Va-PoReg Country Report Burundi 1900–2025

Ruanda and Urundi were two separate kingdoms in the Great Lakes region. The Burundi Kingdom was founded in 1680 under the leadership of the royal clan. On 02/26/1885 at the Berlin Conference, Burundi was defined as part of Germany’s sphere.¹ In 1899, Germany established Ruanda-Urundi as part of the German East Africa protectorate (Cornwell/de Beer 1999: 84). For this time neither LIED nor V-Dem list Burundi in their data. While it was in name a protectorate, according to our coding rules it was a colony.

During World War I Ruanda-Urundi was occupied by Belgian forces, who established military rule in 1916 (Cornwell/de Beer 1999: 84). Both LIED and V-Dem’s PCLI classify political liberties as absent for this period. For the relevant regime period, V-Dem’s JCE is interpreted by us as indicating that judicial constraints on the executive were moderate, while V-Dem’s LCE is classified by us as indicating that legislative constraints on the executive were absent. Based on our observations, no multiparty executive or legislative elections were held during this period, which aligns with the observations of LIED. Therefore, we code this period as a direct occupation regime.

From this date on, Ruanda-Urundi was administrated as a trust territory by Belgium under a League of Nations mandate. In 1946 Burundi became a UN trust territory (Cornwell/de Beer 1999: 84). We code this period as a directly ruled colony. It was not before 1961 that Belgium granted Ruanda-Urundi internal autonomy and agreed on legislative elections to be held after substantial pressure from the UN, accompanied by a promise that the Belgians would prepare the territory for independence (Cornwell/de Beer 1999: 85).² On 08/17/1961, prior to independence, the Belgian administration of the UN Trust territory introduced universal suffrage to women. This right was confirmed at independence.³ Based on our observations, no multiparty executive or legislative elections were held during this period, which aligns with the observations of LIED. For this period, political liberties were coded as absent by LIED and V-Dem’s PCLI. For almost the entire regime period, V-Dem’s JCE is classified by us as indicating that judicial constraints on the executive were moderate, whereas V-Dem’s LCE is classified by us as indicating that legislative constraints on the executive were absent. For the year 1961, V-Dem’s JCE and LCE are both interpreted by us as indicating moderate constraints on the executive.

Because Burundi was granted internal self-governance along with elections and the promise of independence from the international side, this period is coded as quasi-independent. On 07/01/1962, the Kingdom of Burundi restored its complete independence as a constitutional monarchy, where executive power was vested in the mwami, while legislative authority was granted to the parliament.⁴ While the monarchy falls formally in the category constitutional, the monarchy selected and removed governments without parliamentary consultation, expanded its authority through various means, and appointed family members to crucial positions. Governments were structured to maintain a balance between Tutsis and Hutus (Lemarchand 1966:420-423, Lansford 2012e:205, Geddes/Wright/Frantz 2014: 50). While it is classified as a constitutional monarchy it is clearly an autocratic monarchy, being a borderline case between a constitutional and an autocratic monarchy. Based on our observations, only multiparty legislative elections were held during this period, which aligns with the observations of LIED. LIED categorizes the election as not competitive. V-Dem’s EF&FI scores elections during this time as being of ambiguous freedom and fairness. Their CEI scores them as not really clean. In 1962, as per Polity5’s classification, the executive’s authority was subject to minor institutional constraints. Besides, LIED continued to code political liberties as absent. V-Dem’s PCLI changed into a range which we interpret as not really present. Since then, as per Polity5’s classification, the executive’s authority was significantly constrained by institutional checks during this time. For 1962, V-Dem’s JCE is interpreted by us as indicating that judicial constraints on the executive were robust, while V-Dem’s LCE is classified by us as indicating that legislative constraints on the executive were limited. For the rest of the regime period, V-Dem’s JCE is interpreted by us as indicating that judicial constraints on the executive were moderate, while V-Dem’s LCE is classified by us as indicating that legislative constraints on the executive were limited.

A military coup led by Captain Michel Micombero reduced the powers of the monarch. King Mwambutsa was replaced with his youngest son and coup leader Micombero became prime minister. In November 1966, a second coup led by Tutsi military officers, headed by Prime Minister Micombero, led to the official shift from a monarchy to a republic. Micombero assumed control, and the government became predominantly influenced by Tutsi officers and politicians. During this time, Hutus were systematically removed from the officer corps, as well as from the bureaucracy and political sphere, through purges carried out in 1966, 1969, and 1972 (Lemarchand 1974a:75, Lemarchand 1974b:87, Geddes/Wright/Frantz 2014: 50-51). On 11/28/1966 General Michel Micombero overthrew the monarchy and ousted king Ntare V. He appointed himself president of the new republic.⁴ The National Revolutionary Council (NRC) headed by Michel Micombero took control of the government on 29/11/1966. The NRC confirmed the status of Union for National Progress (UPRONA) as the country’s only legal political party on 11/30/1966. Under Micombero’s rule an ideology known as „democratic centralism” was implemented, consolidating control over all national institutions and media, effectively establishing a military dictatorship.⁵ Based on our observations, no multiparty executive or legislative elections were held from 1967 to 1981, which aligns with the observations of LIED. Executive elections were held from 1982 onward and legislative elections from 1984 onward, which weren’t multiparty. Since the center of the rule was the NRC and not UPRONA the regime is classified as a military autocracy. According to Polity5, during this period, the executive held unlimited authority with no institutionalized constraints on decision-making power. For the relevant regime period, V-Dem’s JCE is interpreted by us as indicating that judicial constraints on the executive were moderate, while V-Dem’s LCE is classified by us as indicating that legislative constraints on the executive were absent. According to FH’s classification for the assessed regime period, a score between 11 and 14 makes the country not free, which we also place in the not free category. For the entire period both LIED and V-Dem’s PCLI classified political liberties as not present.

Bagaza was overthrown by Major Buyoya and a military junta was established to rule the country. Buyoya’s government is considered a new regime because he initially dissolved the Tutsi ruling party and established an all­military ruling group of 31 officers, thus changing the identity of the group able to influence policy (Maher 2004:946, Lansford 2012e:205, Geddes/Wright/Frantz 2014: 51). In 1988, under the authoritarian regime of President Buyoya, a commission dedicated to the study of national unity was established, ensuring equal representation from both Hutu and Tutsi groups. The commission’s findings led to the creation of a Charter on National Unity, which received approval through a national referendum in 1991. Following this, a Constitutional Commission was convened to draft a new constitution, which was subsequently endorsed in a 1992 referendum, thereby laying the groundwork for the national elections held in 1993 (Haggard/Kaufman/Teo 2016: 18). Based on our observations, no multiparty executive or legislative elections were held during this period, which aligns with the observations of LIED. Based on Polity5’s evaluation, during this period, the executive faced weak constraints, classified as Intermediate Category 1 between unlimited authority and slight limitations. As classified by FH for this regime period, the country is scored from 11 to 14 as not free, which corresponds to our interpretation of not free. For this time LIED and V-Dem’s PCLI considered political liberties as absent. For the relevant regime period, V-Dem’s JCE and LCE are both interpreted by us as indicating moderate constraints on the executive.

On this date the first multiparty presidential elections took place. They were rated as free and fair by international observers and all candidates accepted the results.⁶ Melchior Ndadaye, leader of the Hutu-dominated Front for Democracy (FRODEBU) defeated (Tutsi) incumbent President Pierre Buyoya. The election marked a turning point for Burundi, signifying the conclusion of the Tutsi-dominated state supported by the military since 1966 since Ndadaye was a Hutu (Maher 2004:946, Lansford 2012e:206, Geddes/Wright/Frantz 2014: 51). Ndadaye was the country’s first Hutu head of state, leading a pro-Hutu government.⁷ He became president on 07/10/1993. Despite his efforts to bridge the ethnic divide, his reforms faced opposition from soldiers in the Tutsi-dominated army.⁸ Based on our observations, multiparty executive and legislative elections were held during this period, which aligns with the observations of LIED. LIED categorizes elections in this period as not competitive. V-Dem’s EF&FI scores the 1993 election as somewhat free and fair. Their CEI scores it as not really clean. LIED classifies political liberties as absent for this period. According to V-Dem’s PCLI political liberties were in an ambiguous state in this period. The reasons for classifying this period as an electoral hybrid regime become more apparent when viewed together with the period from 10/27/1993 onwards, as both periods are connected and were only interrupted by six days of military rule.

On this date, President Ndadaye along with several other government leaders were killed by rebelling Tutsi-dominated military forces who went onto establish the Committee of Public Salvation junta which lasted for only six days. Ngeze was established as head of state by the junta, asserting guardianship legitimation through violent seizure of power and military governance structures rather than procedural facilitation. ⁹ The assassination of elected officials and establishment of a military junta represent explicit guardianship claims that the military possessed the authority to override democratic institutions and determine political outcomes through force. This guardianship legitimation is confirmed by the institutional structure: the Committee of Public Salvation functioned as a military governing body exercising sovereign authority until public unrest and international opposition forced the junta to step down after six days.

Due to public unrest and international opposition Ngeze stepped down with the rest of the junta, on this date, and passed power to the next highest-ranking civilian official who had survived the coup, Sylvie Kinigi, a Tutsi.⁹ He became acting civilian prime minister and served until 02/05/1994.¹⁰ Kinigi’s government was comprising of 15 of the 22 ministers before of the coup. However, the period was characterized by a civil war with ongoing violence between Hutu rebels and the Tutsi majority army, which led to human rights abuses. In early 1994, Cyprien Ntaryamira, a Hutu, was elected president by the parliament after a modification of Article 85 to allow this indirect election. The parliament opposition, however, argued that this amendment was not in line with the constitution. Therefore, the matter was brought before the Constitutional Court, who ruled in favor of the opposition. Subsequently Kingi dismissed all Tutsi judges, the Hutu judges resigned. This led to violence in the capital. Consequently, Kingi reached a compromise with the opposition. Ntaryamira, was sworn in on 02/05/1994 as president and the ethnic Tutsi Anatole Kanyenkiko assumed the prime minister office.¹¹ However, both Ntaryamira and Juvénal Habyarimana, the president of Rwanda, also a Hutu, died together when their plane was shot down in April 1994. This event led to an increase in the number of refugees fleeing to Rwanda. Sylvestre Ntibantunganya, a Hutu and the Speaker of Parliament, assumed the presidency in October 1994. A coalition government, comprising 12 out of the 13 parties, was formed. Kanyenkiko resigned, but remained in power until a new executive was confirmed. The outgoing prime minister and the new president maintained good working relations.¹² Based on our observations, multiparty executive and legislative elections were held during this period, which aligns with the observations of LIED. The 1993 election was not competitive (LIED). Moreover, V-Dem’s CEI scored not really cleanliness. However, the overall election conditions were somewhat free and fair (V-Dem EF&FI). Per FH, for this regime period, the country scores between 11 and 14 as not free, which we also interpret as not free. LIED classifies political liberties as absent for this period. According to V-Dem’s PCLI the presence of political liberties was ambiguous. During this regime period, V-Dem’s JCE is classified by us as indicating that judicial constraints on the executive were moderate, while V-Dem’s LCE is classified by us as indicating that legislative constraints on the executive were also moderate. This regime in this period is classified as an electoral hybrid regime. However, the constitutional line of succession was respected, even during this turbulent period.

With support from the populace and the military, on this date Buyoya overthrew the civilian government in a military coup d’état. It was justified by its lack of progress in stopping ethnic killings. In 1996, Hutu rebels killed over three hundred Tutsis, after which the third Hutu president, Sylvestre Ntibantunganaya, disappeared before a military takeover. Pierre Buyoya, a Tutsi, became the head of the new military government. He had already been in power from 1987 to 1992 and was now named interim president (Haggard/Kaufman/Teo 2016: 74, Maher 2004:948, Lansford 2012e:206, Geddes/Wright/Frantz 2014: 51). It was the fourth such takeover of the Burundian government since the country’s independence in 1962, and the second to result in Buyoya taking power. Despite the coup, the parliament Assembly and political parties continued to operate, albeit under constraints.¹³ In the weeks immediately following the coup, more than 6.000 people were killed in the country.¹⁴ Economic sanctions were also imposed by the international community because of the nature of Buyoya’s return to power but were eased as Buyoya created an ethnically inclusive government. Buyoya established a new partnership in 06/1998 with the National Assembly, which was under the influence of the Hutu-supported Front for Democracy in Burundi (FRODEBU). Subsequently, this move played a role in facilitating the emergence of the Arusha Accords in 2000. These Accords brought out a system of ethnic power-sharing, ultimately leading to the resolution of the Civil War in 2005.¹⁵ According to LIED, no multiparty executive or legislative elections were held during the specified period. As classified by FH for this regime period, the country is scored from 11 to 14 as not free, which corresponds to our interpretation of not free. In addition, LIED codes political liberties as absent and V-Dem’s PCLI as ambiguous. In 1997, as per Polity5’s categorization, the executive experienced minimal limitations on decision-making, placing it in the first intermediate category. Since 1998, as per Polity5’s classification, the executive’s authority was subject to minor institutional constraints during this time. In stark contrast to Polity5 for the relevant regime period, V-Dem’s JCE and LCE are both interpreted by us as indicating moderate constraints on the executive.

On this date a transitional government under President Pierre Buyoya was formally inaugurated following the implementation of the Arusha Peace and Reconciliation Agreement of 08/28/2000. We code this regime change event as a negotiated transition. Out of 26 cabinet portfolios, the Hutu groups got 14 and the Tutsi groups 12 ministries.¹⁶ In accordance with the agreement, President Buyoya transferred power to his Hutu Vice President Domitien Ndayizeye on 04/30/2003. According to LIED, no multiparty executive or legislative elections were held during the specified period. Per FH, for the period until 2002 the country scores between 11 and 14 as not free, which we also interpret as not free. From 2003 onward the country scores from 9 to 10 as not free, which we categorize as rather not free. Political liberties remained absent (LIED) and ambiguous according to V-Dem’s PCLI. Between 2002 and 2004, V-Dem’s JCE and LCE are both interpreted by us as indicating moderate constraints on the executive. For the year 2005, V-Dem’s JCE is interpreted by us as indicating that judicial constraints on the executive were limited, and V-Dem’s LCE is similarly interpreted as indicating that legislative constraints on the executive were also limited. The decisive criterion for the classification is the legitimation pattern: the transitional government derived its authority exclusively from the Arusha Peace and Reconciliation Agreement, a negotiated international framework that explicitly mandated caretaker legitimation through power-sharing between Hutu and Tutsi political parties for a specified three-year period terminating in elections, rather than military guardianship claims of crisis necessity. The Arusha Accords stipulated that Buyoya would serve as transitional president for the first 18 months under binding agreement terms requiring transfer of power to his Hutu Vice President on 04/30/2003, with both phases constituting procedural facilitation toward democratic elections rather than substantive sovereignty assertions. This caretaker legitimation pattern is confirmed by the institutional structure: the transitional government operated through ethnic power-sharing cabinets mandated by the Accords rather than military command hierarchies, Buyoya’s authority derived from the negotiated international framework rather than military position, the power transfer occurred as contractually required demonstrating absence of military override capacity, and the regime successfully transitioned to competitive elections in 07/2005 as prescribed by the Arusha timeline.

On this date president Domitien Ndayizeye, who was the leader of the FRODEBU party, oversaw a competitive parliamentary election as mandated by the Peace Accord. The election resulted in a victory for the National Council for the Defense of Democracy – Forces for the Defense of Democracy (CND-FDD) party (Geddes/Wright/Frantz 2014: 51). The 2005 parliamentary elections were largely peaceful and deemed as generally fair, transparent and free by observers.¹⁷ During the Burundian Civil War, the CNDD–FDD was the most significant rebel group active and became a major political party.¹⁸ On 08/19/2005 indirect presidential elections were held. The National Assembly and Senate members were responsible for selecting the new president, who would serve a five-year term.¹⁹ Pierre Nkurunziza, the only candidate representing the CNDD-FDD party, secured a resounding victory with a vote of 151-9.¹⁹ Based on our observations, multiparty executive and legislative elections were held during this period, which aligns with the observations of LIED. According to LIED elections were competitive during this time. V-Dem’s CEI scores an ambiguous cleanliness, and the elections were somewhat free and fair (V-Dem EF&FI). Per FH’s scoring for 2005, the country is classified as partly free with a score of 8, which we categorize as rather not free. From 2006 onward the country scores from 9 to 10 as not free, which we categorize as rather not free. Regarding the political liberties they were absent according to LIED and ambiguous by V-Dem’s PCLI. According to Polity5, during this period, the executive’s constraints fell into Intermediate Category 3, between substantial limitations and executive parity or subordination. During this regime period, V-Dem’s JCE is classified by us as indicating that judicial constraints on the executive were limited, while V-Dem’s LCE is classified by us as indicating that legislative constraints on the executive were also limited.

Before the second post-war presidential elections were conducted on this date, 250 opposition figures were arrested after the FNL had led a boycott of presidential and legislative balloting. Despite that, the communal elections in May 2010 were deemed free and fair by international observers, but there was no real competition. Due to further boycotts by the opposition, there was only one candidate in the presidential elections, which were different from 2005. The elections were won by incumbent Nkurunziza (CNDD–FDD) with 91.6% of the vote who held office since 2005 and thus was elected for another term (Lansford 2021: 242). Based on our observations, multiparty executive and legislative elections were held during this period, which aligns with the observations of LIED. LIED classifies the elections as not competitive since 2010. Since 2010 V-Dem’s CEI and V-Dem’s EF&FI scores dropped in a range that we interpret in the way that elections were neither free, fair, nor clean. Per FH’s evaluation for the regime period until 2013, the country scores from 9 to 10 as not free, which we categorize as rather not free. From 2014 onward, the country is scored from 11 to 14 as not free, which corresponds to our interpretation of not free. According to LIED political liberties were absent for this period. V-Dem’s PCLI classifies the presence of political liberties as ambiguous from 2010 to 2012, not really present in 2013 and 2014 and as absent from 2015 onward. From 2010 to 2014, according to Polity5, the executive’s constraints fell into Intermediate Category 3, between substantial limitations and executive parity or subordination. Since 2015, as per Polity5’s classification, the executive experienced moderate limitations on authority, placing it in the second intermediate category. For the years 2011 and 2012, V-Dem’s JCE is interpreted by us as indicating that judicial constraints on the executive were limited, and V-Dem’s LCE is similarly interpreted as indicating that legislative constraints on the executive were also limited. For 2013 and 2014, V-Dem’s JCE is interpreted by us as indicating that judicial constraints on the executive were absent, while V-Dem’s LCE is classified by us as indicating that legislative constraints on the executive were limited. For the rest of the regime period, V-Dem’s JCE is interpreted by us as indicating that judicial constraints on the executive were absent, and V-Dem’s LCE is similarly interpreted as indicating that legislative constraints on the executive were also absent. Although constitutional regulations limit a president’s time in office to two terms, Nkurunziza was nominated again by the CNDD-FDD for a third term in 2015 which led to popular uprising and an attempted coup (Lansford 2021: 242). Again, a variety of parties boycotted the polling. Presidential elections took place on 07/21/2015, and Nkurunziza was re-elected for a third term. International observers described the elections as unfair and tainted (Lansford 2021:242). Ahead of the 2020 presidential elections, Nkurunziza endorsed Ndayishimiye as his successor, who was elected president with 71.5% of the vote in balloting on 05/20/2020 and was inaugurated on 06/18/2020 after the sudden early passing of Nkurunziza (Lansford 2021: 242).²⁰ The elections were deemed unfair. There were reports of irregularities on election day.²¹ Despite his military background, there is no reason to believe that this period could be a military autocracy. One of the major traits of electoral politics in Burundi is the prevalence of ongoing incumbency which hinders oppositional parties from developing strong institutions and means to compete with their opponents. Thus, the incumbent party retains control over governance, the elections process and even public opinion (Palmans 2012). The campaign of the parliamentary election, held on 06/05/2025, was fundamentally flawed due to the exclusion of the main opposition party, the National Congress for Liberty (CNL). The CNL’s leader was controversially ousted and barred from participating. This manipulation resulted in an expected but heavily disputed landslide victory for the ruling CNDD-FDD party. The CNDD-FDD won all 108 seats in the National Assembly with 96.51% of the vote, according to the official results.²²

Electoral Autocracy as of 07/01/2025 continued.

Additional sources (Basedau 1999, Kirschke 2007)