under-researched, peace agreement success has been linked primarily to the proper treatment of the parties’ security-related concerns. This study explores why some peace agreements succeed while others fail by using an issue-based approach arguing that issues are an expression of underlying grievances, which have caused the rebel groups to engage in armed conflict. Therefore, peace agreements that do not address the issues, which reflect grievances, will fail.
I tested the hypothesis and the proposed theoretical relationship through the structured focused comparison of three peace agreements: The Lomé Peace Agreement, the Accra Peace Agreement and the Final Agreement National Government – Popular Liberation Army. The method employed in this study comprised first, determining the salience the rebel groups assigned to their issues -for which it was necessary to create a measure for issue salience- and second, examining the peace agreement’s provisions to determine if the rebel group’s issues were addressed.
The results show that peace agreements that included the salient issues of the groups failed; however, peace agreements that did not include them, succeeded. Hence, the findings suggest that the inclusion of the rebel group’s issues in the peace agreement cannot account for the agreement’s success or failure.
Key words: Issues, Peace Agreements, Civil war, Grievances, Sierra Leone, Liberia, Colombia
iii ACKNOWLEDGEMENTS
First, I would like to express my gratitude to Rotary International for granting me a Peace Fellowship that has allowed me to pursue this degree. My life has forever been changed because of it. This thesis would not have been possible without the extraordinary advice and support from my advisor, Ralph Sundberg to whom I would like to express my great appreciation. Special thanks are in order to everyone involved in the WTF project, especially Johan Brosché and Ralph, again, who kindly shared ideas with me and granted me access to the project’s dataset. Advice and encouragement given by Sara Lindberg- Bromley have been important as well for completing this study.
I also thank my family in Guatemala whose loving support throughout this process has nurtured me, especially that from my mom Gladys, my brother Cristian and my cousin Karla, who have always believed in me. I am particularly grateful to my friends Bec Dawson and Paulina Lindén for their uplifting words and for providing solutions when all I could see were problems.
Finally, to my husband Isidro whose constant support has been crucial… amor, we made it.
iv TABLE OF CONTENTS
List of Abbreviations………... vi
List of Figures and Tables………. vii
I. INTRODUCTION……….. 1
II. CONCEPTS, PREVIOUS RESEARCH AND THEORY………. 5
Defining the main concepts……… 5
Why would issues matter? A look into previous research………. 8
Bridging issues and peace agreement success: Theoretical argument……… 14
III. RESEARCH DESIGN………. 18
Method……… 18
Case selection method and selected cases……….. 19
Time period and Data sources………... 23
Operationalization of the variables………... 24
Structure of Empirical Analysis………... 29
IV. THE LOMÉ PEACE AGREEMENT: The case of RUF in Sierra Leone……… 30
RUF’s conflict issues and their salience………. 32
Peace agreements contents………. 35
Peace agreement duration………. 37
V. THE ACCRA PEACE AGREEMENT: The case of LURD And MODEL in Liberia……… 38
LURD and MODEL’s conflict issues and their salience……… 40
Peace agreements contents………... 44
Peace agreement duration……….. 46
VI. THE FINAL AGREEMENT NATIONAL GOVERNMENT – POPULAR LIBERATION ARMY (EPL): The case of EPL in Colombia………. 47
EPL’s conflict issues and their salience………... 49
Peace agreements contents………. 50
Peace agreement duration………... 51
VII. ANALYSIS……… 52
Individual case analyses……….………. 52
Comparison between cases……….……… 54
Alternative explanations……….……… 58
v
VIII. CONCLUSION……….... 60
IX. BIBLIOGRAPHY………... 62
X. APPENDICES………...…….. 68
APPENDIX 1: The WTF project……… 68
APPENDIX 2: Measuring state capacity……… 75
APPENDIX 3: Non-salient conflict issues in the Lomé Peace Agreement……….………. 77
APPENDIX 4: Non-salient conflict issues in the Accra Peace Agreement……….. 80
APPENDIX 5: Non-salient conflict issues in the Final Agreement National Government – Popular Liberation Army………... 82
vi LIST OF ABBREVIATIONS
AFRC Armed Forces Revolutionary Council
AU African Union
ECOWAS Economic Community of West African States ELN National Liberation Army
EPL Popular Liberation Army
EU European Union
EZLN Zapatista National Liberation Army
FARC-EP Revolutionary Armed Forces of Colombia – People’s Army FLRN National Liberation and Reconstruction Front
GOL Government of Liberia.
ICGL International Contact Group on Liberia
INPFL Independent National Patriotic Front of Liberia LURD Liberians United for Reconciliation and Democracy M-19 19th of April Movement
MAQL Quintín Lame Armed Movement MODEL Movement for Democracy in Liberia NPFL National Patriotic Front of Liberia
NTGL National Transitional Government of Liberia PRT Revolutionary Labor Party
RUF Revolutionary United Front UCDP Uppsala Conflict Data Program
UN United Nations
URNG Guatemalan National Revolutionary Unity
vii LIST OF FIGURES AND TABLES
Figure 1. Causal mechanism……….. 15
Figure 2. Process of empirical analysis……….. 29
Table 1. Selected cases and their scope conditions………. 21
Table 2. Selected cases and their values for control variables………. 22
Table 3. Example of FARC-EP’s conflict issues……… 27
Table 4. RUF’s conflict issues and their salience...……… 33
Table 5. RUF’s salient conflict issues and their inclusion in the Lomé Peace Agreement………. 35
Table 6. LURD’s conflict issues and their salience……… 41
Table 7. MODEL’s conflict issues and their salience…..………... 42
Table 8. LURD and MODEL’s salient conflict issues……… 43
Table 9. LURD and MODEL’s salient conflict issues and their inclusion in the Accra Peace Agreement………... 44
Table 10. EPL’s conflict issues and their salience…………...……….. 49
Table 11. EPL’s salient conflict issues and their inclusion in the Final Agreement EPL………. 50
Table 12. Findings of case comparison………... 56
Table 13. Classification of state capacity……… 75
Table 14. State capacity values for Liberia, Sierra Leone and Colombia in the year they signed the studied peace agreement……… 76
1 I. INTRODUCTION
Issues have been identified theoretically as one of the main elements of the definition of “conflict” (Wallensteen, 2002), yet they have remained notably under- researched, especially in relation to conflict resolution and intra-state wars. Studies have found that issues, defined as “a disputed point or question, the subject of a conflict or controversy” (Randle, 1987: 1), have an influence on the behavior of actors and depending on the specific characteristics they have, can lead to the onset of conflict (Diehl, 1992; Holsti, 1991). Further research has looked into those characteristics and found that issues that are considered salient by the parties involved, are more likely to influence the parties’ decision to engage in armed conflict (Vasquez and Mansbach, 1983). However, most of these studies relate to inter-state wars and those that have approached intra-state wars have researched specific issues, such as territory (Hensel and Mitchell, 2005; Hensel et al, 2008).
Moreover, scholarship that has studied negotiated peace settlements as a means of conflict resolution, has also overlooked issues as an explanatory variable for peace agreement success. Studies on peace agreement success have found that there are considerations related to the security of the parties and the upholding of an agreement that affect the main parties’ decisions and actions, and if those are not properly addressed, the parties could withdraw from the settlement, i.e. the peace agreement would fail and the conflict would recur. The success of peace agreements has thus been linked to the proper treatment of the parties’ security-related concerns through the inclusion of certain provisions in peace agreements. The deployment of peacekeeping operations (Walter, 2002), as well as different forms of power-sharing like military or political (Hartzell and Hoddie, 2003; Glassmayer and Sambanis, 2008) are examples of provisions that have had a positive effect on peace agreement success by ensuring the parties that the agreement will be implemented (DeRouen Jr. et al, 2010).
I argue that while those provisions prove important for peace agreement success, there are other elements of the agreements that could provide further explanations for their success and failure. Issues, as an important element of conflict have been overlooked but can provide new insights into conflict resolution. If they are considered important elements for the onset of conflict, they should also be studied in relation to its resolution (Wallensteen, 2002). Consequently, through this study I try to answer the question why do some peace agreements succeed while others fail? by using an issue-based approach
2 that studies peace agreements, by searching for provisions aimed at solving the rebel group’s issues. While I argue that the rebel group’s issues should be recognized and properly treated, I acknowledge that as a product of negotiations, peace agreements may not address all of them. Therefore, I hypothesize that the inclusion of the rebel groups’
salient issues will lead to peace agreement success.
The theoretical thinking behind this is that issues are an expression of underlying grievances, which have caused the rebel groups to engage in armed conflict (Kriesberg and Dayton, 2012). This finds a basis in studies which explain that the parties engage in armed conflict seeking to resolve the issues that were not resolved through other means, which would imply that the parties will continue in conflict until those issues are resolved (Vasquez and Mansbach, 1984; Wallensteen, 2002). And this means that peace processes that do not address those issues lead to peace agreements that do not include solutions to the grievances and will therefore, fail.
If the hypothesis were right, successful peace agreement would include all the rebel group’s issues that were identified as salient, whereas failed peace agreements would not. Were the hypothesis incorrect, there would be successful peace agreements that would not include all the rebel group’s salient issues and failed peace agreements that did include all of them.
I tested the hypothesis and the proposed theoretical relationship through a qualitative study based on the structured focused comparison of three peace agreements:
The Lomé Peace Agreement, signed between the Government of Sierra Leone and RUF;
the Accra Peace Agreement, signed between the Government of Liberia and LURD and MODEL as rebel groups; and the Final Agreement National Government – Popular Liberation Army signed between the Government of Colombia and EPL. The first one represents a failed peace agreement, the second one is a successful peace agreement and the third case, although it was a success as well, was included as a deviant case because previous theories of peace agreement success cannot accurately explain it.
The method employed in this study followed a three-stage structure. The first stage comprised determining the salience the rebel groups assigned to their issues; the second stage involved an examination of the peace agreement’s provisions to determine if the rebel groups’ issues were addressed in the agreement and the third stage involved determining if the peace agreement had been a success. In order to conduct stage one of this study, it was necessary to create a measure for issue salience that applied for the rebel groups. This was done through identifying two attributes of issues: the priority that was
3 assigned to them by the rebel group, which meant that the groups identify an issue as especially important for them and the consistency with which the rebel group raised the issue throughout its active years. An issue was considered to be salient if it possessed the two attributes. In order to define if an issue was included in the agreement, a provision had to be included which addressed the concern, either following the rebel group’s proposal or through other means; also, the responsibility for its implementation had to be clearly assigned. Finally, the third stage involved determining the peace agreement’s success or failure which was done on the basis of how long the rebel group held it. If the rebel group withdrew from the agreement within five years after it was signed, the agreement was considered a failure, otherwise, it was considered a success. The necessary data to conduct this study that are related to the rebel group’s issues came from the dataset of the project “What Are They Fighting For? Conflict Issues and the Resolution of Civil War” from the Department of Peace and Conflict Research at Uppsala University. The data related to the provisions included in the peace agreements they signed with their respective Governments, came from the original texts of the agreements and from the UCDP Peace Agreement Dataset.
The findings suggest that the inclusion of the rebel group’s issues in the peace agreement cannot account for its success or failure. While the Lomé Peace Agreement should have succeeded according to the results of this study, it failed only seven months after it was signed in 1999. In the same vein, the Accra Peace Agreement signed in 2003 should have failed according to the results of this study, but it succeeded and peace has lasted for over 15 years in Liberia. Finally, the peace agreement signed between the Government of Colombia and EPL in 1991 was the only one whose expected outcome matched the actual result, a successful peace agreement that led to the definitive disarmament and demobilization of the rebel group.
I consider that while these results could show the inconsistencies of the theoretical argument proposed in this study, scholarship should keep looking into issues and their relationship to conflict resolution. Possible avenues for future research include (1) looking into polemical issues raised by the rebel group that can thus, be difficult to resolve and how that can impact negotiations; (2) assessing if the rebel groups possess a strategy in relation to what issues to bring up and why and (3) the possible linkages between issues for rebel groups which would help further understand peace agreements’ provisions.
The next section, chapter 2, includes the definition of the most important concepts, a review of previous literature related to issues and peace agreement success, as well as
4 the main theoretical argument that guided this study. This is followed by chapter 3, which includes the description of the research design and operationalization of the main concepts. The next three chapters present the data and some contextual background for the three cases in this study; chapters 4, 5 and 6 are dedicated to the Sierra Leonean, Liberian and Colombian cases, respectively. Chapter 7 contains specific case analysis as well as a between case comparison and provides with alternative explanations and further considerations for this study. Lastly, chapter 8 includes the conclusions and possible ideas for further research.
5 II. CONCEPTS, PREVIOUS RESEARCH AND THEORY
This chapter presents the analytical framework that guides this study. First, conflict issues will be introduced as both a concept and the starting point for an issue- based approach, followed by other important concepts for this study. Then, previous research concerning issues and peace agreement success will be presented. Finally, the theoretical argument that will guide this study will be outlined.
Defining the main concepts
Adopting an issue-based approach to civil war resolution through negotiated peace settlements and more specifically, to peace agreement success, poses certain problems like the scarcity of systematically collected data and previous research based on such an approach. This highlights the need to make the starting point of this study, the definition of the three main concepts: conflict issues, issue salience and peace agreement success.
Conflict issues
Wallensteen defines a conflict as “a social situation in which a minimum of two actors (parties) strive to acquire at the same moment in time an available set of scarce resources” identifying (1) action, (2) actors and (3) incompatibility as the main elements of this definition which can be applied to both inter-state and intra-state conflicts (2002:
16). He further defines incompatibilities as the positions that actors are interested in, and argues that a conflict ends only when the actors “adjust their demands so that there is no longer scarcity”, which makes incompatibilities essential for the onset of conflict but also, for its resolution (Wallensteen, 2002: 15).
Based on this and focusing specifically on intra-state conflicts, Brosché, Sundberg and Wallensteen within the scope of the project “What Are They Fighting For? Conflict Issues and the Resolution of Civil War” (henceforth, WTF project or project), devised the concept conflict issues to approach incompatibilities in intra-state conflicts, namely civil wars and coups1. Within this project, conflict issues are defined as “a stated and concrete
1 Appendix 1 includes various elements from the codebook of the WTF project because of its relevance for this study. Included are: the definition for conflict issues; an explanation on the coding method and a description of the sources used for obtaining the data, and finally descriptions of the conflict issue categories that have been coded for the specific cases included in this study.
6 desire for change in political structures or policies demanded by an organized armed non- state actor” (Brosché, Sundberg and Renvall, 2020: 6-7).
Since this definition is the basis for this study, it is necessary to take a closer look at its components. Conflict issues are considered to be related to concrete types of reforms that can target political structures (such as systems of governance or territory) and the contents of Government directed policies (such as education policies, but also Government or state behavior that although not formalized in terms of policy, still appears to follow the Government’s political lines). Economic issues and changes to other sectors -like security- are also included as part of the political sphere. Conflict issues are collected from statements made by nonstate actors recognized as a warring party in the UCDP Armed Conflict Dataset, leaving out those issues stated by Governments (Brosché, Sundberg and Renvall, 2002: 6-7)2.
Portraying issues as an important factor in civil war resolution moves away from explanations for civil wars as a result of “the feasibility of economic predation and the rational pursuit of economic self-interest” (Cater, 2003: 20) and is more inclined towards an approach that privileges explanations related to the existence of “feelings of dissatisfaction with important aspects of life such as housing, living standard, income, employment, health care, human rights, safety and education”, i.e. grievances, as the main cause for conflict (Klandermans et al., 2001: 42).
However, it is important to clarify, that while the concept of grievances may seem quite similar to conflict issues, the latter is considered to bring support to grievance- related explanations for the onset of conflict (and consequently, its resolution), and should not be used as a synonym for grievances. Unlike them, conflict issues are able to capture concrete demands as well as possible solutions to grievances, and can thus reflect two important factors: the rebel group’s concerns and proposals for change (Brosché, Sundberg and Renvall, 2020: 7-8). The grievance perspective allows the voice of the rebel groups to gain relevance and should therefore, be used as an overarching framework for this study.
Issue salience
When dealing with issues, it is not enough to make an inventory arguing that there are various types of issues and that they matter, without “guidelines for how to identify,
2 For a more detailed description of conflict issues, please refer to Appendix 1, section ‘definition of conflict issues’.
7 categorize, or measure issues and their salience” (Hensel et al., 2008: 119). As will be shown in the previous literature section, issue salience has been studied mostly in inter- state wars. Consequently, its definition has been until now, understood as “the degree of importance attached to that issue by the actors involved” (Diehl, 1992: 333-334). This means that the salience that an issue is assigned depends on the main parties in a conflict.
Since this study takes conflict issues as the key concept and these reflect only the perspective of the rebel groups, the definition of salience needs to reflect that and thus, will be understood here in a slightly different manner. Salience, in the scope of this study, means the importance that the rebel group assigns to a conflict issue of its own.
Peace agreement success and failure
Peace agreements are defined in this study as “formal agreement[s] between warring parties, which address the disputed incompatibility, either by settling all or part of it, or by clearly outlining a process for how the warring parties plan to regulate the incompatibility” (UCDP, 2020a). Hence, they are understood to be the result of a negotiated peace settlement, which is defined as “a war termination in which the Government and rebel group(s) agree to stop fighting and resolve the issue of civil war, and adopt a comprehensive formal treaty about the form of Government and post-war recovery” (Ohmura, 2011: 376). Following this logic, the success or failure of a peace agreement is closely related to the reappearance of the civil war that was deemed as finalized through the signing of said agreement, i.e. the same warring parties that signed the agreement return to fight.
The UCDP Peace Agreement Dataset3 uses the term end when referring to the lack of continuity of a peace agreement, explaining that when one of the parties decides to officially withdraw from the peace agreement, this can no longer be considered to survive because its validity is “contested by one or several of the warring parties that signed” and therefore, it will not be fully implemented (Pettersson, Högbladh and Öberg, 2019). A peace agreement will be considered to have ended when one of the primary parties (1) although officially committed to the agreement, still has “warring militias in field” or (2) openly returns to violence, thus, ceasing to be part of it (UCDP, 2020a).
My definition of peace agreement success and failure is derived from these previous works, but is linked specifically to the rebel groups’ actions. For this study, a
3 The UCDP Peace Agreement Dataset Version 19.1 is available here: https://ucdp.uu.se/downloads/
8 peace agreement classified as ended is considered as a failure when a rebel group was the party that withdrew from it, and it returns to fight (recurrence of the conflict). On the other hand, a peace agreement is classified as successful when the rebel groups that signed it, do not withdraw their support and accordingly, do not return to fight. To define this, it is also necessary to establish a temporality, that is, for how long the peace agreement has to hold in order to be considered successful. Following previous research on the matter (Walter, 1997 and 2004), for this study it has been defined that the rebel group should not withdraw from the agreement for five years, which means that if within the next five years after the agreement was signed the rebel group returns to fight, the agreement failed. In the contrary case, the peace agreement is considered successful.
Why would issues matter? A look into previous research
The importance of issues for conflict resolution in civil wars is, as explained, understudied. In order to grasp why issues would matter for civil war resolution, it is important to look at previous literature that has dealt with issues in inter-state conflicts, as well as literature related to peace agreement success and civil war recurrence.
Literature related to issues
The novelty and specificity of the concept conflict issues makes it necessary to expand the search for previous literature to include the scholarship on issues, as they refer to the incompatibilities between the parties in a conflict. Defined as “a disputed point or question, the subject of a conflict or controversy” (Randle, 1987: 1), issues have been explored by scholars studying inter-state and intra-state conflicts, with the former gathering more literature than the latter. Most of the studies focused on inter-state wars seek to understand the issues over which, states decide to go to war with each other and how these have changed or remained over time.
Vasquez and Mansbach challenged the realist perspective by approaching inter- state war assuming that “at least one of the purposes of politics, both domestic and global, is to raise issues and engage in their resolution”, which meant that war should be understood as more than a power struggle (1983: 258). They found that issues prompted different actor behaviors (Vasquez and Mansbach, 1983: 274).
Also focused in inter-state wars, Holsti (1991) devoted great effort to tracking the issues that generate these conflicts and how they had changed over time, as well as the relation they have to the decision of policy makers about when to use force as a means to
9 solving them. He found that while international peace efforts have attempted (and on various occasions, succeeded) to solve contemporary issues, they have not been able to anticipate new issues (Holsti, 1991: 339).
Diehl (1992) focused on the onset and the escalation of international conflict and argued that the role of issues was an element that needed to be further researched. He looked at the results of different studies and confirmed that policy-related behavior changes depending on the issue area, which means that depending on the issue, actors will decide on different courses of action (Diehl, 1992: 333).
Based on those previous findings, other scholars have shown interest and completed studies on inter-state conflicts that emphasize specific types of issues, territory being the most common (Hensel and Mitchell, 2005 and 2017; Hensel, 2001). And following the same premise, that is, different issues have different effects on war-related behavior, scholars of intra-state conflicts have centered their attention as well, on conflicts that combine some issues like territory, identity and religion (Hassner, 2003; Toft, 2006).
Hassner, for example, focused on conflicts related to ‘sacred space’, that to some extent combine those three issues, and concluded that because of their specific characteristics, they lead to “violent conflict at its most atrocious” (2003: 32).
More recently, Brosché, Sundberg and Wallensteen through the WTF project, are working on providing empirical data on the different issues in civil wars, as well as quantitative and qualitative studies on the matter. However, the dataset and consecutive studies are yet to be released.
Hence, in spite of new contributions that are coming, most scholarship on issues has been dealing with inter-state wars, while studies on intra-state wars have looked into single issues that have shown to lead to more protracted conflicts or with very specific conflicts that have combined issues. But there is one element in common, that is issues have been shown to be important in order to better understand different aspects of conflict, such as actor behavior, that can lead to the onset, escalation or resolution of conflicts.
This stands in stark contrast to how understudied they have been, especially in relation to intra-state wars. According to Diehl, this could be explained by the various difficulties that arise when dealing with issues in conflicts, mainly a lack of systematically collected data on several issues across cases, complications related to salience measures and to the analysis of different types of issues at once (1992: 335).
10 Literature on issue salience
From the previous section it is clear that scholars have not only showed interest on which issues have sparked inter-state or intra-state conflict, but also on why they have done so. The underlying idea in some studies is that certain issues possess qualities that make them more important than others and are thus, more likely to influence the onset of conflict as well as the decisions the actors make within that context. This has motivated scholars to look at what characteristics of issues make them salient for actors.
Although Diehl did not establish a measurement for salience, he pointed out to its importance by explaining that this issue characteristic is likely to “influence the chances of war”, referring to inter-state war (1992: 341). Important for this study is Diehl’s annotation that the fact that one salient issue can influence the onset of conflict, should not lead to the assumption that salience is symmetric for the main parties. In other words, different actors can (and most likely do) assign different salience to the same issue. In accordance with this, Diehl proposes that measures of salience are created separately for each party (1992: 341).
Vasquez and Mansbach (1984) also highlighted the importance of salience as an issue characteristic in their study of inter-state conflicts. They argued that issues can lead to war, but depending on their salience (high or low), the grounds for the onset varies.
They explained that “patterns of cooperation and conflict”, which they named
‘background variables’, can predispose the parties depending on the issues at stake (Vasquez and Mansbach, 1984: 423). So, issue salience becomes important because “low salience issues” can lead to conflict on the basis of those pre-existing patterns, but “high salience issues” may instigate conflict on the basis of the issues’ own characteristics (Vasquez and Mansbach, 1984: 423).
On specific measures for salience, there are other scholars that have focused on studying distinct issues and have thus, developed issue salience measures that reflect dimensions that pertain those particular issues. Since most studies have aimed at understanding territorial issues, measures for salience that have been created are related to that.
The studies from Hensel et al. (2008) and Hensel and Mitchell (2005) are examples of salience measures created for territorial issues. Arguing that previous attempts to determine the salience of territorial issues based on the distinction between tangible and intangible dimensions of an issue did not take into account that both are important and that certain issues can combine them, they focused on studying both
11 dimensions of territorial issues (Hensel et al, 2008: 119). For the intangible dimension, indicators such as the existence of ethnic, religious or other identity ties to the state were taken into account, while for the tangible dimension aspects as the existence of valuable resources (for example water, oil, uranium), were considered (Hensel and Mitchel, 2005:
278).
The scholarship that has been presented accounts for the importance of salience as a characteristic that gives relevance to those issues that possess it. Yet, since theory and empirics on issue salience can only be found in studies on issues and there is not extensive research from an issue-based approach, there have been few attempts to look at salience, especially across different types of issues.
Literature on peace agreement success and civil war recurrence
In order to fully understand the developments in terms of scholarship related to what has been defined for this study as peace agreement success and failure, it is necessary to first delve into the field of civil war recurrence because of the close relationship these two concepts have, as explained previously. There are two main arguments regarding this phenomenon: (1) victories are the only type of civil war termination that guarantees an end to the conflict and (2) when civil wars end via negotiated peace settlements, handling security concerns and implementing peace agreements are key in preventing the recurrence of civil war. As this study is related directly to peace agreement success and failure, only the second argument will be explored in detail, and will be used in consecutive sections. The argument focused on negotiated settlements explains that the implementation of peace agreements is essential to achieve and maintain peace (i.e. civil war does not recur). Credible commitment and security guarantees are found to be important elements in peace processes that can lead to peace agreement success.
Glassmyer and Sambanis (2008) explained that the bargaining power that rebel groups possess diminishes after demobilizing, which makes it possible for the Government to “renege on its promises” (Glassmyer and Sambanis, 2008: 365-367).
According to them, one way of solving this is implementing self-enforcing agreements, which has been attempted through the integration of former rebels into a new national army. Although they explained the complications to arrive to firm conclusions, they found that including military integration in the peace agreements could help the peace last (Glassmyer and Sambanis, 2008: 381).
12 Hartzell and Hoddie (2003) identified security-related concerns as the biggest threat to the stability of civil war settlements and argued that the inclusion of extensive power-sharing provisions in peace agreements is the most effective way to reduce the risk of returning to war. Their study showed that when security concerns are addressed through the inclusion of power-sharing provisions that include “territorial, military, and economic dimensions of state power” the return to war becomes less likely (Hartzell and Hoddie, 2003: 330).
DeRouen Jr. et al (2010) argued that implemented peace agreements are less likely to fail. They focused on state capacity and third-party intervention as important elements for implementation and concluded that stronger states are more capable of implementing peace agreements and in the cases where states are weak, third parties can intervene to achieve implementation and thus, prevent the failure of the peace agreement.
Walter (2010) argued that negotiated settlements can prevent civil war recurrence, when peace agreements are implemented. She claimed that this can be achieved through the political or military (rather than the economic) intervention of third parties, explaining that political interventions create credible guarantees for peace agreement implementation. This argument is linked to her previous work, arguing for third-party security guarantees in the form of “enforcement or verification of the post-treaty implementation period” or power-sharing provisions as a way to clear the concerns of the rebel groups and lead to peace agreement implementation, thus, securing the peace (Walter, 2002: 4).
Jarstad and Nilsson (2008) highlighted the importance of military and territorial power-sharing for achieving peace. They argued that beyond the inclusion of such provisions in a peace agreement, their implementation still needed to be studied to determine the effects it can have on the prospect of peace. They showed that political power-sharing provisions were implemented quickly after the peace agreement, while military and territorial power-sharing provisions took longer, but that it was actually these two that were highly associated with peace because of the high costs they entail, which makes it less likely for peace to fail after a settlement (Jarstad and Nilsson, 2008: 220).
Finally, in terms of temporality of peace agreements, various scholars have considered peace agreements as successful if the conflict does not recur for a determined period of time after it ended. While Doyle and Sambanis (2000) chose a two-year time period, Walter (1997 or 2004) chose a five-year time period. Walter considered that only after five years could agreements be deemed as having “lasting effects” on peace,
13 furthermore, her data showed that peace agreements that broke down before that period were afterwards settled through military victories, which is why she considered that choosing a longer time period was more accurate when studying peace agreement success (1997: 345).
Synthesizing the literature: the research gap
In summary, while scholarship on issues and issue characteristics remains focused on the onset of inter-state and intra-state wars with important contributions for specific issues like territory, scholarship on peace agreement success (and failure) has been explored mainly as the consequence from the lack of credible commitment and the existence of security concerns from the parties involved.
Based on this review of existing research, it is considered that the field related to issues is yet to be thoroughly explored and that important research gaps remain, especially regarding their role in conflict resolution, particularly in the case of civil wars. On the other hand, previous works addressing peace agreement success have recognized the importance of including provisions regarding the parties’ security concerns, but there remains much to be done on the agreement’s contents regarding the solution of issues that motivated the war.
In other words, while literature on issues has identified them as important elements for understanding conflict, literature on peace agreements has established the relevance of the agreements’ contents; but there has not been an attempt to look into the issues addressed in the peace agreements. I argue that this aspect of civil war resolution that has been overlooked, constitutes an important gap and can provide insights into the rebel groups’ decision to adhere to negotiated peace settlements. Moreover, I consider it is relevant to study civil war resolution from an issue-based perspective and to start delving into cases of civil wars related to state structures and policies rather than territory, religion or ethnicity, which have been more widely studied using such a perspective.
In this sense, this study has two purposes. First, to create a measure of issue salience that can be applied to conflict issues, which means, a measure of the salience of issues for rebel groups. And second, determine if solving the rebel groups’ issues in peace agreements is linked to conflict resolution in cases of civil wars related to government as the main incompatibility. By searching for a connection between issues, peace agreements’ provisions and civil war resolution, I consider that this study can prove to be a novel contribution to the fields of issues and conflict resolution.
14 Bridging issues and peace agreement success: Theoretical argument
Previous literature shows that peace agreement failure is closely related to security concerns that come from the parties’ loss of bargaining power, which for rebels happens after handing in their weapons (Glassmyer and Sambanis, 2008; Hartzell and Hoddie, 2003). This displays the importance that guarantees hold for the parties by showing each other their willingness to abide by a peace agreement. However, previous research has not been able to properly address if and how a peace agreement treats the different issues presented by the rebel groups in civil wars, which leaves three important questions unanswered: Were their main concerns addressed in the peace agreement? If they were, how was this done? And finally, did it have an effect on conflict resolution?
When explaining the emergence of conflict, Kriesberg and Dayton (2012) argue that in order for conflict to become manifest, at least one of the parties must identify itself as having a grievance and on that ground, define “goals to change another group’s behavior so that the grievance will be reduced”, believing “that they can indeed bring about the desired change” (2012: 49). In the same line, Vasquez and Mansbach (1983) claim that parties take part in a conflict because they want to settle the issues that prompted the conflict.
Based on this, if rebel groups engage in conflict in order to address a situation that they feel aggravates them -or a larger part of society they claim to represent-, then alleviating those grievances implicates that there would be no more need for armed conflict. Wallensteen (2002) argues that when conflict resolution is attempted via negotiated settlements, the peace agreements become “particularly dependent on the central issues of contention” because the parties act incited by particular grievances and
“thus, the conflict cannot end until such grievances are resolved, ended or at least attended to” (Wallensteen, 2002: 7, 25). This leads me to believe that when studying peace agreement success, beyond their provisions related to security concerns, other agreement’s contents must be studied.
In line with this, conflict issues as a concept allows scholars to capture the rebel groups’ concerns (based on their perceived grievances) and their proposals on how to address them (Brosché, Sundberg and Renvall, 2020: 7-8). If conflict issues reflect grievances and in order to resolve conflicts, grievances need to be addressed, then it follows that the peace agreement that was created in order to solve the conflict, should include those conflict issues, because it would mean that the grievances that fuel the rebel group, are recognized by the Government and measures are taken towards resolving them.
15 Hence, in this study I take an issue-based approach to the question why do some peace agreements succeed while others fail? This study’s main argument is that when approaching civil war resolution via negotiated peace settlements, the conflict issues should be addressed in the peace agreement, therefore attending the rebel group’s concerns, as well as recognizing and solving the underlying grievances.
However, facing the possibility that not all conflict issues will be addressed in an agreement and that in negotiations, neither party will win “all that there is to win, but no one loses all that there is to lose”, I argue that the ones identified as salient are the ones that should be addressed in order to solve the incompatibility, therefore, making it possible to end the conflict (Wallensteen, 2002: 9). Thus, my hypothesis is established as follows: peace agreements that include the rebel group’s salient conflict issues will be successful. The corresponding null hypothesis states that: addressing the salient conflict issues in peace agreements has no effect on their success or failure.
Figure 1. Causal mechanism
Following this, I argue that there are three main aspects to look into. First, the salience of conflict issues needs to be determined. Second, the way in which salient issues are addressed in the peace agreement needs to be examined. Third, the duration of the peace agreement needs to be identified.
Identifying conflict issue salience
Salience has been classified by scholars (Vasquez and Mansbach, 1983; Diehl, 1992) as a defining characteristic of issues that affects the actor’s behavior and has implications for conflicts, both in terms of its onset and its resolution, which means they should bear greater impact on peace agreement success than other issues.
Inclusion of salient conflict issues in peace
agreement
Yes
Rebel group's main concerns are recognized and
grievances are attended to
Peace agreement success
No
Rebel group's main concerns are not
recognized and grievances remain
unattended
Peace agreement failure
16 It was shown in the previous literature section that some scholars have devoted themselves to creating issue salience measures (Hensel and Mitchel, 2005, for example), yet, those efforts have been focused mostly on measuring the salience of a specific type of issue: territorial. This makes such efforts highly valuable for scholarship dedicated to conflicts related to territorial issues, but is limited to that field because of its particularities. In addition, such approaches have studied salience keeping in mind several actors, namely, the main parties to the conflict. These characteristics make previous salience measures inoperative for this study, for conflict issues are defined only by one actor, the rebel group. Consequently, and following Diehl’s (1992) observation that actors assign differentiated salience to the same issue, I decided to create a measure for conflict issue salience. I argue that the salience of conflict issues can be measured through what I call consistency and priority as attributes determined by the rebel groups. For an issue to be considered salient it has to hold both attributes.
Priority is understood as the relevance that the rebel group assigns to certain conflict issues in relation to the others. The fact that groups assign importance to certain issues and not others denotes that they have their own ranking for what changes must be achieved in order for them to stop fighting. Priority can be expressed in different ways, such as referring to it as the ‘only way’ to achieve other goals, the main reason the fighting started, or indicating that it is determinant for the group’s identity. The different ways in which it can be observed that rebel groups assign priority to a conflict issue are further explained in the section operationalization of the variables in the Research Design. This does not mean that only one conflict issue can be considered a priority, what it means is that among the various conflict issues that are brought up by a rebel group, some (at least one) are considered by them to precede the others.
In a way, priority reflects the own rebel group’s internal salience measures, that is, if the rebel group identifies a conflict issue as a priority this means it is more important, hence, salient. However, considering the reluctance there might be towards trusting the rebel groups’ words solely, I considered it necessary to look for another attribute of conflict issues that might reinforce what the rebel groups say, the second attribute is consistency.
Consistency is understood as the stability of a conflict issue being brought up by the rebel group across active years4. The insistence of rebel groups in mentioning a
4 For an account of how the coding is done within the WTF project, please refer to Appendix 1.
17 conflict issue throughout the civil war’s active years is considered to reflect that their concern has not been addressed by the Government (or the Governments that have held office during those years) and that they continue to consider that it should be addressed.
This becomes especially important in civil wars that have lasted for many years because as Governments change, the continuation in bringing up a concern means that it has not been resolved.
As mentioned, for a conflict issue to be considered salient, it has to hold both attributes, i.e. if the rebel group has considered it more important than other throughout the duration of the conflict, I argue it is, in fact, salient.
Inclusion of salient issues and peace agreement success
The second and third part of the theoretical argument entail the verification that the salient issues have been included in the peace agreement and the number of years the agreement has held. For a peace agreement to be considered successful, the rebel group must have held to it and not returned to war for five years. This study’s theoretical argument poses that all the salient conflict issues need to be included in a peace agreement for it to succeed. This decision was made following theoretical considerations.
For the theoretical consideration, Jensen argued that not addressing issues in peace agreements by eliminating them or “deferring their consideration or implementation to a later stage -a stage that may never be reached-” is a dangerous endeavor because not solving issues in a negotiation may result in an agreement that “may become meaningless and unlikely to settle the conflict” (1995: 119). In addition to this, within the scope of this study it was not possible to determine which of the salient conflict issues are more salient than others. These two considerations led me to decide that trying to point out which of the salient conflict issues should be included and which ones could be left unattended, was not possible.
18 III. RESEARCH DESIGN
This chapter presents the research design employed to test the theoretical argument explained in the previous chapter to the selected empirical cases. The method is structured focused comparison, which will be applied to three different peace agreements: The Accra Peace Agreement, the Lomé Peace Agreement, and the Final Agreement National Government – Popular Liberation Army. In this chapter, the following will be described and explained. First, the method, cases and case selection method; second, the sources and third, the structure of the empirical analysis.
Method
The aforementioned scarcity of research on civil war resolution from an issue- based perspective has led to the decision to use a qualitative approach to analyze the research question and proposed hypothesis. I consider that a careful measurement approach to the variables under study is fundamental for an accurate understanding of the theoretical argument that is being proposed. In order to achieve this, qualitative measurements are required. Likewise, the data for this study can only be obtained by looking closely and carefully at the selected sources, for which a qualitative approach is essential.
The specific method that was selected for this study is structured focused comparison, which consists of posing the same set of general questions to each of the selected cases in order to acquire the necessary data, focusing only on clearly defined aspects that are central to the theory (George and Bennet, 2005). This method is considered to allow researchers to do a systematic comparison of the selected cases on the basis of a consistent data collection and when a clear research objective has been defined (George and Bennett, 2005: 67, 70).
Hence, this is a comparative case study that seeks to explain variation in the success of peace agreements (dependent variable) through the inclusion of the salient conflict issues in the agreement’s provisions (independent variable), arguing that addressing the rebel groups’ concerns and proposals (causal mechanism) leads to peace agreement success.
19 Case selection method and selected cases
Case selection in qualitative case studies is an essential process aimed at finding representative cases if the study is to potentially explain a larger population of cases (Seawright and Gerring, 2008: 295). Since the selected method, structured focused comparison, is according to Powner a “between-case design”, at least two cases are needed in order to test the previously stated hypothesis in a credible way (2015: 129).
For the purpose of this study, cases are peace agreements. The scope conditions for selecting them entail the type of conflict it aimed to resolve, the main incompatibility involved, the agreement type and finally, the agreement frame. In terms of the type of conflict it aimed to resolve, selected agreements should have been aimed at resolving a civil war, understood as a contested incompatibility between the Government of a state and at least one other non-state actor, that has resulted in at least 1000 battle-related deaths in one calendar year (UCDP, 2020a). Related to the main incompatibility involved, it should have concerned Government, meaning “[the] type of political system, the replacement of the central Government, or the change of its composition” (Pettersson, Högbladh and Öberg, 2019). In other words, the agreements must have been the result of a negotiated peace settlement of a civil war over control of the Government. This means that those agreements signed to resolve civil wars related to territorial incompatibilities cannot be considered for this study5. Additionally, peace agreements that resulted from civil wars that had major ethnic or religious components will also be left out because of the specific nuances they entail, which may not be able to be captured in such a short study.
The type of the peace agreement refers to the extent to which the incompatibility was settled, while the frame means the point in the peace process at which a peace agreement was signed. Selected peace agreements should be full, meaning that the whole incompatibility was settled (type) and final, meaning that it was signed at the end of the peace process (frame) (Pettersson, Högbladh and Öberg, 2019).
For this study, the most-similar case selection method is used, which means that the cases are selected on the basis of different outcomes (Y) and highly similar background conditions (Z) considered to have an impact on the outcome (Gerring and Cojocaru, 2016: 399; Gerring, 2006: 131). Hence, it is the independent variable, the match
5 A territorial incompatibility is one that concerns “the status of a territory, e.g. the change of the state in control of a certain territory (interstate conflict), secession or autonomy (internal conflict)” (Pettersson, Högbladh and Öberg, 2019).
20 between salient conflict issues and peace agreement’s provisions, the one that remains unknown previous to the analysis. Following this method, two cases are selected on the basis of the dependent variable, the success of the peace agreement, which means I have selected a peace agreement that succeeded and one that failed.
Keeping the scope conditions and the selection method in mind, the following agreements were selected. The failure case is the Lomé Peace Agreement signed between the Government of Sierra Leone and RUF (Revolutionary United Front) in 1999 and which failed in 2000. The success case is the Accra Peace Agreement, signed in 2003 between the Government of Liberia and two rebel groups, LURD (Liberians United for Reconciliation and Democracy) and MODEL (Movement for Democracy in Liberia).
In search of improving this study, I decided to add a deviant case to the analysis, selected on the grounds that it “deviated from an expected causal pattern” following previous literature that explains successful peace agreements (Gerring and Cojocaru, 2016: 399). The third peace agreement that will be analyzed in this study is the Final Agreement National Government – Popular Liberation Army (henceforth, Final Agreement EPL), signed between the Government of Colombia and EPL (Ejército Popular de Liberación) in 1991.
While both the Lomé and Accra peace agreements have similar values for the background conditions, the Final Agreement EPL, also classified as successful, holds none of them (Pettersson, Högbladh and Öberg, 2019), a surprising result which qualifies it as a deviant case (Gerring and Cojocaru, 2016). I chose to include this case on the basis that adding deviant cases to studies such as this can “increase the likelihood of creating a robust set of findings that have relevance in a wide range of contexts” (Wicks, 2010: 290).
Since this peace agreement’s success is not explained by the relevant theoretical approaches listed in the previous literature section (for example, inclusion of military power-sharing provisions or peacekeeping missions), if the salient conflict issues of the rebel group were found to be addressed in the peace agreement, this could pose as an explanation for its success. This would make a very strong argument about the analytical framework of this study posing as a valid explanation for peace agreement success and failure. According to Wicks, the outcomes of studies in which deviant cases have been studied can be stronger because they show that “every piece of data is accounted for”
(2010: 291).
21 Table 1. Selected cases and their scope conditions
Control variables
The fact that selected cases had to show similar values for the background variables as a requirement in order to be selected according to the most-similar case selection method (Gerring and Cojocaru, 2016), was explained in the previous section.
This section will be focused in detailing the four variables included in table 2 that serve as controls for the outcome of the cases. All of them are taken from previous research and have been shown to be important for peace agreement success.
The two first control variables are the inclusion of provisions in the peace agreement, military power-sharing and a peacekeeping operation. Both provisions are aimed at addressing the security concerns related to the upholding of the agreement, the former derived from the explanation that forming a new army, which integrates former rebels, is a way to make an agreement self-enforcing (Glassmyer and Sambanis, 2008) and the latter derived from the argument for third party security guarantees as critical for peace agreement success (Walter, 2002). While the Lomé peace agreement (failure case) and the Accra peace agreement (success case) included both provisions, the Final Agreement EPL did not include either (Pettersson, Högbladh and Öberg, 2019). It should be noted that, also keeping up with Jarstad and Nilsson (2008), the military power-sharing provision included in the Lomé and Accra peace agreements were actually implemented and not just included (Joshi, Quinn and Regan, 2015).
The third control variable is state capacity and it follows from the argument that it has a direct effect on peace because states with a lower capacity might have problems implementing several peace agreement’s provisions (DeRouen Jr. et al, 2010). State capacity is strongly related to the Government’s ability to effectively exercise authority
22 in “economic, political and military matters” which make peacebuilding possible in post- conflict contexts and is defined as “(…) the state’s ability to accomplish those goals it pursues, possibly in the face of resistance by actors within the state” (DeRouen Jr. et al, 2010: 333, 335). While both the Sierra Leonean and the Liberian states were found to have a very low state capacity, the Colombian state was found to have a medium capacity6.
The fourth and final control variable is the rebel group’s strength, defined as “the strength of the rebel forces relative to the Government forces” (Cunningham, Gleditsch and Salehyan, 2012). According to Clayton, “strong rebel groups, those whose military capacity is comparable with the state, are the best equipped to defend themselves in the aftermath of an agreement” (2013: 613). Therefore, I argue that it is important to include this variable because rebel groups with high relative strength could be more likely to withdraw from a peace agreement than those with low relative strength. The value for rebel group’s relative strength for the three cases was found in the Non-state actors in civil wars Dataset version 3.47, indicating that LURD, MODEL and RUF were “weaker”
than their respective Governments, while EPL was “much weaker” (Cunningham, Gleditsch and Salehyan, 2013). The fact that the values in these variables are kept constant in the first two cases, shows that they cannot explain the variation in their success or failure, which means that the inclusion of the salient conflict issues could be a possible explanation. As explained in the case selection section of this study, while two peace agreements were chosen on the basis that they share similar values on the control variables, the third case should not be expected to do so. Its different values make it an interesting case to include.
Table 2. Selected cases and values for the control variables
6 For further explanations on how this variable was operationalized, please refer to Appendix 2.
7 The “Non-state actors in civil wars” Dataset version 3.4 is available here: http://ksgleditsch.com/eacd.html
23 Time Period and Data Sources
The selected time period for determining the success or failure of a peace agreement is five years after the main parties signed it. In other words, a successful peace agreement will hold for at least five years, whereas those agreement that end within that time period are considered to have failed. I argue that the chosen time span (five years) represents an appropriate time period because, as Walter explained, peace agreements that end before that period have not been fully implemented having some of its provisions only partially implemented (1997: 345). It follows that if the group’s salient concerns were not addressed in the agreement, they would not be invested in its implementation and would therefore, be likely to return to war soon after the agreement was signed.
Related to data sources, this study relies on two elements: the conflict issues and the contents of the peace agreements, both of which come from main sources. In the case of the peace agreement’s provisions, this means that the original peace agreements constitute the source material. The UCDP Peace Agreement Dataset was also consulted in order to find information about the time that the peace agreement held, which constitutes the indicator for the agreement’s failure or success. All the content from the peace agreements was considered, including its annexes, if any.
The data on conflict issues were obtained from the dataset of the WTF project.
Two important elements come from said dataset, (1) the conflict issues as they have been coded for a rebel group and (2) the statements used for the coding. The reason for this is that, the statements will be used in order to determine the salience that the conflict issues, already identified by the project, held for the rebel groups8.
For transparency, it is important to briefly explain the data collection method and sources that said project has, since it is essential for this study. Through the project, statements from the rebel groups are collected where the groups have said what it is that they are fighting for. As was explained, conflict issues comprise the concerns of the group together with their proposals, so what the project is looking for in the statements are phrases that explain what the groups want. The conflict issues are identified based on the typology (comprised of 80 categories) that was created for the project. So, if a rebel group states that one of their demands is that a tax reform is implemented by the Government (for example), this is coded under a category that encompasses such demands (in this
8 I was granted access to these data after finishing an internship with the project, during which I coded issues for various rebel groups. This allowed me to get to know the data collection method, the codebook and finally, the different ways the rebel groups communicate in, as well as their concerns.
24 case, said category is ‘Economic reforms’). A group does not have to use the specific name of that category (in fact, they rarely do so) in order for it to be coded under it, so if FARC-EP (Revolutionary Armed Forces of Colombia – People’s Army) has a conflict issue called ‘restructure military forces’, for this study it was necessary to verify two things: (1) the definition of the conflict issue (the explanation of the category, included in Appendix 1) and (2) the exact quote that was used to code that issue. The conflict issue category and the exact quote of the statement are included in a dataset and that was the source for the conflict issues and the quotes that are included in this study. The conflict issues of rebel groups are taken primarily from direct statements that are defined as “direct quotes, speeches made, manifests, and interviews given by leaders and spokespersons”
that were released during the years that the conflict was active according to the UCDP Armed Conflict Dataset (Brosché, Sundberg and Renvall, 2020: 8). Direct statements for this project are considered primary sources9.
With the data that will be used for this study, the concern of bias can be high and could pose an objection towards the study. The two most common biases related to sources are the deliberate misrepresentation of information and the omittance or amplification of information (Höglund and Öberg, 2011: 188). In this case, the data are obtained from primary sources, as was previously explained, which is why the risk of bias coming from an interpretation of what the rebel groups identify as their issues is minimized.
However, I acknowledge that the fact that the information comes from rebel groups could predispose the reader into thinking the rebel groups have several incentives to misrepresent their own issues. It is clear that the rebel groups’ leaders or spokespersons could be drawn to state that grievances are the main reason they engaged in armed conflict even if that may not be the case. In light of this and to those readers, I answer that this study could pose as a test itself for what the groups claim are their issues and ask only that the results be approached keeping the previously stated grievance-based approach in mind.
Operationalization of the variables
The definitions provided in the theoretical argument need to be operationalized in order to test the hypothesis. This will be done in three parts, the first related to the salience
9 For a broader explanation on the coding method and sources for the Dataset, please refer to Appendix 1, sections ‘coding statements’ and ‘direct vs indirect statements’.
