Abstract
This article aims to demonstrate the effects of exclusionary forms of governance on democracy and peace with reference to Burundi’s post-Arusha electoral rivalries. The Arusha agreement was the first in a series of peace agreements that terminated over a decade of civil strife, which was in many respects based on interethnic conflict. The Arusha pacification model resolved Burundi’s past interethnic conflicts and rivalries, and the main cleavage is currently between and within political parties rather than ethnic groups. In explaining the causes of the crisis of democracy and peace, the article gives considerable weight to Burundi’s post-Arusha elections. The first post-Arusha elections held in 2005, won by the former Hutu-dominated rebel movement, the National Council for the Defence of Democracy–Forces for the Defence of Democracy (CNDD–FDD) were regarded as free and fair, but other endogenous and exogenous factors also played a role. The second and third post-Arusha elections held in 2010 and 2015 respectively, largely paved the way for the current crisis of democracy and peace. The CNDD–FDD’s dominant position allowed it to gradually reduce the democratic space in Burundi, leading to the establishment of an authoritarian, one-party system through elections and violence. Burundi’s current crisis of peace is caused by the failure of democracy.
Keywords: Arusha agreement, Burundi, democracy, elections, ethnic conflict, peace
________________________________________________________________________________
1. Introduction
Burundi’s multiparty general elections were held in June 1993 in considerable calm and dignity (Reyntjens, 1993:568) and were portrayed as a model for the rest of Africa (Vandeginste, 2009:64). The Hutu-led Front for Democracy in Burundi (FRODEBU) led by Melchior Ndadaye won the elections, defeating and replacing the incumbent President Pierre Buyoya’s Tutsi-led Union for National Progress (UPRONA) (Vandeginste, 2009:64). However, the newly elected president was assassinated three months after taking office, leaving Burundi in a political and institutional imbroglio, with tens of thousands of people killed in Hutu–Tutsi interethnic massacres, a decade-long civil war and a large number of international and regional conflict resolution initiatives (Burihabwa and Curtis, 2021:1222).
The Arusha peace and reconciliation agreement of August 2000 was the first in a series of peace agreements that resolved the Hutu–Tutsi differences (Vandeginste, 2015:624). As this article demonstrates, judging from this negotiated transition from ethnic conflict to peace, Burundi was regarded as a success story demonstrating the use of power-sharing in resolving internal conflict (Vandeginste, 2015:625). The agreement paved the way for a transitional government, multiparty elections in 2005 and formed the basis of a new Burundian constitution (Burihabwa and Curtis, 2021:1222). While the peace process faced difficulties, it was commonly interpreted as a successful example of internationally and regionally supported peacebuilding and state-building.
However, after a few days, local and international observers were stunned by how rapidly the National Council for the Defence of Democracy–Forces for the Defence of Democracy (CNDD–FDD) regime consolidated its hold on power, violating all the principles of democracy, peace and good governance in the process (Nshimirimana, 2020:163). Contrary to what many feared, conflicts in the post-Arusha Burundi did not lead to the reappearance of the Hutu–Tutsi conflicts, which have historically characterised the Burundian polity. As this article shows, the main cleavages since 2005 have been between and within political parties – rather than ethnic groups – and severely threatened democracy and peace.
2. Democratisation and ethnic conflicts: historical context
Like neighbouring Rwanda, Burundi belongs to one of Africa’s rare “ancient political zones that were real potential nations, comparable to European states in the early nineteenth century” (Twagiramungu, 2016:57). Located in the Great Lakes region of Africa, Burundi is a small, poor, overpopulated state (population estimated at 13,590,102), land-locked between Rwanda, the Democratic Republic of the Congo (DRC) and Tanzania (CIA, 2024). Burundi’s population is composed of about 85% Hutus, 14% Tutsis and nearly 1% Twas (Nsengiyumva, 2024:32). Founded in the early seventeenth century by King Ntare Rushatsi (Twagiramungu 2016:57), Burundi was ruled by a monarch in collaboration with a corps of Hutu and Tutsi representatives until 1966 (Daley, 2006:662–668).
After the 1884–5 Berlin Conference, Burundi was conquered and allocated to Germany by European powers but became a Belgian colony following the German defeat in the wake of the First World War in 1916 (Bentley and Southall, 2005:32; Twagiramungu, 2016:58; Niyukuri, 2020:132).[1] Due to the lack of manpower and resources, the Belgian administration (like Germany) retained the existing political structures but increased the dominance of the Tutsis over the Hutus (Niyukuri, 2020:132). In line with the social Darwinian thinking (Magnarella, 2005:806) and pre-conceptions of racial superiority, the colonial administration concentrated all the privileges of power in the hands of the Tutsis (Twagiramungu, 2016:58). The Tutsi minority, who were considered by the colonisers as higher in status than the Hutu majority, were thus given privileges of education and administration, which frustrated the Hutus (Niyukuri, 2020:132). Belgian colonisers used pseudo-scientific theories to validate the alleged biological evidence pertaining to Tutsis (associated with Caucasians) and Hutus (regarded as true Negroes) as different races (Twagiramungu, 2016:58), which reinforced prejudice against the Hutus. Influenced by the Hamitic theory, they portrayed the Tutsis as distant descendants of Ham, the biblical son of Noah and therefore the founders of the highly civilised “Hamitic kingdom to rule over the savage Hutu Negroes and Twa pygmies” (Twagiramungu, 2016:58).
In the 1950s, the United Nations (UN) trusteeship council’s resolution on democratisation pressed Belgium to change its responsibilities under the League of Nations mandate in Burundi and Rwanda (Reyntjens, 1995:7). However, as it had made little provision for Indigenous political advances, democratisation was seen by Burundian and Rwandan Hutus – like in other nations in Africa – as a struggle by the people for social, political and economic rights (Adejumobi, 2001:163). Like in Burundi, the colonial administration had instituted and strengthened a system of rigid ethnic stratification in Rwanda whereby the Tutsis were the political hegemons. Rwanda’s population is similar to Burundi’s. In this context, the Tutsi minority perceived such a democratic order in which the political power would benefit those capable of securing a majority vote as a serious threat (Bentley and Southall, 2005:40). Consequently, with the rise of the Hutu demands for political and social reforms, the Hutu People’s Emancipation Party (PARMEHUTU) was created in Rwanda. This movement, whose ideology was premised on the very political identities of ‘Hutu’ and ‘Tutsi’ institutionalised by colonial authorities, won Rwanda’s pre-independence elections. After independence, PARMEHUTU changed Rwanda into “a Hutu ethnocracy dressed up as a populist majoritarian democracy that excluded ‘the Tutsi race’ from the political order” (Bentley and Southall, 2005:41). The Tutsis suffered discrimination, persecution and repeated massacres at the hands of the Hutus, which led to hundreds of thousands of Tutsis fleeing Rwanda to neighbouring states, including Burundi (Magnarella, 2005:809).
Following the political changes in Rwanda, the regional dimensions of ethnic conflicts became inseparable from the convoluted course of Burundi’s domestic politics[2]. As Nkurunziza and Ngaruko observe: “Tutsis and Hutus in Burundi started seeing each other through the Rwandese prism” since then (2005:5). With the same ethnic configuration, Burundi could not remain unscathed by the political changes in Rwanda. More specifically, Burundian Hutus could not remain insensitive of the majority rule that was taking place there (Lemarchand, 1996:60). Radical Burundian Hutu leaders believed that “their numerical majority could guarantee them de facto control of political and economic power” and regarded the changes in Rwanda as the model to replicate (Nkurunziza and Ngaruko, 2005:5). In contrast, most Tutsi leaders viewed these changes as the scenario to avoid by all means, arguing that “power in the hands of the Hutus would inexorably lead to their [Tutsis] extermination” (Nkurunziza and Ngaruko, 2005:5). As the Tutsis controlled political institutions, they did everything in their power to prevent similar changes from happening by increased isolation of the Hutus (Nkurunziza, 2015:5).
Burundi’s first move towards democratisation gave birth to two leading nationalist parties – UPRONA and the Christian Democratic Party (PDC) (Daley, 2006:666). Appalled by the developments in Rwanda, UPRONA’s leader, Prince Louis Rwagasore distanced himself from the Burundian elites trained by the colonial administration (Twagiramungu, 2016:59) and opted for a nationalist orientation (Nkurunziza and Ngaruko, 2005:5). He thus joined forces with Hutu and Tutsi people to found UPRONA (Twagiramungu, 2016:59). Among his objectives, Rwagasore aimed at uniting all ethnic groups and to serve the interests of Burundi (Bentley and Southall, 2005:42; Nkurunziza and Ngaruko, 2005:5). Under the slogan ‘unity and progress’, UPRONA’s leadership managed to unite Hutu and Tutsi anti-colonialists (Daley, 2006:666). However, as the Belgian administration feared a mass-based party, it branded Rwagasore a pro-communist and lent its support to the less radical PDC, which was not in favour of immediate self-governance. Despite Belgian propaganda, UPRONA won Burundi’s 1961 legislative elections, and Rwagasore became prime minister (Twagiramungu, 2016:59). However, Rwagasore was assassinated two weeks later by agents of the PDC with the tacit approval of the Belgians (Daley, 2006:666).
Rwagasore’s assassination had severe consequences for Burundi’s early move towards democratisation and the relationship between the conflicting Hutu and Tutsi ethnic groups. Before the end of 1962, UPRONA was so divided that the competition for its leadership fractured it into two rival factions – Monrovia versus Casablanca (Lemarchand, 1996:63–64). For the next three years, the rivalry within UPRONA increasingly pronounced ethnic coloration, with Casablanca becoming a Tutsi bloc and Monrovia siding with the Hutus (Twagiramungu, 2016:59). UPRONA proved too weak to contain the Hutu–Tutsi rivalries after Burundi’s independence from Belgium in 1962 (Reyntjens, 2000:7), but the rift turned bloody afterwards. In January 1965, Pierre Ngendandumwe, a Hutu who had just taken the oath of office as prime minister, was assassinated by a Tutsi refugee from Rwanda (Twagiramungu, 2016:59). Like Rwagasore’s assassination, Ngendandumwe’s murder plunged Burundi into another political crisis.
King Mwambutsa’s refusal to appoint a Hutu to replace Ngendandumwe despite Hutu candidates’ victory after the legislative elections in May 1965 (Ndikumana, 1998:35) exacerbated Burundi’s already fragile political atmosphere. His decision to appoint a Tutsi prime minister enraged Hutu political elites, fearing further political marginalisation and inspired by the Hutu revolts in Rwanda, to push for a “democracy based on ethnic majority” (Nkurunziza and Ngaruko, 2005:5). In October 1965, some Hutu political elites attempted to seize power with the support of Hutu elements in the army and the gendarmerie (Reyntjens, 2000:7). Contemporaneously, a Hutu peasant uprising organised and supported by urban sections of the Hutu gendarmerie broke out, leading to the massacre of hundreds of Tutsi civilians. In revenge, Tutsi troops led by Captain Michel Micombero unleashed a bloody purge of Hutu elements from the security forces and the political class, together with thousands of rural-based Hutus who supported the revolt. In some regions, the pogroms became so widespread, that the World Labour Confederation called for a UN-led international intervention to “stop the genocide against ethnic Hutu” (Twagiramungu, 2016:62). The 1965 interethnic massacres accentuated the magnitude of bitterness and resentment on the already weak relationship between the conflicting Hutu and Tutsi groups. By the end of 1965, and for the next four decades, Hutu people’s exclusion from Burundi’s political life intensified and deepened the divisions.
The successive Tutsi-led military regimes that came to power met Hutu-led resistance and, in response, engineered state-sponsored campaigns of purges and repression (Nkurunziza, 2023). Under Captain Michel Micombero’s regime (1965–76), Burundi’s main power structures (the army, gendarmerie and UPRONA leadership) became an exclusive domain of the Tutsis (Russell, 2015:439). These exclusions and marginalisation fomented the spirit of rebellion among the Hutus. Within this context, some Hutu elites launched an armed rebellion in April 1972 in an attempt to challenge the Tutsi hegemony (Twagiramungu, 2016:63; Nsengiyumva, 2024). Burundi’s interethnic violence had devastating impacts on both Hutus and Tutsis, claiming hundreds of thousands of lives, forcing thousands to flee their homes and drove many others into exile (Russell, 2015:440; Nsengiyumva, 2024).
Burundi’s 1972 interethnic conflicts became significant in shaping the future Hutu–Tutsi relationship, harmony and democratisation efforts. Although not all Tutsis were perpetrators, neither were all Hutus victims. The tragedy had solidified the fear of genocide in both communities. As Oketch and Polzer observe, the 1972 violence “crystallized ethnic tensions in such a way that all subsequent crises have been … [their] consequence” (2002:96). After the 1972 interethnic massacres, Hutus and Tutsis associated every single killing against their ethnic group with genocide and justified later violence against the other group as self-defence. More significantly, the 1972 massacres heralded a culture of impunity, which continued to influence its judicial system, social and political life to the present day (Reyntjens, 2000:7). As none of the perpetrators were prosecuted or punished, the dastardly events emboldened those responsible for massive human rights violations with the resolve to continue to act with impunity, without accountability or the fear of prosecution by either the domestic judicial system or the international community.
Colonel Jean-Baptiste Bagaza’s regime that came to power in November 1976 promised to end military rule and initiate a democracy sensitive to social justice but failed to adequately resolve the question of ethnicity. As Lemarchand observes: “Exhortations to national unity did little to alter the stranglehold of Tutsi elements on the apparatus of power” (1996:108). Under Bagaza, Burundi remained firmly under the control of the Tutsis, and there was no sign of democracy at all. Instead, the army continued to function as an instrument of political and economic domination of the state by the Tutsis (Bentley and Southall, 2005:44). The regime banned all references to ethnic identities, with an additional provision that all references to ethnicity either in public or private spaces would be grounds for charges of incitement to “racial hatred” (Lemarchand, 1996:108). Moreover, while the Tutsis enjoyed full access to education and employment in the state’s organs, the Hutus continued to be marginalised, disempowered and excluded from the government, army, police and the judiciary (Bentley and Southall, 2005:44; Burihabwa and Curtis, 2021:1227). Burundi’s security agencies clamped down on anyone suspected of disloyalty, thus leading to the quick disappearance of the freedom of expression. As Lemarchand notes: “The Bagaza regime’s ‘controlled democratization’ had now given way to a ruthless police state – control everywhere and no democracy” (1996:116). The Bagaza regime’s years of authoritarianism incited the first Hutu rebel movements’ creation – the Party for the Liberation of the Hutu People (PALIPEHUTU) and the Front for National Liberation (FROLINA) – in Tanzania. These Hutu movements were not strong enough to threaten the regime at the time. However, they were well organised, carried out active training in refugee camps and were determined to conduct resistance against the long-standing political and economic dominance under the Tutsis (Reyntjens, 2000:8).
3. Attempts at inclusive democracy and interethnic harmony
Burundi’s political context began to change after another round of interethnic violence that affected thousands of Hutu and Tutsi victims in 1988 (Nsengiyumva, 2024:33), less than a year after Pierre Buyoya overthrew Jean-Baptiste Bagaza in September 1987. On coming to power, President Buyoya pledged to “heal the Hutu–Tutsi rift” (Twagiramungu, 2016:72). His early process of political liberalisation amounted to a policy of political reconciliation and inclusiveness, but without democracy. Buyoya took Hutu people’s grievances seriously and incorporated them into government. However, as they claimed fuller participation in public life, his ambition for a rapprochement between the conflicting ethnic groups faced multiple challenges. For instance, as the reins of power were still in the hands of the Tutsi elites, the Hutus continued to be maltreated, prosecuted and discriminated against. His intention to draw more Hutus into the political elite angered the Tutsi radicals, as well as those in the army who feared that the new reforms would end like the 1959 Rwanda-style anti-Tutsi violence. Consequently, Buyoya’s slow but cautious moves to mollify hardliners within the Tutsi ruling elite and the disgruntled Hutus resulted in uprisings in August 1988 (Twagiramungu, 2016:64). Rumours and fears of a ‘new 1972’ were capitalised on by the PALIPEHUTU activists, who killed several hundreds of Tutsis while many others fled their homes. In response, the Tutsi-dominated armed forces, claiming to “restore order”, randomly killed thousands of Hutu civilians and forced many more into exile (Bentley and Southall, 2005:44).
In contrast to 1972, the international community reacted with shock and horror to the news of Burundi’s 1988 interethnic violence. Human rights organisations, such as Amnesty International, denounced the brutality of the Tutsi-led army and gave detailed clinical accounts of the agonies sustained by the Hutus during the violence (Lemarchand, 1996:128–129). Official reactions in Belgium and Canada, speaking through their Foreign Ministries, requested that an international commission of inquiry be sent to Burundi. The European Parliament, and even the World Bank (despite having no record of addressing human rights issues), expressed concerns through resolutions in which Burundi’s 1988 interethnic massacres were explicitly linked to policies of ethnic discrimination. The United States too, as opposed to its policy in 1972 that “the United States simply should not interfere in any way with the internal affairs of another country”, urged the Burundian regime to “maintain and greatly increase its efforts at national reconciliation” and prosecute “individuals responsible for recent atrocities” (Lemarchand, 1996:129–130).
Under such intense international pressures, President Buyoya resolved in October 1988 to rectify the past ethnic discrimination through policies for ethnic reconciliation and power-sharing (Daley, 2006:670). His first actions towards the restoration of peace and national unity were of considerable psychological value. He showed awareness of the legitimacy of Hutu grievances and created the post of prime minister for a Hutu and established a commission of national unity designed to study ethnic violence with an equal number of Hutus and Tutsis. Between late 1988 and early 1991, Hutu people were increasingly visible in the state’s apparatus, education and civil service (Daley, 2006:670). Nevertheless, as the Tutsi-led army and other state security services feared that the reforms to liberalise Burundi would end in a Rwanda-like anti-Tutsi genocide, they all resisted changes. But when the “winds of change” blew over Africa (Reyntjens, 2000:10), President Buyoya was enticed by the UN to introduce a multiparty democracy in Burundi (Niyukuri 2020:133). As Wimmer observes:
Contrary to the most fervent advocates of exporting democracies across the world, democracy does not automatically produce interethnic harmony. Especially during the early decades of democratisation, tensions along “ethnic” lines may be heightened and lead to violence and finally the abortion of the democratic process itself. … The very nature of democratic legitimacy provides incentives for formulating ethnic and nationalist claims and mobilising followers along these lines (2003:112).
The effectiveness of the Buyoya regime’s policy designed to deal with the ethnic divide through democratic reforms was finally put to the test in the elections of June 1993. Judging from Burundi’s past interethnic atrocities, the presidential race became a two-party affair defined in ethnic terms. While FRODEBU was conceptualised as the defender of Hutu interests, UPRONA was perceived as a Tutsi political party. Resultantly, when sympathisers of the outlawed PALIPEHUTU participated in the activities of FRODEBU, they encouraged the Hutus to join it and portrayed it as the only party that embodied the Hutu causes. Local FRODEBU leaders also embarked on manipulating ethnicity during their campaigns. In this context, some Tutsis who tried to join FRODEBU were menaced by UPRONA supporters who designated them “traitors” and forced them to abandon the “presumed party” (Reyntjens, 2000:10). The ethnic interpretation also gained in cogency from the presidential and legislative elections of June 1993 as each candidate put forward to the electorate benefitted from strong support from either Hutu- or Tutsi-dominated political parties.
Burundi’s additional barriers to democratisation and peace occurred in the legislative elections. Ndadaye’s FRODEBU defeated Buyoya’s UPRONA and Pierre-Claver Sendegeya’s Party for Reconciliation of the People (PRP). For the first time in the history of Burundi’s politics, a civilian, and more substantially, a Hutu won the presidential elections (Ndikumana, 1998:36). However, in the legislative elections, the distribution of seats (80% versus 20%) approximated the ethnic breakdown of Burundi’s population. The belief that FRODEBU’s victory translated demographic majority into political majority, and ethnicity was a significant factor in determining the outcome of the votes raised serious concerns and fears among the Tutsi minority. As Magnarella notes: “Many Burundi Tutsi regarded the election as a Hutu victory, rather than as a majority decision” (2005:814). Consequently, under cover of night on 20–21 October 1993, several units of the army engaged in a military coup during which President Ndadaye and some FRODEBU cadres were killed (Oketch and Polzer, 2002:100) after almost 100 days of his presidential term (Ndikumana, 1998:36).
The aftermath of President Ndadaye’s assassination increasingly radicalised Burundi’s social and political landscapes and handicapped the search for a peaceful solution. Burundi experienced a power vacuum, as the Hutu-led government went into exile and refused to entrust the Tutsi-dominated army with its members’ protection. Moreover, while violence arose in Bujumbura, Hutus in other provinces under the command of FRODEBU local authorities began resistance to the army and reprisals against Tutsi and Hutu supporters of UPRONA. In response, the army and local Tutsis embarked upon attacking and killing Hutu people and followers of FRODEBU. In this context, insurgent attacks conducted by the Hutu rebel movements and reprisals by the Tutsi-led army between 1993 and 2005 left hundreds of thousands of Hutu and Tutsi people killed, with many others displaced or forced into exile. Burundi’s democracy was technically non-existent.
While some FRODEBU leaders engaged in talks during early 1994 (Reyntjens, 2006:117), others resolved in June 1994 to launch a political movement, the National Council for the Defence of Democracy (CNDD) and its armed wing, the Forces for the Defence of Democracy (FDD). They started attacking Burundi from bases in the eastern DRC (Bentley and Southall, 2005:53). In the second half of 1995, other Hutu rebel movements, such as the National Liberation Front (FNL), an armed faction of the PALEPEHUTU, and FROLINA stepped up their actions against the Tutsi-dominated regime in Burundi. By February 1996, Burundi’s institutions were increasingly paralysed by ethnic conflicts and violence. As an attempt to stop the anarchy that prevailed in Burundi, former President Buyoya led a coup on 25 July 1996 (Reyntjens, 2016:68). However, the leaders of Ethiopia, Kenya, Rwanda, Tanzania, Uganda and Zaire (DRC) denounced the coup and decided to impose uniform radical economic sanctions on Burundi. The European Union and the US endorsed the regional initiative, as did the UN.
4. Arusha: successful pacification model, failed democracy
The Arusha peace negotiations for Burundi started from June 1998 under the mediation of former Tanzanian President, Julius Nyerere, and were taken over by former South African President, Nelson Mandela, after Nyerere’s death (Reyntjens, 2016:68). They were inclusive and involved multiple state, non-state, regional and international actors, as well as the conflicting Hutu and Tutsi groups. Pressure from the mediators and regional leaders produced results in August 2000 when the Arusha Peace and Reconciliation for Burundi declaration was signed. The agreement’s main provisions called for a consociational form of power-sharing democracy and were met with hope and optimism as a recipe towards a new era for the Hutu–Tutsi cohabitation, democracy and peace in Burundi. This constitutional formula was not only designed to take Burundi out of the 1993–2005 civil war and accompanying cycles of interethnic massacres but to also ensure that the rights of the Tutsi minority would reconcile with the demands of the Hutu majority (Lemarchand, 2009:158). It thus suggested solutions to the Hutu people’s grievances related to their exclusion from the army, the economy and political power structures and addressed the minority Tutsi’s fears of being outnumbered.
Under considerable pressure from regional leaders and the facilitator, Burundi’s power-sharing agreement was ratified on 6 August 2004 in Pretoria by the Burundian government and the main Hutu-led rebel movement, the CNDD–FDD. This peace agreement also presented many reasons for optimism about Burundi’s future (Falch and Becker, 2008:22). First, it was designed in accordance with requirements of proportional representation to accommodate the Hutu, Tutsi and Twa ethnic groups and political parties. Second, it sought to strike an appropriate balance between the different ethnic groups in the executive and legislative organs of government and the communal councils. It also called for inclusiveness in the composition of the army.
Similarly, the post-transition constitution was consociational and was designed to combine majority rule with minority protection (Reyntjens, 2006:119) and enshrined the complexities of power-sharing with greater attention to minority rights (Lemarchand, 2009:167). For instance, it emphasised that in order to attain an ethnic balance, the president was to be seconded by two vice-presidents: one Hutu from a Hutu-dominated party and another Tutsi from a Tutsi-led party. The composition of the cabinet was to be made up of 60% Hutus, 40% Tutsis and 30% women – as well as in the National Assembly. Furthermore, it stressed a 50/50 equal representation of Hutus and Tutsis in the Senate, the army and the national police. Rooted in South Africa’s model, the Burundian constitution put into practice Lijphart’s model of consociationalism (Vandeginste, 2009:65). Lijphart’s logic is predicated on the assumption that:
Majority rule is a recipe for failure where society is deeply divided by religious, racial, or ethnic fault lines. In such an environment, the existence of group identities needs to be recognized and accommodated through inclusion rather than exclusion (Lemarchand, 2009:167).
The interim constitution was finally approved by referendum in February 2005 by overwhelming support from about 92% of voters, thus paving the way for the local, parliamentary and presidential elections scheduled in June and August 2005 (Reyntjens, 2006:119). Although the peace process in Burundi faced difficulties, it was commonly interpreted as a successful example of internationally and regionally supported peacebuilding and state-building, especially since the region had been so volatile since the 1994 genocide in Rwanda (Burihabwa and Curtis, 2021:1222).
In this context, Burundi’s first post-Arusha elections differed from those held in 1993 and their aftermath in many ways. Burundi’s 1993 elections were bipolar in a dual sense and were contested by two main political parties – FRODEBU and UPRONA – and two ethnic groups – Hutu and Tutsi. In 2005, however, the political landscape became multipolar, heterogeneous and more complex: two main Hutu-led parties, the CNDD–FDD and the FRODEBU, were in competition for the Hutu vote, as were the Tutsi-led parties for the Tutsi vote. As the constitution prescribed political parties to ensure ethnic diversity, the competing parties included Hutu and Tutsi candidates in their ranks. This political context made it difficult for Burundian people’s voting behaviour to be determined by ethnicity. Burundi thus experienced a significant paradigm shift since independence: the old Hutu–Tutsi tensions were replaced by disputes over power between (and within) political parties rather than ethnic groups (Palmans, 2012:7).
Burundi’s 2005 elections differed from those of 1993 in other significant areas. For instance, the political class and civil society adopted a more constructive and less radical approach. Although relations between political parties were not exempt from conflicts, their leaders refrained from taking positions that would result in violent deadlock and used language conducive to keeping communication channels open. Moreover, as opposed to the hate media that prevailed in the 1993–4 period, civil society organisations, particularly the media, eschewed partisan positions. Burundi’s National Defence Forces (FDN) constituted elements of the former Tutsi-led army and integrated combatants of the FDD: its command structures were unified, its troops harmonised and stayed away from politics. External support also explained the peaceful outcome of Burundi’s 2005 elections. Independent radio stations funded by international donors developed into peace media. Regional leaders and the international community played significant roles as well. For instance, the preventive deployments of peacekeepers helped former rebel forces’ integration into Burundi’s army. The UN peacekeepers’ presence in Burundi practically reduced the risk of violent backlash from the former Tutsi-led army and paved the way for the implementation of the power-sharing and peace agreement.
Despite some reported cases of the Hutu-led FNL’s disruptive attacks on the polls in certain areas, Burundi’s 2005 elections proceeded calmly, and the CNDD–FDD won both the municipal and parliamentary elections. The CNDD–FDD’s chairman and former Hutu rebel leader, Pierre Nkurunziza, was elected in a joint congress by the National Assembly and the Senate on 19 August and took the oath of office on 26 August 2005. His inaugural speech contained the following keywords: “Today’s victory is that of all Burundians, the great and the little folk alike … defend peace, tranquillity and developments for all … combat the ideologies of ethnic division and genocide” (Reyntjens, 2006:129). To emphasise the message of inclusiveness, he appointed two vice-presidents – a Hutu and a Tutsi. Nkurunziza’s early words and actions seemed to reflect the rebirth of Burundi’s democracy and hopes for an enduring peace and interethnic harmony.
However, a few days after the 2005 elections, local and international observers were shocked by the new regime’s method of governance. The CNDD–FDD consolidated its hold on power so rapidly and did so by violating all the principles of democracy, peace and good governance. Its leaders took a different approach from that which was envisaged during the Arusha process – rather than inclusion, transparency and liberal principles, the party ruled with an increasingly authoritarian bent (Burihabwa and Curtis, 2021:1223). Almost all senior servants appointed by the government were CNDD–FDD members (Nkurunziza, 2023:120). Although peace prevailed in most provinces of Burundi, the Hutu-led FNL was still fighting against the national army, which retaliated with brutality (ICG, 2006:2). In late 2005 and early 2006, government security forces arbitrarily arrested and committed with impunity extrajudicial executions of suspected FNL combatants and their supporters (HRW, 2006:3). They also targeted the Hutu-led FRODEBU officials, apprehended some because of their political affiliation and others for collaboration with the FNL rebels. While some foreign diplomats in Burundi condemned the regime for its poor record on democracy and freedom of expression, others feared their effects on a state that just emerged from many years of ethnic conflicts and civil war.
In light of the existing conditions, Burundi’s 2010 elections became another challenge to democracy and peace. In the run-up to these elections, the CNDD–FDD repeatedly sought to arrange the legal and institutional context to ensure that it could not lose them. Its leaders used force and intimidation, targeted political opponents in efforts to restrict their political freedom, wrongfully arrested and detained others and used possible strategies to weaken or divide opposition parties. In particular, the CNDD–FDD youth wing’s physical training, war songs and quasi-military organisation raised fears of militia violence and a large-scale intimidation campaign. Such political behaviour led the Hutu-led FNL and FRODEBU to mobilise their youth wings to oppose intimidation tactics used by the regime.
Like in 2005, Burundi’s electoral rivalry in 2010 was inclusive and involved different Hutu- and Tutsi-led political parties (Palmans, 2012:11). Judging from Burundi’s deteriorating socio-economic and political conditions, Hutu and Tutsi-led opposition parties were convinced that the CNDD–FDD would lose the elections. However, the CNDD–FDD won the 2010 elections, and the main opposition parties decided to boycott the electoral process. Given the disagreement over the electoral outcome, Burundi’s democracy and peace were once again at risk. The CNDD–FDD already held a reputation of being an authoritarian regime long before the elections. Also, the fact that most of the opposition parties that were frustrated and felt marginalised had their roots in the armed struggle outside the institutions, the potential for violence in Burundi grew significantly. Consequently, the months after the 2010 elections and throughout the entire period of 2011 were characterised by systematic killings and arrests of opposition party representatives. In such a context where a real parliamentary opposition was absent, civil society and the media had a duty to promote democracy and ascertain transparency in Burundi’s politics. However, the regime started describing them as the opposition in disguise, subjecting their members to regular harassment, apprehensions and even killings. Without any form of opposition, even the 2015 elections were to resemble a mere formality of renewing the CNDD–FDD’s mandate.
Matters came to a head in 2015 when President Nkurunziza decided to run for and won a controversial third term in office. Long before Burundi’s 2015 elections, many in the opposition parties, as well as national and international organisations, church groups and foreign diplomatic representatives expressed their concerns over his ambition to seek another term of office (FIDH and Ligue Iteka, 2016:21). The CNDD–FDD’s political behaviour was becoming intolerable not only for the opposition but even for some within the party itself (Alfieri, 2016:249). Consequently, Nkurunziza’s nomination as the CNDD–FDD’s presidential candidate on 25 April 2015 prompted unprecedented mass demonstrations by thousands of Burundians in the street of Bujumbura. What was evident was the fact that the protesters were fuelled by a much broader set of grievances and frustrations (Van Acker, 2015:7). These were notably the lack of economic and social perspectives and the increased sense of insecurity related to the regime. As personal connections matter more than people’s merits, affiliation with the CNDD–FDD provides access to jobs and other opportunities (BTI, 2024:31). Even so, all its members are not equal and do not benefit in the same way from their association with the party. Such a dissatisfaction was the cause of the creation of the frondeur dissent movement in March 2015 (Van Acker, 2015:6) constituted of CNDD–FDD members who hoped for a change in the party’s leadership (Alfieri, 2016:249). A group of dissent army officers who attempted a coup on 13 May included some CNDD–FDD elements who distanced themselves from the party’s exclusionary policies and violent practices.
On 18 June 2020, General Evariste Ndayishimiye took over from Nkurunziza, and this transition of power seemed to assure many regional and international observers of a new era for democracy and peace in Burundi. Ndayishimiye won the elections of May 2020 in a process that the opposition parties described as unfair (Freedom House, 2024). The appointment in his government of the CNDD–FDD military hardliners who embrace openly partisanship and ethnic nationalism has been another cause of concern. Like in 2010 and 2015, weeks before and after the 2020 elections, Burundi has witnessed a series of government violence and extrajudicial killings, targeting particularly the Hutu-led National Congress for Liberty’s (CNL) followers (HRW, 2020; UNHRC, 2020; Freedom House, 2024). In September 2023, the CNDD–FDD secretary general publicly called on his party supporters to eliminate “wolves” – an act interpreted as a call for violence against opposition party followers (Freedom House, 2024).
Unlike the 2015 elections, Burundi’s 2020 general elections did not plunge the state into chaos but demonstrated how elections can be used for authoritarian consolidation and gradual return towards de facto single-party regimes. To keep its popularity safe and silence any threats from opposition forces, the CNDD–FDD uses its youth, the intelligence service and national police in violence and intimidation to influence people’s political choices (HRW, 2020; Freedom House, 2024). In the period before Burundi’s 2010 elections, the CNDD–FDD involved them in its “plan to wipe out the political opposition” (Immigration and Refugee Board of Canada, 2013:1). During the 2015 protests, these pro-regime forces arbitrarily arrested people, politicised detention, often in inhumane and life-threatening conditions and were involved in multiple acts of violence, including extrajudicial killings and disposed of victims in mass graves (U.S. Department of State, 2023:1). To this day, the CNDD–FDD relies on these forces in its campaign against opposition parties that enjoy near impunity in their abuses (HRW, 2020, 2023).
In this failure to develop a culture of democracy, the CNDD–FDD seeks possible ways to create an environment of self-censorship and minimal political debate in Burundi. Although political party formation is legally allowed, opposition party leaders’ activities are discouraged under the threat of retaliatory violence, repression, arrest, even assassination (Freedom House, 2024). Like after the 2015 electoral violence, the CNDD–FDD unceasingly claims institutional normality in Burundi, but the state’s score (14%) and status on political rights (4 out of 40) and civil liberties (10 out of 60) (Freedom House, 2024) reflect the opposite. Instead, Burundi’s political context demonstrates how political power concentrated in the party and military can detach itself from the citizenry, creating a system that appears to be stable but functions against citizens, not for and with citizens (BTI, 2024:3). The CNDD–FDD’s increased violence against the opposition and restriction of democratic space (including the rising cost of living, a stream of corruption allegations and collapsing public services – e.g., electricity) are more likely to fuel political instability and violence in the 2025 legislative elections.
5. Ethnic dimensions in the post-Arusha conflicts
Despite international fears of genocide and attempts by some political actors to inject ethnicity into the post-Arusha crises, the constitutional formula brought in by the 2000 Arusha agreements still prevail. Ethnic cohabitation is noticeable not only in the armed forces and government but also in informal settings where no quotas are imposed by the agreements. Although political violence is still a major feature of Burundi’s landscapes, Hutu and Tutsi people live side by side from the highest spheres of power to the darkest valleys of poverty. As much as political elites have tried to instrumentalise ethnic antagonism during the 2015 conflicts, for example, growing inter-elite trust and socialisation in the atmosphere of the post-Arusha have trickled down beyond formal state institutions as a driver of political and social reconciliation.
Several features of the electoral crises demonstrate the remarkable resilience of ethnicity in post-Arusha Burundi’s social and political contexts. Political opposition to Nkurunziza’s third-term mandate in 2015 transcended ethnic boundaries in a sense that anti-third-term protesters included many young and middle-aged people, students, women, human rights activists, political dissident members of the CNDD–FDD and some government officials from both Hutu and Tutsi ethnic groups (FIDH and Ligue Iteka, 2016:25). Moreover, despite the coup d’état, the FDN did not fall apart along ethnic lines, as the coup plotters included both Hutu and Tutsi figures (Vandeginste, 2015:632). The fallout of the 2010 elections and pre-electoral crisis also have given rise to a new kind of political bipolarity that is no longer Hutu–Tutsi. The political parties that withdrew from the electoral competition formed a coalition – the Democratic Alliance for Change-Ikibiri (ADC-Ikibiri), comprised of Hutu- and Tutsi-led parties (Twagiramungu, 2016:57). And when the CNDD–FDD started attacking and forcing their leaders to flee from Burundi fearing for their security, the regime did so regardless of their ethnic groups. Although identity politics remains predominant in Burundi today, the main fault line is currently determined by political partisanship rather than ethnic ties. After the 2010 elections, the fault line was adherence or opposition to the CNDD–FDD, and in the run-up to the 2015 elections, the fault line was support or opposition to Nkurunziza’s third term (Vandeginste, 2015:633).
Furthermore, since 2010 when the political space was completely closed, Burundi has become a home for increasing attacks by Hutu- and Tutsi-led armed groups (UNHRC, 2020:6). Initially, Burundi’s political landscape largely opposed two former Hutu-led political movements – the CNDD–FDD and the FNL (Arieff, 2015:5). Efforts by the CNDD–FDD to crush the FNL resulted in torture, disappearances, extrajudicial killings and other forms of massacres of FNL members and supporters (Bouka and Wolters, 2016a:1). As intra-Hutu conflict between the CNDD–FDD and the FNL intensified, the Tutsi-led Movement for Solidarity and Democracy (MSD) emerged as another threat to the regime, leading government violence, targeting the Tutsi youth (Bouka and Wolters, 2016b:6). Due to political and economic marginalisation, Burundi has experienced another paradigm shift symbolised by an alliance between the two arch-enemies during the previous civil war: the Hutu-led FNL and the Tutsi-dominated National Front for the Revolution in Burundi (FRONABU-Tabara) (Bouka and Wolters, 2016a:6). These allied armed groups have been fighting the CNDD–FDD regime since 2015 from bases in the DRC (UNHRC, 2020:6; Aljazeera, 2023; HRW, 2023; Freedom House, 2024). Burundi regularly accuses Rwanda (which rejects the allegations) of hosting and training them and has closed its borders (Aljazeera, 2023). Due to the Great Lakes region’s politics and poor socio-economic prospects, the CNDD–FDD is more likely to restrict democratic space in the lead-up to the 2025 legislative elections, which could worsen insecurity countrywide (ICG, 2024).
6. Conclusion
In Burundi, as in many states in Africa, democracy is flourishing, but proper constitutional liberalism is not. As this study demonstrates, the CNDD–FDD’s initial failure to obtain a military victory on the battlefield has been remedied through the ballot box. Although the CNDD–FDD radical wings maintain ethnic power-sharing, they do so within the de factoone-party system. In this context, the democratic rights that were meant to be exercised by the people and for the people in which the supreme power is vested in the people are almost non-existent in Burundi. Instead, people’s rights have been hijacked to serve authoritarian regimes that claim popular support by pinning democratic labels upon themselves. From a state constantly on the brink of massive interethnic confrontation and even genocide, Burundi has now become a banal dictatorship governed by a group of untouchable people, poor administration at all levels and non-independent institutions.
Contrary to the belief that the forces of democracy are the forces of ethnic harmony and of peace (Zakaria, 1997:35), democracy has not produced peace in Burundi. The Arusha agreement’s “pacification model has prevented electoral violence from transforming into ethnic violence” (Vandeginste, 2015:636). Although it has rightly been praised for solving Burundi’s old Hutu–Tutsi conflicts and remains necessary in order to reassure ethnic minorities and to prevent the state from returning to past tensions, it has not offered a solution to some other governance deficiencies, such as the rise in militarism, the weakness of the rule of law, good governance and the respect for human rights. Burundi has instead become, since 2005, a textbook example of autocratisation through elections. The CNDD–FDD regime’s undemocratic practices have now sown the seeds of renewed grievance-based insurgency in Burundi. Whether after a short or lengthy period of armed confrontations, old recipes will again be called upon to restabilise Burundi’s political landscapes. These new rounds of power-sharing negotiations to restore democracy and peace are more likely to delay the internalisation of democratic electoral competition as the sole legitimate avenue to political power. For a state to reach a lasting peace, the journey needs to be led by all its people without threats, fear and violence to avoid conflicts. As prevention is better than cure, this article concludes with the following recommendations:
- To the international community:
- Increase support to civil society to help it promote the values of inclusive democracy and peace.
- To the government:
- Emphasise peace, unity and development of the state and its people without discrimination.
- Allow opposition parties, the media and civil society to work independently and free from fear and revise its violent approach to political opposition.
- Enhance the respect for people’s rights and constitutional power of citizens.
- Open opportunities to all its people.
- To opposition parties and armed groups:
- Renounce violence and think of other approaches to resolve political differences.
Reference list
Adejumobi, S. (2001) Citizenship, rights, and the problem of conflicts and civil wars in Africa. Journal of Human Rights Quarterly, 23 (1), pp. 148–170.
Alfieri, V. (2016) Political parties and citizen political involvement in post-conflict Burundi: between democratic claims and authoritarian tendencies. Civil Wars, 18 (2), pp. 234–253.
Aljazeera. (2023) Rebel attack in western Burundi kills at least 20. Aljazeera [Internet], 23 December. Available from: <https://www.aljazeera.com/news/2023/12/23/rebel-attack-in-western-burundi-kills-at-least-20> [Accessed 23 December 2023].
Arieff, A. (2015) Burundi’s electoral crisis: in brief. Washington, D.C., Congressional Research Service.
Bentley, A.K. and Southall, R. (2005) An African peace process: Mandela, South Africa and Burundi. In: Bentley, A.K. and Southall, R. eds. An African peace process: Mandela, South Africa and Burundi. Cape Town, Human Sciences Research Council Press. pp. 40–59.
Bertelsmann Stiftung Transformation Index (BTI). (2024) Country report–Burundi. Gutersloh, BTI.
Bouka, Y. and Wolters, S. (2016a) Analysing post-transition violence in Burundi. Pretoria, Institute for Security Studies.
Bouka, Y. and Wolters, S. (2016b) The battle for Burundi: is there a viable solution? Pretoria, Institute for Security Studies.
Burihabwa, N.Z. and Curtis D.E.A. (2021) Post-war statebuilding in Burundi: ruling party elites and illiberal peace. International Affairs, 97 (4), pp. 1221–1238.
Central Intelligence Agency (CIA). (2024) The world fact book: Burundi [Internet]. Available from: <https://www.cia.gov/the-world-factbook/countries/burundi/#geography> [Accessed 17 March 2024].
Daley, P. (2006) Ethnicity and political violence in Africa: the challenge to the Burundi state. Political Geography, 25 (6), pp. 657–679.
Falch, A. and Becker, M. (2008) Power-sharing in peacebuilding in Burundi: power-sharing agreements, negotiations and peace processes. Policy brief, Center for the Study of Civil War (CSCW) Papers. Oslo, The Peace Research Institute Oslo (PRIO).
Freedom House. (2024) Burundi: freedom in the world 2024. Washington, D.C., Freedom House.
Human Rights Watch (HRW). (2006) The Rwandan genocide: how it was prepared. New York, Human Rights Watch.
Human Rights Watch (HRW). (2020) World report 2020: Burundi. New York, Human Rights Watch.
Human Rights Watch (HRW). (2023) World report 2023: Burundi. New York, Human Rights Watch.
Immigration and Refugee Board of Canada. (2013) Ties between the Imbonerakure and the Burundian authorities.Toronto, Immigration and Refugee Board of Canada.
International Crisis Group (ICG). (2006) Burundi: democracy and peace at risk. Nairobi, ICG.
International Crisis Group (ICG). (2024) On the horizon: June–November 2024: Burundi. Nairobi, ICG.
International Federation for Human Rights and the Burundian League for Human Rights (Ligue Iteka). (2016) Repression and genocidal dynamics in Burundi. Paris, FIDH.
Lemarchand, R. (1970) Rwanda and Burundi. London, Pall Mall Press.
Lemarchand, R. (1996) The restructuring of state-society relations. In: Lemarchand, R. ed. Burundi: ethnic conflict and genocide. New York, Woodrow Wilson Centre Press. pp. 108–116.
Lemarchand, R. (2009) Burundi’s endangered transition. In: Lemarchand, R. ed. The dynamics of violence in central Africa. Pennsylvania, University of Pennsylvania Press. pp. 158–179.
Magnarella, J.P. (2005) How could this happen? The background and causes of genocide in Rwanda. Journal of International Criminal Justice, 3 (4), pp. 801–822.
Ndikumana, L. (1998) Institutional failure and ethnic conflicts in Burundi. African Studies Review, 41 (1), pp. 29–47.
Niyukuri, B. (2020) Reconciliation in Burundi crisis: a practical theological approach. In: Nel, M.J., Forster, D.A and Thesnaar, C.H. eds. Reconciliation, forgiveness and violence in Africa: biblical, pastoral and ethnical perspectives. Stellenbosch, African Sun Media. pp. 131–148.
Nkurunziza, J.D. (2015) Timing and sequencing of post-conflict reconstruction and peace-building in Burundi. Working Paper No. 406. Massachusetts, Political Economy Research Institute (PERI).
Nkurunziza, J.D. (2023) The origin and persistence of state fragility in Burundi. In: Bizhan, N. ed. State fragility: case studies and comparisons. London, Routledge. pp. 101–140.
Nkurunziza, J.D. and Ngaruko, F. (2005) Why has Burundi grown so slowly? The political economy of redistribution[Internet]. Available from: <https://www.academia.edu/529104/Why_has_Burundi_grown_so_slowly> [Accessed 10 June 2016].
Nsengiyumva, I. (2024) Perspectives on the 1972–1973 massacre and post-conflict reconstruction efforts in Burundi. African Journal on Conflict Resolution, 24 (1), pp. 29–46.
Nshimirimana, F. (2020) Managing ethnic conflicts, militarisation, and democratic failure in the post-Arusha Burundi. African Journal of Democracy and Governance, 7 (3&4), pp. 151–172.
Oketch, J.S. and Polzer, T. (2002) Conflict and coffee in Burundi. In: Lind, J. and Sturman, K. eds. Scarcity and surfeit: the ecology of Africa’s conflicts. Pretoria, Institute for Security Studies. pp. 85–158.
Palmans, E. (2012) Burundi’s 2010 elections: democracy and peace at risk? Brussels, European Centre for Electoral Support (ECES).
Reyntjens, F. (1993) The proof of the pudding is in the eating: the June 1993 elections in Burundi. The Journal of Modern African Studies, 31 (4), pp. 563–583.
Reyntjens, F. (1995) Burundi: breaking the cycle of violence. London, Minority Rights Group International.
Reyntjens, F. (2000) Burundi: prospects for peace. London, Minority Rights Group International.
Reyntjens, F. (2006) Briefing: Burundi: a peaceful transition after a decade of war? African Affairs, 105 (418), pp. 117–135.
Reyntjens, F. (2016) Institutional engineering, management of ethnicity, and democratic failure in Burundi. Africa Spectrum, 51 (2), pp. 65–78.
Russell, A. (2015) Obedience and selective genocide in Burundi. Africa: The Journal of the International African Institute, 85 (3), pp. 437–456.
Twagiramungu, N. (2016) Burundi: the anatomy of mass violence endgames. In: Conley-Zilkic, B. ed. How mass atrocities end: studies from Guatemala, Burundi, Indonesia, Sudan, Bosnia-Herzegovina, and Iraq. New York, Cambridge University Press. pp. 56–82.
United Nations Human Rights Council (UNHRC). (2020) Report of the commission of inquiry on Burundi (A/HRC/45/32). Geneva, United Nations Human Rights Council.
U.S. Department of State. (2023) Burundi 2017 human rights report. Virginia, Bureau of Democracy, Human Rights, and Labor [Internet]. Available from: <https://www.state.gov/wp-content/uploads/2023/02/415610_BURUNDI-2022-HUMAN-RIGHTS-REPORT.pdf > [Accessed 30 November 2024].
Van Acker, T. (2015) Understanding Burundi’s predicament. Africa policy brief No. 11. Brussels, EGMONT Royal Institute for International Relations.
Vandeginste, S. (2009) Power-sharing, conflict and transition in Burundi: twenty years of trial and error. Africa Spectrum,44 (3), pp. 63–86.
Vandeginste, S. (2015) Briefing: Burundi’s electoral crisis – back to power-sharing politics as usual? African Affairs, 114 (457), pp. 624–636.
Wimmer, A. (2003) Democracy and ethno-religious conflict in Iraq. Survival, 45 (4), pp. 111–134.
Zakaria, F. (1997) The rise of illiberal democracy. Foreign Affairs, 76 (6), pp. 22–43.
Notes
[1] See also Lemarchand’s (1970) work on the international geopolitical implications on Burundi’s ethnic conflicts.
[2] See Lemarchand’s (1996, 2009) extensive coverage of the regional effects on Burundi’s ethnic conflicts.