Social Psychology Paper

African Security, 3:239–272, 2010 Copyright © Taylor & Francis Group, LLC ISSN: 1939-2206 print / 1939-2214 online DOI: 10.1080/19392206.2010.533579

UAFS1939-22061939-2214African Security, Vol. 3, No. 4, Oct 2010: pp. 0–0African Security

Peacemaking in Burundi: Conflict Resolution versus Conflict Management Strategies Peacemaking in BurundiJessica Piombo

Jessica Piombo, PhD

Department of National Security Affairs, Naval Postgraduate School, Monterey, California, USA

ABSTRACT. Peacemakers are faced with a difficult decision when engaging in negoti- ations to end conflict: should they adopt a conflict resolution strategy that seeks to address the fundamental drivers of conflict (root causes) or pursue a more limited strategy of conflict termination that seeks to respond to the symptoms of violence? This article investigates this question through a case study of the Burundian civil war of the 1990s. It analyzes externally facilitated peace negotiations in order to explore the types of issues that were brought into the process, to ask when and why certain issues were or were not considered, and to investigate the effects of the choices made in negotiating strategies. The lessons of the Burundi case suggest that fundamental issues must be addressed if a conflict is to be fully resolved rather than just managed. Delaying the resolution of root causes until after peacemakers have exited the situation can enable powerful groups to avoid addressing the issues. After the peacemaking and negotia- tions process ends, there is less international attention and pressure, so the ability to perpetuate the status quo is enhanced. The difficulty is that the issues that fomented conflict in the first place may prove too sensitive to be introduced into negotiations when the conflict is either ongoing or very recent.

KEYWORDS. Burundi, civil war, peacemaking, peace negotiations, conflict resolution, conflict management.

INTRODUCTION

This paper evaluates a conundrum of whether to focus on root causes or end states when devising policy to deal with violent conflict. It explores these

All views expressed in this paper are the author’s own and do not represent the views or positions of the United States government. Address correspondence to Jessica Piombo, PhD, Department of National Security Affairs, Naval Postgraduate School, Monterey, CA 93943. E-mail: [email protected]

240 Jessica Piombo

issues through a case study of the Burundian civil war of the 1990s, analyzing the foundations of the conflict and evaluating various attempts to resolve or intervene in the conflict. In particular, the paper examines the tradeoffs and consequences of different strategies for conflict mediation: focusing on founda- tions versus end states and how the choices made by interveners affect the course of the conflict and the outcomes of their mediation efforts. Overall, the paper suggests that third-party peacemakers face a near-Faustian bargain when determining the scope of negotiations strategies: including fundamental root causes may be unfeasible or impossible for a number of reasons, but postponing consideration of the drivers of conflict until later stages may prevent their full resolution and foment further rounds of violence.

The Burundian case offers a rich source of material to explore these issues. There were various attempts to resolve the conflict over the years, facilitated by a range of international actors, and each adopted a different approach to negotiations. The case considers the following questions: who were the main mediators at various points in the peace talks to end the conflict that began in 1993, what choices did they make about participants and issues, and what were the ramifications of those decisions?

The lessons of the Burundi case suggest that fundamental issues must be addressed if a conflict is to be fully resolved rather than managed. Issues that fomented conflict in the first place may prove too sensitive to be introduced into negotiations when the conflict is either ongoing or very recent. The Burundian case shows that core issues can still be addressed by subsequent processes if they were excluded from the negotiations but that getting to the point where fundamental root causes can be addressed may take a lot longer if the core issues are not resolved—or even genuinely opened—during the nego- tiations phase. Ignoring key root causes because of political expediency or the opposition of negotiating partners can, and in the case of Burundi did, delay the attainment of a full peace settlement by requiring multiple side negotia- tions and extending the life of the conflict.

The analysis suggests that policy responses to conflict that address only end states may manage a conflict rather than truly settle it. The distinction is between a short-term, crisis-response approach and a long-term, conflict reso- lution approach. Thinking in terms of crisis response versus conflict resolution captures the dilemma facing peacemakers: respond in the short term and address the most pressing effects of conflict (and then exit the situation), or seek to use the mediation effort as a process to open up the fundamental issues that prevent full resolution of the violence. The short-term approach seeks to terminate the conflict and deal with immediate effects; the long-term perspec- tive pursues broad-based peace building and societal reconciliation.1

In the longer term the fundamental issues that drive conflict eventually have to be addressed or else the conflict is likely to break out repeatedly. The short-term motivation of external actors may often lead to end-state-focused

Peacemaking in Burundi 241

solutions, those based on achieving the “negative peace,” or the end of conflict. True conflict resolution, which addresses the drivers of conflicts, requires a more broad-based approach that integrates various instruments. What these instruments are (economic policy, social engineering, political assistance, mil- itary assistance, etc.) will be dictated by the context of the particular conflict that is being addressed by the external actors.

This paper investigates these questions by examining one case in depth: third-party efforts to resolve the Burundian civil war of the 1990s. This war, which began at the end of 1993 and continued in various phases through 2009, has undergone multiple rounds of negotiations facilitated by international actors. In each of these rounds of negotiations, lead facilitators chose different strategies about what issues to include in the negotiations processes. Some negotiators focused on establishing preconditions and agenda setting, others on setting up transitional processes, and still others attempted—though they often failed—to bring fundamental issues into the realm of the negotiations.

The entire cycle of conflict and negotiation in Burundi can serve as a case study to analyze the effects of focusing on different levels. Due to the wide range of negotiation strategies and actors, the Burundi case contains signifi- cant internal variation that allows comparison across different negotiations periods. These comparisons enable us to assess the outcomes of different strat- egies of negotiations. Additionally, compared to other African conflicts where over time a resource logic warped a struggle that began over political issues, the Burundian conflict seems relatively simple. The core issues driving con- flict in 1993 were much the same as those generating tensions and massacres in 1965, 1972, 1988, and 1991. That the Burundians were not able to resolve the issues points to the fact that they were not issues that could be resolved simply. The stability of the issues allows an analysis of the complex process of peace negotiations that does not have to trace ever-changing motivations for the actors. A similar exercise for the war in the Democratic Republic of Congo would be much more complex, as the armed factions multiplied every time an armed group realized that there were substantial material gains to be made from the war enterprise in that mineral-rich country. The Burundi case, therefore, presents a stylized conflict, which simplifies the issues for analysis.

Throughout this paper the language will focus on “process” versus “con- tent” and on the nature of the issues brought into the substantive negotia- tions. When negotiations strategies are analyzed for focusing on process versus content, the discussion refers to whether there were any substantive issues at all raised in the negotiations (content) or whether the discussions focused more on procedures to bring parties together, establish the framework for talks, or set the agenda for more substantive discussions (process). When analyzing substantive negotiations, the analysis then focuses on which issues were put on the table. “Crisis-response strategies” refers to short-term per- spectives that raise issues pertaining to conflict termination and ending

242 Jessica Piombo

violence (securing the negative peace) as well as to those dealing with the immediate effects of conflict. “Conflict-resolution strategies,” in contrast, will be applied when negotiators incorporate discussions of root causes and funda- mental drivers of conflict in the context of the peace talks that they mediate. The phrases root causes and fundamental drivers are used interchangeably in the paper.

PEACEMAKING PROCESSES

Scholars have debated whether, why, and how peacemakers should focus on root causes or complex emergencies, and they have attempted to analyze the consequences of different strategies of negotiations. Debates on the success and failure of negotiated settlements have revolved around several themes: the timing of negotiations, the negotiations process, the nature of the peace agreement, the presence or creation of spoilers during the peace process, and the implementation of peace accords.2 The issues motivating this paper find most resonance in two of these: the nature and process of the negotiations and the nature of the peace agreement.

Scholars have debated whether negotiations and their resultant peace agreements need to address root causes of the conflict or create a space in which the political actors can work out the fundamental issues through some sort of formal process that the peace agreement creates (often through a transitional government that would decide the new rules of the game).3 Is it too difficult to tackle the basic problems that created conflict during the negotiation process and better to focus on securing a negative peace? Or would this strategy merely push resolution of the key problems to a phase where internal forces would dominate and international atten- tion would be less focused, thus only postponing a difficult process to a point where the local power holders would dominate? Are peace agree- ments more likely to fail because of power politics or because they failed to address root causes?4

Nature of the Peace Negotiations The fundamental challenge facing negotiators during the process of a

negotiation or mediation effort is the dual issue of whether peace talks should address the root causes of a conflict or simply attempt to facilitate a cease-fire and allow domestic parties to manage their fundamental differences after the fighting ends. Often this dynamic is viewed in either-or terms, though many experienced negotiators would argue that both need to be considered simulta- neously. This is not an easy feat to accomplish, however, for several reasons. Operating from the outside, international actors may not be positioned to address certain types of fundamental causes of conflict in the short term, or

Peacemaking in Burundi 243

they may consider halting the conflict and addressing humanitarian consider- ations the top priority. Operating from within, parties to the conflict may not be willing to allow discussion of certain fundamental issues. When this is the case, if those parties are still strong at the point of negotiations, they can pre- vent fundamental issues from being placed on the agenda. This is typical of situations like those found in South Africa and in Burundi, where multiple rounds of “talks about talks” were necessary to set the acceptable agenda for substantive peace negotiations.

Experienced mediators argue that both aspects have to be addressed: the crises that finally motivated intervention and/or third-party mediation and the fundamental causes motivating the conflict. As Laurie Nathan cogently argued,

Adding to the burden of peacemaking, the mediator has to focus both on the deep-rooted causes of the conflict and on the crises that arise. The causes may be structural (e.g., the absence of strong institutions of governance); political (e.g., authoritarianism and discrimination); historical (e.g., the colonial demarcation of borders and divide-and-rule policies); and socioeconomic (e.g., deprivation and underdevelopment). . . . Where countries in conflict have weak institutions of governance, the government might be unable to provide credible negotiators who can deliver on their promises, it might lack control over its security forces, and it might be unable to implement the agreements it signs. It seems painfully obvious that deep-rooted conflict cannot be solved quickly or easily. Neverthe- less, mediators and donor governments frequently make the mistake of seeking a quick fix. They have honourable intentions, wanting to stop the destruction and suffering and to provide safe space for humanitarian operations and recon- struction, but they underestimate the complexity of the conflict, overestimate their powers of persuasion and ignore the psychopolitical dynamics of violence. Flouting the imperative that the parties and their constituencies must own the settlement, they push hard for rapid results. This approach can be distinctly counter-productive.5

This passage points to the fundamental contradiction between the desire for short-term solutions that would enable the international community to quickly extradite itself from a situation and the long term, deeper structural issues that have to be addressed in order to prevent conflict from reoccur- ring. The desire to end suffering quickly does not necessarily fit the need to tackle deeper issues motivating the conflict. When a negotiation process is facilitated by third-party actors interested in stabilizing the situation by securing a negative peace it may therefore focus on the superficial level and intentionally avoid opening up consideration of the deeper issues fomenting conflict.

Regardless of the motivation, the result is that there is a choice in negoti- ations strategies: facilitators can focus on process, on getting participants in the talks to agree on a few basic principles and establish procedures for

244 Jessica Piombo

solving their differences once the mediation is over. Alternatively, third-party negotiators can attempt to get participants to delve into the issues that spurred the conflict while they are present and able to help mediate between the parties, a process that is more likely to uncover root causes and to secure the commitment of international actors to helping to resolve those fundamen- tal issues.

Nature of the Peace Agreement What the final text of peace accords includes is also a decisive element in

how the agreements shape future developments, and it is obviously connected to whether negotiations focused on process or content. There is a great deal of evidence that even if difficult to raise during the negotiations phase, the exclusion of root causes from peace talks and negotiations can derail a peace process at multiple points in the cycle.6

When certain issues are too sensitive to include in a negotiations process, even after multiple rounds of agenda-setting meetings, a common approach is for the peace talks to focus on rules and processes. The resultant peace agree- ment most often establishes certain core principles and aims to set up a domestic/transitional government that would then be responsible to process the grievances that caused conflict in the first place. This strategic choice has obvious positive impacts but could also extend conflict in several ways. First, the exclusion of certain key issues could alienate those parties who are most driven by them, compelling these parties to remain outside the negotiations processes where they become spoilers. Second, while focusing on process and conflict termination over root causes may increase participation and facilitate the conclusion of the third-party facilitated negotiations phase, delaying the consideration of fundamental causes leaves the most difficult issues to be sorted through by the agents most personally invested in the peace process and new government. This tradeoff may promote ownership but has the potential downside that the subsequent government is likely to conduct this business outside the lens of international attention. Not only does the absence of international attention make reneging easier, but it might also subject the entire discussion to a credible commitment problem and cause a security dilemma for the weaker parties; they most likely will not trust the stronger parties to openly discuss the real issues motivating a conflict.7 This lack of trust and confidence in domestic processes may then undermine the entire effort.

After peace negotiations are completed and the implementation phase has begun, more problems may arise. Guelke argues that negotiated settlements may derail not only because of what is allowed into discussion but also due to power dynamics that arise during implementation.8 He finds that settlements that do not rest on a normative foundation that is separate from the power

Peacemaking in Burundi 245

politics of the conflict are unlikely to prove durable. Therefore, avoiding con- sideration of root causes can often lead a peace process to break down, most often because of spoilers that arise during the implementation phase. Spoilers can also be created during this phase as groups who are dissatisfied with out- comes of the talks use violence to try to push their issues back onto the agenda.

These considerations compel analysis to focus attention not just on issues of content versus process but also to consider which issues are put onto the negotiation table: power politics and concerns that exist at the point the nega- tions are initiated or the basic root causes that caused conflict in the first place.

CONFLICT IN BURUNDI

The case of the Burundian civil war of the 1990s provides a lens through which to examine the dynamics just discussed. Burundi has experienced cycles of intense violence since achieving independence in 1962, a seemingly endless number of coups and attempted coups, and one major civil war, between 1993 and 2009, when the last significant rebel group ceased hostili- ties and began to transition into a political party (it had signed a cease-fire in 2006).9 International actors attempted to bring an end to the latest manifesta- tion of violence within a month of its initiation, and they remained engaged throughout the conflict. Various parties to the conflict signed multiple peace accords and cease-fire agreements during this time. At the conclusion of a four-year transitional government, the Fourth Republic came into existence in 2005, and elections for the national legislature and president took place in mid-2010.

The issues underlying this cycle of conflict in Burundi are rooted in the country’s history and the political manipulation of traditional authority and ethnic relations. Similar issues spurred violence in 1965, 1972, 1988, 1991, and 1993–2009. In each of these periods, violence was triggered by an attempt to change the social composition of the ruling group, with attacks generating retaliatory counterattacks that escalated violence and further altered the character of the government and security forces. Over time, a fluid political system where clan and regional rivalries dominated politics evolved into a sta- ble system of ethnic domination, with a small Tutsi elite in control of an authoritarian political system buttressed by an ethnically exclusive, repres- sive security apparatus. As a result of three decades of conflict, certain issues became almost nonnegotiable to the power elite, yet were such fundamental issues for the opposition that in order for the conflict cycles to be terminated, the issues had to be addressed.

Despite the differences in the violence and triggers for each episode, the underlying causes and the repercussions not only were the same, they also

246 Jessica Piombo

increased in intensity with each conflict cycle. Hutu periodically attempted to overthrow the Tutsi government in 1972, 1988, and 1991, and each time the same cycle repeated: a Hutu insurgency or coup attempt would inflict initial casualties on Tutsi, followed by large-scale Tutsi retaliation against Hutu. In the new equilibrium following each episode, the Tutsi increased their hold on power, institutionalizing minority rule a bit more each time and creating a more ethnically homogenous military in the wake of each purge. Each of these factors later became one of the root causes of the civil war: Hutu oppression and the increasing institutionalization of Tutsi hegemony; a repressive and ethnically exclusive military that operated to protect the interests of the power elite; and the increasing economic and edu- cational inequalities between Hutu and Tutsi. The early conflicts were never fully resolved, and the ways that the power elite settled the violence in each period exacerbated the fundamental causes that underlined the next round of violence.10

Of all these root causes, the most fundamental was that over the decades of episodic conflict, the government, military, and gendarmerie had become almost entirely Tutsi. Additionally, each successive coup and government reorganization consolidated power under the head of state, creating ever more autocratic and powerful executives. As a result of these changes, the security structures lost all independence and were turned into agents dedicated to uphold- ing the Tutsi ethnocracy. Burundi experts refer to this as the “Tutsification” of the military. The military also assumed an almost entirely inward focus during these cycles, working with the national police to ensure internal security.

Once the corporate identity of the military evolved into being the protector of the Tutsi oligarchy, the composition and mission of the military became in and of itself a cause for future conflict. Military structure, composition, and mission later emerged as one of the core issues that kept the two largest rebel groups fighting long after the rest had signed peace accords. On their side, any attempt to liberalize the polity and/or instill some degree of ethnic parity in the military became threatening to the military’s identity and core mission. This was especially evident in the 1991 conflict cycle.

Finally, each cycle of violence eliminated more Hutu politicians and edu- cated people from the Burundian population in general, who were lost either to assassinations/executions or to refugee flows. Changes to the educational system ensured that only Tutsi would gain sufficient education to enter the civil service, so that over time most of the lucrative government jobs were held by Tutsi. Once the educated Hutu were either killed or driven into exile and those that remained subjected to educational restrictions under Lt. Col. Jean Baptise Bagaza’s twelve-year rule (1976–1987), the Hutu population within the country became increasingly undereducated and unqualified in comparison to the Tutsi. Lucrative, well-paying skilled and government jobs increasingly went

Peacemaking in Burundi 247

to Tutsi, creating an ever-widening gap between the communities. Govern- ment patronage networks excluded Hutu, further enriching Tutsi and widen- ing the gap between the two groups. While the Hutu therefore sought to advance their socioeconomic status and considered this a rightful redress for decades of skewed policies, Tutsi were threatened by this quest.

Because the root causes were made worse in the wake of each conflict cycle rather than addressed in a constructive way, over the course of this 30 year conflict cycle the Hutu radicalized and came to see violent resistance as the only way to achieve power. Multiple movements formed in exile, the largest of which was Palipehutu, which generated and then merged with an armed wing, the Forces Nationales de Libération (National Forces of Libera- tion, FNL), creating Palipehutu-FNL. The repeated cycle of violence created a collective memory within the Hutu population, both those within Burundi and the refugees abroad, cultivating an ever-present fear of another holocaust. For Tutsi at both elite and mass levels, any attempt to liberalize or accommodate Hutu demands for political, economic, or social power would inevitably threaten the Tutsi as a group; therefore, the Tutsi became increasingly wedded to upholding the system that was exacerbating the root causes of Hutu unrest.

In this situation where the root causes of conflict were never addressed— and could never be addressed without fundamentally altering the structure of government, military, and society—Hutu increasingly came to understand that they would never share power peacefully. Therefore each cycle of violence increased the perceptions of injustice felt by the Hutu, both those who remained in the country and those in exile. With each wave of violence, the exile community increased in numbers and radicalization. On the other side, the Tutsi allied their interests with control of the state more and more, and they came to fear the almost inevitable backlash that would occur if the Hutu ever earned any genuine power.

TRIGGERS OF THE CIVIL WAR

Against this background, the Buyoya regime (1987–1993) initiated and com- pleted a political transition that culminated in the Hutu party Front pour la Démocratie au Burundi (Frodebu, Front for Democracy in Burundi) sweeping provincial and national elections in June 1993.11 Frodebu leader Melchior Ndadaye was sworn in as the president in July and formed a cabinet with sig- nificant Tutsi representation. The government attempted power sharing in political appointments and civil service positions and had plans to institute “ethnic parity” in the military as well. Shortly after the new government began to function, however, a faction of the Burundian army assassinated the new president and senior Frodebu leaders during a coup attempt in October. Hutu and Tutsi both engaged in reciprocal attacks, with a death toll that

248 Jessica Piombo

quickly reached into the thousands. The coup itself collapsed after two days, and the chief of staff of the army ordered the military back to the barracks, though Uprona and other parties that had lost in the elections supported the putsch and opposed the international peacekeeping force that the government had requested.12 These events triggered significant unrest throughout the country and the start of the civil war as more groups joined the fighting on each side. Despite repeated efforts at power sharing and another political transition to a civilian government in 2005, the fighting did not completely end until January 2009.

The underlying causes of this war built on the foundations laid since 1965. Prompted by pressures from the international community in the wake of the 1988 massacres and realizing that there would be no end to conflict if things did not change, then-president Buyoya initiated a transition to a new political dispensation. Essentially, the quick unfolding of the transition between 1988 and 1993 threatened Tutsi elites and masses, many of whom supported the coup in October 1993. The rapid pace of reforms stoked Tutsi fears of retalia- tion and repression if the Hutu government were allowed to truly rule. Because multiparty politics had been opened, when the coup-makers unseated the government there were Hutu political organizations operating within the country, not just in exile. These parties linked communities of Hutu internally to those spread throughout the region, most critically the radicalized refugee communities, which were capable of mounting armed resistance.

Roots of the Crisis: The Transition Period, 1988–1993 Following the conflict in June–August 1998, Buyoya inaugurated a

“National Commission to Study the Question of National Unity” and incorpo- rated Hutu into governing structures. The commission published a report that led to sweeping changes, the promulgation of a new constitution, and plans to transition to an elected civilian government. According to Burundi scholar Filip Reyntjens, genuine change had occurred by the middle of this period: Hutu were brought into government at all levels, provided with better educa- tional opportunities, and incorporated into the civil service.13

Tutsi social, economic, and political elites found the pace of the changes too drastic, which created a security dilemma for the Tutsi, felt most acutely by those in elite positions within the government and military. Fears focused most keenly on civil service composition, the potential for land reform, and military restructuring.14 Tutsi feared losing civil service appointments to the hordes of Hutu that they anticipated would return from abroad, while reforms to address economic inequality and land reform threatened to return land to Hutu who had lost their possessions in the wake of the 1972 genocide. There would be an inevitable shift in the distribution of national resources as the government incorporated more Hutu. The situation worsened toward the end

Peacemaking in Burundi 249

of the five-year transitional period, when the pace of refugee returns increased following the Tanzanian government’s decision to force refugees to repatriate to Burundi or to become Tanzanian citizens.

Tutsi masses also feared that pursuing ethnic parity in the military would remove the only guarantor of their physical security, as they regarded the mili- tary their only protection from violent domination by the Hutu majority. Across the Tutsi, from the elites to the masses, a very strong collective fear of the future had manifested. As Sullivan noted, “the Tutsi fear was that if control of the army were given to the Hutu, or even if a move were made in that direction, it would result in the ‘Rwanda syndrome,’ referring to the targeting and mar- ginalization of Tutsi by Hutu in Rwanda after they gained power in 1959.”15

In fact, to many Tusti, the Rwanda syndrome seemed imminent once Frodebu won over 80 percent of the seats in the National Assembly, the presi- dency, and control of almost all regional governments.16 Ndadaye’s plans to implement significant structural change in the composition of the officer and enlisted ranks of the military also generated significant backlash. Entirely under the command of Tutsi officers, most of them from Bururi (a region within Burundi historically associated with internal power rivalries within the Tutsi community), and with only a fraction of Hutu among the troops, the army served as the lynchpin of Tutsi hegemony. When the new president unveiled plans to restructure the military and reorient its focus (toward exter- nal rather than internal activities), it triggered an intervention by certain elements in the military.

On the other side of the fence, the coup attempt further demonstrated to those involved in the Hutu uprising that the system would never change unless several fundamental issues were addressed. The core demands of the Hutu upris- ing in the fall of 1993 were therefore that either Burundi or the international community disband the Tutsi army and rebuild a national army and that an international military force deploy to end the ethnic clashes around the country. Politically, these groups wanted a restoration of the 1992 constitution, and the government that had been elected in June 1993. Finally, they desired to promote Hutu political participation, address resource scarcity, and demilitarize society.17

Some have argued that on top of all these dynamics, the 1993 elections were held too soon and forced on the Burundians by international negotia- tors.18 As a result, when Frodebu polled so well, despite government attempts to prevent the party from fully campaigning, neither side was ready to accept defeat; the situation could only have led to bloodshed.

Posttransition: Failed Power Sharing and Launch of Full-Scale Insurgency After the assassination of Frodebu leaders, the Organization of African

Unity (OAU) sent in a stabilization force and the United Nations (UN) led a

250 Jessica Piombo

mediation effort between Frodebu and Uprona. UN Secretary General Boutros Boutros-Ghali appointed Ahmedou Ould-Abdallah as the Special Representa- tive to the Secretary General (SRSG) in charge of the mediation efforts. Ould- Abdallah brokered an agreement between Uprona and Frodebu, creating a power-sharing government called the “Convention of Government” (September 10, 1994).19 Rather than restoring Frodebu to power, however, the Convention of Government created a system that shared power equally between the Tutsi (Uprona) and Hutu (Frodebu) parties, a compromise with which many Hutu both in government and outside were not happy.

In June 1994, Frodebu split into moderate and hard-line factions: respec- tively Frodebu and the Conseil National Pour la Défense de la Démocratie (CNDD, National Council for the Defense of Democracy). Led by former Minis- ter of the Interior Léonard Nyangoma, the CNDD broke away from Frodebu when it agreed to allow the abrogation of the 1992 constitution and on the prin- ciple of power sharing with Uprona. The CNDD opposed this level of coopera- tion with Uprona and all participation in the convention government and committed itself to winning back power through violent means. The CNDD soon merged with its armed wing, forming the CNDD-FDD (Conseil National Pour la Défense de la Démocratie–Forces pour la Défense de la Démocratie; National Council for the Defense of Democracy–Forces for the Defense of Democracy).20

The CNDD and Palipehutu-FNL both opposed the power sharing govern- ment and became the two largest militant groups opposed to the aborted coup of 1993. There were multiple, smaller, rebel groups that had formed in exile or that had militarized after the elections were overturned. Collectively, these groups did not agree with the basis of the mediated government and felt that the Hutu had once again been deprived of their rightful place in Burundian politics.

Between September 1994 and July 1996, the convention government attempted to rule the country in the face of this ongoing insurgency while the Burundian Army (FAB) waged war against the rebel movements. By 1996, the situation had deteriorated, and the government was unable to stabilize the territory. In that year, Buyoya initiated another coup, overthrowing the convention government and returning the country to military rule. Fighting continued to rage, sending refugees back into neighboring countries and destabilizing the entire region.

THE MEDIATION PROCESS

Since the fighting began in 1993, Burundi has gone through over five separate negotiations periods. This section will focus on the Arusha process, as this was the most comprehensive and enduring of the processes and culminated in the pri- mary peace accord, the Arusha Peace and Reconciliation Agreement for Burundi (the Arusha Accord). While culminating in the agreement that established a

Peacemaking in Burundi 251

transitional government and led to the installation of a permanent govern- ment in 2005, the Arusha process had significant flaws. First of all, Arusha began talks without a cease-fire, which enabled rebel groups to utilize violence as a negotiating tactic. Neither of the two principal rebel groups were involved in the process, and neither signed the Arusha Accord. While the signing of the Arusha Accord was a necessary first step, the exclusive process created a need for separate negotiations with the main rebel groups. Second, the Arusha framework focused on process over content, leaving the main questions about the design of the future government and military to the transitional regime that was selected in 2001.

Regional Framework Arusha Talks, 1996–2000 The real process of peace negotiations began with the regional initiative

to bring the warring parties together under the rubric of the Arusha Talks. The formal multiparty talks took place primarily between 1996 and 2000, first facilitated by Julius Nyrere and then former South African President Nelson Mandela (following Nyrere’s death in October 1999). The formal peace accords were signed in September 2000, without a cease-fire agreement and without the participation of the two main rebel groups at that time, the CNDD-FDD and Palipehutu-FNL. These peace accords focused on laying out principles for the creation of peace in Burundi and on establishing the struc- tures for a transitional government that would oversee the writing of a new constitution and the establishment of a permanent, civilian government. As such, the peace negotiations focused on basic principles and processes rather than on consideration of the deeper issues motivating the crisis. Founda- tional issues were brought in at the later stages, but were not the centerpiece of the talks. The negotiations did not aim to solve the problems in the frame- work of the talks but to create a subsequent government that would address the issues.

The first round of talks, in March–April 1996, involved the mediation team, Frodebu, and Uprona. The aim of this session was to engage the two principals in talks about talks, mainly to get them working with each other. However, these talks quickly collapsed because of the ongoing fighting between the Hutu militia groups and the Burundian army. Frodebu refused to comply with Uprona’s demand that Frodebu condemn the ongoing attacks, which led Uprona and the military to accuse Frodebu of complicity with the militants (primarily the CNDD-FDD). On its side, Uprona rejected Nyrere’s proposal that the government should negotiate with these armed groups.21

The problems of this first meeting pointed to issues that over the following four years deeply affected the peace process: whether groups still fighting should be included in the talks before signing a cease-fire, factionalism within the various groups and how to deal with splinter groups once they arose, and

252 Jessica Piombo

the relationship of parties in the talks to those that remained outside the negotiating framework.

Later in June, the first Regional Heads of State Summit convened in Arusha. This meeting included heads of state from Tanzania (President Benjamin Mkapa) and Uganda (President Yoweri Museveni) as well as Frodebu, Uprona, and several smaller parties. These talks focused on techni- cal issues rather than substantive, such as getting to a cease-fire to allow sub- stantive talks, regional security assistance, and mechanisms to restore and prevent further violence.22 This session spent a great deal of time discussing the possibility of mobilizing an international intervention force, something that Frodebu had been requesting since 1993, but which was finally gaining traction from the regional neighbors.23 Mkapa and Museveni in particular pressured the Convention government of President Sylvestre Ntibantunganya (Frodebu) and Prime Minister Antoine Nduwayo (Uprona) to request an international force. Ntibantunganya welcomed the creation of the force, and while Nduwayo initially bowed to pressure and agreed, he later publicly condemned the idea. According to Mthembu-Salter, this pointed to another recurring dynamic in the negotia- tions processes: the mediators were extremely heavy-handed and forced the par- ticipants through negotiations processes and into agreements, whether or not there was truly any consensus or minimum agreement.24

The Arusha summit set the tone for much of what followed during Burundi’s peace process, with Burundian political representatives being pressured by regional heads of state into highly controversial agreements that it was then left to them to sell to their profoundly skeptical grassroots communities. In this instance, the request for international assistance was welcomed by Frodebu and most other Hutu parties, but vehemently condemned by Uprona and other Tutsi parties. Once back in Burundi, Nduwayo responded to the condemnation by shift- ing his position at a rally and accused Ntibantunganya of a secret agenda to neu- tralise the Burundian army.25

The prospect of this intervention force and the opposition from both Uprona and the Burundian military is considered the final trigger for the coup that Buyoya himself launched in early July.26

Following the coup, the third round of talks (the second Heads of State Summit) opened at the end of July 1996. This session was not a facilitation of talks between the parties to the conflict; instead, this was a meeting of regional actors who had gathered to try to utilize the new network they had formed to reduce the intensity of the fighting. During this meeting, the regional actors decided to impose economic sanctions to restore constitutional order and legitimacy. The Buyoya regime launched a counterinsurgency effort against the rebels and condemned the sanctions.

The fighting continued throughout the rest of 1996 and 1997 without any genuine negotiations that involved the parties to the conflict. The next Arusha

Peacemaking in Burundi 253

framework talks occurred in April 1997. Like the July 1996 meetings, the third Heads of State Summit involved regional actors and again attempted to influence the Burundian conflict through external incentives. Since the fight- ing had not abated and the sanctions had soured relationships between the regional actors and the Burundian military and Uprona, the representatives at these talks decided to ease the sanctions. Relations between Tanzania and Burundi had greatly deteriorated because of the sanctions, which also weak- ened Nyrere’s position as lead negotiator.27 “Tutsi politicians . . . roundly con- demn the initiative as a violation of sovereignty and akin to a declaration of war and most see it as proof of the hostile intent of Nyrere and the Tanzanian government. Many have commented that since then they never again viewed Nyrere as a neutral mediator, but instead as a party to the conflict.”28 Nyrere offered to step down as lead facilitator in May 1997, but the regional actors refused to accept his resignation.

Fourth Arusha Talks: June–July 1998 Substantive talks about resolving the crisis only began in June 1998,

when the parties finally returned to the negotiating table. Nineteen delega- tions participated in the Arusha talks, seventeen from political parties, and one each from the government and national assembly. These talks, however, almost derailed because the CNDD and Palipehutu split into political and mil- itary wings on the eve of the negotiations, which caused a crisis about which of the groups to include in the talks.

Nyrere gave the factions three options: reconcile, put just the old leaders forward, or put the new leaders forward. But this led to deadlock: both camps in each formation considered themselves to be the true voice of the group and would not compromise, while the Burundian government opposed the partici- pation of any groups that refused to cease fighting. Ultimately Nyrere decided that the original leaders, at the head now of political organizations shrift of their armed wings, would represent their parties at the talks.29

Therefore, right from the beginning, the issue of inclusion created serious problems for the talks, a classic situation that causes negotiations to create spoilers. There was a genuine legitimacy issue regarding which camps were represented and which were excluded from the process, and this generated a climate of mistrust that poisoned the discussions. “Keeping Jean Bosco’s faction [the breakaway, armed wing of the CNDD-FDD] outside the formal negotiations has undermined the entire process: the climate of suspicion and mistrust between all parties, and within the parties themselves, has increased.”30

Nyrere’s decree also excluded any groups waging armed conflict, and so CNDD-FDD and Palipehutu-FNL remained outside the negotiations frame- work through October 1999, when the facilitation passed to (former) President

254 Jessica Piombo

Nelson Mandela. Mandela attempted to bring the two into the talks, but for various reasons relating to the weakened position they would find themselves in and their inability to shape the agenda at that point, both remained outside the process.31 The CNDD-FDD and Palipehutu-FNL continued to fight, while the other groups in the conflict attempted to resolve the situation. The final Arusha protocols that were later signed in September 2000 did not include either group, so the South Africans continued separate talks with each group for years after the main peace accord was promulgated.32

Once the talks resumed in mid-June, Nyrere’s facilitation team split the discussions into five separate committees that were tasked with specific topics and agenda items.33 These were, (1) nature of the conflict—this committee was responsible for creating a document that outlined the nature of the con- flict, the problems of genocide and exclusion, and proposing solutions to these; (2) democracy and good governance—to develop strategies for democracy and good governance; (3) peace and security—to devise strategies for achieving peace and security for all; (4) reconstruction and economic development—to develop measures for economic reconstruction and development; and (5) guar- antees for the application of a peace agreement. By splitting the work into smaller, functionally focused groups, the facilitators were able to focus the talks within each issue area and to break down complex issues into manage- able pieces. This framework also helped to neutralize critiques of the domi- nance of Tanzanians in the process.34

Content versus Process These committees were establishing interim, civilian-controlled structures

that would oversee the final transition back to civilian rule. Where there were substantive discussions, therefore, they applied mainly to the formation of the transitional government rather than the permanent one. There are indica- tions that both the decision to focus on technical arrangements and principles during the negotiations and the choice to task subsequent institutions with the burden of responding to fundamental causes may have been the expedient course of action due to the intractability of certain fundamental issues. As dis- cussed earlier, when an issue cannot be discussed constructively in the con- text of the negotiations, the parties to the talks may decide that it is better to postpone the resolution to a later stage.

This is exactly what happened in several of the committees in the Arusha discussions held during the summer of 1999. Based on participant observation of the meetings in June–July 1998, Elisa Renda noted that when the peace and security committee attempted to set the agenda, delegates got so mired in arguments about what to include that the group made no progress in three weeks of discussions.35 When the group tasked with discussing the nature of the conflict experienced a similar problem—participants could not even agree

Peacemaking in Burundi 255

on the definition of genocide—the facilitators shifted the entire conversation to solutions, as this was apparently easier to discuss than the nature of the conflict itself.

Not only did the facilitators shift the conversation, they also shifted the burden for the ultimate resolution to structures that would be established later in the peace process. Renda argued that the facilitators and members of the committee on peace and security agreed that since they could not agree on these basics, they deemed that the best way to proceed was to “give an independent institution the responsibility of unveiling the coun- try’s bloody history.” Similarly, the committee on democracy and gover- nance could not resolve justice issues, and therefore “agreed with the UN Security Council that an international Commission of Judicial Enquiry and the creation of a National Commission of Truth and Reconciliation were the only solutions.”36

These choices may have been necessary to move the negotiations process along, but they came with a price. Two core issues—an honest discussion of the nature of all the conflicts that had preceded the 1993 war and the nature of how past injustices would be addressed—remained fundamental concerns that motivated Palipehutu-FNL to stay fighting long after every other Hutu movement had lain down arms. Eight years later, when Palipehutu-FNL finally engaged in cease-fire negotiations, the organization put both of these issues on the agenda.

Unfortunately even with these coping mechanisms the committees made little progress between June 1998 and January 1999. It was another nine months before any committees showed genuine progress toward concrete pro- posals. When they did draft proposals, these were for the transitional govern- ment, which would then oversee a final transition and drafting of a new constitution. At the time of Nyrere’s death in October 1999, draft agreements had been reached on the necessity of the formation of a democratic regime, the basics of an electoral system for the future, and that a Truth and Reconciliation Commission should be established.

Mandela Takes Over, October 1999 Nyrere’s death opened up a window of opportunity for the Burundian gov-

ernment. It had long been trying to remove the negotiations process from regional control, particularly that of Tanzania. On December 1, 1999, a regional heads of state summit selected former South African President Nelson Mandela as the lead facilitator. Mandela immediately took the negoti- ations in a different direction. His approach was much more heavy-handed than had been Nyrere’s; Mandela personally met with all the parties to the conflict, including the CNDD-FDD and Palipehutu-FNL, and he openly admonished all sides whenever they prevaricated.

256 Jessica Piombo

When Mandela approached the situation, it was without the traditional assumptions and interests of a regional actor. Mandela drew parallels between Burundi and South Africa during apartheid and South Africa’s sub- sequent negotiations process, and used the South African experience as a model for the rest of the negotiations. This influenced his approach: he tended to treat Uprona and Frodebu as the two main negotiating partners in the pro- cess, the way that the National Party and African National Congress had been the real focus of the South African negotiation in the early 1990s. Also, Mandela frequently and publicly likened the Tutsi to white South Africans, drawing parallels between a 15 percent minority population systematically repressing and disenfranchising the 85 percent majority population. While this provided a fresh lens to the negotiations process, it also alienated some Tutsi elements, who resented what they saw as Mandela’s dogmatic mapping of the South African situation onto Burundi.37

Mandela also focused on inclusion, arguing that all stakeholders in the conflict needed to be at the negotiating tables. He made great efforts to bring in CNDD-FDD and Palipehutu-FNL into the process, meeting with the lead- ers of the movements. Nyrere’s exclusion of these groups had continued to threaten any progress made, but including them at this point was also precar- ious. Not only could the CNDD-FDD threaten Frodebu’s claim to be the most legitimate representative of the Hutu posttransition, there was also potential for rivalry between the Hutu politicians who had been participating in the negotiations all along and the armed Hutu rebels, who claimed to be appropri- ate personalities to negotiate security sector reform.38

Many characterize Mandela’s approach as almost authoritarian, noting that he doggedly focused on process, timelines, and sticking to both.39 In order to help participants stick to the final deadline, Mandela “prebooked” interna- tional actors, like U.S. President Bill Clinton, to attend the peace accord sign- ing ceremony that was set for September 2000. He then used the public embarrassment that would be caused to prevent parties from delaying and to attempt to force the process forward. Jan van Eck even called Mandela a bully, arguing that he used his international stature and moral force to push the process past domestic opposition, which van Eck felt no other mediator could have done.40

In terms of process, Mandela’s facilitation team divided the participants into three camps: an extreme pro-Hutu, extreme pro-Tutsi, and a moderate camp. The logic in dividing the groups this way was to allow each group to come up with its own proposals and then through a series of bilateral negotia- tions between the facilitators and each group, to bring the outlier and more radical proposals closer to the position of the moderate camp.41 This technique was designed to sideline extremist elements in the talks and focus on technical issues that prevented grandstanding and politicking. Mandela’s team also removed the five-part committee structure as a way to generate proposals.

Peacemaking in Burundi 257

Instead, his team worked with Burundian parties in the three groups to draft an agreement, which he then presented to all the parties in July 2000. Mandela gave the groups three weeks to come up with counterproposals for the draft, and through this process they worked on revising the draft docu- ment and coming up with the final peace accords that nineteen parties signed in August and September 2000.

Content versus Process Analysts of the Burundian peace process laud the protocols signed in

September for advancing the peace process and inaugurating the transitional government, but note that several aspects of the process generated ripple effects that, further on, endangered the transition. Several of these have already been mentioned, so here we will focus on those that have not been discussed.

Jan van Eck, one of the South African facilitators, has been particularly critical of the Arusha Accord, and of Mandela’s approach in particular. Van Eck faults the Mandela phase of negotiations for violating all the principles of peacemaking that had been developed over decades of practice and research: inclusivity, compromise, ownership of the process, homegrown solution, deal- ing with root causes, and reconciliation. While Mandela himself asserted that this flaunting of the rules was what allowed him to conclude the talks on schedule and with a signed peace agreement, van Eck felt that this may not have allowed for airing of grievances. Many parties felt forced to sign the final accord and did not have time to dialogue about the provisions of the accord.42

Of these principles, the lack of ownership and avoidance of root causes seem to have created the most follow-on difficulties that the peace process encountered after Arusha. For one, a significant number of groups, both sig- natories and nonsignatories, subsequently voiced vigorous opposition to many provisions in the agreement. A number of Tutsi parties had initially refused to sign but under intense pressure were convinced to endorse the accord in early September before the final international ceremony. Simi- larly, the two main rebel groups capitalized on opposition to the agreement to sustain their fighting; the CNDD-FDD acceded to the accord only after it signed the cease-fire agreement in 2003, and one of its splinter factions held out until 2005.43

These groups were then able to use their “principled” stand against the Arusha protocol as campaign tools in the transitional elections in 2005. The CNDD-FDD, for example, capitalized on its opposition to the Mandela phase and the text of the agreement as campaign platforms and used this as a wedge issue to defeat Frodebu in the elections held between June and August 2005. By distancing itself from the Arusha framework negotiations and the final docu- ment, the CNDD-FDD was able to position itself as a more legitimate political

258 Jessica Piombo

force that stood up to its ideals and resisted pressure from the international community.

Overlooking root causes and forcing the process also extended the conflict in this situation. The exclusion of major and minor rebel groups, and then their dissatisfaction with the process-oriented agreement, caused the process to drag on for almost an entire decade more, requiring near constant side negotiations with multiple parties outside the formal talks (1996–2000). This lack of inclusion prolonged both the war and the peace process.

Finally, the accord itself has been faulted for being short on detail and limited in scope.44 The facilitators and main actors engaged in two days of intense negotiations before the signing deadline of August 28 in order to resolve unfinished business, which either resulted in multiple compromises that had not been negotiated or briefed to the wider audience or in incomplete protocols with no substantive agreements.45 Only three of the protocols were complete; the rest were signed on the date but without specifics as the parties could not come to any agreement. The completed protocols pertained to the nature of the conflict, principles of democracy and good governance to be included in a posttransition constitution, and reconstruction and develop- ment. Critically, the negotiating teams were unable to devise concrete propos- als related to the security sector due to the absence of the main rebel groups and the sensitivity of the issues. The protocol assigned the organization and function of the future defense force to national legislation posttransition.

Many have argued that Mandela forced Burundian politicians to sign an incomplete agreement, while Lemarchand criticized the agreement because “much of the agenda inscribed in the protocols did not go beyond the stage of pious intentions.”46 He noted that six years after the protocol had been signed, most had yet to be implemented, outside of the areas of democracy and gover- nance and the protocol on peace and security. The transitional arrangements that Arusha created ignored several root problems fomenting Burundi’s con- flicts: lack of economic opportunities, regionalism, and the role and function of the military.47 None of these issues were incorporated into the interim govern- ment created by the Arusha agreement. Had the agreement been more focused on substance and less on technical arrangements and deadlines, some of these deficiencies could have been remedied.

Despite the drawbacks, there were undeniable positive achievements that did reflect some attention to root causes. The ICG praised the “clear and pre- cise action programme” outlined in the three completed protocols of the agree- ment—on democracy, reconciliation, and reconstruction.48 The transitional structures set a pattern that the constitution drafters elected to retain in the final constitution, and the parallel cease-fire negotiations with the CNDD-FDD helped to create a security establishment that finally included Hutu as well as Tutsi (though the Tutsi remained dominant). Lemarchand concedes that institutionalizing ethnic parity within the army may have been the “most

Peacemaking in Burundi 259

remarkable achievement of the Arusha conference.”49 Other scholars argued that the overrepresentation of Tutsi was not a drawback: while it would have been truly equitable to implement genuine ethnic proportionality in the military, this would have been too much for the Tutsi to accept.

One final aspect of the Arusha agreement that bears on the conflict resolu- tion versus termination debate (addressing root causes or pursuing a negative peace) is whether a cease-fire should have been signed before the peace accords. The Arusha process was supplemented by parallel negotiations with warring parties to obtain a cease-fire arrangement. This left powerful spoilers, organiza- tions with substantial military power outside the political process. Thus, they were in a position to unleash violence to gain concessions in the political arena when the agreements and processes did not suit them.50 The CNDD-FDD and Palipehutu-FNL would not agree to participate in the political process for all the reasons outlined previously, several of which were root-causes issues.

The trade-off chosen by the mediators was to push ahead and focus on one set of root causes, the political process (thereby addressing a root cause), while excluding others, such as the nature and composition of the military. Facilita- tors could not force the spoilers into the peace process, which extended the violent phase of the conflict for several years.

This then raises the dilemma once again: on the one hand, if root causes had been incorporated earlier, would the militant movements have agreed to stop fighting and joined the process? This suggests attending to root causes before end state. On the other hand, the difficulties created by the decision to push ahead with the political process (a root cause) before resolving ongoing hostilities (an end state) suggests that at other times a conflict termination strategy has to be adopted in order to create the room for a true conflict reso- lution process, opening the space to address root causes.

Pretoria Sessions While deserving praise in many respects, the Arusha accord roundly failed

to curtail violence in the country. Violence actually spiked in the four months following the signing of the accord, and conditions worsened in the short term. “Since 28 August 2000, Burundi has experienced a dramatic resurgence of vio- lence, the peace process appears to be paralysed by the refusal of the main rebel groups to take part in it despite the efforts of the South African media- tors, and the daily life of Burundians both inside and outside the country has not been improved at all.”51 The ICG blames the violence on the way that the agreement was signed and the importance of the unresolved issues—the cease-fire and the transition process itself.

Cease-fire negotiations separate from the political discussions of the Arusha process had been initiated in August 2000. Initially with both the CNDD-FDD and Palipehutu-FNL, the talks with the CNDD-FDD were more

260 Jessica Piombo

fruitful than those with Palipehutu-FNL. Both organizations were pressing for issues relating to the demobilization and/or integration of their fighters into the political process and new security forces as well as to the basic struc- ture, role, and composition of the armed forces posttransition. Additionally, they were unhappy with the compromises enshrined in the Arusha agree- ment, especially those pertaining to the treatment of history, status of postconflict justice mechanisms, and the overrepresentation of Tutsi in transi- tional government structures and the posttransition military.

Little progress was made between August 2000 and February 2002, when the South African facilitation team requested Tanzania to help bring the rebels to the negotiating table. During this time, the transitional government and Burundian army were fighting the various rebel movements throughout the Burundian territory. Complicating matters, an internal division within the CNDD-FDD blossomed into a split, with the National Council of Patriots (FDD-CNP) establishing itself as a separate movement in October 2001. Accord- ing to the ICG, the FDD-CNP soon grew stronger than Palipehutu-FNL.52

The Tanzanian government and the CNDD-FDD factions began a dia- logue at the end of July. Palipehutu-FNL and some of the CNDD-FDD fac- tions did not consider the Burundian government selected at the beginning of the transition in November 2001 to be legitimate. Instead they asserted that they could negotiate only with the Burundian army, which was the real power in the country.53 Multiple rounds of talks were held during 2002, though no tangible progress in terms of defining an agenda and obtaining participation was made until the following year. In 2003, the CNDD-FDD factions finally agreed to facilitation by South Africans, led by Jacob Zuma. Palipehutu-FNL did not agree and only began to negotiate for a cease-fire later on.

The final sessions where substantive negotiations and progress took place occurred in three meetings, a Heads of State Summit in Dar es Salaam in September 2003, followed by South African-facilitated principals negotiations in Pretoria, South Africa, in September and November 2003. The Heads of State Summit in Dar es Salaam had derailed because neither the transitional government nor the CNDD-FDD factions would compromise on defense power sharing and military reform.54

When the parties met again in Pretoria, President Thabo Mbeki himself led the mediation effort. Mbeki used proposals that had been created for the September summit to structure the agenda. These cease-fire negotiations focused on the substantive issues that were keeping the CNDD-FDD out of the transitional process: political power sharing, defense power sharing, and security power sharing. This first Pretoria session produced the “Protocol on Political, Defence and Security Power Sharing in Burundi.” The document contained concrete proposals and formulas for the incorporation of CNDD- FDD members in each of these realms, proposed a name change for the military (creating the Burundi National Defence Force, BNDF/NDF), and

Peacemaking in Burundi 261

established power sharing in the police, gendarmerie, militia, and intelligence arms of the security sector.

The participants in the first Pretoria session made significant headway in resolving the issues that had kept the CNDD-FDD out of the process until this point. They were not able to come to agreements on temporary immunity for CNDD-FDD members or the status of the organization as a political party. These issues were therefore the focus of a follow-on meeting in November 2003, which resulted in two further agreements, again both substantive. The “Protocol on Outstanding Political, Defence and Security Power Sharing Issues in Burundi” and the “Forces Technical Agreement” (FTA) resolved the remaining issues. The Protocol on Outstanding Issues granted temporary immunity and agreed that the CNDD-FDD would qualify to register as a polit- ical party once FDD began demobilization. The two-part Forces Technical Agreement (1) specified the size, composition, structure, and organization of the new BNDF and (2) established the BNDF and provided for a disarmament, demobilization, and reintegration process that would unfold in a year-plus time line. Significantly, these provisions also articulated how the transitional government of Burundi would be restructured to incorporate CNDD-FDD members (four ministerial posts, one vice presidency, etc).

The detailed and comprehensive nature of these three agreements created a process that led to the successful disarmament and demobilization of the CNDD-FDD and its transformation into a political force. The organization did not have unresolved issues at the conclusion of these processes, and the fac- tions that had broken off were brought into the process in separate cease-fire agreements in 2005. The movement was able to demobilize and transform enough that when the final phases of the transition unfolded, the CNDD-FDD won control of the National Assembly, the presidency, and many of the provincial governments.

Cease-Fire Accord (CFA) with the FNL The situation for Palipehutu-FNL was much different than for the

CNDD-FDD. Palipehutu-FNL had split into three factions by 2003, which complicated the negotiations process, especially regarding demobilization and disarmament.55 Furthermore, once the FDD had signed its cease-fire, it had begun to cooperate with the Burundian army to fight the FNL. While negotia- tors for the CNDD-FDD cease-fire had often described the FNL as “a small problem,”56 it was nevertheless able to threaten the stabilization that Burundi so desperately needed. By late 2006, observers noted that “in spite of [the] nonstop negotiations [with Palipehutu-FNL] and the holding of democratic elections in August 2005, few Burundians and international role players would be brave enough to claim that the Burundian peace process has been completed and that Burundi has finally achieved durable peace.”57

262 Jessica Piombo

The very nature of the Arusha process created this situation where Burundi had not yet achieved a durable peace, six years after the Arusha Accord had been signed. Ongoing fighting had created a steady decline in the standard of living, which had already been low by international standards in 1993, and by 2006 public services had collapsed across most of the country.58

The political situation was normalizing, with a Hutu rebel movement turned political party finally leading the country and reconstruction and development beginning, but armed conflict plagued western parts of the country where Pal- ipehutu-FNL was still active.

In the summer of 2006, therefore, international negotiators increased the pressure on Palipehutu-FNL to come to a cease-fire agreement with the newly elected government of Burundi. The Tanzanian government particularly helped to push Palipehutu-FNL’s leader, Agathon Rwasa, back to the negoti- ating table. Tanzania wanted to close the remaining refugee camps and either repatriate the Burundians to Burundi or make them citizens of Tanzania. The talks began in June 2006 in Pretoria, facilitated by the South African Minister for Safety and Security, Charles Nqakula. The aim of the discussions was to secure a cease-fire agreement and bring Palipehutu-FNL attacks to an end; Palipehutu-FNL participants presented several core issues for discussion that earlier had been pushed out of the negotiations process.

These negotiations focused on obtaining agreement to a cease-fire rather than addressing issues of fundamental concern to Palipehutu-FNL. Some of the initial reasons that Palipehutu-FNL had gone to war in 1993 had been resolved with the transition in 2005 and the initiation of security sector reform, but there were other issues that had developed over the course of the conflict. These were issues similar in spirit to the ones that originated the war but that had evolved with the changing circumstances. Of particular impor- tance, Palipehutu-FNL had rejected the Arusha Accord, arguing that the nature of the power-sharing arrangements and Uprona’s continuing control over the security establishment merely disguised a perpetuation of Tutsi power. By the time the interim government was transitioning to the perma- nent government, Palipehutu-FNL also desired to negotiate for the status of its members the political realm and FNL fighters in the security forces and demobilization programs after the cease-fire.

This represents a mix of root causes and issues that arose during the course of the conflict. Regarding the composition of the military, the peace agreements signed in 2000 and 2003 had left Uprona with a “significant level” of control over the military, to which Palipehutu-FNL objected.59 In their minds, allowing Uprona a controlling influence in the military would perpetu- ate the political dominance of the old ruling party, as well as Tutsi control over the military, which had been one of the foundational problems causing the war in the first place. This general balance remained even after the CNDD-FDD integrated into the military after 2003, when the Uprona/Tutsi

Peacemaking in Burundi 263

elements were still overrepresented. Palipehutu-FNL leaders wanted to change the ratio and decrease Tutsi overrepresentation. Palipehutu-FNL leaders also did not want to lay down arms without changing the balance in the new Burundian army.

On the government’s side, the CNDD-FDD government initially preferred to pursue a military solution to the Palipehutu-FNL problem and would accept nothing but unconditional surrender.60 The two organizations had been rival rebel movements for years, and the CNDD-FDD did not want to let Palipehutu-FNL score any political points through a substantive peace agree- ment. CNDD-FDD was also concerned with internal divisions that had cropped up since winning the elections and could not afford to be seen as con- ceding to Palipehutu-FNL demands.61 Soon after taking power, therefore, the CNDD-FDD government under President Nkurunziza focused on fighting Palipehutu-FNL and agreed to return to the negotiating table only after Tanzania led a regional initiative to convince both parties to engage in dialogue.

During the negotiations, the integration of the FNL into the defense and security structures proved to be one of the major issues and obstacles to progress. Palipehutu-FNL demanded a complete overhaul of the defense and security sector. Tutsi elements in the security sector obviously opposed any further dilution of their power, while the CNDD-FDD also resisted this demand as its members were now part of the forces that Palipehutu-FNL demanded be reorganized. Furthermore, this issue was more complicated because it would be the second significant overhaul of the security sector since the transition had begun. The incorporation of the CNDD-FDD fighters into the defense and security sectors in 2003 had already necessitated that the transitional arrangements agreed to at the signing of the Arusha Accord be modified to include the CNDD-FDD. Palipehutu-FNL’s demands would now take positions away not just from the Tutsi in the defense and security sectors but also the newly transitioned CNDD-FDD combatants.62 Therefore, not including the CNDD-FDD and Palipehutu-FNL in the initial peace accords made the entire process more difficult and prone to problems: the delicate and negotiated balance between Tutsi and Hutu, former government and former militants, had to be renegotiated every time one of the combatants signed a cease-fire and agreed to demobilize and integrate.

As with the Arusha process that “ended” the war in 2000, certain issues proved too difficult to work out in the context of a negotiating framework that operated with a high degree of international pressure and with a focus on strict deadlines. As occurred with some of the most thorny issues in the ear- lier Arusha negotiations, as the facilitators promoted dialogue between the CNDD-FDD government and Palipehutu-FNL, they eliminated certain issues from discussion in order to reach an agreement. Thus the focus was on technicalities rather than substantive outcomes. Reflecting the nature of

264 Jessica Piombo

the agreement signed on September 7, 2006, the ICG described it as “purely technical agreement [that] seemed more like a surrender than a genuine peace agreement.”63

In this last case, the South African mediators convinced Palipehutu-FNL leaders to sign the cease-fire by promising that the issues would be taken up subsequently. The parties did not devise a particular formula for FNL integra- tion, and the timetable they devised in the second round of negotiations (late June–August 2006) gave only one month for FNL demobilization compared to the thirteen months provided for the same process for the FDD fighters of the CNDD-FDD.64 The problem, once again, was that the Burundian government had made no moves to initiate talks on the remaining issues by January 2007, as the international pressure was off and the government now out of the lime- light. The government and international actors asserted that Palipehutu-FNL should take up these issues with the Joint Verification and Monitoring Mechanism (JVMM) that the June agreement on principles had created, but the JVMM chairman refused to allow Palipehutu-FNL to raise any topics not written into the principles document or the official ceasefire agreement.65

Palipehutu-FNL therefore did not adhere to the provisions, and fighting continued for another two years with little progress toward implementing the cease-fire.

The South Africans facilitated more rounds of negotiations and even got both sides (the CNDD-FDD government and Palipehutu-FNL) to sign a docu- ment “reaffirming” the principles of the 2006 CFA, but this did little to decrease the fighting. The parties made no progress toward implementing the cease-fire until January 2009, when Palipehutu-FNL finally agreed to drop the “Palipehutu” part of the organization’s name and register as the FNL political party.66 Only after this did the organization begin to demobilize and convert into a purely political formation. Three years had passed between the initial signing of the CFA and the beginning of its implementation, four years into the first term of the new civilian government led by the CNDD-FDD, and one and a half years before the next round of elections were scheduled (June 2010).

CONCLUSION: CONFLICT TERMINATION VERSUS CONFLICT RESOLUTION STRATEGIES

Negotiations to end conflict are an iterative process. Burundi, like South Africa and many other conflicts that ended through negotiations, had to go through a complicated and multifaceted peace process. Preliminary talks to establish dialogue between warring parties were followed by talks to set the agenda for subsequent talks that would tackle substantive issues. None of the talks that focused on end states seemed to have a measurable effect on the

Peacemaking in Burundi 265

conflict (UN talks, Carter Center Talks, and the first rounds of the Arusha talks), with the effect that the conflict underwent years of mediation without any decrease in the intensity of fighting. Once substantive talks that addressed some of the core issues finally began in June 1999, the exclusion of certain parties and issues from the agenda meant that the negotiations pro- cess created an incomplete peace accord that had to be supplemented through various side negotiations.

Ultimately the most fundamental of the core issues (the formula for power sharing and the composition, role, and function of the military) were delegated to a subsequent domestic process that unfolded outside the realm of third- party mediation: the transitional government established through the Arusha framework. This transitional government was successful at processing the conflict to some degree, as the country transitioned to a new regime in August 2005. But because third-parties were not able to get the Burundian factions to discuss core issues during the negotiations process, and because the Tutsi-dominated military remained powerful throughout the negotiations and subsequent transitional period during which the final constitution was drafted, the transitional government created a permanent government and military that overrepresents Tutsi. This could potentially reproduce the dynamics that have fueled conflict since 1962. Had the intervening powers directly taken over some of the functions, or had the negotiations taken place within the context of third-party negotiations, it is possible that the Tutsi dominance would have been more effectively diminished. When left to domes- tic processes to resolve, however, the still powerful Tutsi were able to force concessions that maintained a significant degree of their power.

Additionally, pursuing some degree of conflict resolution before conflict termination—engaging in the negotiations while hostilities were still ongoing—meant that the two most powerful rebel groups remained outside of the negotiations process. When they were courted to join the negotiations, it was so late in the process that they declined to enter and demanded that the entire negotiations begin again. Here, the neglect of one critical end state (active rebellion) and the entering into negotiations before cease-fires had been agreed on and adhered to created almost as many problems as ignoring root causes. Negotiating with some but not all of the parties to the conflict enabled the CNDD-FDD and the Palipehutu-FNL to play the role of powerful spoilers, each of which had to be brought into the peace process through paral- lel but separate negotiations processes. This extended the life of the conflict by several years.

Comparing the Negotiations Periods The first rounds of negotiations—those sponsored by the United Nations,

the Carter Center, and the first few rounds of the Arusha talks—addressed

266 Jessica Piombo

end states. The Arusha talks grew out of the regional process initiated by the Carter Center and themselves went through various stages. The first stages, taking place between 1996 and May 1999, attempted to create dialogue between various warring factions and to establish an agenda for substantive discussions. These three attempts all shared a common goal: to end hostilities by getting parties to agree to a cease-fire. None of these talks attempted to address the issues that fueled the conflict; they focused on end states and achieving the negative peace. In Burundi’s case, the conflict would never be sustainably managed, much less fully resolved, without unpacking the core issues that had fueled over thirty years of conflict. At the point where the negotia- tions should have turned to substantive issues (a conflict resolution strategy), however, they encountered difficulties. Many of the core issues that prompted conflict in the first place proved too difficult to process within the framework of externally facilitated peace talks. When the Arusha talks split into five technical committees, each tasked with addressing a different set of issues, several of the committees were unable to agree on basic causes that should be addressed in the context of the discussion. Negotiators therefore utilized strat- egies that focused on technical arrangements and statements of principles, setting up subsequent structures that would process the substantive issues according to formula and principles established during the negotiations phases.

Postponing the fundamental issues to later stages did enable the negotia- tions processes to create peace agreements that parties to the conflict signed. Creating a transitional government that subsequently created a permanent constitution, the core feature of the Arusha Accord, was an undeniable political success. Arusha was an incomplete peace accord, however. Not including certain parties and processing important issues in the early agreements created a host of problems that postponed the full resolution of the conflict. For one, the peace process had to undergo multiple additional rounds of side negotiations and modifications parallel to and outside of the Arusha process. Thus the transitional government established by the Arusha Accord had to be modified once the CNDD-FDD signed a cease-fire, and then these same provi- sions were negotiated again and with much delay and difficulty once Palipehutu- FNL finally signed a cease-fire. Not only did this threaten the nascent govern- ment, which is a significant problem in and of itself, but the side negotiations processes extended the life of the conflict and caused greater economic and human damage as a result.

Second, the flawed peace agreement allowed signatories and nonsignato- ries to reject it later on. Parties like the CNDD-FDD and some of the smaller Tutsi organizations were able to use the forced nature of the signing and the neglect of some key issue areas as political fodder in subsequent electoral campaigns: the CNDD-FDD was able to beat Frodebu in the 2005 elections in part because of the unpopularity of the Arusha Accord at the mass level.67

Peacemaking in Burundi 267

Third, because the core document of the peace process left many critical issues unresolved, subsequent governments have been able to postpone action on important issues and problems. Certain issues, like the overrepresentation of Tutsi in the security sector, still manifest. In 2006, six years post-Arusha and one year into the permanent government, Lemarchand noted that aside from provisions relating to democracy and governance and peace and security, most provisions in the Arusha agreement had yet to be implemented. There had been no progress on rewriting the country’s history and no moves to establish either a truth and reconciliation commission or the International Judicial Commission. Some provisions of the Arusha accord that were left at the state- ment of principles stage, like that of transitional justice, have only recently begun to be enacted by the Burundian government (since 2009).

Ten years after the signing of the Arusha protocols Burundi is making steady progress in almost all sectors and addressing the core issues, even though they were not addressed during the peace negotiations. The military has been reformed to a large degree and the government now shares power between Hutu and Tutsi. In this respect, therefore, the analysis seems to indi- cate that core issues can still be addressed by subsequent processes if they are excluded from the negotiations process.

The trade-off could be that leaving core issues to subsequent or separate processes prevents the issues from being fully addressed. In Burundi, the Tutsi are still overrepresented in the military and government. Carrie Manning would argue that this is because Tutsi remained powerful throughout the negotiations process and the subsequent transitional period and therefore were able to shape the postconflict institutions to their advantage.68 Her work would imply that if the Tutsi had been disempowered through a stronger third-party intervention or mediation effort, or if the fundamental issues had not been processed by a transitional government that operated without exter- nal influence, the Tutsi may not have been able to skew the permanent politi- cal and military institutions in their favor.

Getting to the point where fundamental root causes can be addressed may also take a lot longer if the these issues are not resolved or even genu- inely opened during the negotiations phase. Certain issues may not be appro- priate for negotiation, such as the deep socioeconomic inequalities generated by thirty years of anti-Hutu discriminatory policies. Those types of funda- mental causes cannot be “mediated” away but must be dealt with through steady and consistent government policy in the postconflict period. Others, those that are more explicitly political, can be addressed through the negotia- tions process but may not be, due to political expediency. Yet as just dis- cussed, allowing some root causes to be resolved outside the lens of international mediation may empower the forces that created the problems in the first place. Ignoring key root causes because of political expediency or the opposition of negotiating partners can, and in the case of Burundi did, delay

268 Jessica Piombo

the attainment of a full peace settlement, extending the life of the conflict for another decade.

Ultimately we may be left with another trade-off: including extremely sensitive core issues may prevent progress in peace talks, but excluding those issues and focusing on process and conflict termination creates a peace settle- ment with significant gaps (like the “incomplete” Arusha protocol). These unresolved issues may be addressed through follow-on negotiations or transi- tional processes that prevent them from completely derailing the peace process. There may be no right or wrong approach; it may just be a matter of which trade-off is the necessary one to eventually get to a peaceful outcome.

One last important item to note is that for Burundi, the external imposition of a peace process may have been the primary reason why the road to a stable settlement was so angst-ridden. Van Eck argues that the nature of the exter- nally facilitated negotiations in the Burundian peace process created these problems that continually delayed the full resolution of conflict in Burundi. The external negotiators who facilitated signable documents, particularly the South African delegations (whether led by Mandela, Zuma, or Nqakula) focused so much on tight deadlines and time frames that they forced parties to sign peace agreements without addressing substantive issues that motivated the movements.

The practice of imposing agreements resulted in a virtual minefield of unresolved issues being left behind. Every party that signed against its wishes—from Arusha until today—was promised that it would be able to debate/negotiate its unresolved issues “once they sign and return to Burundi.” These promises remained unfulfilled in each of these cases. The feeling among these parties is that they were defeated and cheated, and the anger and frus- tration that this generates actively contributes to tensions inside Burundi.69

In the end, this analysis suggests that third-party facilitators of peace pro- cesses will continually face a Faustian bargain: including extremely sensitive core issues may prevent progress in peace talks, but excluding those issues and focusing on process and crisis response issues creates a peace process with significant defects. If the short-term crisis response approach is the only one feasible when conflict erupts, then the international policy community will have to acknowledge that it is critical to remain engaged until the fundamental drivers are eventually addressed. Otherwise, sustainable peace may never occur.

NOTES

1. The seminal work in this approach is John Paul Lederach, Building Peace: Sus- tainable Reconciliation in Divided Societies (Washington, DC: USIP Press, 1998).

2. For a few overviews of this literature, see Adrian Guelke, “Negotiations and Peace Processes,” in Contemporary Peacemaking: Conflict, Violence and Peace Processes, eds. John Darby and Roger MacGinty (New York: Palgrave Macmillan, 2003); John Paul

Peacemaking in Burundi 269

Lederach, “Cultivating Peace: A Practitioner’s View of Deadly Conflict and Negotia- tion,” in Contemporary Peacemaking: Conflict, Violence and Peace Processes, eds. John Darby and Roger MacGinty (New York: Palgrave Macmillan, 2003); Timothy Sisk, “Peacemaking in Civil Wars: Obstacles, Options and Opportunities,” in Managing and Settling Ethnic Conflicts: Perspectives on Successes and Failures in Europe, Africa and Asia, ed. Ulrich Schneckener and Stefan Wolff (New York: Palgrave Macmillan, 2004).

3. Sisk, “Peacemaking in Civil Wars.”

4. Guelke, “Negotiations and Peace Processes.”

5. Laurie Nathan, “The Challenges Facing Mediation in Africa,” AFRICA Mediators’ Retreat 2009—The Oslo Forum Network of Mediators. http://innovationfair.spigit.com/ Core/Download/?docid=235 (accessed April 10, 2010).

6. See, for example, Kathleen Collins, “Tajikistan: Bad Peace Agreements and Prolonged Civil Conflict,” in From Promise to Practice: Strengthening UN Capaci- ties for the Prevention of Violent Conflict (Boulder, CO: Lynne Rienner Publishers, 2003), 269.

7. Spoilers are groups that threaten a peace process through the resumption of vio- lence and are typically thought to arise when groups are excluded from the process or when groups that are losing power within the negotiations process resort to violence to increase their bargaining position. See Marie-Joelle Zahar, “Reframing the Spoiler Debate in Peace Processes,” in Contemporary Peacemaking: Conflict, Violence and Peace Processes, eds. John Darby and Roger MacGinty (New York: Palgrave Macmillan, 2003).

8. Guelke, “Negotiations and Peace Processes,” 53.

9. Some would provide dates of 1994 to 2005 for the war; this would represent the time that the CNDD-FDD and Palipehutu-FNL initiated large-scale resistance to a power-sharing government and when the subsequent transitional government of Burundi ceased operating and handed power to an elected government. I am using 1993–2009 because this round of conflict began as interethnic massacres in 1993, and 2009 was when the last rebel group actually began to disarm and demobilize.

10. For the most comprehensive account of this process, see Jean Pierre Chrétien, The Great Lakes of Africa: Two Thousand Years of History, translated by Scott Strauss (Zone Press, distributed by MIT Press, 2003); and for Burundi specifically, René Lemarchand, Burundi: Ethnic Conflict and Genocide (New York: Woodrow Wilson Center and Cambridge University Press, 1994 [1995]). For the political implications of this process in Rwanda, see Peter Uvin, “Prejudice, Cisis and Genocide in Rwanda,” African Studies Review 40 no. 2 (1997): 91–115.

11. For the most comprehensive evaluation of these elections available in English, see Reyntjens, “The Proof of the Pudding is in the Eating: The June 1993 Elections in Burundi,” The Journal of Modern African Studies 31 no. 4 (1993): 563–583.

12. Ibid.

13. Ibid., 564–565.

14. Daniel Sullivan, “The Missing Pillars: A Look at the Failure of Peace in Burundi through the Lens of Arend Lijphart’s Theory of Consociational Democracy,” The Jour- nal of Modern African Studies 43, no. 1 (2005): 75–96.

15. Ibid., 86.

16. This is despite the reality that Ndadaye’s cabinet overrepresented Tutsi, allocating them over one-third of the seats, and that the prime minister was Tutsi as well. See Reyntjens, “The Proof of the Pudding,” 579.

270 Jessica Piombo

17. Gilbert Khadiagalia, “Burundi,” in Dealing with Conflict in Africa: The United Nations and Regional Organizations, ed. Jane Boulden (New York: Palgrave Macmillan, 2003).

18. Jan Van Eck, “Challenges to a Durable Peace in Burundi,” ISS Situation Report (Pretoria, South Africa: Institute for Security Studies, April 2004).

19. For his account of the process, see Ahmedou Ould-Abdallah, Burundi on the Brink: A UN Special Envoy Reflects on Preventive Diplomacy (Washington, DC: United States Institute of Peace Press, 2000).

20. René Lemarchand, “Burundi’s Endangered Transition,” FAST Country Risk Profile. SwissPeace Working Paper No. 5 (Bonn, Switzerland: Swiss Peace Foundation, October 2006). The CNDD-FDD itself spawned numerous factions in the following years, but under Nkurunziza’s leadership became the most powerful of the Hutu rebel forces.

21. Khadiagalia, “Burundi”; Gregory Mthembu-Salter, “Burundi’s Peace Agreement Without Peace,” Track Two 11, nos. 5–6 (October 2002), http://196.25.102.27/archive/ two/11_56/index.html (accessed April 10, 2010).

22. Khadiagalia, “Burundi.”

23. Mthembu-Salter, “Burundi’s Peace Agreement Without Peace.”

24. Ibid.

25. Ibid.

26. Lemarchand, Burundi; Reyntjens; Khadiagalia, “Burundi.” Ironically, before the coup Buyoya had been residing in the United States while writing a book on democracy.

27. Khadiagalia, “Burundi.”

28. Mthembu-Salter, “Burundi’s Peace Agreement.”

29. Ibid.; International Crisis Group, “Burundi’s Peace Process.”

30. Elisa, Renda, “Mediation Efforts in Burundi,” Conflict Trends 3 (2003): 32–39, 38.

31. See International Crisis Group, “The Mandela Effect: Prospects for Peace in Burundi,” Bujumbura/Nairobi: ICG Central Africa Report No. 13 (April 18, 2000).

32. While officially part of the Arusha framework, these separate rounds of talks are most often discussed completely separately from the Arusha political process.

33. ICG, “Burundi’s Peace Process;” Renda, “Mediation Efforts in Burundi;” Kristina Bentley and Roger Southall, An African Peace Process: Mandela, South Africa and Burundi (Cape Town: South Africa Human Sciences Research Council, 2005).

34. Khadiagalia, “Burundi.”

35. Renda, “Mediation Efforts in Burundi,” 37.

36. Ibid.

37. ICG, “The Mandela Effect.”

38. ICG, “The Mandela Effect,” 29. The CNDD-FDD was experiencing several layers of internal divisions at this time as well, further complicating its incorporation into the peace process.

39. See, for example, Devon Curtis, “Transitional Governance in Burundi and the Democratic Republic of the Congo,” in Interim Governments: Institutional Bridges to peace and Diplomacy? eds. Karen Guttieri and Jessica Piombo (Washington, DC: United States Institute of Peace Press, 2007).

Peacemaking in Burundi 271

40. Jan Van Eck, “Challenges to a Durable Peace in Burundi,” ISS Situation Report. Pretoria, South Africa: Institute for Security Studies, April 2004.

41. Bentley and Southall, An African Peace Process.

42. Jan van Eck, “Burundi: An Ongoing Search for a Durable Peace,” African Security Review 16, no. 1 (2007): 113–121.

43. Lemarchand, “Burundi’s Endangered Transition,” 11.

44. Bentley and Southall, An African Peace Process; Khadiagalia, “Burundi;” Curtis, “Burundi and the DRC.”

45. International Crisis Group, “Burundi: Neither War Nor Peace,” ICG Africa Report 25 (December 2000). This ICG report contains an excerpted version of the main ele- ments of the agreement as well as an analysis of the text and the process that created the agreement.

46. Curtis, “Burundi and the DRC;” Lemarchand, “Burundi’s Endangered Transition,” 11; ICG, “Neither War Nor Peace.” These are just a few examples; this is a steady theme in virtually every piece written on the subject.

47. Curtis, “Burundi and the DRC.”

48. International Crisis Group, “Burundi: Neither War Nor Peace.”

49. International Crisis Group, “Burundi’s Endangered Transition,” 11.

50. Curtis, “Burundi and the DRC,” 186.

51. ICG, “Neither War Nor Peace,” 1.

52. International Crisis Group, “The Burundi Rebellion and the Ceasefire Negotia- tions,” ICG Africa Briefing, August 6, 2002.

53. International Crisis Group, “The Burundi Rebellion.” This particular ICG briefing has an excellent, concise overview of the main rebel factions, their histories, and the issues that motivated their continued struggles.

54. Henri Boshoff and Jean Marie Gasana, “Mapping the Road to Peace in Burundi: The Pretoria Sessions,” ISS Situation Report: African Security Analysis Programme (Pretoria, South Africa: Institute for Security Studies, 2003), 1.

55. Ibid.

56. Ibid., 6.

57. Van Eck, “Burundi: An Ongoing Search,” 2.

58. Lemarchand, “Burundi’s Endangered Transition,” 9.

59. Henri Boshoff, Jean Marie Gasana, and Richard Cornwell, “Burundi: The End of the Tunnel?” ISS Situation Report (Pretoria, South Africa: Institute of Security Studies, 2009), 2.

60. International Crisis Group, “Burundi: Finalising Peace with the FNL,” ICG Africa Briefing 131 (August 2007).

61. International Crisis Group, “Burundi: Finalising Peace with the FNL.”

62. Curtis, “Burundi and the DRC.”

63. International Crisis Group, “Finalising Peace with the FNL,” 2.

64. International Crisis Group, “Finalising Peace with the FNL,” 6.

65. Van Eck, “Ongoing Search for a Durable Peace.”

272 Jessica Piombo

66. Boshoff and Gasana, “Burundi: The End of the Tunnel?”

67. Lemarchand, “Burundi’s Endangered Transition.”

68. Carrie Manning, “Interim Governments and the Construction of Political Elites,” in Interim Governments: Institutional Bridges to Peace and Diplomacy? eds. Karen Guttien and Jessica Piombo (Washington, DC: United States Institute of Peace Press, 2007).

69. Van Eck, “Ongoing Search for a Durable Peace,” 6. Emphasis added.

Copyright of African Security is the property of Routledge and its content may not be copied or emailed to

multiple sites or posted to a listserv without the copyright holder's express written permission. However, users

may print, download, or email articles for individual use.