Peacebuilding in Burundi and the Great Lakes

AI

Peacebuilding in Burundi has produced one of the most instructive records anywhere in Africa: a comprehensive negotiated settlement, a decade of cautious consolidation, and then a crisis that stripped the model down to its weakest joints. Between 1993 and the mid-2000s, civil war killed an estimated 300,000 people in a country of then roughly seven million. The settlement that ended it, the Arusha Peace and Reconciliation Agreement of August 2000, became a reference point for mediators far beyond Central Africa.

Burundi cannot be read in isolation. Its war was fed by, and fed into, the wider conflicts of the Great Lakes region of Africa: the Rwandan genocide of 1994, the two Congo wars, and refugee movements that crossed every border in the neighbourhood. Any honest account of Burundi’s peace process therefore runs on two tracks at once, the domestic bargain struck at Arusha and the regional architecture built around it.

This analysis traces both tracks: what the Arusha framework contained, how power-sharing actually worked, where land, gender and regional economics entered the equation, and what the 2015 crisis revealed about the guarantees that were supposed to hold the settlement together.

The Arusha framework and its legacy

Burundi’s descent began in October 1993, when soldiers assassinated Melchior Ndadaye, the country’s first democratically elected president and its first Hutu head of state, barely four months into his term. The killing triggered massacres and a civil war that ground on for over a decade, layered on earlier episodes of mass violence, above all the killings of 1972.

Negotiations opened in Arusha, Tanzania, in 1998 under the facilitation of Julius Nyerere. After Nyerere’s death in 1999, Nelson Mandela took over and pressed the parties hard. On 28 August 2000, nineteen Burundian parties signed the Arusha Peace and Reconciliation Agreement, a text whose protocols covered the nature of the conflict, democracy and good governance, peace and security, reconstruction and development, and guarantees for implementation. The full text is archived by the United Nations at UN Peacemaker.

Arusha had a conspicuous gap. The two main armed movements, the CNDD-FDD and the FNL, were not signatories, and fighting continued for years after the ceremony. The CNDD-FDD joined the process through the Global Ceasefire Agreement of November 2003; the FNL signed a ceasefire in 2006 and only demobilised fully in 2009. Even so, the agreement’s core prescriptions survived the wait. They were written into the 2005 constitution, endorsed by referendum, and governed the elections that brought Pierre Nkurunziza and the CNDD-FDD to power that year.

Power-sharing as conflict management

The heart of the Arusha peace agreement for Burundi is a set of ethnic quotas designed to take demography off the table as an existential stake. The 2005 constitution capped the government and National Assembly at 60 percent Hutu and 40 percent Tutsi, split the Senate evenly, and reserved three co-opted seats for Twa deputies. The army and police were to be composed half and half, regardless of the fact that Hutu make up roughly 85 percent of the population.

The military provision is widely counted as the settlement’s single clearest success. Former government soldiers and former rebels were merged into a new National Defence Force, and the integrated army held together well enough that Burundi became one of the largest troop contributors to the African Union mission in Somalia from 2007 onward. An institution that had been the instrument of successive coups acquired a stake in the new order. Comparable reform efforts next door achieved far less, as the record of security system reform in the DRC shows.

The quotas also changed electoral incentives. Because candidates run on ethnically mixed lists, parties compete for votes across the ethnic divide rather than by mobilising one group against the other. The result is striking: since 2005, Burundi’s political violence has run mainly along partisan lines, ruling party against opposition, rather than Hutu against Tutsi. Power-sharing managed the ethnic conflict it was designed for. It offered no protection against single-party dominance, and that is precisely where the settlement later cracked.

The regional dimension: ICGLR and Great Lakes cooperation

Burundi’s war never respected its borders. Rebel movements operated from bases in the DRC and camps in Tanzania; refugee flows moved in every direction; arms and finance circulated along the same corridors as trade. The regional response was the International Conference on the Great Lakes Region (ICGLR), established in 2004 under United Nations and African Union auspices. Its member states adopted the Pact on Security, Stability and Development in December 2006, a framework built around four areas of action: peace and security, democracy and good governance, economic development and regional integration, and humanitarian and social issues. The pact entered into force in 2008, and the ICGLR’s executive secretariat sits in Bujumbura, a deliberate signal of Burundi’s centrality to the regional project.

Field research conducted across Rwanda, the DRC and Burundi in 2008 found that the ICGLR’s chief value was as a room where estranged governments could talk. In the absence of formal diplomatic relations between Kinshasa and Kigali, both used the conference to discuss issues of mutual concern. The same research warned that without member-state ownership and a trimmed agenda, the body risked becoming a diplomatic talking shop with little to show on the ground. Donor contributions had passed US$10 million by 2006, yet only two member states had paid their own dues at the time of the study.

That ambivalence runs through the whole regional file, examined at length in this journal’s survey of regional cooperation in the Great Lakes. Regional institutions gave Burundi’s peace process a scaffolding, but the scaffolding was donor-built, and the violence in the eastern DRC kept testing it.

Land, return and reintegration

No issue reaches deeper into Burundian society than land. The country is among the most densely populated in Africa, and the great majority of its people live from subsistence agriculture on small family plots. Every wave of violence, in 1972, in 1993 and after, pushed hundreds of thousands off their hills, and every return created competing claims between refugees and the people who had settled or bought their land in the meantime.

Arusha’s fourth protocol addressed reconstruction and the return of refugees directly, and the scale of the challenge soon became clear. After the 2000 agreement and the 2003 ceasefire, more than half a million Burundians came home, most from Tanzania, in returns facilitated by UNHCR. Some had been in exile for over thirty years and found their parcels occupied by two generations of other families. The government created the National Commission on Land and Other Assets (CNTB) in 2006 to arbitrate such claims. It has processed tens of thousands of cases, and land disputes remain the single largest category of litigation before Burundian courts.

The land question is peacebuilding in its most concrete form. Where restitution was handled transparently, return held; where decisions were seen as politicised, grievances persisted and occasionally turned violent. The 2015 crisis then produced a new exodus and, since 2017, a new cycle of organised returns, feeding the same disputes back into the same overloaded system.

Gender dimensions of Burundi’s peace process

Burundi is also a leading case for the study of gender issues in peace processes. The nineteen delegations at Arusha were overwhelmingly male, so in July 2000 the All-Party Burundi Women’s Peace Conference brought women from every negotiating party to Arusha with support from the facilitation team and UN partners. Most of their recommendations, on women’s representation, on the rights of returnee women, on protections against sexual violence, made it into the final text. The 2005 constitution then set a minimum of 30 percent women in the government and National Assembly, a threshold Burundian institutions have exceeded ever since.

Later EU-funded research on gender and peacebuilding, synthesised in 2010, used Burundi as one of its central field cases. Its most cited finding concerned community security. A community policing programme financed by the UN Peacebuilding Fund and implemented by Dushirahamwe, a Burundian women’s organisation, showed that increasing communication between police and women eased communication between police and society as a whole. A network of more than twenty Burundian civil society organisations built on that experience, pressing for the protection of women and girls to be integrated into the French and Kirundi versions of police training manuals and for gender-disaggregated data in security programming.

The same body of research warned against easy assumptions: seats in parliament do not automatically translate into influence, and women legislators can face the same elite incentives as their male colleagues. The full argument is set out in the journal’s review of the gender and peacebuilding synthesis. Burundi’s record supports both halves of the conclusion, real institutional gains and real limits on what quotas alone deliver.

The 2015 crisis and lessons for guarantors

In April 2015, the CNDD-FDD nominated Pierre Nkurunziza for a third presidential term. Opponents read this as a breach of the two-term limit rooted in Arusha; the constitutional court, under visible pressure, disagreed. Protests broke out in Bujumbura, a coup attempt failed in May, and the state answered with sustained repression. By 2017, judges of the International Criminal Court authorising an investigation cited estimates of at least 1,200 people killed, and more than 400,000 Burundians had fled to Tanzania, Rwanda, Uganda and the DRC.

The response of the settlement’s guarantors is the sobering part of the story. The African Union announced a 5,000-strong protection mission, MAPROBU, in December 2015, then dropped the plan within weeks when Bujumbura declared that any deployment without consent would be treated as an invasion. The UN Security Council authorised 228 police officers under resolution 2303 in 2016. They never deployed. The East African Community’s mediation, led by Uganda’s Yoweri Museveni with former Tanzanian president Benjamin Mkapa as facilitator, ran for years without producing an agreement, while regional governments split over how hard to press a fellow head of state. Burundi then withdrew from the International Criminal Court in October 2017, the first state ever to do so, and a 2018 constitutional referendum extended presidential terms to seven years, loosening several of Arusha’s political safeguards while keeping the ethnic quotas.

The lesson is structural rather than personal. Arusha’s guarantee provisions rested on the consent of the government being guaranteed. When that government withdrew consent, no guarantor was willing to act without it, and each institution deferred to the next. Detailed reporting on this sequence is available from the International Crisis Group. Nkurunziza died in June 2020, weeks before the end of his term; his successor, Évariste Ndayishimiye, has reopened doors to donors and neighbours, but the enforcement gap exposed in 2015 remains unfixed in the settlement’s design.

Regional economic cooperation as peace infrastructure

One thread of Burundi’s peacebuilding story runs through power lines rather than peace tables. The Economic Community of the Great Lakes Countries (CEPGL), founded by Burundi, the DRC and Rwanda in 1976, collapsed institutionally in the mid-1990s. Yet its most tangible creation kept working: SINELAC, the joint electricity company running the Ruzizi II hydropower station on the border, continued to deliver power to all three countries through years when their governments had no diplomatic relations at all. Researchers studying the region in 2008 and 2009 treated this as evidence of the resilience of technical cooperation, and as a warning against reading too much politics into it, since the shared dam never produced reconciliation at the level of heads of state.

“Reviving the Economic Community of Countries in the Great Lakes (CEPGL) seems like an avenue for not only offering solutions to economic and social problems of these countries but also for restoring peace and security in the sub-region.”
— International Conference on the Great Lakes Region, regional programme of action, 2006

The stakes were real. UN energy statistics for 2006 put Burundi’s net electricity availability at roughly 123 million kWh for some 8.5 million people; Sweden, with a comparable population, had over 130,000 million. The European Commission committed €50 million to relaunch the CEPGL secretariat in 2007, and Burundi joined the East African Community the same year, betting its economic future on regional markets. But the 2009 research concluded that peacebuilding effects do not follow automatically from economic cooperation: they must be designed in, with conflict-sensitivity, local consultation and political dialogue linked to the technical work. Donor-driven institutions with absent member states deliver neither peace nor much electricity. Similar dynamics, where shared resources can either bind neighbours together or set them against each other, are examined in the journal’s study of peacebuilding around Lake Albert.

Peacebuilding in Burundi: what worked and what did not

Judged against the depth of the war it ended, Burundi’s peace process delivered more than sceptics predicted in 2000, and less than its drafters hoped.

• What worked: the ethnic quotas defused the Hutu-Tutsi security dilemma; the integrated army held and became a regional peacekeeping asset; the women’s quota created durable representation; refugee return happened at scale; and the ICGLR gave the region a standing forum it previously lacked.
• What did not: the guarantee architecture failed its first serious test in 2015; term-limit safeguards proved unenforceable; the special tribunal envisioned at Arusha was never established, leaving transitional justice to a domestic truth commission created only in 2014; land arbitration was repeatedly accused of politicisation; and regional economic bodies remained donor-driven, with thin peace dividends for ordinary households.

“…it seems doubtful that the CEPGL can be effectively restored without a fundamental change of political positions of its member states.”
— Stevens, Hoebeke and Vlassenroot, Egmont Institute, 2008

The deeper pattern is that Burundi’s settlement was strongest where it changed everyday incentives, in the army, in party lists, in a village land hearing, and weakest where it relied on external actors to punish defection. That asymmetry, not any single event, explains both the decade of stability after 2005 and the speed of the unravelling in 2015. Analysts at the Institute for Security Studies made a version of this argument throughout the crisis years, and subsequent events have not contradicted it. Researchers and practitioners will keep returning to the case, and further country studies in this journal’s analysis section apply the same questions elsewhere in the region.

Frequently asked questions

What was the Arusha Peace and Reconciliation Agreement for Burundi?

It was the comprehensive settlement signed in Arusha, Tanzania, on 28 August 2000 by nineteen Burundian parties. Its protocols covered power-sharing between Hutu and Tutsi, reform of the army and police, transitional institutions, refugee return, reconstruction and implementation guarantees. Its core provisions were written into Burundi’s 2005 constitution.

Who mediated the Arusha peace process?

Former Tanzanian president Julius Nyerere facilitated the talks from 1996 until his death in 1999. Nelson Mandela then took over and pushed the parties to signature in August 2000, backed by regional heads of state and international donors who pledged reconstruction funding at the signing.

Why did civil war break out in Burundi in 1993?

Soldiers assassinated Melchior Ndadaye, Burundi’s first democratically elected and first Hutu president, in an attempted coup in October 1993, four months after he took office. The killing set off massacres and a civil war that lasted more than a decade, layered on earlier mass violence, including the killings of 1972.

How does power-sharing work in Burundi?

The 2005 constitution caps the government and National Assembly at 60 percent Hutu and 40 percent Tutsi, splits the Senate evenly, reserves three co-opted seats for Twa deputies, and requires the army and police to be composed half Hutu and half Tutsi. Candidates run on ethnically mixed party lists, so parties must compete across the ethnic divide.

Did the CNDD-FDD sign the Arusha Agreement in 2000?

No. The CNDD-FDD, the largest rebel movement, stayed outside the 2000 agreement and joined the process through the Global Ceasefire Agreement of November 2003. It entered the transitional institutions, won the 2005 elections, and its leader Pierre Nkurunziza became president. The FNL fought on until its 2006 ceasefire and demobilised in 2009.

What is the ICGLR and what does it do?

The International Conference on the Great Lakes Region was established in 2004 under UN and African Union auspices. Its members adopted the Pact on Security, Stability and Development in 2006, in force since 2008, covering peace and security, governance, economic integration and humanitarian issues. Its executive secretariat is based in Bujumbura, Burundi.

What caused the 2015 crisis in Burundi?

President Pierre Nkurunziza’s nomination for a third term in April 2015, which opponents considered a breach of the two-term limit derived from the Arusha framework. Protests in Bujumbura, a failed coup in May 2015 and a sustained government crackdown followed, ending a decade of relative post-war stability.

How many people fled Burundi after 2015?

More than 400,000 Burundians fled to Tanzania, Rwanda, Uganda and the DRC in the two years after April 2015, according to UNHCR figures. Judges of the International Criminal Court, authorising an investigation in 2017, cited estimates of at least 1,200 people killed during the crisis. Organised voluntary returns have been under way since 2017.

What role did women play in Burundi’s peace process?

The All-Party Burundi Women’s Peace Conference of July 2000 brought women from all nineteen negotiating parties to Arusha, and most of their recommendations entered the final agreement. Women’s organisations such as Dushirahamwe later became central to community security work, including a UN-funded community policing programme that improved relations between police and local communities.

What is the gender quota in Burundi’s constitution?

The 2005 constitution sets a minimum of 30 percent women in the government and the National Assembly. Burundian institutions have consistently met or exceeded this threshold since the 2005 elections, giving Burundi one of the stronger records of women’s formal representation in the region.

Why are land disputes so central to peacebuilding in Burundi?

Burundi is one of Africa’s most densely populated countries and most households depend on small farm plots. Successive waves of displacement in 1972 and 1993 created overlapping claims between returning refugees and those who occupied their land in the meantime. Land disputes remain the largest category of cases before Burundian courts.

What is the CNTB?

The National Commission on Land and Other Assets, created in 2006 to arbitrate property claims arising from refugee return. It has handled tens of thousands of cases between returnees and current occupants. Its rulings have at times been criticised as politicised, which has fed local grievances rather than settling them.

What is the CEPGL?

The Economic Community of the Great Lakes Countries, founded in 1976 by Burundi, the DRC and Rwanda. It collapsed during the wars of the 1990s and was relaunched in 2007 with a €50 million commitment from the European Commission. Its joint electricity company SINELAC kept the Ruzizi II hydropower station running throughout the regional wars.

Did the 2018 constitutional referendum end the Arusha framework?

Not formally, but it weakened it. The 2018 revision extended presidential terms to seven years and loosened several power-sharing safeguards, while retaining the ethnic quotas. Most analysts read it as a departure from the spirit of the Arusha settlement, whose two-term principle had already been breached in 2015.

Is Burundi still a member of the International Criminal Court?

No. Burundi’s withdrawal took effect on 27 October 2017, making it the first state ever to leave the ICC. The court retains jurisdiction over crimes committed while Burundi was a member, and the investigation authorised in 2017 covers the violence that followed the 2015 crisis.

What lessons does Burundi hold for guarantors of peace agreements?

That guarantees which depend on the consent of the guaranteed government fail exactly when they are needed. In 2015-16 the African Union shelved its planned protection mission, UN police authorised under resolution 2303 never deployed, and regional mediation stalled. Enforcement provisions need agreed triggers and costs of defection, not just goodwill.

Does the Ruzizi II power station still operate?

Yes. Ruzizi II, managed by SINELAC on behalf of Burundi, the DRC and Rwanda, has operated since the late 1980s and continued producing electricity through periods when the three governments had no diplomatic relations. It is regularly cited as evidence that technical cooperation can survive political rupture in the Great Lakes region.